Pattern recognition in diagnostic imaging Peter Corr MBChB, FFRad (D) SA, FRCR ...

Pattern recognition
in diagnostic imaging


Peter Corr
MBChB, FFRad (D) SA, FRCR


Professor of Radiology
Nelson R Mandela School of Medicine


University of Natal,
Durban


South Africa


In collaboration with
Wilfred Peh, Wong Siew Kune, Leonie Munro, William Rae, Fei Ling Thoo, Lai Peng Chan,


Lesley A. Goh, Lawrence Hadley, Malai Muttarak, Swee Tian Quek.


Medical Artist: Merle Conway
Photography: NV Chetty, S Ezikiel


Diagnostic Imaging and Laboratory Technology
Essential Health Technologies


Health Technology and Pharmaceuticals
WORLD HEALTH ORGANIZATION


GENEVA




WHO library Cataloguing-in-Publication Data


Pattern recognition in diagnostic imaging; Peter Corr; in collaboration with Wilfred Peh ... [et al.]


1.Diagnostic imaging- methods 2.Pattern recognition 3.Radiography, Thoracic -methods 4.Musculoskeletal system
-radiography. 5.Radiography, Abdominal-methods 6.Manuals I.Corr, Peter II.Peh, Wilfred.


ISBN 92 4 154632 8 (NLM classification: WN 180)


This publication is a reprint of material originally distributed as WHO/DIL/01.2


© World Health Organization 2001


Reprinted 2003 (Twice)


All rights reserved. Publications of the World Health Organization can be obtained from Marketing and
Dissemination, World Health Organization, 20 Avenue Appia, 1211 Geneva 27, Switzerland (tel: +41 22
791 24 76; fax: +41 22 791 4857; email: bookorders@who.int). Requests for permission to reproduce or
translate WHO publications - whether for sale or for noncommercial distribution - should be
addressed to Publications, at the above address (fax: +41 22 791 4806; email: permissions@who.int).


The designations employed and the presentation of the material in this publication do not imply the ex-
pression of any opinion whatsoever on the part of the World Health Organization concerning the legal
status of any country, territory, city or area or of its authorities, or concerning the delimitation of its
frontiers or boundaries. Dotted lines on maps represent approximate border lines for which there may not
yet be full agreement.


The mention of specific companies or of certain manufacturers' products does not imply that they are
endorsed or recommended by the World Health Organization in preference to others of a similar nature
that are not mentioned. Errors and omissions excepted, the names of proprietary products are distin-
guished by initial capital letters.


The World Health Organization does not warrant that the information contained in this publication is
complete and correct and shall not be liable for any damages incurred as a result of its use.


Designed by minimum graphics in New Zeland
Typeset by Best-set in Hong Kong
Printed in Malta By lnterprint Ltd.




iii


Contents


-- JAN 200~


Preface v
Foreword vii
Definitions X


Part 1. Technique, quality control and radiation protection 1
Chapter 1. Image quality optimisation and control 3
Chapter 2. Radiation protection in radiological practice 16
Chapter 3. Contrast media in imaging 21
Chapter 4. Digital imaging and telemedicine 23


Part 2. Chest imaging patterns 27
Chapter 5. The normal chest radiograph 29
Chapter 6. Pulmonary infection 34
Chapter 7. Lung cancer 40
Chapter 8. Pulmonary hypertranslucency and cystic lungs 45
Chapter 9. Pleural and extra pleural disease 51
Chapter 10. Rib lesions 56
Chapter 11. Chest trauma 59
Chapter 12. Pulmonary AIDS 65
Chapter 13. Paediatric chest 68
Chapter 14. Cardiac disease 73
Chapter 15. Mediastinal masses 77
Chapter 16. Diaphragm lesions 79
Chapter 17. Pneumoconiosis 80


Part 3. Musculoskeletal patterns 83
Chapter 18. Approach to focal bone lesions 85
Chapter 19. Periosteal reactions 91
Chapter 20. Extremities trauma 99
Chapter 21. Fractures-classification, union, complications 114
Chapter 22. Spinal trauma 124
Chapter 23. Facial and pelvic trauma 133
Chapter 24. Bone infections 141




PATTERN RECOGNITION IN DIAGNOSTIC IMAGING


iv


Part 4. Gastrointestinal and urinary tract patterns 147
Chapter 25. Plain abdominal radiographs 149
Chapter 26. The acute abdomen 151
Chapter 27. Gastrointestinal contrast studies 159
Chapter 28. Paediatric abdomen 178
Chapter 29. Urinary tract imaging 183


Acknowledgements 205




Preface


As modern, high technology based diagnostic imaging is moving increasingly into
therapeutic medicine, and molecular imaging is becoming daily routine, it is impor-
tant to remember that thousands of hospitals and medical institutions worldwide do
not even have possibilities to perform the most basic examinations. Today, few other
areas of medicine experience such a rapidly growing gap between what might be
technically possible, e.g., what can be done in highly developed, rich countries com-
pared to what is the reality in many less fortunate areas of the world.


As the ultimate target for the World Health Organization is to provide Health For
All, it is with great pleasure and sincere gratitude to Professor Carr, his staff and co-
authors that this book on Pattern Recognition in Diagnostic Imaging is now being
published and distributed. It aims in a simple, but precise way at assisting medical
professionals doing a tremendous work to save lives and reduce suffering in countries
where diagnostic imaging has not yet reached the stage of molecular imaging.


We would warmly recommend that this book should not be put on a shelf or into
a locker, but be used by everybody whose obligation it is to prescribe, perform, or
interpret simple, but often life-saving diagnostic imaging procedures especially in
locations where the presence of qualified and fully trained specialists would be a rare
exception.


The book is developed and published as a WHO Document under the umbrella of
the Global Steering Group for Eduction and Training in Diagnostic Imaging. For
further information, please contact:


Team for Diagnostic Imaging and Laboratory Technology,
World Health Organization
20, Avenue Appia
CH-1211 GENEVA 2 7, SWITZERLAND


Fax: +41 22 791 4836; e-mail: ingolfsdottirg@who.int


Geneva, 30 June 2001
Harald Ostensen, MD


y






Foreword


Imaging is currently being performed and interpreted by radiographers/technologists
and primary care physicians/hospital medical officers in many developing countries.
Many primary care physicians have had little or no training in the interpretation of
images, both radiographic and sonographic. Radiographers are trained in producing
images but often do not have the background in medicine to interpret images with
confidence. This book seeks to bridge this gap by providing images of common
pathologies seen in many developing countries in a pattern format. The pattern
recognition format has been used successfully by both national and international
radiographic societies to educate and train radiographers working in regions where
radiology advice or services are unavailable.


We hope this book serves you well in your daily work which involves imaging.


Peter Corr
Durban 2001


vii






Authors


Lai Peng Chan MBBS, FRCR
Registrar, Department of Diagnostic Radiology, Singapore General Hospital, Singapore


Peter Corr MBChB, FFRad (D) SA, FRCR
Professor of Radiology, Nelson R Mandela School of Medicine, University of Natal,


Durban, South Africa


Lesley A Goh MBBS, FRCR
Registrar, Department of Diagnostic Imaging, Tan Tack Seng Hospital, Singapore


Lawrence Hadley MBChB, FRCS (Edin)
Professor of Paediatric Surgery, Nelson R Mandela School of Medicine, University of


Natal, Durban, South Africa


Wong Siew Kune MBChB, FRCR
Associate Consultant, Department of Diagnostic Radiology, Singapore General Hospital,


Singapore


Leonie Munro Nat Dip Radiography (D), MA (Unisa), Dip Public Admin-postgrad
(UDW), Cert for trainers (Unisa)


Chief Tutor School of Radiography, King Edward VIII Hospital, Durban, South Africa


Malai Muttarak MD
Professor of Radiology, Department of Radiology, Chiang Mai University Medical


School, Chiang Mai, Thailand


Wilfred C G Peh MBBS, MD, FRCR, FRCPE, FRCPG
Senior Consultant, Department of Diagnostic Radiology, Singapore General Hospital,


Singapore


Swee Tian Quek MBBS, FRCR
Consultant, Department of Diagnostic Imaging, National University of Singapore,


Singapore


William Rae MBChB (Wits) PhD (OFS)
Senior Medical Physicist, Addington Hospital, Durban


Fei Ling Thoo MBBS, FRCR
Consultant, Department of Radiology, Changi General Hospital, Singapore


ix




X


Definitions


ALARA keeping radiation dose 'as low as reasonably achievable'


AP anteroposterior means patient is facing the X-ray tubejbeam (see PA)
Atelectasis radiographic pattern to describe (i) incomplete expansion of lungs at birth, or


(ii) collapse of adult lung usually with limited re-expansion
Baud number of bytes transmitted in one second in telemedicine


Bit smallest unit of digital information


Byte a group of 8 bits used to transmit a value of character


Collapse radiographic pattern of partially or completely airless lung due to some form of
obstruction


Consolidation a region of lung opacification following pneumonia with air bronchograms.
Strictly a pathological term for lobar pneumonia.


CTR cardia-thoracic-ratio is the ratio of the measurement of widest transverse diameter
of the heart on a chest radiograph versus the widest transverse ratio of the thoracic
cage


Decubitus view patient lying on either left or right side and radiograph is taken using a
horizontal X-ray beam at right angles to the cassette placed either behind the patient
(PA decubitus) or in front of patient (AP decubitus)


DICOM a standard allowing interfacing of digital imaging devices with other digital
devices


Digitise process to convert analogue data or images into digital data


Effusion fluid in a cavity, e.g. pleural cavity


FFD focal film distance, i.e. distance from source of X-ray beam to the film


Horizontal beam/shoot-through film taken using horizontal X-ray beam at right angles to the
cassette; patient can be supine, prone, semi-erect, lateral


ISDN integrated system digital network


IVU intravenous urography


KUB plain-film-radiograph of abdomen; i.e. kidneys to bladder region


Lossless compression there is no alteration of original image after reconstruction in digital
imaging


Osteopaenia decreased bone density on a radiograph.


PA posteroanterior view with X-ray beam entering from behind the patient and
emerging through anterior part because patient is positioned facing cassette


Sclerosis increased bone density or opacity on the radiograph




PART 1


TECHNIQUE, QUALITY CONTROL AND
RADIATION PROTECTION




j
j
j
j
j
j
j
j
j
j
j
j
j
j
j
j
j
j
j
j
j


J


j
j
j
j
j
j
j
j
j
j
j


j
j
j
j
j
j


j
j
j


j


j
j
j
j




CHAPTER 1
- Rll!l!!!ill


Image quality optimisation and control
Leonie Munro


Introduction
What is pattern recognition in imaging and what are the factors that impact on this
recognition? Pattern recognition may be defined as being able to recognise normal
anatomical and physiological appearances on an image and those variations of
appearances, which may indicate pathology. This implies that certain criteria should
be met, to be competent in pattern recognition. Firstly a person who performs
pattern recognition should have a fair amount of expertise in medical imaging and
knowledge of radiographic anatomy and normal variants so as to identify variations that
may indicate pathology. This is the overarching aim of this book hence the many
aspects of pattern recognition are fleshed out in the other chapters. This chapter
concentrates on factors that impact on image quality. It is not easy to clearly define
optimal image quality. In theory optimal image quality allows one to make accurate
diagnosis. This is an ideal as we also should consider dose to patients in keeping with
the ALARA principle (as low as reasonably achievable). There are times when an image
is suboptimal but not unacceptable. In other words slight deviations in image quality
may not have a significant impact for pattern recognition. Unacceptable images may
cause one to miss a fracture or a destructive lesion. To repeat or not to repeat depends
on the reasons for the examination and whether one can confidently perform pattern
recognition to make a diagnosis. Such a decision is usually based on experience and a
set procedure when evaluating images. Examples of unacceptable radiographs are
included in this chapter to highlight the importance of optimal image quality. If an
image is unacceptable then the radiation received by the patient has no benefit. Thus
there are some important factors andjor basic tests that should be considered at all
times.


It would be difficult to confidently perform pattern recognition if the image quality
of a dynamic image or hard copy is not of an acceptable standard. There is consensus
that optimal image quality entails meeting medico-legal requirements, such as each
image to contain the patient's details, date of examination, anatomical marker, and
adequate visualisation of radiographic anatomy /signs. This means that patient
positioning should be correct for each projection, that the images are not blurred and
that optimal image density is visualised. Image quality thus depends on correct
radiographic techniques being used for each projection, correct selection of exposure
factors and use of suitable imaging systems which are of an optimal standard. Just to
mention that it is usually necessary to do two projections/views, usually at right angles.
The patient/area of interest should be in accordance with recommended projections to
ensure that all relevant anatomical parts are visualised (fig 1.1 a). For example in skull
radiography the patient's head should be straight to allow one to comment of symmetry
of the skull bones (fig 1.1 b). Chest radiographers should always be exposed on full
inspiration to prevent incorrect diagnosis due to unacceptable radiography (fig 1.2).
Apart from positioning criteria the following are considered to have an impact on image
quality.


3




4


Fig 1.2
Poor inspiration PA


chest as only shows
7 ribs. Arrow points


to crimp mark
caused by poor film
handling (fingernail


pressure).


PATTERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 1. TECHNIQUE, QUALITY CONTROL AND RADIATION PROTECTION


Fig 1.1a
Poor radiographic technique.
Ankle and knee joints not on
film.


Fig 1.1b
Poor patient positioning as skull rotated.


Care and maintenance of imaging equipment and
accessories, and some QA tests
Care and maintenance of imaging equipment
is normally being promoted by national
authorities to ensure that staff, patients,
and members of the public do not receive
unnecessary radiation doses. Well maintained
equipment benefits service delivery because
repeat radiographs, due to malfunctioning
units, poorly calibrated generators, and so
forth, decrease. This means that unnecessary
dose to patients is also reduced. Many checks
can be done by radiographers whilst others
require sophisticated test tools which are
usually expensive and/or require the exper-
tise of physicists (Borras, 1997). Reject


analysis should be carried out regularly to determine reasons for poor quality
radiographs. Some basic tests can be done to minimise rejects.


Safe/ight tests
Unwanted film blackening is fog which reduces radiographic contrast. It is important
that darkroom safelighting does not fog unexposed and/or exposed films. Safelight
tests should be done at least every six months to ensure that safelights are in proper
working order.




CHAPTER 1. IMAGE QUALITY OPTIMISATION AND CONTROL


Equipment for the tests
An acceptable film/screen light-tight cassette; black paper one-half the size of the
cassette (2 sheets of black paper needed), clock/timer with second hand, box of
unopened radiographic film, and general X-ray unit capable of selection of low mAs.


Step 1: Switch off all lights in the darkroom and cover lights on the processor. In total
darkness place an unexposed film in the light-tight cassette containing intensifying
screens.


Step 2: Expose the loaded cassette to radiation to obtain approximately 1 density.
Suggested exposure factors-half mAs of finger exposure (see Annexure 2-these
factors are appropriate for use with 400 speed system. Should a slower system be
used then mAs adjustments to be made.) Working tip: If it is not possible to select
low mAs then increase FFD using the inverse square law principle to determine mAs
as per FFD changes.


Step 3: In total darkness in the darkroom open the cassette and using the black paper
block off half of the exposed film. NB this section of the film to remain covered during
the test. The density of the covered part of the film is used to see if the safelights are
functioning correctly. Place the other sheet of paper on the other half of the film and
move the paper down to uncover part of the unexposed film. Switch on a safelight
and expose the uncovered film portion for 60 seconds. Move this sheet of paper off
the film and expose the remaining uncovered film half to a further 60 seconds.


Step 4: Process the film which has half exposed to radiation only and the other half to
radiation plus light from the safelight.


Step 5: Check the density of the film (fig 1.3). The film- half that was covered
throughout the test should have a density of approximately 1. Acceptable density
limits when comparing the half not exposed to the safelight and the side exposed to
the safelight should not exceed 0,02 for 60 seconds (Barber & Thomas, 1983: 176).
The above steps to be repeated to check each safelight in each darkroom. Should


there be unacceptable density /fogging then check safelight position: correct height from
working surface should not be less than 120
centimeters. Safelight filters and wattage of
safelight bulbs to be checked so that the fault
is then corrected. The indicator light on
the processor should be checked as per the
above steps. Indicator light to be uncovered
and safelights switched on. Density should
not exceed 0,05 for a 2 minute exposure to
indicator light plus safelights.


Careful film-handling and film storage
Note that exposed film is more sensitive to
light thus care should be taken when handling
film. Crimp marks from excessive handl-
ing may be of concern. A problem that
has recently surfaced is that of darkroom
personnel/radiographers using cell phones
whilst handling film in the darkroom (fig 1.4).


Film should always be handled with clean
hands and in a dust free environment. Film to


Fig 1.3
Safelight test:


5


1 = film-half only
exposed to
radiation to obtain
approximately
density 1. Sections
2 & 3= film
exposed to safelight
plus radiation.
Section 2 exposed
to safelight for 2 x
60 seconds, section
3 for 60 seconds.
Both also exposed
to radiation.
Obvious increased
density due to safe-
light; this exceeds
0,02 density limit
compared to 1.




6


Fig 1.4
Example of film-


fogging caused by
LED of cell phone in


a darkroom.


Fig 1.5
Sensitometric strips


showing steps 1
to 21.


PATIERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 1. TECHNIQUE, QUALITY CONTROL AND RADIATION PROTECTION


be stored in boxes in cool room with good air
circulation. Boxes of film must never to be
stacked on top of each other as this will cause
marks on the films.


Processor control
This quality control method should be done
for all processors to reduce unnecessary
repeats caused by processing factors, such
as exhausted chemistry, incorrectly mixed
chemistry, and incorrect replenishment.
Monitoring film quality due to processing
factors means that assessment of film
contrast, film speed and base fog is done
based on an objective method.


Equipment required
A sensitometer to expose film to different
light intensities in steps (fig 1.5); a den-
sitometer to measure optical density of
selected sensitometric steps, a non-mercury
thermometer to manually check the tem-
perature of the chemistry, and a box of
unexposed film and sheets of processing
control charts or graph paper.


Step 1: Under safelight conditions expose one
film to the sensitometer. It is important that
blue light be used for monochromatic (blue
sensitive film) and green light for
orthochromatic film. Select appropriate
switch on the sensitometer. Note that some
sensitometer only produce blue light thus
can only be used with blue sensitive film.
Process film after checking temperature of
chemistry as per step 2.


Step 2: Temperature of chemistry to be taken
using the thermometer. Temperature gauge


readings to be recorded to check with thermometer readings. This is done to ensure
that the gauge is accurate.


Step 3: Process exposed film.


Step 4: Using a densitometer read densities of each step.


Step 5: The sensitometric step with the density closest to 1, 20 (mid-density) to be used
to determine speed index. If sensitometric step 9 has density closest to 1,20
(including base fog) then all subsequent readings for speed index to be at this
sensitometric step (ie step 9) for all future films used for processor control.


Step 6: On chart record temperature, date and base fog reading.


Step 7: Plot position for speed index obtained in the above step.




8


PATTERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 1. TECHNIQUE, QUALITY CONTROL AND RADIATION PROTECTION


PROCESSOR PERFORMANCE LOG
MONTH:


PROCESSOR TYPE:


PROCESSOR LOCATION:--------------


DAY


Speed
Index


Contrast


+ 0,15


-0,15


Index .• _,5


Gross
Fog


-0,15


Developer
Temperature


.


Fig 1.6a


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31


1 2 3 4 5 6 7 8


,..,_


L


+;t+-sCl"·l,i
~4 +


-+


,-~ _; ~ -·~.:;: ,. ~ M ;


,·.,._.-. r.


: ~_;- ::.;_t_ :_ f.··--


+ ·+'- + ..
- 1-l ::-r+:


~+


Bt + ttc:~l::::t


9 10 11 12 13


tt
14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31


Example of blank graph paper for processor monitoring.


Steps for the test
Place the cassette on the table and centre primary beam to center of cassette. Open light~
beam diaphragms to set measurements-suggest 30 x 30 ems (note readings on the
scale). Ensure FFD is at 100 ems or higher. Record this measurement. Place a coin in
each corner of the light square and place a metal corner clip/allan key exactly at the
edge of each corner. Position L or R marker at center of beam. Expose the cassette:
suggest 50 kVp and 4 mAs for 200 speed system. Process the film (fig 1.8). The
suggested performance criterion is -+2% of source to image distance. If FFD is 100 ems
then upper limits of difference between light field edges and edges of primary beam
radiation visualised on the film should not exceed 2% of 100 ems .ie 2 ems (Borras,
1997: 253). This will entail measuring the area of light beam based on position of the




CHAPTER 1. IMAGE QUALITY OPTIMISATION AND CONTROL


Step 8: For 5 consecutive days repeat the preceding steps to obtain average density at
sensitometric step 9. This value will be the control speed index against which all
future sensitometric films will be compared.


Step 9: Plot average speed index on chart/graph paper as obtained over the five days.


Step 10: Draw 2 lines above and below the speed index. One parallel line to be +0, 15 of
speed index and the other to be at -0, 15 from the speed index. Deviations outside
these 2 lines means that there is an unacceptable processor problem. For example
replenishment may have been decreased, the temperature may have decreased/
increased, and so forth.


NB: Base fog reading of each film, the date the test was done, and temperature of
chemistry to be recorded on the chart. Working tip: Always do processor control at
the same time.


Some firms supply pre-exposed sensitometric films but it is important to only use the
film within a given period because with time film fog increases. If a densitometer is not
available then do visual checks by placing the strips on an illuminator but ensure that
all extraneous light is masked off and ambient (overhead) light switched off. Visual
comparisons are a crude method but preferable to no checks at all. Some film suppliers
have facilities to read film for customers. Suggestion: find out from film supplier
whether such facilities are available. Arrangements could be made to post batches of
film strips to the supplier. Recommendations: density readings to be written next to
each sensitometric step on each film then this should allow one to compared with visual
comparisons. However the proper method of processor control should where possible
be used for valid objective testing to enhance processor control results which enables
speedy problem solving (fig 1.6a, 1.6b).


• Film-screen contact test
It is essential that images be obtained with good film-screen contact. Poor film-screen
causes loss of information which may cause inaccurate pattern recognition. The film-
screen contact test tool is readily available but a bit expensive. To perform the test place
the contact mesh tool (wire mesh encased in perspex) on the top of the suspect cassette
containing an unexposed film. Centre to center of cassette, collimate to cover cassette.
Make sure that table on which the cassette is placed and the central ray are at right
angles. Expose the test tool using approximately 55-60 kVp and 4 mAs (for 200 speed
system) and 100cms Focal film distance. Process the film and view at a distance of 150-
180 ems to evaluate the sharpness of the wire mesh. Poor contact is seen as a "blurred"
outline (fig 1. 7). Poor film-screen contact usually occurs when a cassette gets dropped
when excessive force is used during handling.


• Collimator-beam alignment test
This test should be done at least every month to check proper alignment of collimator
and primary beam as daily use of the collimator contributes to poor alignment of the
light beam and primary beam. This in turn causes suboptimal positioning as it may be
difficult to accurately centre as per routine techniques.


Equipment for the test
35 x 43 cassette or smaller lloaded with unexposed film. Four coins/steel washers and
metal clips/allan keys. Lead markers (L & R).


7




10


Fig 1.9
Arrows indicate


middle of film; cross
and film-centre not


aligned.


PATTERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 1. TECHNIQUE, QUALITY CONTROL AND RADIATION PROTECTION


Fig 1.7
Film-screen contact test: arrows show areas of poor
contact.


Fig 1.8
Two fields do not match; arrows show actual light-
beam edges. Difference within the acceptable range
of not more that 2% of FFD.


difference in densities is not very
marked; we perceive shades of grey.
This means that at times a short
scale contrast image may be needed
for diagnostic purposes. A typical
example is soft tissue thighs to
visualise low density calcification
for query cysticercosis. Kilovoltage
selection of approximately 65-75 is
usually recommended for visuali-
sation of renal size, shape, and
position. On the other hand
>llOkVp is required to adequately
visualise the gastric-intestinal-tract


when doing contrast studies. Chest radiography should always be performed using
moderately high to high kVp (fig l.lOa-l.lOg).


Image sharpness refers to the amount of detail that we can see when viewing an
image. Good film-screen contact enables high image sharpness. Detail screens are used
to visualise fine detail but this results in additional dose to the patient due to mAs
adjustments (fig 1.11 and fig 1.12}.
• There are several factors that effect contrast and sharpness. Image contrast is broadly


divided into three categories: subjective, subject/patient, and objective. Image
sharpness is dependent upon geometric factors, movement factors, and systems
factors. The most important factor influencing contrast is kilovoltage. Some
radiographers make use of a fixed Kv and adjust mAs to suit different types of
patients and/or pathology. Most modern units have automatic exposure devices
which are programmed for the various anatomical regions. What one has to consider
is dose to patients thus it should be borne in mind that high mAs selections result
in the patient receiving additional radiation. Selection of the fastest film-screen




CHAPTER 1. IMAGE QUALITY OPTIMISATION AND CONTROL


DAY
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31


Speed
Index + 0,15


. 0,15


1:::1


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31


Contrast
Index


Gross
Fog


Fig 1.6b


+ 0,15


~ 0,15


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31


Parallel lines above and below speed and contrast indices.


coins and metal clips and then comparing this measurement with total area of film-
blackening to establish performance criteria.


• Test to check alignment of beam
This simple test can be done by placing a cassette loaded with an unexposed film on the
table. Reduce lateral diaphragms to a slit. Close the other diaphragms. Centre to cassette
and expose using 60 kVp and 4 mAs. Close these diaphragms and open the others to a
slit and expose again. Process the film and bend in half to check that exposed "cross"
(fig 1.9) is in the middle of the film. If there is not proper alignment check that the tube
is straight so that the primary beam is vertical at 90 degrees to the cassette/table top.
This test can also be used alignment of central ray to the bucky tray.


Factors relating to contrast and sharpness of the image
Contrast refers to the difference in density (film blackening) of two areas. To put it
simply an image that only has two densities/tones will have high contrast as it only
has a short scale such as a black/white image. Long scale contrast occurs when the


9




12


Fig 1.11
Example of different


densities of
calibrated step-


wedge. Step-wedge
of Right not same
as the other two.
Note ball-bearing


taped on step 8 to
make it easy to


align similar density
steps. On right step


4/5 aligned to
match steps of the


other two. Right
step-wedge of a


fast system thus if
system on Left is


used then mAs has
to be adjusted
as per scale in


Figure 1.13.


PAITERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 1. TECHNIQUE, QUALITY CONTROL AND RADIATION PROTECTION


Fig 1.10f, g
PA chest of same patient. Film (f) taken using 75 kV. Film (g) taken using 110 kV and 4 mAs. Note improved
visualisation of entire lung-fields and good inspiration.


combinations will reduce dose but one
may not see fine detail. The deciding factor
should be the reason for the examination
paying attention to ALARA_ Careful posi-
tioning should be practiced with appropriate
exposure factors to produce an image with
film blackening within the useful density
range. A dark image is unacceptable as is a
pale image (fig 1.12).
Factors influencing mAs selection include:
FFD as per the inverse square law, speed of
imaging system, collimation of the primary
which reduces scatter/secondary radiation
thus contrast is improved, and use of
secondary radiation grids. It is necessary to
increase mAs when reducing the area of the
primary beam (coning) to compensate for loss
of film blackening caused by scatter. Grid
exposures require an increase in mAs based
on the grid ration. For example a 8: 1 grid


ratio has a grid factor of 4 (half of 8) thus the mAs has to multiplied by 4 when moving
from non-grid to grid technique. Grids improve contrast but reduce image sharpness as
the distance between the patient and film is increased. The output of the generator also
influences mAs selection, eg single phase unit requires more mAs than a high output
generator. A simple method to establish difference is to perform a stepwedge test as
described in this chapter (see page 14). Contrast is high when images are produced
using single phase units as the overall kV is less than the selected kV. The benefits of
higher output generators far outweigh the loss of contrast as such units enable use of
low mAs settings, short exposure times, and a range of kV selections allowing the use
of moderately high to high kv techniques to reduce dose to patients in keeping with
the ALARA principle.




CHAPTER 1. IMAGE QUALITY OPTIMISATION AND CONTROL


Fig 1.10a
Lateral view of femur a bit
over-penetrated for
visualisation of soft tissue
but image acceptable.


Fig 1.10b
AP supine abdomen (control) of 10 year male
patient for barium enema. Note lack of gonad
protection. Exposure selection not based on ALARA
as low kV and high mAs (70kVp, 40mAs) used.


Fig 1.10e


11


Fig 1.10c
AE film of same patient; again no gonad protection
and inappropriate exposure selection.


Too much mAs used for the chest hence lung
markings not demonstrated.


Fig 1.10d
Delayed AE of same patient. Radiograph taken by
different radiographer who applied ALARA principles,
gonad protection, 88 kV used and mAs decreased
accordingly to 5 mAs. Note improved visualisation of
large bowel patterns due to long-scale contrast
image.




CHAPTER 1. IMAGE QUALI1Y OPTIMISATION AND CONTROL


NUMBER FOR EXPOSURE (mAs) ADWSTMENT


13


Fig 1.12a, b, c
Range of
unacceptable
images. Insufficient
kV used for lateral
spine (a); poor
patient positioning
for peg view as C1/
2 not demonstrated
due to overlying
occiput (b); not true
lateral view for
scapula plus
insufficient kV used
(d).


I 2 3 4 5 6 7 8 9 10 I ' I ' I ' I I I ' ' " ' I ' I ' I ' I " d " 'II I " " d d 'I, I II II 1111111111 ' 11,1' " ''"'I, I II I, I ' ,I, IIIII I' I lflljll Tl ljiiii,IIIIJII '/lflljllllllfjllll//1/ljl 'I""Jif llfjllli/IIIIJI fTIIIJIIII
.0 .2 .3 .4 .5 .6 .7 .8 .9 .0


Fig 1.13


MANTISSA OF LOGARITHM


Practical use of a stepwedge to determine mAs adjustment when comparing 2 different speed
systems. Note if 4 step difference then log= 0.304 then antilog = 2.


Scale for determining logarithms to calculate adjustments to mAs for different speed systems.




14
PAITERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 1. TECHNIQUE, QUALITY CONTROL AND RADIATION PROTECTION


• Subjective contrast includes how tired one is when viewing an image, the brightness
of the illuminator (viewing box), and the brightness of surrounding light. Where
possible illuminators should be kept clean and have the same intensity and light
colour. Images should be viewed in a room which has subdued lighting.


• Subject/patient contrast includes selection of kVp (exposure factors), the thickness of
the part being imaged, use of contrast media, and scatter-secondary radiation. If we
require an image with high contrast/short scale then low kVp factors are selected but
when a long scale image is important then high kVp should be used.


Determining mAs adjustments: stepwedge tool
A calibrated aluminium stepwedge is a basic useful tool as it can be used to establish
mAs adjustments when using different speed systems.


• Method
Expose each system with same film type if determining different speed of intensifying
screens. The same factors should always be used. For example 70 kVp and 5 mAs at
120cms FFD. NB never use less than 0,1·second exposure time to allow the screens to
respond to the quanta to emit light (ie to minimise reciprocity failure). Process the films.
Place the films edge to edge on a horizontal view box and mask all unwanted light
areas. Visually compare the steps and adjust to align with a selected step. It makes it
easier to do if a metal balnearing is taped to the middle step of the wedge (see fig 1.11}.
If there are 4 step differences obtained when using a calibrated step-wedge with a
calibration factor of 0,076 at 70 kVp, 5 mAs and at 120 centimetres FFD. then the
logarithm value is 4 x 0,076 = 0,304 (fig 1.13) The exposure factor for mAs adjustment
can be determined by reference to a table of logarithms, by use of a slide rule (D and L
scales) or by reference to the scale at the back of the step-wedge (if supplied) or by the
above figure. Note that the antilogarithm of 0,3 is 2. The mAs would have to be halved
if moving from a slow to faster system or vice versa.


The step wedge method may be used to check output of units. ideally. the resultant
stepwedges should be the same density for a specific unit. Should there be differences
and if the processing factors are optimal then this information could be useful when
communicating with the firm/repair technician. Making use of a calibrated stepwedge
to regularly check output of units saves money and time. This tool pro.duces crude
results but is inexpensive.


Optimal image quality requires careful positioning, regular maintenance and care of
equipment, and careful selection of exposure factors. The acid test being that one is
confident when engaged in pattern recognitimi.


References
1. Ballinger PW. 1998. Merrill's atlas of radiographic positions and radiologic procedures. St


Louis: Mosby Year Book.


2. Bontrager KL & Anthony BT. Textbook of radiographic positioning and related pathology.
Second Edition. StLouis: CV Mosby Company.


3. Clifford MA & Drummond AE. Radiographic techniques related to pathology. Third Edition.
London: Wright PSG.


4. Freeman M. 1988. Clinical imaging: an introduction to the role of imaging in clinical practice.
New York: Churchill Livingstone.


5. Goldman M & Cope D. Radiographic index. Eighth Edition, London: Wright PSG.




CHAPTER 1. IMAGE QUALITY OPTIMISATION AND CONTROL


6. Swallow RA, Naylor E, Roebuck EJ & Whitley AS. Clark's positioning in radiography.
Eleventh Edition. London: William Heinemann Medical Books, Ltd.


7. Barber TC & Thomas JM. 1983. Radiologic quality control manual. Reston: Prentice-Hall
Company.


8. Bluth EI, Havrilla M & Blakeman C. 1993. Quality improvement techniques: value to improve
the timeliness of the preoperative chest radiographic report. AIR, 160: 995-998.


9. Borras C. Editor. 1997. Organisation, development, quality assurance and radiation
protection in radiology services: imaging and radiation therapy. Washington DC: PAHO/
WHO.


10. Gould R & Boone JM. 1996. Syllabus: categorical course in physics: technology update and
quality improvement of diagnostic imaging equipment. Oak Brook: RSNA Learning Center.


11. Pizzutiello RJ & Cullinan JE. 1993. Introduction to medical radiographic imaging. Rochester:
Eastman Kodak Company.


12. Smit KJ. 1996. Proposed regulations and the role of quality control in diagnostic radiology.
Bellville: Directorate Radiation Control, Department of Health, South Africa.


13. Thornhill PJ. 1987. Quality assurance in diagnostic radiography: are we using it correctly and
what is the future? Radiography, Vol 53, No 609: 161-163.


14. Ball J & Price T. 1990. Chesneys' radiographic imaging. Oxford: Blackwell Scientific
Publications.


15. Burns EF. 1992. Radiographic imaging: a guide for producing quality radiographs.
Philadelphia: WE Saunders Company.


16. Curry TS 111, Dowdey JE & Murry RC. 1993. Christensen's introduction to the physics of
diagnostic radiology. Philadelphia: Lea & Fabiger.


17. Stockley SM. 1988. A manual of radiographic equipment. London: Churchill Livingstone.
18. Wilkes R. 1985. Principles of radiological physics. London: Churchill Livingstone.


15




16


CHAPTER 2


Radiation protection in
radiological practice
William Rae


X-ray production
The radiation emitted from X-ray units while taking X-rays is photon radiation, and
these photons are known as X-rays. They are generated when high energy electrons,
accelerated by a high voltage potential difference, strike a target in the X-ray tube and
their energy is converted to photons which radiate out from the target. The energy of
the electrons, and hence the resultant photons, is expressed in terms of thousands
of electron volts (keV). Photon energies used in diagnostic radiology are in the range
20 keV to 150 keV. These photons have enough energy to cause ionisation, resulting
in deposition of energy in the irradiated material. This energy deposition results in
a reduction in the number of photons in the beam, and the beam is said to be
"attenuated" by the absorbing material. The amount of energy absorbed differs for
materials of different density or atomic composition. The differential absorption of
X-rays between different structures, allows the creation of the contrast that is seen
on X-ray films, and is diagnostically useful.


When a patient is exposed to an X-ray beam a large amount of radiation is also
produced in other directions. Much of the radiation entering the patient is scattered and
exits the patient in all directions. Some of the photon energy is lost during the scattering
process, so the scattered photons are of a lower energy than the primary photons. For
most radiographic procedures only about 1 to 10% of the primary beam emitted from
the X-ray tube actually interacts with the detector system. (This excludes the photons
absorbed by the casing and collimators of the X-ray tube). The rest of the energy is lost
due to scatter or absorption in the patient. With newer, and more efficient, detector
systems, less radiation is required to produce diagnostic quality images. This reduces
exposure to the patient.


Biological effects of X-ray radiation
The damaging (negative) effects of X-rays were noticed soon after their discovery. Early
workers were initially unaware of the associated risks and thus took no precautions
against being exposed to the X-ray beam whilst imaging patients. Skin damage and
induced cancers were soon attributed to the extreme overexposures experienced by
early radiation workers. As a result, efforts to understand and limit the negative
biological effects have been made since the very early years of diagnostic radiology.
Most of the information about the negative effects of radiation comes from nuclear
power plant accidents, atomic explosions, radiotherapy patients, and radiation workers.


The biological effects of X-ray radiation are due to the ionisation that occurs when
photon energy is deposited in living tissue. Intracellular water is ionised producing free
radicals that can damage either the genetic material of the cell in the DNA of the
chromosomes, or the intracellular organelles. The resultant effects are related to the
amount of radiation absorbed by the tissues.




CHAPTER 2. RADIATION PROTECTION IN RADIOLOGICAL PRACTICE


Negative effects associated with X-ray radiation
The intracellular damage that takes place after exposure results in two main categories
of effects. Either the damage to the cells results in immediate effects, which may result
in progressively worsening function and eventually cell death, or the damage to the
cell's genetic content allows it to live and reproduce, but ultimately result in cancer
after some delay period. The risk of seeing both these types of effects increases with
increasing radiation exposure to the individual. The direct damage effects are only seen
above some relatively high threshold level of exposure and they worsen as the dose
increases. The chance of getting cancer though is low, but can be induced by low
amounts of radiation. Cancer induction is an all or nothing effect, and there is a
statistical chance that it will occur. The chance of occurrence is proportional to the
amount of exposure.


Rapidly dividing cells are more sensitive to radiation effects, and the most sensitive
tissues are thus gastrointestinal tract, gonads, bone marrow, and skin. The tissues most
susceptible to radiation induced malignancy are bone marrow, bowel mucosa, breast
tissue, gonads and lymphatic tissues.


The foetus is most susceptible to radiation at about 20 to 40 days post conception,
and microcephaly and mental retardation are the most likely effects, followed by an
increased incidence of cancer in childhood.


Quantification of radiation dose
The amount of radiation delivered to an object (or person, or tissue) can be quantified
as the energy (joules), deposited per unit mass (kilogram), of whatever is being exposed
to the radiation. The specific SI unit for radiation dose is the gray (Gy), which is defined
as one joule per kilogram. Different biological tissues respond differently to different
types of radiation and to account for this a biological weighting factor is used. The SI
unit of biologically effective dose is the sievert (Sv). This is also measured in terms of
joules per kilogram, but accounts for the biological response of the particular tissue
being irradiated.


Radiation protection regulations
Regulations have been developed internationally over many years to control the amount
of exposure that is allowed, and thus to minimise the incidence of radiation effects. The
latest relevant publications are ICRP Publication 60, printed in 1990, and titled 1990
Recommendations of the International Commission on Radiological Protection, and the
International Basic Safety Standards for Protection Against Ionizing Radiation and for
Safety of Radiation Sources jointly published in 1996 by the WHO, IAEA and other
international Organizations.


The "Aiara" principle
The guiding principle used in all these documents is that radiation doses to the public,
and to people who work with radiation, must be kept As Low As Reasonably Achievable
(ALARA principle). The effect of radiation at very low doses is still debated and the
ALARA principle is adopted to avoid radiation exposure as much as possible, knowing
that the risk of negative effects from small amounts of radiation approach those seen in
the general public for those same negative effects.


Protection regulations
Equivalent dose is the sum of all doses from different types of radiation to an organ in an
exposed person. Effective dose is the sum of the weighted equivalent doses, and is the


17




18
PATIERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 1. TECHNIQUE, QUALI1Y CONTROL AND RADIATION PROTECTION


Table I. Occupational dose limits
Application


Effective Dose


Equivalent Dose: Eyes
Equivalent Dose: Skin
Equivalent Dose: Hands


Pregnant Women
Workers


Occupation dose limit


20 mSv per year, averaged over 5 yrs,
and not more than 50 mSv in any year.


150 mSv per year
500 mSv per year
500 mSv per year


2 mSv to the surface of the woman's
abdomen for remainder of pregnancy


Public dose limit


1 mSv per year.


15 mSv per year
50 mSv per year


As for members of
the public


dose that gives an indication of the overall effect of the exposure to the person. This is
the dose that is limited by the regulations. A radiation worker is defined as a person who
works with radiation, and may potentially exceed 30% of the prescribed dose limits.


The regulations impose limits on the radiation doses that may be received by
radiation workers and the general public. The dose limits are all set to a level that
will reduce the risk of effects to below some arbitrary acceptable level. This level is
conservative and the result is that the radiation profession is one of the safest fields of
work, if the rules are properly followed.


Dose monitoring of radiation workers
Radiation workers should be monitored at all times when working. The reason for
monitoring is to ensure that the practices being followed by the workers in their daily
routine are safe and do not result in high doses being received. If monitoring is not done
then unsafe practices will not be noticed and excessive exposures to staff may result.


Personnel monitoring is normally the responsibility of national authorities. They
should supply pre-packed thermoluminescent dosimeters, which are returned for
automatic readout on a monthly cycle. The dosimeter should be worn on the torso and
under any protective lead clothing being worn. All radiation workers should wear their
own dosimeter at all times during working hours. If the monthly limit exceeded, the
responsible radiation protection officer should be informed and a report should be
required.


It is good practice to set a local action limit at some lower level then the national level
so that the practices in the department are appropriate to the local conditions. Doses
above the allowed local limit could be followed up in an attempt to rectify any practices
resulting in increased doses to staff. A record should be kept of all radiation doses
recorded for all radiation workers.


Recommendations for pregnant radiation workers
Pregnant radiation workers should not work in areas where there is a risk of getting
more than 30% of the allowed whole body limits for radiation workers. They should
wear an alarm radiation monitor at all times. They should not be allowed to work with
fluoroscopy, theatre radiography, mobile X-ray units, or interventional radiography.


X-ray doses to patients
The skin dose delivered during X-ray investigations has been reduced over the years
with the introduction of newer technology equipment using more sensitive detectors
and better shielding, but the ability to deliver large doses of radiation very quickly and
easily during examinations has also increased. Generally the doses given are lower than




CHAPTER 2. RADIATION PROTECTION IN RADIOLOGICAL PRACTICE


Table 2. Typical effective dose equivalent values for radiological
examinations


Procedure Effective dose equivalent in mSv


PA chest X-ray 0.01-0.05


Skull X-ray 0.1-0.2


Abdominal X-ray 0.6-1.7


Barium enema 3-8


Head CT 2-4


Body CT 5-15


Nuclear Medicine 2-10


WaitingArea ~ n...J\ _d'\
(Occupancy Facto~ ~ ~ ~


Protecti~e barrier with
lead glass window for
viewing patients.


Leakage radiation
(limited by tube
casing)


Short access and e)(jt route
(Avoid line of sight from focus
out of room if possible)


Workload of tube
(mA.min/week)


Scattered Radiation
Wall thickness needed to effecti~ely
limit the primary beam is in practice,
equivalent to about 1.6mm lead.


(Direction is energy dependent)


Toilet
(Occupancy Factor =0 .25)


Fig 2.1


Office Space
(Occupancy Factor =1)


A typical layout of a X-ray suite, showing some of the safety aspects that must be considered when designing a room
to be used for taking X-rays. Barrier thicknesses are designed to ensure that exposures are within acceptable limits.
Workloads and occupancy factors are used in the calculation of the required thickness of the protective barriers.


the threshold for deterministic effects, but interventional radiology, and similar long
procedures, can result in epilation (>3000 mSv), and erythema (>5000 mSv). The doses
to patients should be minimised wherever possible, and the following good practices
are advised:


1. Only image patients if there are good clinical grounds for doing the procedure.


2. Only image the area required. Proper collimation minimises risk.


19




20


PATTERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 1. TECHNIQUE, QUALITY CONTROL AND RADIATION PROTECTION


3. Gonadal shielding should be used in people of reproductive age.


4. Imaging of pregnant patients should be avoided when medically possible. If the last
menstrual period has been missed then the patient should be assumed pregnant.


5. Minimise repeat examinations by using good radiographic practice.


6. Increase the focus skin distance to reduce the entrance dose.


In general the risk of radiation injury from X-ray examinations is far less important than
the clinical benefit being derived from doing the examination. The availability of X-ray
units has resulted in a marked increase in the medical radiation dose being delivered to
the population as a whole. This medical exposure should be limited if possible.


X-ray suite design
The design of a X-ray room takes into account the expected doses in the room and
surrounding areas. The dose calculation takes into account occupancy, workload, X-ray
energy, beam direction, shielding materials used, and other relevant factors.


Practical ways to minimise radiation doses
Radiation dose increases with decreasing distance from the source (ocl/distance2), time
of exposure (octime), intensity of the radiation beam (ocintensity), and the inverse of the
thickness of any absorbers between the source and the exposed person (ocexp-f!thickness).


To reduce the dose to radiation workers the following practices should be adhered to:


1. The distance from the source of the radiation must be increased as much as possible.
One way to encourage this is to mark distances from the source on the floor of the
X-ray room.


2. The time of exposure should be decreased, and workloads should be shared as much
as possible. If it is not required to be present during exposure then leave the room.


3. Protective shields should be used and worn by workers during exposure. Standard
lead aprons, lead gloves and thyroid shields substantially reduce the effective dose
for most diagnostic examinations. Lead glass shields can also be used to protect the
eyes.


4. The primary source of radiation should be collimated as much as possible. For
example the smallest fields possible should be viewed when doing fluoroscopy.


5. Controlled access to areas where radiation exposure may be taking place is required.
Suitable radiation warning signs should be displayed at entrances to rooms and on
any radiation source.


If the above simple principles are applied and all attempts are made to keep the
radiation doses to staff and patients "as low as reasonably achievable" (ALARA), then
the risks from exposure to radiation in an X-ray department should be minimal.




CHAPTER 3


Contrast media in imaging
Peter Carr


Purpose of contrast media
Contrast media are used in imaging to opacify normal structures including the vascular
system, collecting system of the kidneys and the lumen of the gastrointestinal system to
obtain further diagnostic information about focal lesions in the body.


How do they work?
Vascular contrast media contain SO% by weight of molecular iodine which absorbs
X-rays via the photoelectric effect and appears white on X-ray film. Oral agents
consisting mainly of barium works on the same principle.


What are they?
Vascular contrast agents are iodinated organic compounds that are very hydrophilic and
have a low lipid solubility and low binding affinity for proteins and membranes. Most
agents have a molecular weight of less than 2000 (1). On intravascular injection they are
rapidly distributed into the extravascular space but do not enter the intracellular spaces.
They are rapidly excreted by normal kidneys some 90% within two hours. They do not
cross the blood brain barrier.


What do I need to know to use them safely?
Vascular contrast media are not drugs like antibiotics and are pharmacologically inert.
However even though they are very safe when injected intravenously or intraarterially,
they can have side effects and complications. Before you use them you must be familiar
with their side effects and how to manage them.


Side effects can be classified into allergic idiosyncratic reactions and non
idiosyncratic reactions (1). Allergic reactions are the most serious and unpredictable
reactions to contrast media. Reactions occur immediately or within 5 minutes of con-
trast injection. Patients with a history of allergy and atopy, for example hay fever or
asthma, are 8 times more likely to have allergic reactions than non-allergic patients.
These reactions are not dose dependent and are due to a release of vasoactive molecules
such as histamine and kinins.


Non idiosyncratic contrast reactions are due to direct contrast toxicity which is dose
dependent. Patients with renal failure or renal impairment from dehydration, diabetes
or multiple myeloma are especially susceptible. Newborns and elderly patients are less
able to excrete contrast media hence are more likely to have nephrotoxic complications.


Complications of contrast media
Although idiosyncratic reactions are unpredictable, prevention is the best policy.
Whenever a contrast injection is performed a resuscitation trolley should be close by in
the same room. It must have an "Ambu bag" for ventilation, airways, ECG monitor,
oxygen cylinder as well as the following drugs: adrenaline, hydrocortisone, IV fluids,


21




22
PATIERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 1. TECHNIQUE, QUALI1Y CONTROL AND RADIATION PROTECTION


chlorpheniramine and bronchodilator spray. It is mandatory that the trolley is checked
weekly and that all the drugs are in stock. Do not use intravenous contrast agents
without being fully trained in cardiorespiratory resuscitation.


Complications
Complications are divided into: minor, intermediate, major or life threatening and
death.


• Minor complications include nausea, facial flushing or a warm sensation and
urticaria. These complications usually disappear within 15 minutes and only require
reassurance. It the symptoms persist an injection of lOmg of an antihistamine
intramuscularly, such as chlorpheniramine, should cure the allergic effects.


• Intermediate complications include bronchospasm and hypotension. These com-
plications respond to reassurance and an inhaled bronchodilator such as salbutamol,
intravenous hydrocortisone 100 mg bolus and intramuscular adrenaline 0.3-1.0 mls
of 1 in 1000 solution.


• Severe life threatening reactions include seizures, severe bronchospasm and
laryngeal oedema, pulmonary oedema and cardiovascular collapse. These reactions
require urgent treatment. The airway must be secure and intravenous access
established. Adrenaline 0.3-1.0 ml of a 1 in a 1000 solution by intravenous injection
is the most effective drug to treat anaphylaxis. Death following contrast injection is
extremely rare.


Reference
1. Grainger R. Intravascular Contrast Media. In: Diagnostic Radiology: A textbook of medical


imaging. 1997. Eds Grainger R, Allison D. Churchill Livingstone, Edinburgh.




CHAPTER 4
- -Digital imaging and telemedicine
Peter Carr


Digital imaging
A digital image consists of a matrix of numbers or digits that when processed by a
computer will produce an image on a monitor. Digital information is stored as bits, with
8 bits forming a byte that represents a value or character.


Digitization is the process of acquiring or converting analogue images into a digital
format. Many imaging modalities acquire the image initially in this format, for example
with CT, MR and ultrasound. All images today can be converted into digital format. The
advantages of digital imaging are the ease of storing images and the ease of transmitting
images and manipulating the images during image interpretation. You no longer have to
rely on finding the radiographs! It is important to be aware of the disadvantages. Digital
imaging hardware is expensive to purchase and to maintain. Long term storage of
digital images is especially expensive. Given these challenges, there is no doubt that
as computer processors and storage devices become less expensive, many hospitals in
developing countries will use digital imaging in the future. Each medical image is stored
as a file on the computer. The file can be compared to the X-ray packet of a conventional
radiograph. The files vary greatly in their size or number of bytes they contain. Chest
radiographs when in digital format consist of 2 Mbytes (2 million bytes) while an
ultrasound or CT scan may be 10 times smaller at 200 Kbytes in size. Generally plain
analogue radiographs when in digital format have much larger files than more modern
imaging investigations, such as ultrasound or CT imaging.


Teleradiology and telemedicine
Telemedicine is the electronic transmission of medical images from one site to another
for interpretation and consultation. The concept of telemedicine is not new and was
first used in the 1950s. However with the development of more reliable and cheaper
electronic communication and computers, telemedicine is becoming more accessible to
many developing countries.


Goals of telemedicine
The goals are threefold:


i. to provide consultation and interpretation of images in regions of need,


ii. to provide specialist services in hospitals without specialist support


iii. to promote educational opportunities for doctors working in rural hospitals.


Advantages and disadvantages
There are many advantages. Specialist advice is available without the patient having to
travel to the regional or city hospital. Better utilization of specialist resources is made
at the regional hospital. Travel and accommodation costs are reduced for patients who
are less likely to be referred to the regional hospital after telemedical consultation.


23




24
PATTERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 1. TECHNIQUE, QUALITY CONTROL AND RADIATION PROTECTION


Telemedicine can be used to provide continual medical education programmes to
doctors working in rural areas. Disadvantages of telemedicine include: high initial
capital costs of hardware, staff training, requires a good telecommunication network,
and patient confidentiality is more difficult to maintain.


Applications in telemedicine
Telemedicine has been successfully used in radiology, ultrasound, surgery, opthal-
mology, pathology and dermatology. In imaging it has been used for plain radiographs,
CT, MR, ultrasound, angiograms and nuclear medicine.


Image acquisition
Analogue images such as radiographs have to be digitised using a digitiser which
currently is the most expensive component of the system. Most radiographs such as a
chest radiograph produce large files of up to 2MB in size which takes a few minutes to
digitise. The data are usually compressed using a lossless algorithm to reduce the
transmission time.


Image transmission
Conventional telephone lines found in many developing countries have very slow
transmission rates however but are inexpensive to transmit data (around 12 kbjsec).
This means that a chest radiograph will take 3 minutes to transmit. Integrated services
digital network (ISDN) lines which are available in certain countries are twenty times
faster than conventional copper telephone lines, in the region of 256kbjsec. Satellite
communication is obviously wireless technology and is very fast but very expensive and
not freely available in many countries.


Image display
To read images at the receiving station high resolution monitors are recommended.
The American College of Radiology recommends 2000 x 2000 x 12 bit resolution as a
standard (1). These monitors are very expensive and not freely available. The monitors
must be sufficiently bright to see all levels of grey scale in medical images. Most
images can be interpreted using 1000 x 1000 x 8 bit resolution which are much
cheaper.


Image files
Each image is kept in its own file. Static ultrasound, CT and MR images are relatively
small compared to radiographs: 100 kilobytes versus 2 megabytes. The larger the file the
longer transmission time.


Problems with teleradiology
Most teleradiology systems will have limited spatial resolution and subtle lesions in
the lungs and fine bone fractures can easily be missed unless the original radiographs
are reviewed later (2). As faster computer systems and digital telephone lines
become available limited spatial resolution should become less of a problem. The high
capital costs of equipment and limited opportunities to train health professionals in
some countries are a barrier to the development of telemedicine services in some
developing countries (3). WHO is looking into the development of telemedicine services
as a way of providing imaging services to rural clinics and hospitals in developing
countries.




CHAPTER 4. DIGITAL IMAGING AND TELEMEDICINE


LEARNING POINTS: DIGITAL IMAGING AND TELEMEDICINE


• Digital imaging is the process of acquiring, storing, transmitting and interpreting medical images in
a digital format


• A digital format is when images are stored as a matrix of numbers or digits and can be processed by
a computer to produce a medical image on the image


• Files contain an image in digital format
• Files are measured in the amount of digital data they contain in bytes
• Plain radiographs contain the largest amount of data while CT and ultrasound contain the least


data
• Transmission of digital images depends on the transmission rate of the communication system


used (in bauds)
• Good spatial resolution of the monitors is necessary to interpret images is extremely important


References
1. American College of Radiology: Telemedicine Standards, 1994. Reston, Virginia.


2. Corr P. Teleradiology in KwaZulu Natal: a pilot project. SAMJ, 1998;88:48-49.
3. Blignault I, Kennedy C. Training in Telemedicine. J Telemed Telecare, 1999;5:5112-4.


25






PART2


CHEST IMAGING PATTERNS






CHAPTER 5


The normal chest radiograph
Peter Corr


Understanding the anatomy of the chest is critical in interpreting chest radiographs.
Only by reading many normal chests will you be able to detect abnormalities. It is
important to develop a routine system and to keep to it.


Soft tissues
The soft tissues of the chest consist of the skin, muscle, fat and fascial planes of the
chest wall. There are a number of "companion" shadows to bones such as the clavicles.
The breast shadows and axillary folds should be symmetrical (figs S.la, S.lb & 5.2).


) l


29


Fig 5.la
Normal adult male
chest.


Fig 5.lb
Normal adult
female chest.




30


Fig 5.2
Lateral adult chest.


Fig 5.3
Lateral sternum.


PATTERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 2. CHEST IMAGING PATTERNS


Skeleton
There are 12 ribs which can be traced from their posterior attachment to the spine
to their anterior costochondral junctions. The anterior cartilaginous region ossifies
especially in women and must not be confused with calcified lung lesions. The 11th and
12th ribs are shorter and do not articulate anteriorly with the sternum. Imaging of
ribs requires both PA or AP chest views plus localised oblique views. The clavicles
articulate medially at the sternoclavicular junction. This joint is difficult to visualise on
PA films. Dislocation or subluxation can therefore be easily missed. Laterally the
clavicle articulates with the acromion of the scapula. The sternum consists of two
bones; the manubrium superiorly and the body of the sternum which articulate at the
manu brio sternal junction. The sternum is best imaged obliquely or laterally (fig 5 .3).




CHAPTER 5. THE NORMAL CHEST RADIOGRAPH


The cervical spine normally has 7 vertebrae, the thoracic 12 vertebrae. Normally the
spine has a slight thoracic kyphosis and cervical lordosis. The lower cervical and
thoracic spine is visualised best using a high Kvp technique (>120Kvp).


Mediastinum
To understand the anatomy it is best to think of the mediastinum in the following
compartments:


Superior mediastinum
This is the compartment superior to a horizontal plane passing through the aortic
arch. It includes the trachea, oesophagus, lymph nodes, superior vena cava and
brachicephalic artery on the right, left subclavian artery, recurrent laryngeal, phrenic
and vagal nerves. With age there is widening of the superior mediastinum as the arteries
dilate, however the contour is maintained.


Anterior mediastinum
This region is bounded superiorly by the superior mediastinum and posteriorly by the
middle mediastinum. It is best visualised on the lateral chest radiograph. It includes
the thymus in infants which appears like a sail, lymph nodes and mediastinal fat. The
thymus is important to recognise and not to confuse with a pathological lesion.


Middle mediastinum
This region contains the heart, pericardium, lymph nodes, tracheobronchial tree and
carina, and the hila regions. The right atrium and right ventricle comprise the anterior
half of the heart, with the posterior half composed of the left atrium and ventricle. The
carina is where the trachea bifurcates into the left and right main bronchi with the right
main bronchus being much steeper than the left. Surrounding the carina and in the hila
are numerous lymph nodes which have a maximum diameter of 10 mm. The hila are
slightly different on each side. The left hilum is located 2 em superior to the right. The
reason for this is the left main pulmonary artery ascends and curves over the superior
border of the left main bronchus while on the right, the right main pulmonary artery
is located anterior to the right main bronchus. The pulmonary veins enter the hila
posteriorly. It is important to understand the hila anatomy so as not to be confused
between pulmonary aneurysms and tumours in this region.


Posterior mediastinum
This compartment contains the descending thoracic aorta, the oesophagus, nerves,
lymph nodes and paraspinal fascia. The descending aorta is located anterior and to the
left of the thoracic spine with the oseophagus situated between them.


Heart (figs 5.1a, 5.1b & 5.2)
The heart is situated within the middle mediastinum with in the pericardium. On P A
and AP chest radiograph one third of the heart is to the right of the thoracic spine, two
thirds to the left of the spine. On a lateral chest the anterior border of the heart is
comprised of the right atrium and right ventricle; the posterior border is comprised of
the left atrium and ventricle. It is important to be able to recognize the normal frontal
contour of the heart.


31




32
PATIERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 2. CHEST IMAGING PATIERNS


Lungs
It is important to remember than the right lung with its three lobes, is different from the
left with two lobes. The greater or oblique fissures separate the right upper and middle
lobe from the lower lobe and the left upper from the lower lobe. The fissures are best
seen on a lateral chest radiograph as thin white lines, the right fissure being steeper than
the left. The lesser or horizontal fissure separates the right upper from the right middle
lobe and extends from the right hilum to the chest wall. The pulmonary arteries and
veins extend out from the hila and are visible to the outer one third of the lungs. The
veins tend to be more lateral than the arteries but often cannot be distinguished apart on
plain radiographs.


Diaphragm
The diaphragm consists of three parts: the right hemidiaphragm, the central tendon,
and the left hemidiaphragm. The right hemidiaphragm is 3 em superior to the left due to
the presence of the liver inferiorly. The hemidiaphragms are muscles under control of
the phrenic nerves. The diaphragm inserts peripherally into the costal margin and
thoracic wall at the costophrenic angles. The diaphragm may be scalloped as a normal
variant.


Pleura
The pleura is a thick fibrous layer consisting of a parietal pleura and visceral pleura. The
visceral pleura covers the lungs while the parietal pleura covers the inner surface of
the chest wall. Usually the pleural space is a potential space only. The normal pleural
surface cannot be visualised using radiographs.


Chest radiography
Good radiographic technique is critical for producing diagnostic chest X-rays.
Important points to remember are:


• Exposure factors-a high kV > 120 technique is important to improve visualization of
the soft tissue planes of the mediastinum and tracheobronchial tree. The pulmonary
vessels are well visualised with this technique.


• Size and shape of the chest-exposures will vary depending on the size and shape of
the chest.


• Good inspiration is critical. You should aim to visualise at least 11 ribs posteriorly
above the diaphragm. Poor inspiration will result in difficulties in measuring heart
size and assessing the lungs.


• Patient positioning-the PA position is best. AP and supine films will result in
difficulties in assessing cardiac size and pulmonary vasculature. Check that the
patient is not rotated by checking that the medial edges of the clavicles and the
spine are equidistant.


Technique
The patient should stand erect with the anterior chest wall flat against the bucky I
cassettte with the hands on the hips and the elbows rotated forwards, The X-ray tube
should be more than 1m from the cassette and centred at the T3 level.




CHAPTER 5. THE NORMAL CHEST RADIOGRAPH


How to read a chest radiograph
Try to develop a systematic method and keep to it. Start peripherally and read towards
the centre of the chest


1. Soft tissues: compare both sides. In females check both breasts shadows are present.
Look for focal soft tissue calcification and subcutaneous gas.


2. Skeleton: count all ribs. Check for focal lesions such as metastases (lytic or sclerotic)
and fractures. Check clavicles, shoulders, cervical and thoracic spines.


3. Lungs: compare both sides. Divide the lungs into three zones: upper, middle and
lower and compare both sides.


4. Diaphragm: the right hemidiaphragm is 2 em superior to the left. Compare the shape
and position. Look for free air beneath the diaphragm.


5. Hilar regions: the left is 2 em superior to the right. Check position, contour and
density


6. Mediastinum: check the position with two thirds of the transverse diameter of the
heart to the left of the spine and one third to the right. In the superior mediastinum
the trachea should be central anterior to the thoracic spine.


7. Heart: check size (normally <50% CTR), position and contour.
8. Pleura: normally invisible. Check costophrenic angles for pleural fluid and


pneumothorax.


Chest patterns
Clinical information
To improve diagnostic specificity always take a relevant history from the patient. Ask
the following questions:


How long have there been symptoms, such as cough?


Is there haemoptysis?


Is there chest pain?


Is there shortness of breath?


Important clinical information includes:


Does the patient have a fever?


Is the patient immunosuppressed or HIV positive?


Is the patient taking any drugs? eg. antibiotics or chemotherapy?


Is the patient exposed to any occupational dusts?


33




34


Fig 6.1
Right upper lobe
pneumonia from


strep. pneumoniae
infection with lobar


opacification and
air bronchograms


(arrow).


CHAPTER 6


Pulmonary infection
Peter Corr


Most pathological processes involving the lungs will cause increased density of the lung
and appear white or appear as focal opacities.


Pneumonia patterns
Pneumonia is air space consolidation in which the alveolar air space is filled with
inflammatory exudate from the infection. Pneumonias can be classified both
anatomically and by aetiology. Anatomical classification is useful as certain patterns
have specific causes, for example lobar pneumonia is often due to streptococcal
pneumoniae. Age is also important to consider as childhood pneumonias have a
different appearance and cause from adult infections. The presence of immuno-
suppression from human immunodeficiency virus (HIV) infection has complicated
these patterns in many developing countries.


Lobar pneumonia pattern
Pneumonia is airspace opacification of a lung lobe. The alveolar air spaces are filled
with inflammatory exudate while the bronchi and bronchioles remain patent. The cause
is often strep. pneumoniae. The pattern to identify is opacification of the pulmonary
lobe with the presence of air bronchograms which appear like the branches of a leafless
tree (fig 6.1). Air bronchograms are the patent air containing bronchial tree, which is
surrounded by airspace opacification. Once you see bronchograms these are diagnostic
of lobar pneumonia. The important differential diagnosis is lobar atelectasis where there
are no air bronchograms as the bronchus is usually obstructed and the air in the distal
bronchus is reabsorbed (fig 6.2).




CHAPTER 6. PULMONARY INFECTION


35


Fig 6.2
Left lower lobe
atelectasis from
obstruction of the
left lower lobe
bronchus. Note
absent air
bronchograms and
slight volume loss.


Fig 6.3a
Right middle lobe
pneumonia
demonstrates loss
of the right heart
border (arrow). This
is called "loss of
the silhouette" sign.


Fig 6.3b
Lateral chest
demonstrates the
right middle lobe
pneumonia.




36


Fig 6.4
Cavitating right


upper lobe
pneumonia from


klebsiella infection.


Fig 6.5
Chest radiograph


demonstrates
"ground glass"


opacification in an
HIV positive child


with pneumocystis
carinii infection.


PATIERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 2. CHEST IMAGING PATIERNS


To localize a lobar pneumonia anatomically, the loss of the silhouette sign can be
used. Right middle lobe pneumonias will cause the right heart border to disappear
(figs 6.3a, 6.3b) and lingula left upper lobe pneumonias result in loss of visualisation
of the left heart border. In lower lobe pneumonias either hemidiaphragm will be not be
visualised.


Bronchopneumonia pattern
In this pattern there is multifocal peribronchial opacification bilaterally. This pattern is
common in childhood infections. The cause of the infection is often viral or following
mycoplasm infection.


Cavitating or necrotising pneumonia (fig 6.4)
Necrotising pneumonia pattern occurs when there is extensive necrosis of lung tissue.
Cavities form, which may have multiple fluid levels. Common causes are infections
from klebsiella, bacteroides and pseudomonas bacteria. The clue to this pattern is
the presence of a cavity within the pneumonia. The differential diagnosis includes
cavitating cancer (usually a squamous primary or secondary) and tuberculosis.


"Ground glass" pneumonia pattern (fig 6.5)
This pattern is often difficult to recognize initially, however the clue is the pulmonary
vessels appear ill defined or "fuzzy" and the lung appears slightly opaque. No air




CHAPTER 6. PULMONARY INFECTION


bronchograms are found. This pattern is found in pneumocystis cannu pneumonia
infections in patients who are immunosuppressed especially from AIDS, mycoplasma
infection and, viral infections.


Pulmonary tuberculosis patterns
The appearance of pulmonary tuberculosis is changing in many developing countries
with the spread of HIV I AIDS. It is therefore very important to establish whether the
patient is immunosuppressed.


Primary pulmonary TB pattern (fig 6.6)
This is usually a small focus of opacification in the lung (Chon focus) with hilar
adenopathy and mediastinal adenopathy on the same side. Often the primary
pulmonary focus is not detected only the hilar or mediastinal adenopathy which may
cause tracheobronchial airway compression in young children.


Secondary or post primary TB pattern (figs 6. 7, 6.8)
In patients with normal immunity this pattern is:


cavitation-usually in the upper lobes or lung apices;


37


Fig 6.6
Chest radiograph of
a child with primary
tuberculosis. Note
right hilar and
para tracheal
lymphadenopathy.


Fig 6.7
Chest radiograph
in an adult with
post primary or
secondary
tuberculosis
demonstrates right
upper lobe
infiltrates with
cavities which have
spread to the right
and left lower lobes.




38


Fig 6.8
Chest radiograph


demonstrates
healed tuberculosis


with focal
calcification in the
left mid zone and


scarring in the right
apex.


Fig 6.9
Chest radiograph of


a patient with
miliary tuberculosis


demonstrates a
micronodular


infiltrate in both
lungs.


PATIERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 2. CHEST IMAGING PATIERNS


small nodules/infiltrates in the upper lobes. The lesions heal by calcifying with
fibrosis of the surrounding lung.


Miliary TB pattern (fig 6.9)
This is a very important pattern to detect as the disease is fatal if untreated. Patients
often have non-specific symptoms and signs. Multiple small nodules 2-5 mm in size are
detected throughout both lungs from blood borne spread of TB. It is therefore extremely
important to exclude miliary tuberculosis in any patient with a miliary infiltrate
pattern. The differential diagnosis includes metastates in adults from melanoma,
carcinoma of the prostate, pancreas and thyroid, pneumoconioses, sarcoid and
lymphoma.




CHAPTER 6. PULMONARY INFECTION


LEARNING POINTS: PULMONARY INFECTIONS
• Lobar pneumonia: air bronchograms in the presence of lobar opacification.
• Broncho pneumonia: diffuse peribronchial opacification.
• Necrotising pneumonia: cavitation in the presence of pneumonia; can progress to a lung abscess.
• Ground glass opacification: in pneumocystis carinii pneumonia, mycoplasma, CMV infection.
• Primary PTB: focal pulmonary opacification and unilateral hilar adenopathy (Ghon focus).
• Secondary TB: upper lobe cavitation, nodular(acinar) infiltrates.


Reference
1. Chapman S, Nakielny R. (eds). In: Aids to Radiological Differential Diagnosis. 1995,


Saunders, London.


39




40


Fig 7.1
Chest radiograph in
a patient presenting


with haemoptysis
demonstrates a left


mid zone solitary
bronchial


carcinoma.


CHAPTER 7
--Lung cancer
Peter Carr


Lung cancer is a serious public health problem in many developing countries due to
increasing cigarette smoking especially amongst young women (1). Chest radiographs
are important to establish the diagnosis. The common patterns are:


solitary pulmonary nodule


hilar mass


lobar atelectasis


multiple pulmonary nodules.


Solitary pulmonary nodule pattern (fig 7.1)
The commonest cause of a solitary pulmonary mass over 3 em diameter is a carcinoma.
There are usually no specific features to suggest cancer apart from the size of the lesion.
The presence of focal "pop corn" type calcification suggests a benign cause such as a
hamartoma. The differential diagnosis includes granulomas such as tuberculomas and
fungal infections, such as cryptococcoma, benign tumours; and a solitary metastasis.
The diagnosis can be confirmed by percutaneous fine needle biopsy using fluoroscopy
or bronchoscopic biopsy.


Hilar mass pattern (fig 7.2)
A hilar mass is a common presentation as many cancers involve the proximal bronchi
with associated tracheobronchial lymphadenopathy. They present as masses distorting
the normal contour of the hilum or causing increased density to the hilum.




CHAPTER 7. LUNG CANCER


Pancoast's tumour (apical sulcus tumour) (figs 7.3a, 7.3b)
This tumour involving the apex of the lung is often difficult to detect and may often be
confused with pleural thickening at the lung apex. Clues to the diagnosis are presence of
erosion or destruction of the first three ribs and the presence of a bulging convex
inferior to the border to the mass. An apical or lordotic projection is extremely useful to
demonstrate this region. The patient may present with pain down the arm from brachial
plexus involvement and or involvement of the sympathetic chain with a Horner's
syndrome on clinical examination.


a


41


Fig 7.2
Chest radiograph of
a patient with
haemoptysis
demonstrates a
large right hilar
mass with volume
loss of the right
lung.


Fig 7.3a, 7.3b
Chest radiograph
of a Pancoast
tumour in the right
apex. Note the
elevated right
hemidiaphragm
from right phrenic
nerve palsy from
tumour infiltration
of the right phrenic
nerve. The apical
view demonstrates
the tumour better
(3b).




42


Fig 7.4a, 7.4b
Chest radiograph


demonstrates a left
hilar mass and left


upper lobe
opacification and


volume loss caused
by a left main


bronchus
obstruction from a


carcinoma.


PATTERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 2. CHEST IMAGING PATTERNS


Pulmonary atelectasis pattern (figs 7.4a, 7.4b)
There is a common presentation where there is opacification of a lobe or segment of a
lobe with volume loss with no air bronchogram as the airway is obstructed by the
tumour with reabsorption of the air distal to the obstruction. This pattern must be
differentiated from segmental or lobar pneumonia where there is minimal or no volume
loss and air bronchograms are usually present. Where chest infections or pneumonias
do not resolve after two weeks of treatment, a follow up chest radiograph is recom-
mended to exclude the possibility of endobronchial obstruction from a tumour or
foreign body (in a child).


a


b


Multiple masses pattern (figs 7.5a, 7.5b)
Multiple lung masses or nodules >2 em diameter are most likely due to metastases or
granulomas such as tuberculomas or sarcoid. Metastases commonly originate from the
breast, primary lung cancer, colon, kidney, pancreas, thyroid, testis or sarcomas from
bone or soft tissue. It is often impossible to differentiate between metastasis and




CHAPTER 7. LUNG CANCER


granulomas, although metastases are of variable sizes and have smooth borders while
granulomas are the same size and have irregular borders.


a


b


Multiple micronodular or miliary pattern (fig 7.6)
This is a very important pattern to recognise. These small pulmonary nodules are easy
to miss if the radiograph is too dark (over-penetrated film). Patient's symptoms are
often non-specific especially with miliary tuberculosis: loss of weight and cough (2).
The correct diagnosis is only suspected once the chest radiograph is reviewed.
Micronodules are less than 1 em in diameter. This pattern is also called the "miliary"
pattern and is found in metastatic lung disease, lymphoma, miliary tuberculosis, and
pulmonary sarcoid. Causes of "miliary" pulmonary metastases include: thyroid,
prostate, breast, pancreas, bronchial carcinomas. Miliary tuberculosis micronodules are
usually very discrete without associated hilar lymphadenopathy and are distributed
diffusely throughout both lungs. On treatment the pattern resolves very slowly
compared to the patient's rapid clinical improvement. A common cause in children with
AIDS is lymphoproliferative pneumonitis. This is a peribronchial lymphoid reaction to
the HIV virus and is common in seropositive children and young adults.


43


Fig 7.5a, 7.5b
Chest radiographs
of two patients with
multiple pulmonary
metastases.
In 5a from
choriocarcinoma, in
5b from carcinoma
of the breast. Note
absent left breast
following a
mastectomy and
destoyed left 5th rib
from metastatic
disease.




44


Fig 7.6
Chest radiograph of


a patient with
multiple


micronodular
metastases.


PATTERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 2. CHEST IMAGING PATTERNS


LEARNING POINTS: LUNG CANCER
• Solitary pulmonary nodule >3 em is most likely cancer. Other causes are tuberculomas, benign


tumours
• Hilar mass is a common presentation for cancer.
• Pulmonary atelectasis is a common presentation; there is opacification of a segment or lobe with


volume loss and absent bronchograms.
• Multiple masses: metastases and granulomas are commonest cause. Cancers are from breast,


lung, thyroid, colon, kidney, pancreas, testis and soft tissue sarcoma.
• Micronodular or miliary infiltrates: metastatic disease and TB are common causes. Cancers


originate from the thyroid, stomach, prostate, and melanomas.


References
1. Combating the tobacco epidemic. In: World Health Report 1999. Geneva WHO.


2. Chapman S, Nakielny R. In: Aids to Radiological Differential Diagnosis. 1995. Saunders,
London.


3. Hussey G, Chisholm T, Kibei M. Miliary tuberculosis in children: a review of 94 cases, Pediatr
Infect Dis 1, 1991;10:832-6.




CHAPTER 8


Pulmonary hypertranslucency
and cystic lungs
Peter Corr


Hyperinflated lung pattern (figs 8.1 & 8.2)
Assessment of lung hyperinflation on chest radiograph is often difficult. Hyperinflation
is present if the posterior margins of at least 11 ribs are detected above the diaphragm,
which is flattened. An accurate assessment of airways obstruction requires lung


45


Fig 8.1
Chest radiograph of
a patient with
hyperinflation and
upper lobe bulla
from emphysema.


Fig 8.2
Chest radiograph in
a patient with
emphysema with
recurrent chest
infection. Note
upper lobe bulla
with mid zone
opacification from
pulmonary
infection.




46


Fig 8.3
Chest radiograph of


a patient with
chronic cystic


bronchiectasis
demonstrates


multiple cysts in the
lower lobes many of


which containing
fluid levels from


infected fluid.


Fig 8.4
Localised view of


the left lower lobe
of a patient with


bronchiectasis
demonstrating the


multiple dilated
bronchi.


PATTERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 2. CHEST IMAGING PATTERNS


function tests, not chest radiographs. Hyperinflation is found in airway obstruction from
asthma and chronic obstructive airways disease following cigarette smoking. The
presence of focal thin walled cysts in the lung (called bulla) and hyperinflation are seen
in emphysema. Patients may have radiographic signs of pulmonary hypertension with
prominent central arteries and thin peripheral arteries.


Cystic lung pattern (figs 8.3 & 8.4)
Lung cysts appear as focal translucencies in the lung contained by thin cyst walls which
are usually 2 mm or less in thickness. The most common cause is cystic bronchiectasis
where there is cystic dilatation of the bronchi usually appearing as multiple basal ring
shadows. Often air fluid levels are present if there is superimposed infection. The
diagnosis is best made on high resolution CT of the lungs where the bronchial dilatation
is better defined. Lung cysts must be differentiated from cavities. Hydatid disease can
appear "cystic" if the fluid drains into the bronchi.




CHAPTER 8. PULMONARY HYPERTRANSLUCENCY AND CYSTIC LUNGS


Cavities (figs 8.5-8. 7)
Cavitities occur commonly in pulmonary tuberculosis, necrotising pneumonias or
abscesses and cavitating tumours. With cavities the wall of the cavity is irregular and
much thicker than 2 mm, usually in the region of 1 em thick. In necrotising pneumonias
or lung abscesses there is often an associated air fluid level within the cavity.


47


Fig 8.5
Chest radiograph of
a patient with
active pulmonary
TB with a large left
mid zone cavity.
Note the wall is
much thicker than a
bulla.


Fig 8.6
Chest radiograph of
a patient with
pulmonary hydatid
disease. The right
lower hydatid cyst
contents have
drained into the
bronchial tree
leaving a thicked
walled cyst with a
nodular border.




48


Fig 8.7
Chest radiographs


of an epileptic
patient who


developed two
abscesses in the


right lung to/owing
aspiration during a


seizure.


Fig B.Ba & B.Bb
Chest radiograph of


a left mid zone
cavity containing a


mycetoma which
moves to the


dependent region of
the cavity on


positioning the
patient in the left


decubitus position
(fig 8.9).


PATIERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 2. CHEST IMAGING PATIERNS


a




CHAPTER 8. PULMONARY HYPERTRANSLUCENCY AND CYSTIC LUNGS


b


Mycetomas (figs 8.8 & 8.9)
These are inflammatory masses within prior lung cavities usually from chronic TB.
They are due to fungal infections, especially aspergillosis (so called "aspergillomas").
They are not attached to the cavity wall and move on changing the patient's position
to the most dependent position within the cavity. These masses are important to detect
as they are an important cause of haemoptysis. Thickening of the lateral wall of the
cavity is a good predictor of the presence of a mycetoma. The main differential
diagnosis is an intracavitatory haematoma. Surgical resection is the treatment of choice.


LEARNING POINTS: HYPER TRANSLUCENCY & LUNG CYST PATTERNS


• Hyperinflated lungs are found in airways obstruction from asthma and chronic obstructive airways
disease following cigarette smoking.


• The presence of a bulla and hyperinflation is found in pulmonary emphysema.
• Unilateral hyperlucent lung may be due to technical factors such as chest rotation, a previous


mastectomy, or a hyperinflated lung from Swyer James syndrome (bronchiolitis obliterans) or a
"ball valve" effect from partial bronchial obstruction by a foreign body especially in young children
or rarely an endobronchial tumour.


• Hydatid cysts can mimic bulla.
• A cavity has a thick wall >2 mm which is irregular and may contain an air fluid level if there is a


communication with the bronchial tree.
• Cavities are found in lung abscess/necrotising pneumonia, tumours especially squamous cancers


and pulmonary tuberculosis.
• Mycetomas are intracavitatory inflammatory masses of aspergillus. They cause haemoptysis.


References
1. Chapman S, Nakielny R. eds. In: Aids to Radiological Differential Diagnosis. 1995, Saunders,


London.


2. Sansom HE, Baque-Juston M, Wells AU, Hansell DM. Lateral cavity wall thickening as an
early radiographic sign of mycetoma formation. Eur Radiol2000;l0:387-90.


49


Fig 8.9a & 8.9b
Chest radiograph of
the same patient as
shown in fig. 8.8,
but taken in the left
decubitus position.






CHAPTER 9
-----Pleural and extra pleural disease
Peter Carr


Pleural effusions (figs 9.1-9.3)
Pleural fluid is first detected in the costophrenic angles and subpulmonic pleural spaces
(fig 9.la). There may be actually 250mls of fluid present before it is detected
radiologically. Subpulmonic fluid collections are common particularly in trauma where


51


Fig 9.1a, 9.1b
Chest radiograph of
a patient with a
right subpulmonic
pleural collection.
Note the pseudo
elevated
hemidiaphragm
with a more lateral
apex than normal.
On the right side
down decubitus film
the pleural fluid is
now demonstrated
~ laterally. Note the '"~-----J presence of a small


a


b


medial
pneumothorax.




52


Fig 9.2a, 9.2b
Chest radiograph of
a patient in cardiac


failure
demonstrating a
"pseudomass" in


greater and lesser
right fissures from


encysted pleural
fluid.


PATIERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 2. CHEST IMAGING PATIERNS


the apparent hemidiaphragm appears elevated and the apex is displaced laterally. To
confirm the diagnosis a decubitus chest radiograph will detect the subpulmonic
collection (fig 9.1 b). Encysted collections are common due to pleural adhesions and can
appear like masses on radiograph. Fluid encysts in the fissures and can also appear as a
pseudo mass (fig 9 .2). The diagnosis becomes more evident on the lateral chest film
where the fluid loculates in a "aeroplane propeller" configuration. Common causes of
pleural effusions are:


tuberculosis


malignancy (either primary or secondary cancer).
Ultrasound using a high frequency transducer (>5 MHz) will confirm the presence of
fluid as opposed to a mass. Ultrasound is very useful to direct aspiration of encysted
effusions.


a


b




CHAPTER 9. PLEURAL AND EXTRA PLEURAL DISEASE


Empyema (figs 9.4a & 9.4b)
Empyema is a collection of pus in the pleural space. The pleura is thickened and bulges
outwards. It may not be possible to differentiate between a pleural effusion and
empyema on the chest radiograph alone. Empyemas appear echogenic on ultrasound.
Ultrasound is excellent in guiding a needle into the the collection for aspiration.


a


53


Fig 9.3
Chest radiograph
demonstrates a
large left pleural
effusion with
mediastinal shift to
the right.


Fig 9.4a, 9.4b
Chest radiograph of
a patient with a
right empyema.
Note the anterior
convex margin of
the pleural
collection
suggesting the pus
in the pleural space
is under pressure.




54


Fig 9.5
Chest radiograph of


a patient with
breast cancer and


pleural metastases.
Note right


mastectomy from
the absent breast


shadow and the left
peripheral pleural


metastases.


PATIERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 2. CHEST IMAGING PATIERNS


b


Pleural mass (figs 9.5-9.7)
Pleural masses are either benign or malignant. Malignant pleural masses include
metastases from adenocarcinoma often from breast cancer, lung primary, bowel, ovary
or a primary cancer such as a mesothelioma from asbestosis exposure. Benign masses
include inflammatory masses, fibrosis, benign fibromas and pleural plaques from
asbestos exposure. Inflammatory lesions include tuberculosis. It is not possible to
differentiate benign from malignant masses. Rib destruction is more common with
malignant tumours but is also detected with granulomatous infections such as
tuberculosis.


Extra pleural masses
Extrapleural masses which may be associated with rib destruction. Metastatic
involvement of the ribs often has an associated soft tissue mass. Similarly chronic
infections of the rib such as TB, actinomycosis can be associated with an abscess or
granulomatous mass. Plasmacytomas causes localised rib expansion and a soft tissue
mass. To differentiate an extrapleural mass from a pleural mass, look at the medial
border of the mass. If it is convex medially it is probably extrapleural while if it is
concave medially it is probably pleural in origin.




CHAPTER 9. PLEURAL AND EXTRA PLEURAL DISEASE


LEARNING POINTS: PLEURAL DISEASE
• Subpulmonic effusions result in an apparent elevation of the hemidiaphragm, the diagnosis can be


confirmed on an ipsilateral decubitus chest radiograph.
• Common causes of unilateral pleural effusions are malignancy (primary and secondary) and


tuberculosis.
• Encysted pleural collections resemble pleural masses; ultrasound can differentiate between a mass


and fluid.
• Pleural masses are most commonly adenocarcinoma metastases, mesothelioma, inflammatory


masses or benign fibromas.
• Extrapleural masses often have associated rib destruction.


Reference
l. Chapman S, Nakielny R. Aids to Radiological Differential Diagnosis. 1995 Saunders, London


55


Fig 9.6
Chest radiograph of
a patient with a left
pleural
mesothelioma. Note
mediastinal shift to
the left and
opacification of the
left pleural space.


Fig 9.7
Chest radiograph of
a patient with left
lower zone pleural
fibrosis following a
haemothorax. Note
the crowding of the
left lower ribs and
the pleural
calcification.




56


Fig 10.1
Chest radiograph of


a patient with
multiple myeloma.


There is a soft
tissue mass in the


right lower zone
with destruction of


the lower right ribs.
A lytic lesion in the


right clavicle is
noted.


CHAPTER 10
-Rib lesions
Peter Carr


It is always important to scan the bones of the chest for both focal and diffuse
abnormalities. Often the presence of bone lesions will suggest the correct diagnosis of
the pulmonary abnormality.


Focal rib lesions
Lytic lesions are usually due to tumours especially metastases or myeloma. They can be
difficult to detect initially if the individual rib contours are not carefully traced out from
posterior to anterior. The patterns of these lesions appear as if the ribs have been
"erased" with a rubber eraser with an associated extrapleural soft tissue mass (fig 10.1).
Bronchial carcinoma, especially Pancoast's tumour at the apical sulcus of the lung,
invades the pleura and overlying brachial plexus to cause early lytic destruction of
the first and second ribs. Occasionaly pleural tuberculosis involves the overlying ribs.
With myeloma it is particularly useful to assess other flat bones such as the clavicles
and scapulae for similar lesions as the presence of such lesions will suggest the
diagnosis.


Sclerotic rib lesions in which bone density is increased are especially common with
prostate and breast metastases. Occasionally all the ribs become diffusely sclerotic
(fig 10.2).




CHAPTER 10. RIB LESIONS


Rib erosions usually destroy the superior or inferior rib cortex. Inferior rib notching or
erosions are often due to erosion from neurofibromas or collateral vessels in aortic
coarctation (fig 10.3). In neurofibromatosis the ribs may appear thin and gracile
because of the associated mesodermal dysplasia. Superior cortical erosions are
detected in collagen vascular disorders such as rheumatoid arthritis, systemic lupus
erythrematosis and scleroderma. Erosions of the lateral margins of the clavicles are
common in advanced rheumatoid arthritis.


Diffuse rib lesions
Metabolic bone disease, such as osteomalacia, hyperparathyroidism, Paget's disease
and rickets in children, may demonstrate bone softening and rib deformity.
Microfractures with exuberant callus may be detected in osteomalacia, Cushing's
syndrome and Paget's disease.


Diffuse osteopenia or decreased bone density occurs in osteoporosis, osteomalacia,
and Cushing's syndrome. Diffuse bony sclerosis occurs in Paget's disease, sclerotic
metastases from breast and prostate cancer.


57


Fig 10.2
Chest radiograph of
a patient with
diffuse sclerotic
metastases in the
ribs from carcinoma
of the prostate.


Fig 10.3
Chest radiograph of
a patient with
coarctation of the
aorta. Note left
ventricular
hypertrophy, small
aortic arch and
bilateral inferior rib
notching of the ribs.




58
PATTERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 2. CHEST IMAGING PATTERNS


LEARNING POINTS: RIB LESIONS


• Common cause of focal lytic rib lesions are lytic metastases for breast and lung carcinoma, and
multiple myeloma


• Common cause of focal and diffuse sclerotic rib lesions are sclerotic metastases from prostate and
breast carcinoma, and Paget's disease


• Common causes of rib notching and erosion include coarctation of the aorta, collagen vascular
disorders, and neurofibromatosis


• Common causes of decreased bone density are osteoporosis, hyperparathyroidism, osteomalacia,
and rickets in children


Reference
1. Chest wall, pleura and diaphragm. Wilson AG, Flower CDR, Verschakelen JA. In: Diagnostic


Radiology: a textbook of medical imaging. 1997. Eds Grainger RG, Allison DJ. Churchil
Livingstone, Edinburgh




CHAPTER 11
-Chest trauma
Peter Carr


Good quality radiographs are critical for evaluation of chest trauma patients. An
erect chest radiograph is important to detect pneumothorax and haemothorax. If the
patient is unable to sit or stand a decubitus chest is useful to detect a pneumothorax and
or subpulmonic haemothorax.


Blunt trauma
Bony injury (fig 11.1)
Ribs usually fracture laterally after blunt trauma; the lower six are commonly involved.
Fractures in two places may lead to a flail chest. Fractures of the upper four ribs are
usually associated with severe blunt injury and vascular injury must be excluded if
there is widening of the superior mediastinum. Sternal injuries are detected on lateral
coned views of the sternum. It is important to check the thoracic spine carefully for
associated fractures in these patients.


Pneumothorax (figs 11.2a, 11.2b)
This is an important diagnosis and can be detected as a fine white line of the visceral
pleura with absent peripheral lung markings. Pneumothoraces can also be medial
or subpulmonic in position. A decubitus chest radiograph will identify a shallow
pneumothorax. If the pneumothorax is large and there is mediastinal shift away from


59


Fig 11.1
Chest radiograph of
a patient following
chest trauma. Note
fractures of the left
81h and gth ribs.




60


Fig 11.2a
Chest radiograph of


a patient who
sustained a


penetrating chest
wound. Note the


shallow left
pneumothorax.


Fig 11.2b
Chest radiograph of


a patient with a
large left


haemopneumothorax,
with mediastinal
shift to the right


and surgical
emphysema in the


chest wall.


PAnERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 2. CHEST IMAGING PAnERNS


the side of the injury, the pneumothorax is considered to be under tension. This is
extremely important to recognise because if left undetected this can be rapidly fatal. A
clue to the presence of a pneumothorax is gas in the soft tissues, which appears as low
density streaks (surgical emphysema).


Haemothorax (fig 11.3)
Blood collects inferiorly in the pleural spaces beneath the lung to form a subpulmonic
haemothorax. The hemidiaphragm appears elevated with the apex of the dome
displaced more laterally than normal. A decubitus chest radiograph with the affected
side dependent will confirm the presence of blood in the pleural space.




CHAPTER 11. CHESTTRAUMA


Pulmonary lesions (fig 11.4)
Haematomas within the lungs appear as pulmonary masses that resolve over days. They
can be differentiated from other masses by reviewing serial radiographs. Contusion
appears as opacification of the lungs.


Vascular injuries (figs 11.5a, 11.5b)
Aortic arch injuries are important to detect especially after deceleration injuries to
the chest and mediastinum. The aorta tears transversely at the level of the ductus
arteriosus. The important signs on the chest radiograph are:


e A widened mediastinum >8 em wide on an erect film


• Loss of the normal mediastinal outline


61


Fig 11.3
Chest radiograph of
a patient with a
right subpulmonic
haemothorax. Note
the "pseudo
elevated" right
hemidiaphragm.


Fig 11.4
Chest radiograph of
a patient with a left
upper lobe
pulmonary
haematoma
following blunt
trauma.




62


Fig 11.5
Chest radiograph of
a patient following


blunt trauma to the
sternum,


demonstrates a
markedly widened


superior
mediastinum with
loss of the normal
aortic arch outline


from a false
aneurysm of the


aortic arch and a
rupture of the left


hemidiaphragm and
herniation of the


stomach in the
chest. Fig 11.5b


demonstrates the
false aneurysm of


the proximal
descending aorta.


PATIERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 2. CHEST IMAGING PATIERNS


b


• Fracture of the upper ribs


• Apical cap from haematoma


If these signs are present the patient should be transferred for aortography to confirm
the diagnosis. It is important to measure the superior mediastinum on an erect chest
radiograph if possible. Note a supine radiograph may falsely suggest a wide
mediastinum.


Trauma to the tracheobronchial tree
Injury to the trachea or bronchi leads to a pneumomediastinum and pneumothorax and
often a collapsed lung.


a




CHAPTER 11. CHESTTRAUMA


Diaphragmatic injury (fig 11.6)
Rupture of the diaphragm usually involves the left hemidiaphragm. Diaphragmatic
tears can often be asymptomatic untill bowel herniation and incarceration occur. The
diagnosis is easily missed and this can be fatal. Contrast studies of the stomach and left
colon demonstrate the position of these structures in relation to the diaphragm and, to
detect herniation.


Penetrating trauma
Penetrating trauma often causes haemopneumothorax and pneumomediastinum.
Vascular injury will be detected by a widened mediastinum; the patient will require
arteriography to confirm the diagnosis. Injuries may involve the oesophagus with
perforation (fig 11.7) A swallow using water-soluble contrast will detect the leak.


63


Fig 11.6
Chest radiograph of
a patient following
blunt trauma
demonstrates a
large left traumatic
diaphragmatic
hernia and free
intraperitoneal air
below the right
hemidiaphragm.


Fig 11.7
Barium swallow of a
patient who
sustained a knife
wound to the
superior
mediastinum
demonstrates an
oesophageal
perforation and leak
of contrast into the
mediastinum.
Important: Use
water-soluble
contrast medium
when perforation is
suspected!




64
PATIERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 2. CHEST IMAGING PATIERNS


Injuries of the lower chest may cause a defect in the diaphragm with the possibility of
bowel herniation and incarceration. This is very dangerous and needs to be detected
early when asymptomatic by checking the diaphragm carefully in a patient with a
history of penetrating injury.


LEARNING POINTS


• Pneumothorax is best detected on erect chest or decubitus chest radiograph.
• Mediastinal shift away from the injured side indicates a tension pneumothorax
• Subpulmonic haemothorax causes apparent elevation of the hemidiaphragm with a laterally


displaced dome. Diagnosis confirmed on a decubitus chest radiograph.
• Aortic injury sugested by a wide mediastinum >8 em dianmeter, loss of normal mediastinal outline,


apical cap. Requires aortography to confirm the diagnosis.
• Diaphragmatic injury is often missed, it commonly involves the left hemidiaphragm. Requires


contrast study of stomach and left colon to confirm the diagnosis.
• Always attempt to assess mediastinal widening on an erect chest radiograph if possible.


Reference
1. Chest trauma Flower CDR. In: Diagnostic Radiology-a textbook of medical imaging. Eds


Grainger RG, Allison DJ, 1997, Churchill Livingstone, Edinburgh.




CHAPTER 12


Pulmonary AIDS
Peter Corr


There has been a rapid increase in AIDS in many countries in Africa and Asia over the
last ten years. An understanding of common presentations of pulmonary opportunistic
infections and AIDS related neoplasms is important as 80% of AIDS patients will have
respiratory disease (1).


Tuberculosis patterns in AIDS (fig 12.1)
The human immunodeficiency virus is synergistic with tuberculosis (2). Tuberculosis
infection is progressive, often extra pulmonary and multifocal in patients with AIDS (2).
Adults present with a pattern similar to primary TB in children with unilateral hilar
or mediastinal adenopathy and lower lobe opacification. Cavities are uncommon.
Progressive disease follows spread along the bronchial tree in addition to blood borne
spread. Patients often have skeletal or abdominal tuberculosis as well (2).


Ground glass pattern (fig 12.2)
The cause is pneumocystis carinii and or cytomegalovirus infection. Patients have a
"ground glass" opacification of the lungs that obscures the pulmonary vessels. It starts
around the hila and spreads outwards. It is easy to miss in early cases but the clue is the
lung vessels appear blurred and ill defined. Lymphadenopathy is rare in pneumocystis
infection. However not all patients have the typical "ground glass" pattern, a large
number also have consolidation, nodules, cavitating masses, subpleural cysts and
spontaneous pneumothorax (3).


65


Fig 12.1
Chest radiograph of
an AIDS patient
with pulmonary TB.
Note the
predominantly lower
lobe distribution of
the opacification
with perihilar
cavities.




66


Fig 12.2
Chest radiograph of


a patient with
pneumocystis


carinii infection.
Note the perihilar


"ground glass"
opacification.


Fig 12.3
Chest radiograph of


an AIDS patient
with pulmonary


Kaposi's sarcoma.
Note the hilar


lymphadenopathy
and the thickened


bronchovascular
markings in the


perihilar regions.


PATTERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 2. CHEST IMAGING PATTERNS


Bacterial infection patterns
Bacterial pneumonias are very common in AIDS patients especially streptococcus
pneumoniae, haemophilus influenza, pseudomonas infection because the HIV infected
lymphocytes cannot destroy bacteria with capsules. A destructive pneumonia or
bronchiectasis often results from inadequate treatment especially in children.


Viral infection patterns
Cytomegalovirus infection and herpes simplex cause a bronchioloitis and pneumonitis
with a "ground glass" pattern identical to pneumocystis carinii infection.


Neoplasm patterns (fig 12.3)
Kaposi's sarcoma is a vascular tumour that occurs in up to 25% of all patients with
AIDS. It is caused by the herpes type 8 virus and causes masses in the skin, bowel and
lungs ( 4). In the chest, nodules are noted spreading from the hila peripherally along the
pulmonary vessels and bronchi in "tongue" like projections.


Non Hodgkin's lymphoma presents in the lungs of these patients with pulmonary
nodules or cavitating masses, however lymph node involvement in the hila and
mediastinum is uncommon.




CHAPTER 12. PULMONARY AIDS


Lymphoproliferative interstitial pneumonia (LIP) pattern (fig 12.4)
Benign lymphocyte proliferation is a common cause of the miliary or small nodular
pattern in AIDS patients (5). It is a benign response by lymphocytes to the HIV virus and
is found in children and young adults. It is important to consider this pattern in the
differential diagnosis of miliary tuberculosis.


LEARNING POINTS: PULMONARY AIDS


• Tuberculosis is progressive, multifocal and often extrapulmonary in AIDS patients. Cavities are
uncommon in severe immunosuppression. Lower lobe infiltrates and hilar adenopathy are common
in adults.


• Pneumocystis pneumonia is common in severe immunosuppression. Ground glass opacification is
common in both pneumocystis and cytomegalovirus infections.


• Bacterial infection is very common, and leads to destructive pneumonia and bronchiectasis.
• Viral infections especially cytomegalovirus and herpes simplex infections are common.
• Kaposi's sarcoma is the commonest neoplasm. Multiple nodules along the bronchovascular


bundles are noted on chest radiographs and CT.
• Lymphocytic interstitial pneumonia (LIP) is a benign lymphocytic response to HIV. It causes


multiple small pulmonary nodules. Main differential is miliary TB.


References
1. Fauci A. The AIDS epidemic-considerations for the 21'' century. NEJM 1999;341:1046-1050.


2. Havlir D, Barnes P. Current concepts: Tuberculosis in patients with the human immuno-
deficiency virus infection. NEJM 1999;340:367-373.


3. Boiselle PM, Crans CA, Kaplan MA. The changing face of pneumocystis carinii pneumonia in
AIDS patients. AIR 1999;172:1301-1309.


4. Antman K, Chang Y. Kaposi's sarcoma-review article. NEJM 2000;342:1027-1033.


5. Berdan WE, Mellins RB, Abramson SJ, Ruzal-Shapiro C. pediatric HIV infection in its second
decade-the changing pattern of lung involvement. Clinical, plain film, and CT findings. Rad
Clin N America 1993;31:453-463.


67


Fig 12.4
Chest radiograph of
an AIDS patient
with lympocytic
interstitial
pneumonitis. Note
the miliary nodules
in the lungs.




68


Fig 13.1
Chest radiograph of
a premature infant


with hyaline
membrane disease.
Note the pulmonary


opacification and
small volume lungs.


CHAPTER 13
- Uil


Paediatric chest
Peter Corr


Children present challenges for the radiographer. The child will often not keep still for
the radiograph so immobilization is important. This requires an assistant wearing a lead
rubber apron and gloves to hold the child by the outstretched arms. Chests are taken in
the AP position. You must be aware of the radiation exposure to the gonads of the child,
and a lead strip must be placed over the gonadal area.


NEONATAL CHEST PATIERNS
Ground glass pattern in premature babies (fig 13.1)
A bilateral "ground glass" appearance to the lungs is found in premature babies with
hyaline membrane disease and viral infections. Hyaline membrane disease is due to
insufficient surfactant. Lung volumes are small and air bronchograms are commonly
detected.


"Bubbly" or cystic lungs
Cystic lesions in the lungs of a neonate may be due to congenital diaphragmatic
herniation of bowel that requires urgent surgery. This is due to a congenital defect in
the left hemidiaphragm (fig 13.2). There is usually mediastinal shift to the opposite
side. Usually both the compressed lung and contralateral lung are hypoplastic. Cystic
adenomatoid malformation of the lung is a congenital malformation of the lung that
usually presents as a cystic mass. There is usually no mediastinal shift away from the
lesion.




CHAPTER 13. PAEDIATRIC CHEST


Meconium aspiration pattern (fig 13.3)
The lungs are hyperinflated with perihilar opacities in these full term babies who have
aspirated meconium into their lungs during labour. There are streaky linear opacities in
both lungs.


INFANT CHEST PATIERNS
Bilateral air trapping pattern
Hyperinflated lungs with peribronchial infiltrates are found in bronchiolitis from viral
infections especially the respiratory syncytial virus and adenovirus infections. It is
important to exclude pneumonia, which may require antibiotic treatment.


69


Fig 13.2
Chest radiograph of
an infant with
respiratory distress
from a congenital
diaphragmatic
hernia in the left
hemithorax.


Fig 13.3
Chest radiograph of
a neonate with
respiratory following
meconium
aspiration. Lungs
are hyperinflated
with perihilar
infiltrates.




70


Fig 13.4
Chest radiograph of


a child who
aspirated a ball


point pen tip
(foreign body) into
his left lower lobe
bronchus causing


lower lobe
atelectasis.


PATIERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 2. CHEST IMAGING PATIERNS


Unilateral lung hyperinflation pattern
Unilateral lobar or segmental hyperinflation should suggest a foreign body in the
bronchus causing a "ball valve" effect with air trapping in the lobe. By taking chest
radiographs in inspiration and expiration the hyperinflation can be accentuated.
Similarly lobar atelectasis in a child is due to an impacted foreign body until proven
otherwise (fig 13 .4).


Bronchopneumonia pattern
This is a very common pattern of peribronchial opacification of both lungs from viral
and or bacterial infection. It is important not to confuse this pattern with cardiac failure
where the heart is enlarged and there are often pleural effusions.


Pulmonary atelectasis pattern (figs 13.4a & 13.4b)
Unilateral volume loss and opacification of a lobe or segment of the lung may be due to
bronchial obstruction from a foreign body, mucus plug as seen in pertussis, asthma, or
enlarged hilar nodes in tuberculosis.




CHAPTER 13. PAEDIATRIC CHEST


CHILDHOOD PNEUMONIA PATIERNS
Round pneumonia (fig 13.5)
This pattern is common in childhood infection and can mimic a mass in the lung. The
key to this pattern is the presence of air bronchograms within the opacification. Round
pneumonias occur because infection spreads easily through the interalveolar foramina.


Staphylococcal pneumonia pattern (figs 13.6a & 13.6b)
This is a serious pneumonia in children. There are features of a bronchopneumonia
with multiple cavities or cysts. It is important to recognise this pattern because
untreated or treated with the incorrect antibiotic will lead to very serious complications
such as pneumothorax and empyema.


a


b


71


Fig 13.5
Round pneumonia.


Fig 13.6a, 13.6b
Chest radiograph of
an infant with
staphylococcal
pneumonia in the
right lower zone
with cavitation.
A tension
pneumothorax with
collapse of the right
lung is a recognised
complication 6b.




72


Fig 13.7
Chest radiograph of


an infant with
primary TB


demonstrates right
hilar and


para tracheal
lymphadenopathy.


PATIERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 2. CHEST IMAGING PATIERNS


Primary TB pattern (figs 13.7a & 13.7b)
This pattern is detected in primary tuberculosis where there are large unilateral hilar
and mediastinal lymph nodes which may cause bronchial obstruction and lobar
atelectasis. The primary lung lesion is often not detected as an area of focal pulmonary
opacification (Chon focus).


a


b


LEARNING POINTS: PAEDIATRIC LUNG PATTERNS
• Hyaline membrane disease presents in premature babies with small volume lungs and ground glass


opacification.
• Bubbly lungs in a neonate must suggest congenital diaphragmatic herniation if there is mediastinal


shift to the opposite side.
• Meconium aspiration syndrome results in hyperinflated lungs with a streaky appearance in full term


infants following meconium aspiration.
• Bilateral air trapping in infants should suggest bronchiolitis.
• Unilateral hyperinflation or lobar atelectasis of a lung or lobe should suggest an impacted foreign


body in the bronchial tree.
• Round pneumonias in children can mimic lung masses but they often have bronchograms.
• Staphylococcal pneumonia presents with cavitating bronchopneumonia or cyst formation. Serious


complications of pneumothorax and empyema can occur if not recognized early and treated.




CHAPTER 14


Cardiac disease
Peter Carr


Cardiac size (fig 14.1)
The heart size is normally 50% or less of the thoracic diameter in adults as measured
from the inner borders of the lower ribs. It is important to use an erect chest film,
preferably a P A film, with a good inspiration. A. film focal distance of 180 em is
important for cardiac assessment. A poor inspiration and or a supine film is suboptimal
as it results in the heart having an apparently increased transverse diameter and
apparent cardiomegaly. In children up to 12 years of age the maximum heart size is 60%
or less of the internal thoracic diameter.


Cardiac silhouette enlargement (fig 14.2)
Enlargement of the cardiac silhouette may be due to cardiomegaly or pericardia!
effusion. Pericardia! effusion gives the heart a globular appearance with loss of the
normal cardiac contour (fig 14.3). The diagnosis of pericardia! effusion can be made
by ultrasound examination using a 2 MHz transducer. There is a hypo echoic fluid
collection in the preicardium surrounding the heart. Common causes of pericardial
effusions are:


• infection-viral, pyogenic and tuberculous


• neoplastic-metastatic disease and direct invasion from breast and lung cancer


• trauma-haemopericardium or sympathetic effusion


Tuberculous pericarditis can be identified by detecting strands of fibrin in the
pericardia! effusion. As the effusion resolves pericardia! constriction may occur. This
can be identified as faint calcification of the pericardia! lining on the P A chest
radiograph.


73


Fig 14.1
Chest radiograph of
a patient with
cardiomegaly.




74


Fig 14.2
Chest radiograph of


a patient with a
pericardia/ effusion.


Note the globular
contour of the


cardiac shadow.


Fig 14.3
Chest radiograph of


a patient with
hypertension with


left ventricular
cardiomegaly. Note
how the left cardiac


border slopes
towards the left


costophrenic angle.


PATTERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 2. CHEST IMAGING PATTERNS


Mitral valve disease
Mitral valve disease causes enlargement of the left atrium and right ventricle following
either mitral valve stenosis or incompetence as the end result of rheumatic heart
disease. Enlargement of the left atrium is detected as a double shadow behind the right
atrium on the P A chest radiograph with splaying of the carina of the trachea and
posterior displacement of the oesophagus on barium swallow.


Aortic valve disease
Aortic valve stenosis causes left ventricular hypertrophy that is recognized as left
displacement of the cardiac apex on the PA chest radiograph. Aortic valve incomptence,
usually following rheumatic fever or infective endocarditis leads to left ventricular
dilatation.


Cardiac failure pattern (figs 14.4 & 14.5)
Left ventricular failure initially causes pulmonary venous distension in the upper lobes
and constriction of the pulmonary veins in the lower lobes. As the venous pressure




CHAPTER 14. CARDIAC DISEASE


rises, there is perihilar oedema, detected as blurring of the hilar vessels and perihilar
opacification. Pleural effusions develop at the costophrenic angles, then septal lines
form at the CP angles.


Septal lines are 1-2 em long and 1 mm thick in the region of the costophrenic angles.
They represent distended lymphatics and are an early sign of interstitial pulmonary
oedema. As the cardiac failure progresses there is perihilar opacification ("bats
wing" distribution) which represents alveolar pulmonary oedema. Pulmonary oedema
usually resolves rapidly over hours with diuretic and anti failure treatment. This pattern
can be used to differentiate alveolar oedema from other air space opacification from
pneumonia or pulmonary haemorrhage.


Pulmonary plethora pattern
Plethora means increased blood flow and is usually due to left to right intracardiac
shunts such as an atrial septal defect, ventricular septal defect and patent ductus
arteriosus. Pulmonary arteries are dilated from the hila out to the periphery of the lungs.
This pattern should not be confused with pulmonary venous dilatation that involves
only the upper lobes.


Pulmonary oligaemia pattern
This pattern is often difficult to recognise unless you have a history of cyanotic heart
disease to suggest the correct diagnosis. The lungs have diminished pulmonary blood


75


Fig 14.4
Chest radiograph of
a patient with
interstitial
pulmonary oedema.
Note the septal
lines and prominent
upper lobe
pulmonary veins.


Fig 14.5
Patient in acute
pulmonary oedema
demonstrates
pulmonary
opacification and
dilated upper lobe
veins.




76


Fig 14.6
Patient with


pulmonary
hypertension


from an atrial
septal defect
(Eisenmenger


syndrome). Note
cardiomegaly, with
dilated pulmonary
arteries centrally


and peripheral
pruning.


PATTERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 2. CHEST IMAGING PATTERNS


flow however the bronchial arteries often enlarge to compensate for the pulmonary
oligaemia and the lungs may appear plethoric. Pulmonary oligaemia is detected in the
"3 T's": Tetralogy of Fallot, truncus arteriosus, and transposition of the great vessels.
It is also seen in reversed left to right intracardiac shunts (ASD with pulmonary
hypertension: Eisenmenger's syndrome fig 14.6).


LEARNING POINTS: CARDIAC DISEASE


• Cardiomegaly >50% cardio thoracic ratio in adults, >60% in children less than 12 years.
• Remember cardiac silhouette enlargement may be due to cardiac dilatation, hypertrophy and


pericardia! effusion (globular contour). Pericardia! effusion can easily be diagnosed by ultrasound.
• Signs of cardiac failure are upper lobe pulmonary vein distention, blurring of the perihilar vessels


and pleural effusions and septal lines.
• Signs of pulmonary oedema are septal lines and perihilar opacities that often resolve within hours


on anti-failure treatment.




CHAPTER 15


Mediastinal masses
Peter Carr


The mediastinum can be divided into superior, anterior, middle, and posterior
compartments.


Superior mediastinum
The superior mediastinum is superior to the level of the aortic arch. Masses in this
region include retrosternal thyroid masses and aneurysms of the ascending aorta and
great vessels. Retrosternal goitres cause displacement of the trachea, extend inferiorly
from the neck, and may have focal calcification within them. The diagnosis can be
confirmed by ultrasound of the neck or by using A common cause of a superior
mediastinal mass on the right is a dilated brachiocephalic artery in an elderly patient.
Here the diagnosis can be easily confirmed by ultrasound. The important point to
remember is that this is a normal aging phenomenon and not to do any further
investigations on these patients.


Anterior mediastinum
This region is between the heart and ascending aorta and the sternum and inferior to the
aortic arch. It is best visualized on a well penetrated lateral chest radiograph. The region
contains lymph nodes, the thymus in children and the thymic remnant in adults, and
the pericardium. Masses include: enlarged lymph nodes, thymic masses, teratomas or
dermoids, and aneurysms of the ascending aorta. Often it is not possible to differentiate
between these masses on radiographs however if the mass has a lobulated contour
it suggests enlarged lymph nodes. Focal calcification is detected in dermoids and
teratomas. Linear calcification may be detected in aortic aneurysms from syphilis or
Takayashu's disease.


Middle mediastinum
This region is between the anterior and posterior mediastinum and contains the heart,
great vessels, pulmonary vessels and lymph nodes. Masses in this region include:
enlarged hilar and tracheobronchial lymph nodes, bronchogenic cysts, aortic arch
aneurysms, and bronchial carcinomas.


Posterior mediastinum
The posterior mediastinum extends from the middle mediastinum to the spine and
paraspinal gutters. It includes the descending thoracic aorta, lymph nodes, nerves and
the oesophagus. Masses in this region: neurogenic tumours, paravertebral masses,
tuberculous abscesses, oesophageal tumours, hiatus hernia, and dilated oesophagus. It
is very important when a posterior mediastinal mass is detected to clearly visualise
the spine, so as to detect early bony destruction or erosion from tuberculosis or bony
metastases.


77




78
PATIERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 2. CHEST IMAGING PATIERNS


LEARNING POINTS: MEDIASTINAL MASSES
• Common superior mediastinum masses: retrosternal thyroid goitre, aneurysms of great vessels
• Common anterior mediastinum masses: 3T's teratomas or dermoids, retrosternal thyroid, thymoma


and ascending aortic aneurysm
• Common middle mediastinum masses: lymphadenopathy, bronchogenic cysts, bronchial


carcinoma
• Common posterior mediastinum masses: lymphadenopathy, neurogenic tumours, paraspinal


abscess especially TB, oesophageal tumours


References
1. Chapman S, Nakielny R. Aids to Radiological Differential Diagnosis. 1995, Saunders, London.
2. Armstrong P. The Mediastinum. In: Diagnostic Radiology-Textbook of Medical Imaging.


Grainger RG, Allison DJ. 3rct edition 1997 Churchill Livingstone, Edinburgh.




CHAPTER 16


Diaphragm lesions
Peter Carr


Diaphragmatic elevation
Eventration is a congenital deficiency of the muscle in the hemidiaphragm usually the
left, resulting in non-function and resultant elevation. The diagnosis is confirmed on
fluoroscopy j screening when there is limited or no movement on deep inspiration and
expiration. If fluoroscopy is unavailable perform one chest film using a double exposure
in inspiration and an expiration at SO% normal mas, this will show movement of the
diaphragm. Eventration must be distiguished from phrenic nerve palsy.


Phrenic nerve palsy results in paralysis of diaphragmatic movement on one side resulting
in elevation of the hemidiaphragm and limited or paradoxical movement on deep
inspiration and expiration. There are many causes, however the commonest causes is
carcinoma of the bronchus.


Subphrenic masses, liver enlargement and abscesses will elevate the hemidiaphragm. On
the chest radiograph with a liver or right subphrenic abscess, the right hemidiaphragm
is elevated and its outline becomes "fuzzy" or ill-defined with associated linear
atelectasis in the right lower lobe. This is a very important pattern to recognise as you
need to perform an ultrasound examination of the right subphrenic space and liver to
detect an abscess. Pseudo elevation is seen with subpulmonic pleural effusions and
haemothoraces. The apex of the "elevated hemidiaphragm" is usually displaced more
laterally that what is normally seen. A decubitus chest film with the affected side
dependent will confirm the diagnosis.


Diaphragmatic injury is commonly missed especially following penetrating trauma. Injury
is nine times more common on the left. The diagnosis can be difficult to make. On the
left, detection of bowel loops above the hemidiaphragm or a soft tissue mass contiguous
to the diaphragm is suspicious. A contrast study of the stomach and colon will confirm
the diagnosis. If there is incarcerated bowel present is will be opacified by contrast
above the diaphragm.


79




80


Fig 17.1
Patient with


silicosis. Note
dense bilateral


pulmonary nodules
in mid zones of the


lungs.


Fig 17.2
Patient with


progressive massive
fibrosis with


bilateral perihilar
masses centrally.


CHAPTER 17


Pneumoconiosis
Peter Corr


Occupational dust exposure is a common cause of lung disease in many developing
countries. You must always ask whether the patient has been exposed to dust at
work.


Small nodules
This is the commonest occupational dust exposure pattern and results from inorganic
dust exposure such as silica. The nodules are around 1-2 mm size and dense com-
pared to the granulomas of tuberculosis (fig 17.1). Hilar nodes may be enlarged and
have peripheral calcification ("egg shell" calcification). A complication of silicosis is
conglomeration of the nodules into large masses of fibrosis called progressive massive
fibrosis (fig 17.2).




CHAPTER 17. PNEUMOCONIOSIS


Asbestos exposure
The most common appearance are pleural plaques which become easier to see once
they calcify giving a "holly leaf" appearance. Diaphragmatic calcification is also
common. These findings indicate asbestos exposure (fig 17 .3). Pulmonary asbestosis
presents with a linear or reticular fibrosis of the lower lobes and represents pulmonary
fibrosis from asbestos exposure. Mesothelioma is a pleural malignancy that spreads
circumferentially along the pleura. The diagnosis is usually made by biopsy. The
differential diagnosis is encysted pleural fluid that can be detected by using a 5 MHz
ultrasound transducer on the chest wall.


LEARNING POINTS: PNEUMOCONIOSIS


• Always ask the patient about dust exposure
• Consider silicosis in the small nodular lung pattern especially when the nodules appear very white


or dense
• Consider asbestosis esposure when there are calcified pleural plaques and or calcification of the


diaphragm


Fig 17.3
Patient with
asbestos dust
exposure
demonstrates
multiple calcified
plaques on the
paritetal pleural
surface.


81






PART3


MUSCULOSKELETAL PATTERNS






CHAPTER 18


Approach to focal bone lesions
Fei-Ling Thoo & Wilfred C. G. Peh


Focal bone lesions can generally be divided into benign and malignant bone lesions.
The malignant group can be further subclassified into primary and secondary tumours.
The secondary tumours can arise from transformation of benign conditions or from
metastatic lesions.


Clinical information
The patient's age and determination of whether a lesion is solitary or multiple are
important approaches in the diagnosis of bone tumours. Aneurysmal bone cysts rarely
occur beyond 20 years of age. Giant cell tumour usually occurs after the closure of the
growth plate. Metastases tend to be multi-focal and are more common in the older
age group. The rate of tumour growth may be an additional factor in differentiating
malignant tumours (usually rapid growing) from benign lesions (slower growing). It is
also important to know if a lesion is an incidental finding or is symptomatic. If painful,
the lesion requires attention regardless of its imaging appearance. Some benign osseous
tumours may undergo sarcomatous transformation and this should also be considered
in a patient who presents with pain and a lesion that appears benign.


Imaging modalities
In the evaluation of bone tumours, plain radiographs are the standard imaging study.
The choice of the imaging technique is dictated by the type of the suspected tumour
and also by equipment available. Imaging modalities for bone tumours include bone
scintigraphy, CT scan, MRI and angiography. CT is superior to MR for the detection
of calcification in the tumour matrix, cortical erosions and periosteal reaction. If
the radiographs suggest cortical destruction and soft tissue mass, MRI would be
the preferred as it provides excellent soft tissue contrast and can determine the
extraosseous extent of tumour much better than CT.


Site of the lesion
The bone tumour can be epiphyseal, metaphyseal or diaphyseal in location. There is
predilection of some bone tumours for specific sites in the bone.


Skeletal predilection of benign osseous neoplasms include
Enchondroma-short tubular bones (fig. 18.1)
chondroblastoma-epiphyses


giant cell tumour-articular ends of the femur, tibia and radius chondromyxoid


fibroma-tibial metaphyses


simple bone cyst-proximal humerus and femur


osteoid osteoma-femur and tibia


85




86


Fig 18.1
Shows fracture


through an
enchondroma
located in the


proximal phalanx of
the little finger. The
tubular bones of the


fingers are typical
sites for


enchondroma.


PAITERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 3. MUSCULOSKELETAL PAITERNS


Skeletal predilections for malignant osseous tumours include
chordoma-sacrum, clivus, C2


multiple myeloma-pelvis, spine and skull


parosteal osteosarcoma-posterior cortex of posterior femur


chondrosarcoma-epiphyseal lesion of femur and humerus


adamantinoma-tibia, fibula


Lesions that have a predilection for flat bones such as the scapula body and the
iliac wing or the diaphyses of long bones include Ewing sarcoma, lymphoma and
Langer hans cell histiocytosis (eosinophilic granuloma). Ewing sarcoma would be more
likely in a younger age group. Lymphoma can be seen in any age but peaks later in life.
Langerhans cell histiocytosis can be seen at any age.


Borders of the lesion
Slow growing lesions are usually benign and have sharply outlined sclerotic borders
(narrow zone of transition). An example of a benign lesion (non ossifying fibroma) is
shown in fig 18.2. Aggressive or malignant lesions typically have indistinct borders (a
wide zone of transition) with either minimal or no reactive sclerosis (fig 18.3). This is
seen in fig 18.4 of a child with osteogenic sarcoma. Treatment can alter the appearance
of malignant bone tumours; they may exhibit sclerosis as well as a narrow zone of
transition.




CHAPTER 18. APPROACH TO FOCAL BONE LESIONS


87


Fig 18.2
Well defined
osteolytic lesion
with sharply
outlined sclerotic
border and a
narrow zone of
transition in the
distal right tibia due
to a non-ossifying
fibroma.


Fig 18.3
Expansive osteolytic
lesion in the
subarticular region
of the distal femur
in keeping with a
giant cell tumour.




88


Fig 18.4
Plain radiograph of


an osteogenic
sarcoma in the


proximal/eft
humerus of a child.


There is an
osteolytic lesion in


the diametaphyseal
region which shows


a poor zone of
transition. Adjacent
sunburst periosteal


reaction and soft
tissue mass are


present.


Fig 18.5
Plain radiograph


shows calcification
within the medullary


cavity from a
chondroid tumour.


PATTERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 3. MUSCULOSKELETAL PATTERNS


Type of matrix
Osteoblastic and cartilaginous matrix can be recognized radiographically. Osteogenic
sarcomas can form cloud-like dense osteogenic matrix. Cartilaginous tumours can be
identified by popcorn-like, punctate, annular or comma-shaped calcifications. Cartilage
tumours tend to grow in lobules and can be identified by their lobulated growth
(fig 18.5).


Type of bone destruction
The type of bone destruction can be described as geographic, moth-eaten or permeative.
A benign process tends to show a geographic-uniformly destroyed area with sharply




CHAPTER 18. APPROACH TO FOCAL BONE LESIONS


defined border. A likely malignant process shows moth-eaten areas of destruction with
ragged borders. It should be kept in mind that non-neoplastic lesions like osteomyelitis
may also appear as an aggressive destructive bone lesion with moth eaten areas. The
clinical presentation, radiographic findings of cloaca and sequestrum and uninterrupted
periosteal reaction would be helpful in the diagnosis of osteomyelitis. The permeative
bone destruction is indicative of a more aggressive (malignant) process with ill-defined
area spreading through the bone marrow. This is seen in Ewing's sarcoma and bone
lymphoma.


Periosteal reaction
The type of periosteal reaction is one of the most important features in determining the
aggressiveness of the bone lesion. Periosteal reactions are discussed in the next chapter.


Soft tissue extension
Benign tumours do not exhibit soft tissue extension. Few exceptions are giant cell
tumours, aneursysmal bone cysts, osteoblastomas and desmoplastic fibromas. It should
be kept in mind that non-neoplastic conditions such as osteomyelitis also show a soft
tisssue component, the involvement of the soft tissue is usually poorly defined, with
obliteration of the soft tissue layers. In a malignant process, the soft tissue component is
sharply defined, extending through the destroyed cortex.


If a bone lesion is associated with a large soft tissue mass, round cell tumours should
be a consideration. These include metastatic neuroblastoma (seen in infancy), Ewing's
sarcoma, primitive neuroectodermal tumour (paediatric patients), lymphoma (most
common in adults) and plasmacytoma (seen in the middle-aged to elderly population).


Multiplicity of lesions
i) Multiple malignant appearing lesions usually indicate metastatic disease, multiple


myeloma or lymphoma.


ii) Benign lesions with multifocal presentations include polyostotic fibrous dysplasia,
multiple enchondromas, enchondromatosis, histiocytosis, haemangiomatosis, and
Paget's disease.


Management: biopsy versus "do not touch lesions"
There are certain features of a bone lesion on a radiograph, which help distinguish
between benign and malignant lesions. Benign lesions usually have well defined sclerotic
borders, a geographic type of bone destruction, an uninterrupted solid periosteal reaction
and no soft tissue mass. Malignant lesions have poorly defined borders with a wide zone
of transition, a "moth eaten" or permeative pattern of bone destruction, an interrupted
periosteal reaction of a "sun burst" or "onion skin" type and an adjacent soft tissue mass.
The analysis of a lesion involves clinical and radiological information. A decision must
be made whether the lesion is definitely benign and not to be biopsied ("a do not touch"
lesion) but rather monitored or whether a biopsy is required.


The following is a list of "do not touch" lesions
Tumour and tumour-like lesions


Fibrous cortical defect


Non ossifying fibroma


Cortical desmoid


89




90


Solitary fibrous dysplasia


Pseudotumour of haemophilia


Intraosseus ganglion


PATTERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 3. MUSCULOSKELETAL PATTERNS


Enchondroma of a short tubular bone


Non neoplastic processes
Stress fracture


Avulsion fracture


Bone infarct


Bone island


Myositis ossificans


Degenerative or post traumatic cysts


Brown tumour of hyperparathyroidism


Discogenic vertebral sclerosis


LEARNING POINTS: FOCAL BONE LESIONS


• Helpful clinical data are:
(a) age of the patient
(b) duration of symptoms and
(c) growth rate of the tumour


• Key radiographic features:
(a) site of tumour
(b) border of the lesion
(c) type of matrix
(d) type of bone destruction
(e) type of periosteal reaction
(f) the presence of absence of soft tissue extension


• A lesion is slow growing (likely to be benign) when it shows:
(a) geographic bone destruction
(b) sclerotic margin
(c) solid, uninterrrupted periosteal reaction, or no periosteal response
(d) no soft tissue mass


• A lesion is aggressive (likely to be malignant) when it shows:
(a) poorly defined margins
(b) moth-eaten or permeative type of bone destruction
(c) interrupted periosteal reaction
(d) soft tissue mass


• A lesion is likely to represent a cartilage tumour when it shows:
(a) lobulation (endosteal scalloping)
(b) calcifications in the matrix




CHAPTER 19


Periosteal reactions
Lai-Ping Chan & Wilfred C. G. Peh


The periosteum is a thick layer of fibrous tissue that covers the surface of the bone. The
periosteum has abundant neurovascular supply and the cells in the deeper layers are
able to form bone. The periosteum is not normally s~en on imaging but when it
responds to various bony insults, the resultant periosteal reaction is seen as projections
of bone arising from the bony cortex. There are several patterns of periosteal reaction,
and they can be due to benign or malignant conditions. Careful perusal of the
underlying bone will give important clues about the etiology of the periosteal reaction.


Patterns of periosteal reactions
Periosteal reactions can be solid or interrupted. Four types are described:


SOLID PERIOSTEAL REACTIONS
Thin undulating periosteal reaction (diagram 19A)


This appears as an undulating bony margin around the shafts of the
bone, sparing the epiphysis. This type of periosteal reaction tends
to be bilateral and symmetrical and is commonly due to systemic
disease rather than a localized process. This pattern, when due
to vascular insufficiency (either arterial, venous or lymphatic) is
usually found in the legs with soft tissue swelling. Another cause
of an undulating periosteal reaction is hypertrophic pulmonary
osteoarthropathy (HPOA). In these patients, the underlying bone
may appear osteoporotic. Patients with HPOA have painful
swelling of the joints, especially the wrists and ankles. They may
also have clubbing of the fingers. HPOA is often secondary to
chronic pulmonary disease, heart disease,
inflammatory bowel disease, liver disease
and malignancy. It is important to do a chest
X-ray to exclude thoracic causes of the


..._.,_. ~ disease.
Diagram 19A


Thick solid periosteal reaction (diagram 19B)
This periosteal reaction tends to merge with the cortex of
the bone. The cortex of the bone can appear sclerotic and
thickened. This can be due to an osteoid osteoma, a benign
neoplasm of the bone. This occurs in young patients who
usually present with pain. The lesion tends to occur in the long
bones such as the femoral neck, the proximal tibia, fibula, and
humerus. The lesion appears as a radiolucent area within the


Diagram 198


91




92


Fig 19.1
Osteoid osteoma of


the shaft of the
tibia with a thick


solid periosteal
reaction and


thickened cortex.
Note the small


radiolucent nidus.


Fig 19.2
Brodie's abscess of


the proximal tibia
demonstrates a
thick periosteal


reaction with
surrounding


sclerosis.


PATIERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 3. MUSCULOSKELETAL PATIERNS


thickened bony cortex with a central sclerotic nidus (fig 19.1). A Brodie's abscess can
have a similar appearance (fig 19 .2). This is a subacute osteomyelitis that is most
commonly due to staphylococcus aureus. The tibial metaphysis is the most common
site and the lesion appears as a lucent area surrounded by dense sclerosis. The bony
cortex may be thickened. The lesion can be cortical or intra-medullary in location. A
Brodie's abscess can be differentiated from an osteoid osteoma if a sinus track or
channel can be detected.


0




CHAPTER 19. PERIOSTEAL REACTIONS


Cloaking periosteal reaction (diagram 19C)
This type of periosteal reaction is very abundant and covers
almost the entire bone. If it occurs in a single bone, this is
usually due to chronic osteomyelitis. There is usually associated
thickening and sclerosis of the underlying bone (fig 19 .3). There
may be underlying radiolucent areas present due to involuted
bone and also dead detached cortical bone, which appears very
dense (sequestrum). There is often underlying bone destruction
and new bone formation. There may be soft tissue swelling


Diagram 19C


present. A sinus track leading to the skin
may be present within the soft tissue. If
this type of periosteal reaction occurs in
multiple bones in young infants, the
diagnosis can be congenital syphilis
whose hallmark is a bilateral sym-
metrical osteomyelitis involving multiple
bones (fig 19 .4). The underlying bone


will show evidence of bony destructionin the metaphysis, which
appears as lucent bands adjacent to a widened epiphyseal plate.
The metaphysis may also appear frayed. Bilateral and sym-
metrical focal bone destruction in the medial aspects of the
proximal tibial metaphyses is known as Wimberger's sign and is
almost pathognomonic of congenital syphilis.


Lamellated periosteal reaction (diagram 19D)
In this pattern, the periosteum has an appearance like an "onion
skin" with many layers around the shaft of the bone. This can


Diagram 19D


93


Fig 19.3
Chronic
osteomyelitis of the
humerus
demonstrates
marked cortical
thickening with a
thick periosteal
reaction.




94


Fig 19.4
Cloaking periosteal
reaction along the


shaft of the tibia in
keeping with


congenital syphilis.


Fig 19.5
"Onion skin'


periosteal reaction
in Ewing's sarcoma


of the clavicle.


PATTERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 3. MUSCULOSKELETAL PATTERNS


be due to tumours such as osteosarcoma and Ewing's sarcoma (fig 19. 5), or chronic
osteomyelitis.


In children, rickets can also give this appearance as
uncalcified subperiosteal osteoid separates the periosteum
and the ossified cortex of the bone. The underlying bone
appears poorly mineralised and there may be deformities
such as bowing. The epiphyseal appears widened due to an
abundance of unossified osteoid and there may be cupping
and fraying of the metaphysis. In scurvy, subperiosteal
haemorrhages can occur and during healing, the
periosteum can calcify, giving the appearance of lamellae
around the bone (fig 19 .6).


INTERRUPTED PERIOSTEAL REACTIONS
Four types are discussed:


"Hair-on-end" periosteal reaction (diagram 19E)
These appear as straight projections of bone that are
parallel to the bony cortex, like hairs standing on-end.


~,
--,
,-.


,---(


, __


1--


1-
c==__
1---
c:==
~
;-
1--
(-
r:::::::=-
I


Diagram 19E




CHAPTER 19. PERIOSTEAL REACTIONS


This is usually due to an aggressive process such as a bone tumour or acute
osteomyelitis. Ewing's sarcoma is a tumour that typically causes this appearance. There
is usually associated destruction of the bony cortex. Ewing's sarcoma occurs in the
young, with most cases occuring in patients less than 20 years of age. The long bones
are affected in 60% of patients and the flat bones in the other 40%. This tumour tends
to involve the diaphysis and there are mottled "moth-eaten" destructive changes in the
underlying bone. There may also be soft tissue swelling but unlike osteomyelitis, the
soft tissue planes are usually preserved. The radiographic changes of acute
osteomyelitis begin usually after a few days (initial radiographs tend to be normal).
There is soft tissue swelling adjacent to the bone and the fat planes are obliterated.
There may be elevation of the periosteum associated with the periosteal reaction
(especially in children). The underlying bone will show destructive changes. In
children, the common site to be affected is the bony metaphysis. As the disease
progresses, involucrum and sequestrum can develop. Growth disturbances to the bone
may occur, either shortening of the bone due to epiphyseal destruction or premature
maturation of the epiphysis due to hyperemia. A "hair-on end" appearance can also be
seen in the skull vault (sparing the occipital bone) and this is usually due to chronic
haemolytic anaemias such as sickle cell anaemia and thalassaemia (fig 19. 7). In these
patients, the bone marrow expansion within the skull causes this appearance.


"Sunray" periosteal reaction (diagram 19F)
In this pattern, spicules of bone radiate from the bone in a divergent manner, just like
the rays of the sun (fig 19 .8). This type of periosteal reaction typically occurs in
osteosarcoma. This common malignant primary bone tumour has a peak age of
involvement from 10 to 25 years of age, and another peak after 60 years. The tumour
tends to involve the metadiaphyseal region of long bones and is especially common
around the knee. The tumour can be seen usually expanding and destroying the


95


Fig 19.6
Large calcified
subperiosteal
haematoma around
the distal femur in
a patient with
scury.




96


Fig 19.7
"Hair on end"


periosteal reaction
of the skull in a


child with
thalasemia major.


PATIERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 3. MUSCULOSKELETAL PATIERNS


underlying bone. At the margins of the lesion,
periosteum is lifted up by the tumour and this
elevation is termed a "Cadman's triangle". There is
usually a soft tissue mass and abnormal tumourous
new bone formation is a prominent feature. A
pathological fracture may also be present. Cadman's
triangle is not pathognomonic of osteosarcoma, and
can occur in any condition that elevates the
periosteum. At the periphery of the lesion, the elevated
periosteum calcifies, forming a triangle with the cortex
of the bone. Though not specific, Cadman's triangle
does tend to occur in more aggressive lesions such as
tumour or infection. Bony metastases, especially those
from colonic tumours, can also cause a sunray
periosteal reaction.


Interrupted amorphous periosteal reaction (diagram 19G)
This type of periosteal reaction is discontinuous and appears
irregular. It is usually a result of reaction to a localised process.
A healing fracture is a common cause of this pattern and careful
study of the underlying bone will show a fracture line. There
may also be thickening of the cortex of the bone. In some
instances, the periosteal reaction is due to stress fracture and
common sites being the postero-medial cortex of the proximal
tibia and in the 2nd metatarsal. There is usually associated
cortical thickening (fig 19.9).


This pattern of periosteal reaction can sometimes be seen
in early acute osteomyelitis and there will be associated
destructive lytic changes in the adjacent bone. At later stages,
the periosteal reaction may become more lamellar in ap-
pearance and extend parallel to the shaft of the bone (fig 19 .10).


Diagram 19F


Diagram 19G




CHAPTER 19. PERIOSTEAL REACTIONS


97


Fig 19.8
"Sun burst"
periosteal reaction
due to
osteosarcoma of
the distal femoral
metaphysis.


Fig 19.9
Healing fracture of
the 2'd metatarsal
with surrounding
callus.




98


Fig 19.10
Patient with acute


osteomyelitis of the
proximal femur
demonstrates a


fluffy irregular
periosteal reaction.


PATTERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 3. MUSCULOSKELETAL PATTERNS


LEARNING POINTS: PERIOSTEAL REACTIONS
• Periosteal reactions are result of bone reacting to various insults. The same disease may produce


several patterns of periosteal reaction.
• An aggressive lesion will usually be associated with destruction of the adjacent bone.
• Bilateral periosteal reactions are usually due to systemic diseases or syndromes.
• Obliteration of the soft tissue planes tends to favour an infective process.
• Looking at the underlying bone for an expansile mass, destruction, bone mineralisation and


fracture will give important clues for diagnosis




CHAPTER 20


Extremities trauma
Siew-Kune Wong & Wilfred C. G. Peh


Introduction
Importance of radiographs
The primary aim is to diagnose the presence of a fracture or dislocation. It is also
important to assess the position of the bone ends before and after treatment. Follow-up
radiographs are subsequently needed for bony union and complications.


Principles of radiographic examination
a. It is essential to take radiographs in at least 2 planes, preferably at right angles to each


other. This will ensure that a fracture will not be missed and the bony alignment can
be accurately assessed.


b. The joint above and below the fracture should be included in the radiograph. This is
to assess for associated dislocation especially in paired bones such as those in the
forearm and leg.


c. Due to bone resorption, a fracture line will become more visible about 2 weeks after
an injury. Callus formation may also be present. Hence serial examinations may be
required if a fracture is clinically suspected but is not visible immediately after injury
(fig 20.1).


99


Fig 20.1
Callus formation is
seen around a
stress fracture of
the 4th metatarsal
bone.




100


Fig 20.2
Neutral (a) and


inversion (b) views
of the ankle


demonstrate a
latera/ligament


injury.


PATIERN RECOGNITION IN DIAGNOSTIC IMAGING: PART 3. MUSCULOSKELETAL PATIERNS


d. Comparison views of the opposite limb may be required in the immature skeleton
before epiplyseal closure. This will help to confirm if a bony fragment is an accessory
ossicle, unfused ossified epiphysis, or a fracture.


e. Stress views are useful to assess for ligamentous injury, especially to the ankle and
knees. These views help to accentuate the abnormal widening of the joint space
associated with laxity or injury to the supporting ligaments (figs 20.2a & 20.2b).


a


b




Copyright © 2017 Engineering World Health. All Rights Reserved.