a novel pharmacological approach. Pharmacol Rev 2000;52:179, 205. 23. Kawai T, Hiroi S, Torikata C. Expression in lung carcinomas of platelet-derived growth factor and its receptor. Lab Invest 1997;77:431, 436. 24. Watt FM, Brigid LM. Out of eden: stem cell and their niches. Science 2000;287:1427, 1430.
JWBK089-19 JWBK089-Boulton April 22, 2006 15:5 Char Count= 0
19 Radiology and Magnetic Resonance Imaging of the Diabetic Foot Richard W. Whitehouse
Clinicians managing diabetic patients’ foot problems make frequent use of the radiology de- partment, particularly where trauma or infection is suspected.
Whilst this chapter emphasises the imaging appearances of diabetic foot complications, this should be tempered with the re- alisation that many articles (this chapter included) illustrate the imaging of the complications with ‘textbook’ examples.
The true value of a diagnostic test in practice should be based on a Bayesian statistical approach.
For infection, for example, a clinical finding of an ulcer that can be probed to bone is as predictive of the presence of osteomyelitis as any imaging test, whilst a warm, well-perfused, swollen foot with intact skin is much more likely to be neuropathic than osteomyelitic.
Imaging tests would need extremely high specificity to significantly alter the pre-test probability in each of these scenarios.
Thus, a plain film examination might con- firm neuropathic changes in the latter case but may be inappropriate to confirm superadded osteomyelitis.1 Current literature tends to underemphasise the plain radiograph in the assessment of the foot in patients with diabetes, nuclear medicine scans utilising a variety of radiopharmaceu- ticals, computed tomography, angiography and magnetic resonance (MR) imaging (and MR angiography) being at the fore.
These latter techniques are powerful tools in the evaluation of the diabetic foot; they are, however, time consuming to perform and interpret, expensive, of limited availability and, for nuclear medicine in particular, of relatively high radiation dose.
Plain radiographs are relatively cheap, quick, widely available and usually can be provided immediately to the requesting clinician, rather than awaiting a radiologist’s report.
They have also undergone considerable technological development in the last decade.
Computerised ra- diography (CR) and direct digital radiography (DDR) both produce electronic renditions of the radiographic image.
The image data from such acquisitions undergo computer manipulation to optimise the greyscale, widen the latitude and emphasise tissue planes before being presented to the viewer.
If viewed at an appropriate workstation rather than a hardcopy (film), then further image manipulation is possible.
This technology markedly reduces the effect of radiographic overexposure and also allows a degree of underexposure to provide acceptable images.
Many of the limitations of conventional radiography do, however, remain.
Limited soft tissue contrast
The Foot in Diabetes, 4th Edition. Edited by Andrew J.M. Boulton, Peter R. Cavanagh and Gerry Rayman. C 2006 John Wiley & Sons, Ltd. ISBN: 0-470-01504-7
JWBK089-19 JWBK089-Boulton April 22, 2006 15:5 Char Count= 0
RADIOLOGY AND MAGNETIC RESONANCE IMAGING OF THE DIABETIC FOOT 223
resolution, superimposition of complex structures and adequate evaluation of cortical bone being restricted to those bone edges depicted in profile are the main limitations of radiography.
In addition, visible changes on radiographs tend to lag behind the evolution of disease, both in its progression and in its resolution.
The computer manipulations that ‘bring out’ soft tissue appearances in the image also tend to reduce the contrast of bony structures; consequently, the assessment of bone density, already subjective on conventional films, is even more difficult.
In practice, radiographs remain valuable for the initial diagnosis and day-to-day manage- ment of most diabetic patients’ foot pathology, and most primary care physicians rely heavily on radiographic findings, despite the known limitations of sensitivity and specificity for this technique.2 It is recommended that plain radiography should be the initial imaging procedure for suspected osteomyelitis in the diabetic patient.3 Further imaging investigations are then appropriate in the minority of patients where radiography and other clinical tests have not provided sufficient information for satisfactory clinical management.
The role of a radiologist in providing interpretation of the radiograph, along with suggestion of further imaging where appropriate, is valuable, but prior to this, the need for interested and motivated radiographers cannot be overemphasised.
Conventional radiography of the foot includes the routine dorsi-plantar and oblique views, but many more specialised views, such as lateral views, weight-bearing views, views of the toes, sesamoids (Figure 19.1), forefoot, subtalar joint, heel and ankle, can be performed.
These views, when appropriate, have the advantage of placing the region of greatest clinical interest into the most appropriate projection for the suspected pathology; they allow the radiographic exposure to be optimised to that specific region.
The smaller imaged volume from these views results in reduced X-ray dose and consequently also reduced X-ray scatter, which improves image quality.
Whilst originally developed in the absence of any alternative imaging technique, these views are still useful when appropriately performed.
Despite the value of an initial radiograph and of serial follow- up radiographs where appropriate, as described in this chapter, it is concerning that even hospital admission for infected diabetic foot ulceration may not precipitate a radiographic examination, with 33% of such patients not being radiographed in one study.4 Radiological manifestations of diabetes in the foot are due to a combination of trauma, neuropathy, infection and vascular disease, which between them affect all the tissues of the foot (bone, muscle, blood vessels, connective tissues and skin).
Neuropathy and vascular disease can be thought of as intrinsic complications of diabetes, whilst infection requires ingress of an extrinsic component (microorganisms), usually through a skin ulcer or perforation but also
Figure 19.1 Sesamoid view, demonstrating the sesamoid bones beneath the great toe metatarsal head
JWBK089-19 JWBK089-Boulton April 22, 2006 15:5 Char Count= 0
224 RADIOLOGY AND MAGNETIC RESONANCE IMAGING OF THE DIABETIC FOOT
Figure 19.2 (a) Patient with a warm, swollen foot, thought to be cellulutis, is found to be radiograph- ically normal. (b) Three weeks later, repeat radiography demonstrates a neuropathic LisFranc fracture dislocation
potentially by haematogenous spread.
Features of infection in the diabetic foot are almost invariably superimposed on pre-existing neuropathy and vascular disease.
Trauma to the diabetic foot may go unnoticed or its severity under-appreciated by the patient and carers.
Simple fractures are common in the feet of people with diabetes and diagnosis may be delayed.
In the author’s view, all diabetic patients presenting with trauma or unexplained swelling or deformity of the foot, however minor, should have foot radiography performed as part of their assessment, and if this is normal, it should be repeated within 2, 4 weeks (Figure 19.2).
RADIATION EXPOSURE The consequences of ionising radiation exposure are divided into deterministic and non- deterministic (stochastic or random) effects.
Deterministic effects (for example hair loss or skin erythema) always occur if a threshold dose of radiation is exceeded.
Stochastic effects may or may not occur but the probability of an effect occurring does increase with increasing radiation dose, the induction of malignant disease being the best known stochastic effect.
The threshold doses for significant deterministic effects are unlikely to be exceeded by foot radio- graphy, even when frequently repeated.
However, there is uncertainty about a threshold dose below which there is no risk of inducing malignant disease.
The ‘linear no threshold’ (LNT) hypothesis extrapolates the recognised and quantified risk of malignant disease induction at high radiation exposure back to a risk of zero at a dose of zero.
Although there is actually epi- demiological and experimental evidence suggesting a beneficial effect (the radiation hormesis hypothesis) from low doses of ionising radiation (doses below 200 millisieverts (mSv)), the LNT hypothesis indicates that any radiation exposure, however small, carries a finite risk of
JWBK089-19 JWBK089-Boulton April 22, 2006 15:5 Char Count= 0
NORMAL APPEARANCES 225
causing malignant disease.
Legislation for radiation protection is currently based on the LNT hypothesis.
In the United Kingdom, this legislation is embodied in the Ionising Radiation (Med- ical Exposures) Regulations , IR(ME)R.
This legislation requires all medical exposures to be justified, the justification to be provided by the person requesting the examination and based on his/her knowledge of the risks of radiation and benefit of the examination.
The justification has to be accepted by the person performing the examination, who can then authorise and perform it.
The typical effective radiation dose from foot radiography is less than 5 μSv, whilst the background radiation in the United Kingdom averages 2.4 mSv per annum.
The effective radi- ation dose from foot radiography is thus less than the background radiation received in a single day from living in the United Kingdom.
The radiation risk from foot radiography against which to balance the benefits of the examination should therefore be considered to be negligible.
NORMAL APPEARANCES Variation in the appearance of the foot and ankle on radiographs due to radiographic projections and exposure, the complexities of normal anatomy, variations of normal anatomy between in- dividuals and the sometimes gross abnormalities that may occur in diabetic feet all contribute to the challenge of interpretation of radiological images. Additional (accessory) ossicles are common in the foot and ankle (Figure 19.3), with over 20 recognised and named accessory