Diagnostic Performance of Technetium-99m Leucocyte SPECT/CT in Suspected Diabetic Foot Osteomyelitis – Comparison with Bone Biopsy

INTRODUCTION

Diabetes mellitus is a widely prevalent non-communicable disease in both developed and developing countries. It is one of the fastest-growing global health emergencies, and the global prevalence of Diabetes mellitus is approximately 11.1%.^[1] Around 18.6 million people are affected by a diabetic foot ulcer across the globe every year,^[2] with 19% to 34% of the diabetic population developing a foot ulcer during their lifetime.^[3] Among those with a foot ulcer, about one-third develop osteomyelitis over time. Peripheral vascular disease and foot bone osteomyelitis are significant predictive factors for limb surgeries, either salvage procedures or amputation, in diabetic patients.^[4,5] The incidence of osteomyelitis in patients with diabetic foot ulcers ranges between 10–15% in moderate infections and approximately 50% in severe infections.^[6,7] Osteomyelitis in a diabetic foot ulcer is the most common cause of non-traumatic amputation.^[8] Early prompt diagnosis of diabetic foot osteomyelitis helps in deciding appropriate management and hence better prognosis.

Clinical examination and inflammatory markers, such as the erythrocyte sedimentation rate and C-reactive protein, are nonspecific and not very reliable in establishing the diagnosis of osteomyelitis. Imaging can guide far better in diagnosing osteomyelitis of the underlying bone. Radiological investigations such as plain radiograph, ultrasonography, computed tomography, and magnetic resonance imaging are widely available and used more frequently in clinical practice. Nuclear medicine imaging, such as bone scan, bone marrow scan, Fluorodeoxyglucose positron emission tomography (FDGPET) imaging, and labelled leucocyte scintigraphy, is relatively specific but, compared to other radiologic imaging modalities, it is used less commonly.^[9,10] Acute osteomyelitis changes are not visualised on plain radiographs, as it takes a few weeks for radiological changes to occur. Neuroarthropathic changes in chronic osteomyelitis may cause difficulty in interpretation. Although there are certain patterns of involvement which are useful in differentiating osteomyelitis from neuroarthropathy, reliability is still questionable.^[11] Magnetic resonance imaging is the imaging standard in the evaluation of diabetic foot osteomyelitis due to its superior soft tissue resolution. The high sensitivity and specificity of magnetic resonance imaging (MRI) are due to its capability to delineate soft tissue and skeletal abnormalities accurately. The bone marrow changes in T1 images with low signal and its continuity with the sinus tract, whenever present, help in the diagnosis of osteomyelitis. Also, MRI helps to precisely demarcate the extent of osteomyelitis, thereby guiding the surgery.^[12,13] Bone scan is reasonably sensitive and useful in diabetic foot osteomyelitis. FDG PET scans provide better resolution compared to bone scans and labelled leucocyte scintigraphy.^[14] Leucocytes move to the site of infection and inflammation by diapedesis and chemotaxis. This is because injury causes vasodilation and increased permeability of vessels. Local chemical mediators like chemokines and chemotaxins are released and attract leucocytes.^[15] Leucocytes migrate from the intravascular space to the interstitial space to exert their immune function. This property of leucocytes is the basis for labelled leucocyte scintigraphy.^[16] Amongst the large gamut of above-mentioned imaging options to identify bone infection in diabetic foot ulcers, it is pivotal to recognise the most helpful modality in this setting.

MATERIAL AND METHODS

The prospective study was carried out between October 2015 and June 2017 in a tertiary care hospital in South India and approved by the institute's ethics committee (JIP/IEC/2015/16/596). Patients who presented to the Department of Surgery, Medicine, Diabetic or Endocrinology outpatient departments with a foot ulcer were clinically evaluated.

The inclusion criteria required that the foot ulcer be present for at least two weeks, with a surface area of 2 cm² or greater and a depth exceeding 3 mm.Additionally, the clinical evaluation had to demonstrate either a positive “probe-to-bone” test or primary bone exposure within the ulcer. Patients with a prior known diagnosis of osteomyelitis, life-threatening systemic infections, or gangrenous ulcers were excluded from the study. Furthermore, a technical requirement for the labelling process was a sufficient white blood cell count; therefore, patients with circulating leucocyte levels of less than 2 × 10^6/mL were also excluded. The study participants underwent random blood sugar, complete hemogram, and radiograph of the affected foot, followed by labelled leucocyte scintigraphy. Since the aim of the study was to compare the labelled leucocyte scintigraphy with bone biopsy and the study involved a significant number of investigations, a conscious decision was taken to exclude MRI for the study participants.

Scan procedure

Scans were acquired on hybrid SPECT-CT system (Symbia T6: Siemens, Erlangen, Germany). After a waiting period of 30 minutes post-injection, patients were taken up for imaging. Spot images of the thorax in anterior and posterior views were acquired for an in vivo quality check of leucocyte labelling. After 3 to 4 hours, spot images of the region of interest were acquired for a duration of 10 minutes. SPECT-CT of the region was also acquired subsequently. Delayed images of the region were also acquired at around 24 hours for a duration of 10 minutes. We observed three patterns of tracer distribution in the lungs at 30 minutes after injection: uniform distribution in both lungs, uniform washout from both lungs and a few foci of uptake, probably due to leucocyte clumping. The representative images are as depicted in Fig 1.

Close

Fig 1: Pattern of labelled leucocyte distribution at 30 minutes post administration thorax images: (a) Physiological distribution in spleen more than liver with uniform washout of tracer from bilateral lungs; (b) Diffuse tracer distribution in the bilateral lungs; (c) Uniform washout of tracer from bilateral lungs with foci of tracer uptake in right lung suggestive of leucocyte clumping

Export to PPT

RESULTS

Median and interquartile range for patient demographics and ulcer characteristics were calculated. The images acquired at 4 hours, along with SPECT-CT and at 24 hours, were interpreted for focus of infection/inflammation and compared with the histopathologic examination of the underlying bone specimen. The sensitivity and specificity of the combined labelled leucocyte scintigraphy SPECT-CT were calculated with histopathology as the gold standard. Positive leucocyte scintigraphy is defined as the presence of a focal uptake in both the planar and SPECT-CT images at 4 hours and persistence of uptake in the 24-hour delayed image.

A total of 30 patients, who fulfilled the inclusion and exclusion criteria, were included in the study, of which 11 patients were female, and 19 were male; these patients had a total of 35 ulcers. All patients were diagnosed to have type 2 Diabetes mellitus, out of which nine were on only oral hypoglycaemic agents, whereas 21 of them were on both insulin injection and oral hypoglycaemic agents. Patient demographics, ulcer characteristics and biopsy results are depicted in Table 1. The median injected dose of labelled Tc-99m HMPAO leucocytes was 10.48 mCi (IQR: 7, 14.5). The median labelling efficiency in our study was 42.3% (IQR- 27.15, 57.24).

The median time of acquisition of delayed images was 4 hours (IQR- 3.5, 4) and 23.75 hours (IQR- 22, 24) after injection. The 30-minute scans of the thorax of four patients showed multiple foci of increased tracer distribution in bilateral lung fields. This could be attributed to the formation of white blood cell clumps during the labelling procedure. However, the images of these patients were still good enough for interpretation. There was significant patient motion resulting in image misregistration encountered in SPECT-CT images of five patients. Registration could be aligned satisfactorily in three patients and failed in two patients.

Among the 35 ulcers, 22 ulcers showed tracer uptake at the 4-hour image, and 19 ulcers showed tracer uptake at the 24-hour image. SPECT showed tracer uptake in 26 of the 35 ulcers. 4 ulcers that did not show any tracer uptake in 4-hour planar images were found to be positive on SPECT. SPECT was positive in a greater number of patients than planar images. However, none of the lesions noted only on SPECT were localised in bone.

After reviewing the delayed planar images and SPECT-CT, it was found that 4 ulcers were categorised as having underlying osteomyelitis [Fig 2]. The rest of the 31 ulcers either did not have any infection or had only soft tissue infection. Five of the ulcers did not show tracer uptake in any of the delayed images, as well as on SPECT-CT, suggesting that there was no infection. The remaining 26 ulcers had surrounding soft tissue infection.

Close

Fig 2: Planar images at (a) 4 hours and (b) 24 hours of bilateral feet show a focal increase in tracer uptake along the medial aspect of the right forefoot; (c) SPECT-CT at 4 hours shows the lesion localised to the head of the first metatarsal bone. SPECT-CT: Single photon emission computed tomography - computed tomography

Export to PPT

A biopsy could be performed in 17 ulcers. Biopsy specimens were sent for both microbiologic examination (including culture & sensitivity) and histologic examination based on HOES scoring (Histopathological Osteomyelitis Evaluation Score).^[17] Out of 17 bone biopsy specimens sent for culture and sensitivity, culture was positive in 16 ulcers and was sterile in one ulcer. Staphylococcus aureus was the most common organism encountered. The other organisms were Pseudomonas, Enterococcus faecalis, Escherichia coli, Klebsiella pneumoniae, Proteus mirabilis, Coagulase-negative staphylococcus, and Acinetobacter baumanii.

The number of true positives and true negatives was 5 and 12, respectively, based on histopathological examination findings for confirmation of osteomyelitis. The sensitivity of our study in diagnosing osteomyelitis was 60%. We were able to rule out osteomyelitis in 12 of 17 ulcers that were biopsied. The specificity of labelled leucocyte scintigraphy and SPECTCT was 100%.

DISCUSSION

The first line imaging in clinically suspected diabetic foot infections is plain radiography owing to its wide availability. The classical radiograph triad of osteolysis, periosteal reaction, and bone destruction occurs at a later stage of the disease process, limiting its use in the early diagnosis. Moreover, the inherent complications of diabetes, such as vascular insufficiency, Charcot’s neuroarthropathy, and soft tissue oedema, make it further difficult to diagnose with certainty. In addition, it is arduous to precisely localise the small bone involvement in the foot on the two-dimensional radiograph. The sensitivity and specificity of plain radiographs is 76%. However, the combination of probe-to-bone test and plain radiography has improved sensitivity (94%) and specificity (83%) in the diagnosis of osteomyelitis, but is not very helpful in the early stage.^[18]

Magnetic resonance imaging of the foot is the radiologic gold standard for the diagnosis of diabetic foot osteomyelitis, thanks to the excellent soft tissue contrast. The marrow oedema, soft tissue infection and the sinus tracts can be characterised better. The pooled sensitivity and specificity of MRI in diagnosing diabetic foot osteomyelitis are 96% and 84%, respectively.^[10] However, accurate demarcation of sterile Charcot’s neuroarthropathy and osteomyelitis is quite challenging.

Triple-phase skeletal scintigraphy and fluorodeoxyglucose PET-CT have reasonable availability thanks to the development of Nuclear Medicine facilities. Bone scan is highly sensitive, with a specificity of about 95%, but the specificity is less than 50%, particularly when the disease has co-existing complications like neuroarthropathy.^[19] FDG is a non-specific glucose metabolism marker; nevertheless, thanks to superior resolution and hybrid nature, CT provides comprehensive morphologic and functional information with a sensitivity of about 95% and better specificity than MRI, about 92%.^[20] One of the significant practical challenges with FDG PET is fasting glucose control, which is difficult in diabetic patients. Labelled leucocyte scintigraphy is the most specific imaging modality and the Nuclear Medicine gold standard for localising infective pathologies. The pooled sensitivity and specificity of labelled leucocyte scintigraphy are 91% and 92%, respectively.^[14]

Biopsy of the suspected bone is the gold standard in diagnosing osteomyelitis of the underlying bone and was performed in 17 ulcers (48.5%). One patient, though originally consented for the procedure, did not turn up for the biopsy, and the procedure was abandoned in another patient due to difficulty in obtaining the biopsy. In the other ulcers, biopsy was omitted due to non-involvement of bone as shown on the leucocyte scan. The sensitivity and specificity of combined labelled leucocyte scintigraphy with SPECT/CT in diagnosing diabetic foot osteomyelitis in our study was 60% and 100%, respectively.

Przybylski et al. did a retrospective study in 14 patients with diabetic foot ulcers.^[21] They found a sensitivity of 87.5% for combined labelled leucocyte scintigraphy with SPECT-CT. The sample size was only 14, but they had more cases of osteomyelitis in the study group, and this could be the reason for increased sensitivity. In a study by Filippi et al., 19 ulcers that had clinical suspicion of osteomyelitis were analysed.^[22]They found 8 ulcers to have underlying osteomyelitis, all of which were detected by labelled leucocyte scintigraphy and SPECT-CT. Though this was a prospective study, a biopsy was not performed in all the ulcers to prove the presence of osteomyelitis.

Most of our patients presented at late stages of ulcer. The median duration of ulcers was 3 months. A leucocyte scan is known to be good for acute infection, since granulocytes are the predominant white blood cells in areas of acute inflammation. Chronic ulcers may not have significant granulocyte accumulation and hence are prone to show low activity on the leucocyte scan. One of the patients showed low-grade uptake on planar scan, but did not show uptake on SPECT/CT. However, biopsy of the bone revealed features of chronic osteomyelitis. One observation was that the leucocyte scan helped in identifying the site of bone to be biopsied when there was uptake on the scan. Without a guide, the surgeon could miss the actual site of infection.

Labelling of leucocytes with radionuclides is not a new procedure, but is less commonly practised. The major hindrances to its wide usage are the lack of technical skills, inadequate facilities for labelling, cost, time required for labelling, and lack of awareness. There were some minor problems during labelling, for example, difficulty and choice of method to separate plasma from red blood cells (use of hetastarch), and red blood cell contamination. Contamination of red cells occurred due to incomplete sedimentation of a small proportion of red blood cells, which remained suspended in plasma. This could result in high background and poor image quality. But the amount of red blood cells in plasma was not significant and was noted on visual analysis in only five patients. It did not result in any significant degradation of image quality.

Median labelling efficiency was 42.3% (IQR 27.5,57.24), and the median injected activity was 10.48 mCi (IQR- 7, 14.5). Overall image quality was good in all the patients, and additional time was required to acquire an adequate number of counts when a low amount of activity was injected. Clumps of cells were noted on lung images acquired at 30 minutes in 4 patients. However, no significant degradation of image quality was noted, and hence, the scans could be interpreted.

Patient motion during the SPECT procedure was an issue in 5 patients, and could not be corrected in two of them. The acquisition time duration for SPECT is around 20 minutes, and could be one of the possible reasons for patient motion despite attempted immobilisation of the feet. The predominant site of patient motion was the digits. Such misregistration could not be completely realigned using a software algorithm in two patients.

Filippi et al. studied the usefulness of SPECT-CT and its added value to leucocyte scintigraphy.^[22] They found that SPECT-CT was not helpful when planar scans showed no increased uptake. In contrast to their conclusion, we found that SPECT could detect focal uptake in four ulcers when planar images did not show any significantly increased trace uptake due to high background activity at 4 hours. SPECT images have improved contrast and hence can detect lesions in a high background.

Co-registered CT played the crucial role in ruling out or identifying bone involvement in cases of positive planar images. This was possible from the information on the accurate location of uptake provided by CT. Bony changes seen on CT were instrumental in confirming bone involvement. Labelled leucocyte SPECT-CT has excellent specificity in the diagnosis of osteomyelitis. It offers additional diagnostic value over MRI in selected complex clinical situations such as Charcot’s neuroarthropathy and/or metallic implants.

Bone biopsy is not routinely practised in day-to-day surgical practice in many resource-stretched institutions. It is invasive and poses a risk of infection. Also, bone biopsy is not frequently performed for benign indications. As a result, the technique of obtaining a needle biopsy of bone was less familiar among surgeons. In diabetic patients, the biopsy procedure itself causes injury to the bone that may take a longer time to heal. It is very difficult to biopsy a foot bone owing to its small size and thick cortex. Bone biopsy of foot bones warrants technical expertise and skill. The specimen yield was inadequate at times. In patients whose scan was negative for osteomyelitis, identifying the bone for biopsy was difficult. Thus, good skills for obtaining a bone biopsy is a deciding factor to procure an adequate specimen for analysis. If the biopsy is not performed meticulously, the reliability of the histopathology report in diagnosing osteomyelitis is in question. Though it is still considered a reference standard, Meyr et al. studied 39 histologic specimens with four pathologists who read it independently. He reported that there were significant differences of opinion among the pathologists in the histologic diagnosis of osteomyelitis, with only 33.33% agreement.^[23] The interobserver variation is so significant that the author concludes histopathology can no longer be considered as the gold standard for diagnosing diabetic foot osteomyelitis. We found that obtaining a biopsy was a far more strenuous procedure compared to labelling of leucocytes, and the histopathology report depends on how accurately a biopsy is taken.

LIMITATIONS

The sample size is not very large. The inclusion and exclusion criteria that we had chosen were very stringent, and this excluded most patients who presented in the advanced phase and had obvious evidence of osteomyelitis. However, leucocyte scintigraphy would have been redundant in such a scenario. A smaller sample size could have impacted the calculated sensitivity and specificity. One of the reasons for low sensitivity could be due to the late stages of ulcer or chronic ulcer. Another contributing factor could be the low labelling efficiency in some patients, although there was no discernible degradation of image quality.

CONCLUSION

Combined labelled leucocyte scintigraphy and SPECT-CT is helpful in differentiating soft tissue from bone infection in cases of infected foot ulcers with suspected osteomyelitis. It is highly specific in this setting and is feasible, particularly in patients with complex neuroarthropathic changes and ambiguous results from conventional radiological imaging. It is less invasive compared to a biopsy and may help avoid unnecessary bone biopsies in many cases. In case of scan-positive osteomyelitis, images can direct the site of biopsy.