99mTc-Nanocolloid scintigraphy for assessing osteomyelitis in diabetic neuropathic feet

Clinical Radiology (1998) 53, 120-125

99mTc-Nanocolloid Scintigraphy for Assessing Osteomyelitis in Diabetic Neuropathic Feet D. REMEDIOS, J. VALABHJI*, R. OELBAUM*, P. SHARP* AND R. MITCHELL

Departments of Clinical Radiology and *Endocrinology, Northwick Park Hospital, Harrow, Middlesex, UK Distinguishing osteomyelitis from neuropathic osteoarthropathy in diabetic feet is a common and difficult clinical problem with no highly accurate descriminatory investigation. This study assesses the novel use of marrow scintigraphy and compares it with magnetic resonance imaging (MRI) for the diagnosis of osteomyelitis in neuropathic osteoarthropathic diabetic feet. Nine diabetic patients with chronic foot ulcers were prospectively assessed independently using 99mTc-nanocolloid scintigraphy and MRI. Those patients showing features of osteomyelitis underwent percutaneous bone biopsy or surgical ray excision for histological confirmation. Other patients were followed up clinically for a minimum of 6 months to exclude osteomyelitis. Marrow scintigraphy, in agreement with MRI, demonstrated all four cases of biopsy proven osteomyelitis and excluded three cases with neuropathic osteoarthropathy alone. One case of suspected osteomyelitis of the ankle on marrow scintigraphy, but not MRI, was not confirmed clinically. One case of suspected osteomyelitis on both imaging modalities was shown on biopsy to demonstrate changes of avascular necrosis but not osteomyelitis. In this study 99mTc-nanocolloid scintigraphy shows a sensitivity of 100% and specificity of 60%. An important false positive result is seen with avascular necrosis, both on marrow scintigraphy and on MRI. Although larger studies are needed to evaluate this technique, 99mTcnanocolloid marrow scintigraphy may be an alternative to MRI for assessing diabetic feet for osteomyelitis. Remedios, D., Valabhji, J., Oelbaum, R., Sharp, P. & Mitchell, R. (1998). Clinical Radiology 53, 120-125. 99rnTc-Nanocolloid Scintigraphy for Assessing Osteomyelitis in Diabetic Neuropathic Feet

Accepted for Publication 4 June 1996

Diabetic patients spend more time in hospital for foot problems than for any other complication [1]. Approximately 15% of diabetics will develop foot ulcers during their life-time and early detection of osteomyelitis is crucial to optimize treatment [2]. The differentiation of osteomyelitis from neuropathic osteoarthropathy is a major clinical problem as symptoms and signs are similar. These patients all present with hot, erythematous feet. The differentiation between these two entities on imaging is difficult, particularly in cases of rapidly progressive, noninfected neuro-osteoarthropathy [3,4].

PATIENTS AND METHODS Nine diabetic patients with peripheral neuropathy, chronic foot ulcers and clinical signs compatible with osteomyelitis were prospectively recruited. All had had plain radiography of the symptomatic foot and had already been imaged with 99mTc-methylene diphosphate (99mycMDP) bone scintigraphy. In this study, patients all underwent examination with 99mTc-nanocolloid (99mTc-NC) marrow scintigraphy and magnetic resonance imaging (MRI) of the affected foot. Three phase marrow scintigraphy was performed using 400 MBq of intravenous 99mTc-NC. Images were taken with Correspondence to: Dr D. Remedios, Department of Clinical Radiology, Northwick Park Hospital, Watford Road, Harrow, Middlesex HA1 3UJ, UK. 9 1998 The Royal College of Radiologists.

a large field-of-view gamma camera equipped with a low energy, high resolution collimator using a 20% window centred at 140 keV. Vascular phase images were taken over the first 3 min, the blood pool image starting at 5-10min and the static delayed image at 60-90 rain. Plantar images were taken in all studies with additional lateral images in those with mid- or hind-foot lesions. Studies were considered to be positive for osteomyelitis if static images showed significantly more focal activity than corresponding blood pool images. This focal activity was compared with 99mTcMDP scans for better anatomical resolution. Images were interpreted by two radiologists (RM and DR) with a consensus opinion. MRI of the symptomatic foot was performed using a 0.5 T superconductive magnet employing a head coil. T1weighted spin-echo sagittal, T2-weighted spin-echo or fast spin-echo sagittal, STIR or fast STIR (fat suppressed) sagittal and T2-weighted spin-echo or fast spin-echo axial images were acquired. Slice thickness was 5 mm. MRI was performed within 2 weeks of marrow scintigraphy in all but one case (3 weeks). Studies were considered to be positive for osteomyelitis if there was evidence of reduced marrow signal on T1 images and increased marrow signal on STIR or T2 images, particularly associated with adjacent deep ulceration [5,6]. Images were interpreted without reference to marrow scans by two radiologists (RM and DR) with a consensus opinion. In those cases with imaging criteria for osteomyelitis on both modalities who were not due to undergo surgery, fluoroscopically guided percutaneous bone biopsy using a

121

99MTc-NANOCOLLOID SCINTIGRAPHY FOR ASSESSING DIABETIC FEET

Table 1 - Patient data and summary of imaging findings

Case

Age~sex

Plain film

Bone scan

MR1

Nanocolloid scan

Histology

DB

65/M

Several toes hot

R 2MT oedema

PC

67/F

Hot ankle

Unremarkable

Hyperaemia, hot R 2 MT and L 3 MT Hyperaemia only

Osteomyelitis R2 and L 3 MT (ray excision) --*

MF

60/F

Hot ankle

Ankle effusion and sclerosis

Hyperaemia and hot ankle

--*

AF

25/M 70/M

Oedema in 1 MT and P Oedema in 2 MT and P

Hyperaemia hot 1 MTP Hyperaemia hot 2 MTP

DL

60/F

Oedema in 1-3 NIT heads

58/F

Hyperaemia hot i - 3 MTP joints Hyperaemia only

Osteomyelitis

KT RW

53/F

KW

59/M

Hot 1 MTP joint Hot 1 and 2 MTP joints Hot 1 - 4 MT heads and 1 TMT joint Hot ankle and midfoot Hot ankle and midfoot Hot ankle

Osteomyelitis

RJ

Lysis in R2 and L3 metatarsals Sclerotic ankle and rnidfoot Sclerotic, deformed ankle and midfoot Lysis 1 MTP joint Lysis 1 and 2 MTP joint Lysis 1 - 4 metatarsal heads Sclerotic ankle and midfoot Sclerotic ankle and midfoot Sclerotic ankle and midfoot

HyperaemJa hot ankle Hyperaemia only

Avascular necrosis --*

Ankle effusion only Tarsal oedema Sclerosis around ankle

Osteomyelitis

--*

MT, metatarsal; MTP, metatarso-phalangeal; TMT, tarso-metatarsal. *All patients without histology were followed up clinically for at least 6 months and showed no evidence of osteomyelitis.

route remote from the ulcer (usually dorsally) was performed with an 11 G Jamshidi trephine needle. Biopsy cores and surgical excision specimens were examined histologically and microbiological]y. A positive diagnosis for osteomyelitis was taken as either microbiological and/or histological evidence of bone infection.

RESULTS Nine patients were recruited with a mean age of 57 years and a sex ratio of 4:5 (M:F). Pedal ulcers were all on the plantar aspect, mostly related to the metatarsal heads and os calcis. Imaging findings for all patients are shown in Table 1. Plain films were abnormal with all showing hypertrophic changes. In addition, some also demonstrated lyric changes in the fore-foot. The 99mTc-MDP bone scans showed focal osseous uptake for all nine patients and did not descriminate between osteomyelitis and neuropathic osteoarthropathy. Five patients had histological correlation, four following percutaneous bone biopsy and one after surgical ray excision. Osteomyelitis was proven in four cases. Diagnosis was made only on histological examination as all these cases were already undergoing antimicrobial therapy at the time of biopsy. The four patients who did not undergo biopsy or operation were followed up clinically for a minimum of 6 months and showed no complications that suggested osteomyelitis. 9 9 m T c - N C m a r r o w scintigraphy like MRI showed all four cases of proven osteomyelitis (Fig. 1). Both techniques also excluded osteomyelitis in three cases (Fig. 2). One case of suspected osteomyelitis of the ankle on 99mTc-NC but not MRI was not confirmed clinically. Plain films in this case showed gross neuropathic changes with disorganization of the ankle and subtalar joints. One case of suspected osteomyelitis of the mid-foot on both marrow and MRI scans showed changes of avascutar necrosis but not osteomyelitis on biopsy (Fig. 3). Only one marrow scan had to be repeated due to patient movement. Several MRI sequences required repeating for the same reason. 9 1998 The Royal College of Radiologists, Clinical Radiology, 53, 120-125.

DISCUSSION To date, no highly descriminatory investigation has been found to distinguish between osteomyelitis and neuropathic osteoarthropathy. MRI is favoured by some [5,7-9] but lacks specificity (Table 2), is expensive and is not ubiquitous. In clinically unsuspected osteomyelitis, MRI also lacks sensitivity [10]. A number of MRI abnormalities have been described in diabetic feet but none are truly specific [6]. Nuclear medicine has been widely used to image osteomyelitis [2]. The variety9 of radiopharmaceuticals is extensive and includes (1) 9mTc-labelled: phosphates, Table 2 - Estimated sensitivity and specificity of imaging techniques used for evaluating diabetic pedal osteomyelitis. Data are from this study and other reported series (adapted from Gold et al. [1])

Study

Number of patients

Sensitvity (%)

Specificity (%)

13 35 77 14 51 35 16 34

75 100 100 80 79 89 100 93

89 83 38 54 78 69 67 83

Yuh et al. [9] Wang et al. [16] Weinstein et al. [8] Newman et al. [10] Morrison et al. [30] Reported series

24 50 32 16 27 9

100 98 100 29 82 100

89 81 81 78 80 80

111In-human IgG scan Oyen et al. [31]

16

79

84

9

100

60

111]n leucocyte scans Maurer et al. [24] Schauwecker et al. [25] Keenan et al. [26] Seabold et al. [3] Larcos et al. [27] Newman et al. [28] Newman et al. [10] Crerand et al. [29] MR1

99mTc-nanocolloid scan

Reported series

122

CLINICAL RADIOLOGY

9<3H 'PC MDP

99M TC HAHOCOLL

(b)

(a)

(c)

(d)

Fig. 1 - Biopsy proven osteomyelitis in the right first metatarsal and proximal phalanx of the great toe in a 25-year-old man with a 10-year history of diabetes and a chronic ulcer under the first metatarsal head. The plain radiograph (a) shows erosions and soft tissue swelling around the first metatarso-phalangeal joint (arrowheads). (b) 99mTCMDP bone and 99mTc-nanocolloid blood pool and delayed scans showing time dependent accumulation of tracer in the first metatarsal and phalanx. (c) Tl-weighted and (d) STIR (fat suppressed) MRI scans showing low T 1 and high STIR signal in the first metatarsal and proximal phalanx (solid arrows). Note that the normal fatty marrow with high T1 and low STIR signal is seen in the cuneiform bone proximally (open arrows).

nanocolloid, non-specific polyclonal human immunuglobulin G (IgG), hydroxymethylpropyleneamine oxime (HMPAO) tagged leucocytes, colloid ingested leucocytes, anionic phospholipid liposomes and antigranulocyte antibodies; (2) rll In labelled: leucocytes, human IgG, 111In chloride and (3) 67Ga citrate. The most widely used 99mTcMDP bone agent is sensitive but lacks specificity [1,7]. More sophisticated labelled leucocyte studies, although

more specific, are time consuming, expensive and lack resolution [1,2]. Marrow scintigraphy with 99mTc-nanocolloid has been shown to be of particular use in bone and joint infections [11-14]. It is cheaper than MRI (approximately half the price) and can be performed in any nuclear medicine department. It has been reported to demonstrate skeletal sepsis with a sensitivity of 87% and specificity of 93% [12]. 9 1998 The Royal College of Radiologists, Clinical Radiology, 53, 120-125.

123

99MTc-NANOCOLLOID SCINTIGRAPHY FOR ASSESSING DIABETIC FEET

L~

(a)

.....

(b)

, j

(c)

(d)

Fig. 2 - A 59-year-old diabetic man with long-standing left pedal ulcers and neuropathic osteoarthropathy. The 99n~Tc-MDP bone scan (a) shows time dependent accumulation of tracer in the mid tarsal joints bilaterally. The 99mTc- nanocolloid scan (b) shows hyperaemia in the mid-foot on the blood pool phase but no further increase in activity on the delayed images, i.e. a negative scan. (c) T1- and (d) T2-weighted MRI scans showing gross disorganization of the midtarsal joint with flattening of the arch, destruction and disappearance of the navicular and subhixation of the anterior talus. Sclerotic changes are indicated by the hypointense trabeculations in the talus and cuneiform on both sequences (arrowheads). The underlying marrow signal is unchanged, hence there is no evidence for osteomyelitis in this neuroarthropathic foot, Clinical follow-up showed no evidence for osteomyelitis.

Nanocolloid is derived from microaggregated human serum albumin with a particulate size of < 80nm. Following intravenous injection, 15% to 20% of the dose is taken up by the bone marrow, the rest being distributed mainly to the liver and spleen [15]. The postulated mechanism of action is extravasation of tracer through the basement membrane of capillaries, particularly those which are damaged. Particles are phagocytosed or adsorbed by macrophages and granulocytes [11,15]. The advantages over labelled leucocyte scans are fourfold: lower dose (estimated by deSchrijer et al. [11] as one ninth of that for 111In leucocyte scintigraphy), shorter examination time (1-2 h compared with 24-48 h), a simple one step labelling procedure (leucocyte labelling is very demanding technically and timewise) and the optimal emission characteristics associated with 99mTc. In our small study, the sensitivity of 100% for both marrow scintigraphy and MRI is probably an overestimate. The specificity for marrow scintigraphy of 60% is comparable with that for labelled leucocyte scans (Table 2) but less than the 80% for MRI in our study and 81% to 89% in others [8,9,16]. Refinements in technique may well improve specificity. Two problems encountered in our study were, firstly, the long imaging time to acquire sufficient photon counts and, secondly, the high soft tissue uptake of 9 9 m T c NC associated with the hypervascularity of the neuropathy 9 1998 The Royal College of Radiologists, Clinical Radiology, 53, 120-125.

and frequent cellulitis, combined with venous stasis. This high background activity was often seen in the mid- and hind-foot where neuropathic changes are common in contradistinction to osteomyelitis, which, in this study and like others [8], is frequently found in the fore-foot. Static images at 60 rain in these patients clearly showed residual soft tissue activity. Later in the series, we performed delayed static images at 90 min rather than 60 min to improve the marrow to soft tissue ratio. Further delay to 120rain with a higher injected activity (up to 700 MBq) may be beneficial. The high soft tissue uptake may be the reason for one false positive marrow scan. The other false positive was caused by biopsy proven avascular necrosis (osteonecrosis), a common complication in diabetic arteriopathic and neuropathic feet. Osteonecrosis is a recognized cause for 99mTc-NC uptake and, indeed, is an indication for this investigation [15]. Bone marrow oedema detected by MRI may be inflammatory or osteonecotic in nature as well being due to a variety of other causes, including neoplasm, trauma, transient osteoporosis, transient bone marrow oedema or even altered biomechanics [17-20]. These false positives will require bone biopsy for further elucidation. We, like others [1,21], favour the use of image guided bone biopsy following a route remote from the ulcer. In addition, like others [22], we have found histological confirmation of infection essential, and certainly more

124

CLINICAL RADIOLOGY

m.t~F,~~~2 . ~: ~. . . . . ~ .....

(a)

(c)

(b)

Fig. 3 - A 53-year-old diabetic woman with a right mid-foot chronic ulcer. (a) 99mTc-nanocolloid scan showing accumulation of tracer in the right mid-foot suggesting osteomyelitis. (b) T1 and (c) STIR MRI scans showing the talar and navicular marrow return low T1 and high STIR signals, indicating oedema and suggesting osteomyelitis (arrowheads). Normal fatty marrow is seen in the calcaneus (open arrows). Note the disorganized midtarsal joint typical of a neuro-arthropathic foot and also the oedema in the plantar fascia and muscles related to the ulcer. Bone biopsy showed avascular necrosis but no infection in this false positive case.

useful than microbiology, beause most patients would have been on antibiotics for soft tissue sepsis prior to bone biopsy. In addition, bone infection is likely to be chronic [23] and often low grade. Frequently, the infective agent may have been isolated from other cultures, e.g. wound swabs or blood cultures. Estes and Pomposelli [32] reported that 56% of diabetic feet referred for vascular surgery were infected with l 1% showing osteomyelitis. The distinction of osteomyelitis from uncomplicated neuropathic osteoarthropathy is critical as the former requires a prolonged course of antimicrobial therapy, often for several months, with surgery playing an important role [33], whereas the latter is often treated in a total contact plaster cast for several months [34]. 99mTc-NC scintigraphy may be an alternative to MRI in the assessment of osteomyelitis in diabetic neuropathic feet. This relatively cheap technique may be useful for patients unable to undergo MRI either because of contra-indications, physical limitations or non-availability. We would welcome further studies to evaluate its use.

Acknowledgements.

We are grateful to Lister Bestcare and Novo Nordisk for supporting this study. We thank Miss Jeanette Ryder in the Nuclear Medicine section and the MRI staff of Lister Bestcare, Northwick Park Hospital for excellent technical work.

REFERENCES 1 Gold RH, Tong DJF, Crim JR, Seeger LL. Imaging the diabetic foot. Skeletal Radiology 1995;24:563-571. 2 Rosenthall L. Radionuclide investigation of osteomyelitis. Current Opinion in Radiology 1992;4:62-69.

3 Seabold JE, Flickinger FW, Kao SCS et al. Indium-Ill-leucocyte/ technetium-99m-MDP bone and magnetic resonance imaging: difficulty of diagnosing osteomyelitis in patients with neuropathic osteoarthropathy. Journal of Nuclear Medicine 1990;31:549-556. 4 Vestring T, Fiedler R, Greitemann B, Sciuk J, Peters PE. The diabetic foot. Radiologe 1995;35:447-455. 5 Beltran J, Campanini DS, Knight C, McCalla M. The diabetic foot: magnetic resonance imaging evaluation. Skeletal Radiology 1990:19:37-41. 6 Moore DE, Yuh WTC, Kathol MH, E1-Khoury GY, Corson JD. Abnormalities of the foot in patients with diabetes mellitus: findings on MR imaging. American Journal of Roentgenology 1991;157:813816. 7 Brower AC. What is the preferred method for diagnosing osteomyelitis in the foot of a patient with diabetes? American Journal of Roentgenology 1994;163:471-472. 8 Weinstein D, Wang A, Chambers R, Stuart CA, Motz HA. Evaluation of magnetic resonance imaging in the diagnosis of osteomyelitis in diabetic foot infections. Foot and Ankle 1993;14:18-22. 9 Yuh WTC, Corson JD, Baraniewski HM et al. Osteomyelitis of the foot in diabetic patients: evaluation with plain film, 99mTc-MDP bone scintigraphy, and MR imaging. American Journal of Roentgenology 1989;152:795-800. 10 Newman LG, Waller J, Palestro CJ et al. Leukocyte scanning with i11in is superior to magnetic resonance imaging in diagnosis of clinically unsuspected osteomyelitis in diabetic foot ulcers. Diabetes Care 1992;15:1527-1530. 11 De Schrijver M, Streule K, Senekowitsch R, Fridrich R. Scintigraphy of inflammation with nanometer-sized colloidal tracers. Nuclear Medicine Communications 1987;8:895-908. 12 Streule K, De Schrijver M, Fridfich R. 99Tcm-labelled HSA-nanocolloid versus l llln oxine-labetled granulocytes in detecting skeletal septic process. Nuclear Medicine Communications 1988;9:59-67. 13 Flivik G, Sloth M, Rydholm U, Herrlin K, Lidgren L. Technetium-99mnanocolloid scintigraphy in orthopedic infections: a comparison with indium-I 1 I-labeled leukocytes. Journal of Nuclear Medicine 1993;34:1646-1650. 14 Vome M, Lantto T, Paakkinen S, Salt S, Soini I. Clinical comparison of 9 1998 The Royal College of Radiologists, Clinical Radiology, 53, 120-125.

99MTc-NANOCOLLOID SCINTIGRAPHY FOR ASSESSINGDIABETIC FEET

15 16 17

18

19 20 21 22

23 24

25

99Tcrn-HMPAO labeUed leucocytes and 99Tcm-nanocolloid in the detection of inflammation. Acta Radiologica 1989;30:633-637. Reske SN. Recent advances in bone marrow scanning. European Journal of Nuclear Medicine 1991:18:203-221. Wang A, Weinstein D, Greenfield L et al. MRI and diabetic foot infections. Magnetic Resonance Imaging 1990;8:805-809. Vande Berg BE, Malghem JJ, Labalsse MA, Noel HM, Maldague BE. MR imaging of avascular necrosis and transient marrow edema of the femoral head. RadioGraphics 1993;13:501-520. Kroon HM, Bloem JL, Holscher HC et aI. MR imaging of edema accompanying benign and malignant bone tumors. Skeletal Radiology 1994;23:261-269. Hayes CW, Conway WF, Daniel WW. MR imaging of bone marrow edema pattern: transient osteoporosis, transient bone marrow edema syndrome or osteonecrosis. RadioGraphics 1993;13:1001-1011. Schweitzer ME, White LM. Does altered biomechanics cause marrow edema? Radiology 1996;198:851-853. Newman LG. Imaging techniques in the diabetic foot. Clinics in Podiatric Medicine and Surgery 1995;12:75-86. White LM, Schweitzer ME, Deely DM, Gannon F. Study of osteomyelitis: utility of combined histologic and microbiologic evaluation of percutaneous biopsy samples. Radiology 1995;197:840-842. Milgram JW. Osteomyelitis in the foot and ankle associated with diabetes mellitus. Clinical Orthopaedics and Related Research 1993;296:50-57. Maurer AH, Millmond SH, Knight LC et al. Infection in diabetic osteoarthropathy: use of indium-labeled leucocytes for diagnosis. Radiology 1986;161:221-225. Schauwecker DS, Park HM, Burr RW, Mock BH, Wellman HN.

9 1998 The Royal Collegeof Radiologists,ClinicalRadiology, 53, 120-125.

26 27

28

29

30

31

32 33 34

125

Combined bone scintigraphy and indium-111 leukocyte scans in neuropathic foot disease. Journal of Nuclear Medicine 1988;29:1651-1655. Keenan AM, Tindel NL, Alavi A. Diagnosis of pedal osteomyelitis in diabetic patients using current scintigraphic techniques. Archives of Internal Medicine 1989;149:2262-2266. Larcos G, Brown ML, Sutton RT. Diagnosis of osteomyelitis of the foot in diabetic patients: value of u 1in_leukocyte scintigraphy. American Journal of Roentgenology 1991;157:527-531. Newman LG, Waller J, Palestro CJ et al. Unsuspected osteomyelitis and diabetic foot ulcers. Diagnosis and monitoring by leukocyte scanning with indium l n l l l oxyquinoline. Journal of the American Medical Association 1991;266:1246-1251. Crerand S, Dolan M, Liang P, Bird M, Smith ML, Klenerman L. Diagnosis of osteomyelitis in neuropathic foot ulcers. Journal of Bone and Joint Surgery 1996;78-B:51-55. Morrison WB, Schweitzer ME, Wapner KL, Hecht PJ, Gannon FH, Behm WR. Osteomyelitis in feet of diabetics: clinical accuracy, surgical utility and cost-effectiveness of MR imaging. Radiology 1995;196:557-564. Oyen WJG, Netten PM, Lernmens JAM et al. Evaluation of infectious diabetic foot complications with indium-11 I-labeled human nonspecific immunnglobulin G. Journal of Nuclear Medicine 1992;33:13301336. Estes JM, Pomposelli FB Jr. Lower extremity arterial reconstruction in patients with diabetes mellitus. Diabetic Medicine 1996;13:$43-$57. Van der Meer JWM, Koopmans PP, Lutterman JA. Antibiotic therapy in diabetic foot infection. Diabetic Medicine 1996;13:$48-$51. Klenerman L. The Charcot joint in diabetes. Diabetic Medicine 1996;13:$52-$54.