9 The Scintigraphic Diagnosis of Osteomyelitis Donald S. Schauwecker1 Review Article Osteomyelitis is a serious health problem that results in multi- pie limb amputations annually. This article reviews the current scintigraphic procedures used in the diagnosis of osteomyelitis and discusses some of the newer radiopharmaceuticals now being developed. The goal is to understand the strengths and weaknesses of each method so that the procedure most effective for specific clinical settings can be selected. In general, the three- phase bone scan is the procedure of choice if the suspected osteomyelitis is not superimposed on another disease that causes increased bone remodeling (i.e, findings on the radio- graph are normal). If the suspected osteomyelitis is superim- posed on a disease that causes increased bone remodeling, the combined 1111n-labeled Ieukocyte-”Tc bone scan is the proce- dure of choice in the non-marrow-containing skeleton and the 1111n-labeled leukocyte and mTc bone marrow scans are the procedures of choice in the marrow-containing skeleton. Use of scintigraphy in the diagnosis of osteomyelitis was established in 1975 when three landmark articles [1-3] ap- peared. First, Duszynski et al. [1] discovered that a bone scan could be used to diagnose osteomyelitis several days before radiographs showed any abnormalities. Second, Gilday et al. [2] added the blood-pool image to the bone image as a forerunner of the current three-phase bone scan. Finally, Deysine et al. [3] proposed the use of gallium-67 for the study of chronic and postoperative osteomyelitis. Three-Phase Bone Scan Localization of 99mTc-diphosphonate in bone is related to both osteoblastic activity and skeletal vascularity. Commonly, the bone scan will show abnormal uptake of the radionuclide 1 0-1 4 days before loss of bone mineral is great enough to be seen on plain radiographs [4]. For example, 17 of 19 sites of osteomyelitis in children that were imaged within 3 days after the onset of symptoms showed abnormal uptake on the bone scan, whereas only one abnormal site was seen on the radiograph [5]. In fact, if antibiotic therapy is begun early enough to arrest the loss of bone mineral, the characteristic radiographic findings may never develop [6]. The three-phase bone scan is the routine nuclear medicine procedure for diagnosis of osteomyelitis. The first phase, the nuclear angiogram or flow phase, consists of serial 2- to 5- sec images of the area of suspected osteomyelitis that are obtained during injection of the radiopharmaceutical. The second phase, the blood-pool image, is obtained within 5 mm after injection. In areas of inflammation, capillaries dilate, causing increased blood flow and blood pooling. The third phase, the bone image, is obtained about 3 hr later, when urinary excretion has decreased the amount of the radio- nuclide in the soft tissues. Classically, with cellulitis, diffuse increased uptake occurs in the first two phases, but uptake is normal or diffusely increased in the third phase (Fig. 1). If present, diffuse increased uptake in the third phase is prob- ably due to regional hyperemia caused by the cellulitis [7] (Fig. 1). Osteomyelitis causes focally increased uptake in all three phases (Fig. 2). Recently, some have proposed adding a 24-hr image to the three-phase bone scan to create the four-phase bone scan [8, 9]. The amount of radionuclide in the lesion vs the amount in normal bone should continue to increase during the fourth Received April 19, 1 991 ; accepted after revision July 1 9, 1991. This work was supported in part by National Institutes of Health grant AR 36460. 1 Department of Nuclear Medicine, Indiana University School of Medicine, Wishard Memorial Hospital, 1 001 W. 1 0th St. , Indianapolis, IN 46202. Address reprint requests to D. S. Schauwecker. AJR 158:9-18, January 1992 0361 -803X/92/1 581-0018 © American Roentgen Ray Society Downloaded from www.ajronline.org by 27.79.75.39 on 02/11/23 from IP address 27.79.75.39. Copyright ARRS. For personal use only; all rights reserved
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
The scintigraphic diagnosis of osteomyelitis.Review Article
Osteomyelitis is a serious health problem that results in multi- pie limb amputations annually. This article reviews the current scintigraphic procedures used in the diagnosis of osteomyelitis and discusses some of the newer radiopharmaceuticals now
being developed. The goal is to understand the strengths and weaknesses of each method so that the procedure most effective for specific clinical settings can be selected. In general, the three- phase bone scan is the procedure of choice if the suspected osteomyelitis is not superimposed on another disease that causes increased bone remodeling (i.e, findings on the radio- graph are normal). If the suspected osteomyelitis is superim- posed on a disease that causes increased bone remodeling, the combined 1111n-labeled Ieukocyte-”Tc bone scan is the proce- dure of choice in the non-marrow-containing skeleton and the 1111n-labeled leukocyte and mTc bone marrow scans are the procedures of choice in the marrow-containing skeleton.
Use of scintigraphy in the diagnosis of osteomyelitis was established in 1975 when three landmark articles [1-3] ap- peared. First, Duszynski et al. [1] discovered that a bone scan could be used to diagnose osteomyelitis several days before radiographs showed any abnormalities. Second, Gilday et al. [2] added the blood-pool image to the bone image as a forerunner of the current three-phase bone scan. Finally, Deysine et al. [3] proposed the use of gallium-67 for the study of chronic and postoperative osteomyelitis.
Three-Phase Bone Scan
Localization of 99mTc-diphosphonate in bone is related to both osteoblastic activity and skeletal vascularity. Commonly,
the bone scan will show abnormal uptake of the radionuclide 1 0-1 4 days before loss of bone mineral is great enough to
be seen on plain radiographs [4]. For example, 17 of 19 sites of osteomyelitis in children that were imaged within 3 days after the onset of symptoms showed abnormal uptake on the bone scan, whereas only one abnormal site was seen on the radiograph [5]. In fact, if antibiotic therapy is begun early enough to arrest the loss of bone mineral, the characteristic radiographic findings may never develop [6].
The three-phase bone scan is the routine nuclear medicine procedure for diagnosis of osteomyelitis. The first phase, the nuclear angiogram or flow phase, consists of serial 2- to 5- sec images of the area of suspected osteomyelitis that are obtained during injection of the radiopharmaceutical. The second phase, the blood-pool image, is obtained within 5 mm after injection. In areas of inflammation, capillaries dilate, causing increased blood flow and blood pooling. The third phase, the bone image, is obtained about 3 hr later, when urinary excretion has decreased the amount of the radio- nuclide in the soft tissues. Classically, with cellulitis, diffuse increased uptake occurs in the first two phases, but uptake is normal or diffusely increased in the third phase (Fig. 1). If present, diffuse increased uptake in the third phase is prob- ably due to regional hyperemia caused by the cellulitis [7] (Fig. 1). Osteomyelitis causes focally increased uptake in all three phases (Fig. 2).
Recently, some have proposed adding a 24-hr image to the three-phase bone scan to create the four-phase bone scan [8, 9]. The amount of radionuclide in the lesion vs the amount in normal bone should continue to increase during the fourth
Received April 19, 1 991 ; accepted after revision July 1 9, 1991. This work was supported in part by National Institutes of Health grant AR 36460. 1 Department of Nuclear Medicine, Indiana University School of Medicine, Wishard Memorial Hospital, 1 001 W. 1 0th St. , Indianapolis, IN 46202. Address reprint
requests to D. S. Schauwecker.
AJR 158:9-18, January 1992 0361 -803X/92/1 581-0018 © American Roentgen Ray Society
D ow
nl oa
de d
fr om
w w
w .a
jr on
lin e.
or g
by 2
7. 79
.7 5.
39 o
n 02
/1 1/
23 f
ro m
I P
ad dr
es s
27 .7
9. 75
.3 9.
C op
yr ig
ht A
R R
S. F
or p
er so
1L1 A
Fig. 1.-Three-phase bone scan in plantar projection shows classic findings of cellulitis. A, Four frames from dynamic phase show diffuse increased blood flow to right medial forefoot. B, Second phase shows increased uptake of radionuclide in blood pool throughout right foot, especially in same right medial forefoot area shown in A. C, Delayed image shows no area of focal increased activity corresponding to findings of first two phases (A and B). Diffuse increased activity throughout
right foot compared with left foot is caused by hyperemia from cellulitis [7].
Fig. 2.-A-C, Three-phase bone scan shows classic findings of osteomyelitis. Flow images (A), blood-pool image (B), and delayed image (C) all show intense focal localization in left fifth metatarsal.
phase, if the lesion is osteomyelitis [8, 9]. The four-phase bone scan is based on solid theoretical considerations. Uptake of ggmTcmethylene diphosphonate stops at about 4 hr in lamellar bone (normal skeleton) but continues for about 24 hr in woven bone (abnormal bone around osteomyelitis and bone tumors) [9]. The four-phase bone scan has proved useful for the evaluation of lesions of the feet in patients with peripheral vascular disease [8]. Often, bone detail is very poor on 3-hr delayed images in these patients, but is greatly improved by 24 hr. To date, the four-phase bone scan has not been widely accepted, perhaps because of the inconvenience of having the patient return for the fourth phase. False-positive results have been reported for degenerative diseases [8] and metas- tases [9]. Guan et al. [1 0] reported that the increased 24-hr/ 4-hr uptake ratio can be used to differentiate benign from malignant bone lesions.
In adults who have normal findings on radiographs (i.e., who have no lesions that cause increased bone turnover), the three-phase bone scan has high sensitivity and specificity (Table 1). When the results in Table 1 are added together, sensitivity is 94% (1 82/1 94) and specificity is 95% (360/380). In routine clinical practice, most cases of osteomyelitis can be detected by using the three-phase bone scan.
In neonates, the sensitivity of the three-phase bone scan decreases. Neonates with osteomyelitis have falsely normal or cold defects on three-phase bone scans 22-68% of the time [14-16].
There are two possible explanations for these results. First, these studies were performed in the late 1970s; gamma cameras have improved since then. More recently, Bressler et al. [1 7] studied 33 neonates less than 6 weeks old, and 13 (87%) of 1 5 had classical findings for osteomyelitis on three-
D ow
nl oa
de d
fr om
w w
w .a
jr on
lin e.
or g
by 2
7. 79
.7 5.
39 o
n 02
/1 1/
23 f
ro m
I P
ad dr
es s
27 .7
9. 75
.3 9.
C op
yr ig
ht A
R R
S. F
or p
er so
AJR:158, January 1992 SCINTIGRAPHY OF OSTEOMYELITIS 11
TABLE 1: Results of Three-Phase Bone Scans in Osteomyelitis: Optimal Conditions
Aeference Sensitivity Specificity
Gilday et al. [2] 70/71 65/65 Howie et al. [5] 55/62 152/165 Majd and Frankel [6] 22/23 42/42 Lisbona and Rosenthall [1 1 ] 1 6/1 7 21/23
Kolyvas et al. [1 2] 7/8 NA Maurer et al. [13] 12/13 80/85
Note-NA = not available.
phase bone scans. The remaining two patients had cold defects. The authors stressed the importance of optimal- quality scintigrams of the region of the epiphysis.
Second, Allwright et al. [1 8] found that the cold lesion on the third phase of the bone scan was caused by a subperios- teal abscess. In children, a large portion of the blood supply to the bone comes from periosteal vessels, and these are apparently disrupted by the subperiosteal abscess [1 8]. When clinical findings strongly suggest osteomyelitis, but findings on the three-phase bone scan are normal, 67Ga-citrate has been recommended for the second study [19]. This does not necessarily mean that 67Ga-citrate is more sensitive for neo- natal osteomyelitis than the three-phase bone scan is. After further evolution of the lesion, the second study would more likely be diagnostic.
The three-phase bone scan is less specific when bone remodeling is increased (Table 2). When the results in Table 2 are added together, sensitivity is 95% (250/262) and spec- ificity is 33% (1 1 0/330). In these situations 67Ga-citrate or 1 111n-Iabeled leukocytes have been used to try to increase specificity.
Gallium-67
Results of studies evaluating the sensitivity and specificity of 67Ga-citrate scans in osteomyelitis are variable (Table 3). When the results of these studies are added together, sensi- tivity is 81 % (209/257) and specificity is 69% (1 88/272). 67Ga- citrate localizes in areas of osteomyelitis by granulocyte or bacterial uptake, and by binding to lactoferrin at the site of infection [40]. However, 67Ga-citrate was originally proposed
as a bone scanning agent [41 ], and it shows increased uptake in areas of increased bone remodeling, such as bones with neuropathic changes [42, 43] or pseudarthrosis [32] (Fig. 3).
To overcome this problem, some authors have proposed comparing the 67Ga-citrate uptake in the lesion with the uptake on a bone scan; disparate distribution of uptake or increased intensity would constitute osteomyelitis [44, 45]. Unfortunately, as seen in Table 3, 67Ga-citrate can be sensitive or specific in complicating cases, but it is difficult to obtain high sensitivity and high specificity simultaneously [28]. When 67Ga-citrate and 111n-labeled leukocytes have been directly compared in the same patient, the labeled leukocytes are usually significantly better for the diagnosis of osteomyelitis [28, 37, 38]. Currently, 67Ga-citrate is rarely used for the diagnosis of osteomyelitis when 11n-labeled leukocytes are available.
TABLE 2: Results of Three-Phase Bone Scans in Osteomyelitis: Complicating Conditions
Reference Sensitivity Specificity
Park et al. [20] 20/24 9/12 Modic et al. [21 ] 20/22 11/14 Lewin et al. [22] 5/8 13/23 Maurer et al. [23] 3/4 5/9 Unger et al. [24] 9/1 1 13/20 Sugarman [25] 37/37 30/95 Magnuson et al. [26] 50/50 9/48 Splittgerber et al. [27] 3/3 0/3 Schauwecker et al. [28] 32/32 0/25 AI-Sheikh et al. [29] 10/10 3/12
Ivancevic et al. [30] 15/1 5 0/2 Ruther et al. [31] 13/13 0/19 Hadjipavlou et al. [32] 2/2 0/9 Keenan et al. [33] 31/31 17/39
TABLE 3: Results of 67Ga-Citrate Scans in Osteomyelitis
Reference Sensitivity Specificity
Lisbona and Rosenthall [1 1 J 17/1 7 23/23 Shafer et al. [34j 1 5/1 6 21/22
Esterhai et al. [351 13/1 3 0/11 Merkel et al. [361 1 1/23 6/7 Al-Sheikh et al. [29] 8/10 10/12 Modic et al. [21] 12/13 7/7 Lewin et al. [22] 8/8 16/23 Sugarman [25] 25/26 15/50 Seabold et al. [37] 2/9 16/16 Schauwecker et al. [281 21/21 2/8 Ivancevic et al. [30] 12/1 5 2/2 Handmaker and Giammona 13/1 7 NA
[19]
Merkel et al. [381 19/35 70/80 Hadjipavlou et al. [32] 2/2 0/9 Borman et al. [39] 31/32 0/2
Note-NA = not available.
Indium-i 1 1-Labeled Leukocytes
Leukocytes labeled with 1In have been used to overcome certain limitations of the 67Ga-citrate scan. Labeled leukocytes accumulate in areas of infection. Localization of the cells is specific for infection when suspected osteomyelitis is super- imposed on processes that cause increased bone remodeling [23, 28, 46, 47]. The sensitivity and specificity of 11n-labeled leukocyte scans are reviewed in Table 4. When the results are added together, sensitivity is 88% (422/480) and specific- ity is 85% (389/457).
Most of the reported studies of 1In-labeled leukocyte scintigrams are retrospective. Studies by Seabold et al. [37] and Esterhai et al. [49] are particularly important as they are prospective studies that used 1ln-labeled leukocytes to eval- uate osteomyelitis complicating fracture nonunion. Their gold standard was open biopsy and culture in all patients. Esterhai et al. [49] had 15 true-positive findings and five true-negative findings. Seabold et al. [37] reported 1 6 true-positive, 28 true- negative, one false-positive, and three false-negative results. Together, these prospective studies had a combined sensitiv-
D ow
nl oa
de d
fr om
w w
w .a
jr on
lin e.
or g
by 2
7. 79
.7 5.
39 o
n 02
/1 1/
23 f
ro m
I P
ad dr
es s
27 .7
9. 75
.3 9.
C op
yr ig
ht A
R R
S. F
or p
er so
Fig. 3.-Neuropathic osteopathy can be difficult to differentiate from osteomyelitis. Three-phase bone scan showed focal increased activity in tarsal region on all three phases.
A, Bone Image of three-phase bone scan shows increased uptake of radionuclide in tarsal region. B, ‘TGa-citrate Image on day 5 is virtually identical to delayed bone image (A). C, 24-hr “In-labeled leukocyte study shows no accumulation of radionuclide. At surgery, no evidence of osteomyelitis was seen. An ulcer was skin
grafted, and foot healed normally.
a Diabetic foot.
TABLE 4: Results of 1111n-Labeled Leukocyte Scans in Osteomyelitis
Reference Sensitivity Specificity
Schauwecker et al. [28] 24/32 24/25 Merkel et al. [36] 19/23 6/7 Ouzounian et al. [48] 13/1 4 39/41 Wukich et al. [47] 20/21 11/29 Esterhai et al. [49] 10/1 0 5/5 lIes et al. [50] 16/1 9 27/33 Magnuson et al. [26] 44/50 35/48 Seabold et al. [37] 16/1 9 28/29 Al-Sheikh et al. [29] 8/1 0 9/12 Maurer et al. [2318 3/4 8/9 Splittgerber et al. [2T] 3/3 3/3 Schauwecker et al. [51 ja 1 7/1 7 15/18
McCarthy et al. [52] 27/28 9/11
Schauwecker [53] 1 79/206 140/151
Mulamba et al. [54] 12/1 3 17/17 Keenan et al. [33]$ 1 1/1 1 13/19
ity of9l % and specificity of97% for diagnosis of osteomyelitis complicating fracture nonunion.
The diagnosis of osteomyelitis in patients with diabetic osteopathy is a serious clinical problem. Although the results in diabetics appear promising (Table 4), important difficulties still exist. Determining that infection is either present or absent does not always answer all the clinical questions. Maurer et al. [23] found that 111ln-labeled leukocyte scintigraphy had good sensitivity and specificity for the detection of infection in diabetic foot disease, but the scans were not useful for differentiating infection in the bone from that in the adjacent soft tissue. More recent work uses the combined “Tc bone
1 1ln-labeled leukocyte study to determine if the leuko-
cyte collection is in the bone or soft tissue (Figs. 4 and 5). Using this combined study, Schauwecker et al. [51 ] correctly localized the infection to the bone or soft tissue 89% of the time in patients with neuropathic foot diseases.
Unfortunately, even with these advances, the diagnosis of osteomyelitis superimposed on diabetic neuropathic osteo- pathy remains a challenging clinical problem. Seabold et al. [55] found that rapidly progressing noninfected neuropathic osteoarthropathy of recent onset can be indistinguishable from osteomyelitis with either the combined 99mTc bone scan- 111In-labeled leukocyte study or MR imaging.
One study [53] retrospectively reviewed 485 patients with suspected osteomyelitis and showed some of the limitations of using 11n-labeled leukocytes. Acute osteomyelitis can be readily diagnosed anywhere in the body, whereas chronic osteomyelitis, with its lower influx of labeled leukocytes, is seen accurately in the peripheral skeleton only. The 11n- labeled leukocytes accumulate in active bone marrow, which reduces the sensitivity for detection of chronic osteomyelitis in the central skeleton [53]. Low sensitivity for vertebral
osteomyelitis when using labeled leukocytes has also been reported by Whalen et al. [56] and Ruther et al. [31].
A cold defect in the central skeleton on a 1ln-labeled leukocyte study is often confusing (Fig. 6). The defect occurs in 1 0-40% of such studies for suspected osteomyelitis in the central skeleton [53, 56-59]. Although the cold defect could represent osteomyelitis, it also is seen in metastases, frac- tures, Paget disease, surgical defects, and intervals after irradiation [53, 56-59]. The cold defect is consistent with, but not diagnostic of, osteomyelitis. Palestro et al. [60] performed serial studies in patients with vertebral osteomyelitis and found a progression from hot to cold as the lesions became more chronic. Whalen et al. [56] showed that antibiotic ther- apy was correlated with a high incidence of cold defects.
D ow
nl oa
de d
fr om
w w
w .a
jr on
lin e.
or g
by 2
7. 79
.7 5.
39 o
n 02
/1 1/
23 f
ro m
I P
ad dr
es s
27 .7
9. 75
.3 9.
C op
yr ig
ht A
R R
S. F
or p
er so
AJR:158, January 1992 SCINTIGRAPHY OF OSTEOMYELITIS 13
Fig. 4.-A-F, Combined three-phase bone scan-”ln-labeled leukocyte study for diagnosis of osteomyelitis in a patient with obvious cellulitis. Images A + B = C and 0 + E = F.
A and B, Left lateral 1111n-labeled leu- kocyte scan (A) and bone image (B).
0 and E, Plantar 1111n-labeled leuko- cyte scan (0) and bone image (E).
C and F, Superimpositions of A on B (C) and 0 on E (F), respectively. 111ln- labeled leukocyte activity is localized to ventral and dorsal soft tissues and not within bone. This represents cellu- litis without evidence of osteomyelitis.
Fig. 5.-A-F, Combined three-phase bone scan-’111n-labeled leukocyte study for diagnosis of osteomyelitis. lmagesA +B=CandD+E=F.
A and B, Plantar ‘11ln-labeled leuko- cyte scan (A) and bone image (B).
D and E, Left lateral ‘111n-labeled leu- kocyte scan (0) and bone image (E).
C and F, Superimpositions of A on B (C) and D on E (F), respectively. On both projections, activity clearly lies within, and adjacent to, bone. Osteo- myelitis and adjacent cellulitis are clearly present.
L .
D E
Either of these mechanisms could result in a more indolent infection with decreased leukocyte accumulation.
One way to improve specificity in the central skeleton is to perform a 99mTc bone marrow scan In-labeled leukocyte
study [61 -63]. The 99mTc sulfur or antimony colloid will clearly delineate the extent of the bone marrow. Any incongruity of the bone marrow and 1111n-labeled leukocyte images would be considered significant (Fig 7). Recently, this study has
D ow
nl oa
de d
fr om
w w
w .a
jr on
lin e.
or g
by 2
7. 79
.7 5.
39 o
n 02
/1 1/
23 f
ro m
I P
ad dr
es s
27 .7
9. 75
.3 9.
C op
yr ig
ht A
R R
S. F
or p
er so
se o
nl y;
a ll
ri gh
ts r
es er
ve d
14 SCHAUWECKER
been found to increase both the sensitivity and specificity for complicating osteomyelitis in the central skeleton [64]. Thus, the “Tc bone marrow scan and 111InIabeled leukocyte study are the procedures of choice for diagnosing complicating osteomyelitis in the marrow-containing skeleton.
AJR:158, January 1992
False-positive and false-negative results of 1In-labeled leukocyte studies have multiple causes. Some authors have suggested that antibiotic therapy may lower the sensitivity of the 1111n-labeled leukocyte study [56]. Other authors [65, 66] have found no significant antibiotic effect. False-positive re- suits have been reported with rheumatoid arthritis [67], heal- ing fractures [68, 69], noninfected prosthesis [70], and met- astatic carcinoma [53, 71 ]. Abreu [72] recently reviewed the literature and listed many causes of skeletal uptake of 11n- labeled leukocytes. Usually the processes that give false- positive results for infection have fainter accumulation of labeled leukocytes than is seen with osteomyelitis.
The clinician often is asked to diagnose suspected osteo- myelitis superimposed on healing fractures. In a canine model, 1 1 11n-labeled leukocytes accumulated appreciably during the first month but were seen only faintly when the dogs were imaged later [73]. This finding appears to agree with the results of Van Nostrand et al. [69]. A healing fracture should have faint uptake and be seen for only a few weeks after injury, compared with fracture complicated with osteomyelitis.
Other Radiopharmaceuticals 1…
Osteomyelitis is a serious health problem that results in multi- pie limb amputations annually. This article reviews the current scintigraphic procedures used in the diagnosis of osteomyelitis and discusses some of the newer radiopharmaceuticals now
being developed. The goal is to understand the strengths and weaknesses of each method so that the procedure most effective for specific clinical settings can be selected. In general, the three- phase bone scan is the procedure of choice if the suspected osteomyelitis is not superimposed on another disease that causes increased bone remodeling (i.e, findings on the radio- graph are normal). If the suspected osteomyelitis is superim- posed on a disease that causes increased bone remodeling, the combined 1111n-labeled Ieukocyte-”Tc bone scan is the proce- dure of choice in the non-marrow-containing skeleton and the 1111n-labeled leukocyte and mTc bone marrow scans are the procedures of choice in the marrow-containing skeleton.
Use of scintigraphy in the diagnosis of osteomyelitis was established in 1975 when three landmark articles [1-3] ap- peared. First, Duszynski et al. [1] discovered that a bone scan could be used to diagnose osteomyelitis several days before radiographs showed any abnormalities. Second, Gilday et al. [2] added the blood-pool image to the bone image as a forerunner of the current three-phase bone scan. Finally, Deysine et al. [3] proposed the use of gallium-67 for the study of chronic and postoperative osteomyelitis.
Three-Phase Bone Scan
Localization of 99mTc-diphosphonate in bone is related to both osteoblastic activity and skeletal vascularity. Commonly,
the bone scan will show abnormal uptake of the radionuclide 1 0-1 4 days before loss of bone mineral is great enough to
be seen on plain radiographs [4]. For example, 17 of 19 sites of osteomyelitis in children that were imaged within 3 days after the onset of symptoms showed abnormal uptake on the bone scan, whereas only one abnormal site was seen on the radiograph [5]. In fact, if antibiotic therapy is begun early enough to arrest the loss of bone mineral, the characteristic radiographic findings may never develop [6].
The three-phase bone scan is the routine nuclear medicine procedure for diagnosis of osteomyelitis. The first phase, the nuclear angiogram or flow phase, consists of serial 2- to 5- sec images of the area of suspected osteomyelitis that are obtained during injection of the radiopharmaceutical. The second phase, the blood-pool image, is obtained within 5 mm after injection. In areas of inflammation, capillaries dilate, causing increased blood flow and blood pooling. The third phase, the bone image, is obtained about 3 hr later, when urinary excretion has decreased the amount of the radio- nuclide in the soft tissues. Classically, with cellulitis, diffuse increased uptake occurs in the first two phases, but uptake is normal or diffusely increased in the third phase (Fig. 1). If present, diffuse increased uptake in the third phase is prob- ably due to regional hyperemia caused by the cellulitis [7] (Fig. 1). Osteomyelitis causes focally increased uptake in all three phases (Fig. 2).
Recently, some have proposed adding a 24-hr image to the three-phase bone scan to create the four-phase bone scan [8, 9]. The amount of radionuclide in the lesion vs the amount in normal bone should continue to increase during the fourth
Received April 19, 1 991 ; accepted after revision July 1 9, 1991. This work was supported in part by National Institutes of Health grant AR 36460. 1 Department of Nuclear Medicine, Indiana University School of Medicine, Wishard Memorial Hospital, 1 001 W. 1 0th St. , Indianapolis, IN 46202. Address reprint
requests to D. S. Schauwecker.
AJR 158:9-18, January 1992 0361 -803X/92/1 581-0018 © American Roentgen Ray Society
D ow
nl oa
de d
fr om
w w
w .a
jr on
lin e.
or g
by 2
7. 79
.7 5.
39 o
n 02
/1 1/
23 f
ro m
I P
ad dr
es s
27 .7
9. 75
.3 9.
C op
yr ig
ht A
R R
S. F
or p
er so
1L1 A
Fig. 1.-Three-phase bone scan in plantar projection shows classic findings of cellulitis. A, Four frames from dynamic phase show diffuse increased blood flow to right medial forefoot. B, Second phase shows increased uptake of radionuclide in blood pool throughout right foot, especially in same right medial forefoot area shown in A. C, Delayed image shows no area of focal increased activity corresponding to findings of first two phases (A and B). Diffuse increased activity throughout
right foot compared with left foot is caused by hyperemia from cellulitis [7].
Fig. 2.-A-C, Three-phase bone scan shows classic findings of osteomyelitis. Flow images (A), blood-pool image (B), and delayed image (C) all show intense focal localization in left fifth metatarsal.
phase, if the lesion is osteomyelitis [8, 9]. The four-phase bone scan is based on solid theoretical considerations. Uptake of ggmTcmethylene diphosphonate stops at about 4 hr in lamellar bone (normal skeleton) but continues for about 24 hr in woven bone (abnormal bone around osteomyelitis and bone tumors) [9]. The four-phase bone scan has proved useful for the evaluation of lesions of the feet in patients with peripheral vascular disease [8]. Often, bone detail is very poor on 3-hr delayed images in these patients, but is greatly improved by 24 hr. To date, the four-phase bone scan has not been widely accepted, perhaps because of the inconvenience of having the patient return for the fourth phase. False-positive results have been reported for degenerative diseases [8] and metas- tases [9]. Guan et al. [1 0] reported that the increased 24-hr/ 4-hr uptake ratio can be used to differentiate benign from malignant bone lesions.
In adults who have normal findings on radiographs (i.e., who have no lesions that cause increased bone turnover), the three-phase bone scan has high sensitivity and specificity (Table 1). When the results in Table 1 are added together, sensitivity is 94% (1 82/1 94) and specificity is 95% (360/380). In routine clinical practice, most cases of osteomyelitis can be detected by using the three-phase bone scan.
In neonates, the sensitivity of the three-phase bone scan decreases. Neonates with osteomyelitis have falsely normal or cold defects on three-phase bone scans 22-68% of the time [14-16].
There are two possible explanations for these results. First, these studies were performed in the late 1970s; gamma cameras have improved since then. More recently, Bressler et al. [1 7] studied 33 neonates less than 6 weeks old, and 13 (87%) of 1 5 had classical findings for osteomyelitis on three-
D ow
nl oa
de d
fr om
w w
w .a
jr on
lin e.
or g
by 2
7. 79
.7 5.
39 o
n 02
/1 1/
23 f
ro m
I P
ad dr
es s
27 .7
9. 75
.3 9.
C op
yr ig
ht A
R R
S. F
or p
er so
AJR:158, January 1992 SCINTIGRAPHY OF OSTEOMYELITIS 11
TABLE 1: Results of Three-Phase Bone Scans in Osteomyelitis: Optimal Conditions
Aeference Sensitivity Specificity
Gilday et al. [2] 70/71 65/65 Howie et al. [5] 55/62 152/165 Majd and Frankel [6] 22/23 42/42 Lisbona and Rosenthall [1 1 ] 1 6/1 7 21/23
Kolyvas et al. [1 2] 7/8 NA Maurer et al. [13] 12/13 80/85
Note-NA = not available.
phase bone scans. The remaining two patients had cold defects. The authors stressed the importance of optimal- quality scintigrams of the region of the epiphysis.
Second, Allwright et al. [1 8] found that the cold lesion on the third phase of the bone scan was caused by a subperios- teal abscess. In children, a large portion of the blood supply to the bone comes from periosteal vessels, and these are apparently disrupted by the subperiosteal abscess [1 8]. When clinical findings strongly suggest osteomyelitis, but findings on the three-phase bone scan are normal, 67Ga-citrate has been recommended for the second study [19]. This does not necessarily mean that 67Ga-citrate is more sensitive for neo- natal osteomyelitis than the three-phase bone scan is. After further evolution of the lesion, the second study would more likely be diagnostic.
The three-phase bone scan is less specific when bone remodeling is increased (Table 2). When the results in Table 2 are added together, sensitivity is 95% (250/262) and spec- ificity is 33% (1 1 0/330). In these situations 67Ga-citrate or 1 111n-Iabeled leukocytes have been used to try to increase specificity.
Gallium-67
Results of studies evaluating the sensitivity and specificity of 67Ga-citrate scans in osteomyelitis are variable (Table 3). When the results of these studies are added together, sensi- tivity is 81 % (209/257) and specificity is 69% (1 88/272). 67Ga- citrate localizes in areas of osteomyelitis by granulocyte or bacterial uptake, and by binding to lactoferrin at the site of infection [40]. However, 67Ga-citrate was originally proposed
as a bone scanning agent [41 ], and it shows increased uptake in areas of increased bone remodeling, such as bones with neuropathic changes [42, 43] or pseudarthrosis [32] (Fig. 3).
To overcome this problem, some authors have proposed comparing the 67Ga-citrate uptake in the lesion with the uptake on a bone scan; disparate distribution of uptake or increased intensity would constitute osteomyelitis [44, 45]. Unfortunately, as seen in Table 3, 67Ga-citrate can be sensitive or specific in complicating cases, but it is difficult to obtain high sensitivity and high specificity simultaneously [28]. When 67Ga-citrate and 111n-labeled leukocytes have been directly compared in the same patient, the labeled leukocytes are usually significantly better for the diagnosis of osteomyelitis [28, 37, 38]. Currently, 67Ga-citrate is rarely used for the diagnosis of osteomyelitis when 11n-labeled leukocytes are available.
TABLE 2: Results of Three-Phase Bone Scans in Osteomyelitis: Complicating Conditions
Reference Sensitivity Specificity
Park et al. [20] 20/24 9/12 Modic et al. [21 ] 20/22 11/14 Lewin et al. [22] 5/8 13/23 Maurer et al. [23] 3/4 5/9 Unger et al. [24] 9/1 1 13/20 Sugarman [25] 37/37 30/95 Magnuson et al. [26] 50/50 9/48 Splittgerber et al. [27] 3/3 0/3 Schauwecker et al. [28] 32/32 0/25 AI-Sheikh et al. [29] 10/10 3/12
Ivancevic et al. [30] 15/1 5 0/2 Ruther et al. [31] 13/13 0/19 Hadjipavlou et al. [32] 2/2 0/9 Keenan et al. [33] 31/31 17/39
TABLE 3: Results of 67Ga-Citrate Scans in Osteomyelitis
Reference Sensitivity Specificity
Lisbona and Rosenthall [1 1 J 17/1 7 23/23 Shafer et al. [34j 1 5/1 6 21/22
Esterhai et al. [351 13/1 3 0/11 Merkel et al. [361 1 1/23 6/7 Al-Sheikh et al. [29] 8/10 10/12 Modic et al. [21] 12/13 7/7 Lewin et al. [22] 8/8 16/23 Sugarman [25] 25/26 15/50 Seabold et al. [37] 2/9 16/16 Schauwecker et al. [281 21/21 2/8 Ivancevic et al. [30] 12/1 5 2/2 Handmaker and Giammona 13/1 7 NA
[19]
Merkel et al. [381 19/35 70/80 Hadjipavlou et al. [32] 2/2 0/9 Borman et al. [39] 31/32 0/2
Note-NA = not available.
Indium-i 1 1-Labeled Leukocytes
Leukocytes labeled with 1In have been used to overcome certain limitations of the 67Ga-citrate scan. Labeled leukocytes accumulate in areas of infection. Localization of the cells is specific for infection when suspected osteomyelitis is super- imposed on processes that cause increased bone remodeling [23, 28, 46, 47]. The sensitivity and specificity of 11n-labeled leukocyte scans are reviewed in Table 4. When the results are added together, sensitivity is 88% (422/480) and specific- ity is 85% (389/457).
Most of the reported studies of 1In-labeled leukocyte scintigrams are retrospective. Studies by Seabold et al. [37] and Esterhai et al. [49] are particularly important as they are prospective studies that used 1ln-labeled leukocytes to eval- uate osteomyelitis complicating fracture nonunion. Their gold standard was open biopsy and culture in all patients. Esterhai et al. [49] had 15 true-positive findings and five true-negative findings. Seabold et al. [37] reported 1 6 true-positive, 28 true- negative, one false-positive, and three false-negative results. Together, these prospective studies had a combined sensitiv-
D ow
nl oa
de d
fr om
w w
w .a
jr on
lin e.
or g
by 2
7. 79
.7 5.
39 o
n 02
/1 1/
23 f
ro m
I P
ad dr
es s
27 .7
9. 75
.3 9.
C op
yr ig
ht A
R R
S. F
or p
er so
Fig. 3.-Neuropathic osteopathy can be difficult to differentiate from osteomyelitis. Three-phase bone scan showed focal increased activity in tarsal region on all three phases.
A, Bone Image of three-phase bone scan shows increased uptake of radionuclide in tarsal region. B, ‘TGa-citrate Image on day 5 is virtually identical to delayed bone image (A). C, 24-hr “In-labeled leukocyte study shows no accumulation of radionuclide. At surgery, no evidence of osteomyelitis was seen. An ulcer was skin
grafted, and foot healed normally.
a Diabetic foot.
TABLE 4: Results of 1111n-Labeled Leukocyte Scans in Osteomyelitis
Reference Sensitivity Specificity
Schauwecker et al. [28] 24/32 24/25 Merkel et al. [36] 19/23 6/7 Ouzounian et al. [48] 13/1 4 39/41 Wukich et al. [47] 20/21 11/29 Esterhai et al. [49] 10/1 0 5/5 lIes et al. [50] 16/1 9 27/33 Magnuson et al. [26] 44/50 35/48 Seabold et al. [37] 16/1 9 28/29 Al-Sheikh et al. [29] 8/1 0 9/12 Maurer et al. [2318 3/4 8/9 Splittgerber et al. [2T] 3/3 3/3 Schauwecker et al. [51 ja 1 7/1 7 15/18
McCarthy et al. [52] 27/28 9/11
Schauwecker [53] 1 79/206 140/151
Mulamba et al. [54] 12/1 3 17/17 Keenan et al. [33]$ 1 1/1 1 13/19
ity of9l % and specificity of97% for diagnosis of osteomyelitis complicating fracture nonunion.
The diagnosis of osteomyelitis in patients with diabetic osteopathy is a serious clinical problem. Although the results in diabetics appear promising (Table 4), important difficulties still exist. Determining that infection is either present or absent does not always answer all the clinical questions. Maurer et al. [23] found that 111ln-labeled leukocyte scintigraphy had good sensitivity and specificity for the detection of infection in diabetic foot disease, but the scans were not useful for differentiating infection in the bone from that in the adjacent soft tissue. More recent work uses the combined “Tc bone
1 1ln-labeled leukocyte study to determine if the leuko-
cyte collection is in the bone or soft tissue (Figs. 4 and 5). Using this combined study, Schauwecker et al. [51 ] correctly localized the infection to the bone or soft tissue 89% of the time in patients with neuropathic foot diseases.
Unfortunately, even with these advances, the diagnosis of osteomyelitis superimposed on diabetic neuropathic osteo- pathy remains a challenging clinical problem. Seabold et al. [55] found that rapidly progressing noninfected neuropathic osteoarthropathy of recent onset can be indistinguishable from osteomyelitis with either the combined 99mTc bone scan- 111In-labeled leukocyte study or MR imaging.
One study [53] retrospectively reviewed 485 patients with suspected osteomyelitis and showed some of the limitations of using 11n-labeled leukocytes. Acute osteomyelitis can be readily diagnosed anywhere in the body, whereas chronic osteomyelitis, with its lower influx of labeled leukocytes, is seen accurately in the peripheral skeleton only. The 11n- labeled leukocytes accumulate in active bone marrow, which reduces the sensitivity for detection of chronic osteomyelitis in the central skeleton [53]. Low sensitivity for vertebral
osteomyelitis when using labeled leukocytes has also been reported by Whalen et al. [56] and Ruther et al. [31].
A cold defect in the central skeleton on a 1ln-labeled leukocyte study is often confusing (Fig. 6). The defect occurs in 1 0-40% of such studies for suspected osteomyelitis in the central skeleton [53, 56-59]. Although the cold defect could represent osteomyelitis, it also is seen in metastases, frac- tures, Paget disease, surgical defects, and intervals after irradiation [53, 56-59]. The cold defect is consistent with, but not diagnostic of, osteomyelitis. Palestro et al. [60] performed serial studies in patients with vertebral osteomyelitis and found a progression from hot to cold as the lesions became more chronic. Whalen et al. [56] showed that antibiotic ther- apy was correlated with a high incidence of cold defects.
D ow
nl oa
de d
fr om
w w
w .a
jr on
lin e.
or g
by 2
7. 79
.7 5.
39 o
n 02
/1 1/
23 f
ro m
I P
ad dr
es s
27 .7
9. 75
.3 9.
C op
yr ig
ht A
R R
S. F
or p
er so
AJR:158, January 1992 SCINTIGRAPHY OF OSTEOMYELITIS 13
Fig. 4.-A-F, Combined three-phase bone scan-”ln-labeled leukocyte study for diagnosis of osteomyelitis in a patient with obvious cellulitis. Images A + B = C and 0 + E = F.
A and B, Left lateral 1111n-labeled leu- kocyte scan (A) and bone image (B).
0 and E, Plantar 1111n-labeled leuko- cyte scan (0) and bone image (E).
C and F, Superimpositions of A on B (C) and 0 on E (F), respectively. 111ln- labeled leukocyte activity is localized to ventral and dorsal soft tissues and not within bone. This represents cellu- litis without evidence of osteomyelitis.
Fig. 5.-A-F, Combined three-phase bone scan-’111n-labeled leukocyte study for diagnosis of osteomyelitis. lmagesA +B=CandD+E=F.
A and B, Plantar ‘11ln-labeled leuko- cyte scan (A) and bone image (B).
D and E, Left lateral ‘111n-labeled leu- kocyte scan (0) and bone image (E).
C and F, Superimpositions of A on B (C) and D on E (F), respectively. On both projections, activity clearly lies within, and adjacent to, bone. Osteo- myelitis and adjacent cellulitis are clearly present.
L .
D E
Either of these mechanisms could result in a more indolent infection with decreased leukocyte accumulation.
One way to improve specificity in the central skeleton is to perform a 99mTc bone marrow scan In-labeled leukocyte
study [61 -63]. The 99mTc sulfur or antimony colloid will clearly delineate the extent of the bone marrow. Any incongruity of the bone marrow and 1111n-labeled leukocyte images would be considered significant (Fig 7). Recently, this study has
D ow
nl oa
de d
fr om
w w
w .a
jr on
lin e.
or g
by 2
7. 79
.7 5.
39 o
n 02
/1 1/
23 f
ro m
I P
ad dr
es s
27 .7
9. 75
.3 9.
C op
yr ig
ht A
R R
S. F
or p
er so
se o
nl y;
a ll
ri gh
ts r
es er
ve d
14 SCHAUWECKER
been found to increase both the sensitivity and specificity for complicating osteomyelitis in the central skeleton [64]. Thus, the “Tc bone marrow scan and 111InIabeled leukocyte study are the procedures of choice for diagnosing complicating osteomyelitis in the marrow-containing skeleton.
AJR:158, January 1992
False-positive and false-negative results of 1In-labeled leukocyte studies have multiple causes. Some authors have suggested that antibiotic therapy may lower the sensitivity of the 1111n-labeled leukocyte study [56]. Other authors [65, 66] have found no significant antibiotic effect. False-positive re- suits have been reported with rheumatoid arthritis [67], heal- ing fractures [68, 69], noninfected prosthesis [70], and met- astatic carcinoma [53, 71 ]. Abreu [72] recently reviewed the literature and listed many causes of skeletal uptake of 11n- labeled leukocytes. Usually the processes that give false- positive results for infection have fainter accumulation of labeled leukocytes than is seen with osteomyelitis.
The clinician often is asked to diagnose suspected osteo- myelitis superimposed on healing fractures. In a canine model, 1 1 11n-labeled leukocytes accumulated appreciably during the first month but were seen only faintly when the dogs were imaged later [73]. This finding appears to agree with the results of Van Nostrand et al. [69]. A healing fracture should have faint uptake and be seen for only a few weeks after injury, compared with fracture complicated with osteomyelitis.
Other Radiopharmaceuticals 1…
Related Documents
