Imaging plays an important role in the diagnosis of acute pyogenic osteomyelitis. It should always start with plain radiographs of the affected area.3, 4, 5
On the basis of the route of infection, acute osteomyelitis can be classified as hematogenous or exogenous (see the images below). Hematogenous osteomyelitis is predominantly seen in children and involves the highly vascular long bones, especially those of the lower limb. In adults, hematogenous spread is more common to the lumbar vertebral bodies than elsewhere. In neonatal osteomyelitis, isotopic bone scans are reportedly normal in most patients.
Before puberty, infection starts in the metaphyseal sinusoidal veins. Because bones are relatively rigid structures, focal edema accumulates under pressure and leads to local tissue necrosis, breakdown of the trabecular bone structure, and removal of bone matrix and calcium. Infection spreads along the haversian canals, through the marrow cavity, and beneath the periosteal layer of the bone. Subsequent vascular damage causes the ischemic death of osteocytes, leading to the formation of a sequestrum. Periosteal new-bone formation on top of the sequestrum is known as involucrum.
Osteomyelitis may be acute, subacute, or chronic. With acute osteomyelitis, the presenting complaint is usually local pain, swelling, and warmth. These often occur in association with fever and malaise.
Differentiating acute osteomyelitis from bone infarction in patients with sickle cell disease is a major challenge. The 2 conditions must be differentiated on the basis of clinical findings and imaging studies because both are common in patients with sickle cell disease. The 2 diseases are managed differently.
Fine-needle aspiration (FNA) or needle biopsy may be used under ultrasonographic, fluoroscopic, or CT guidance to obtain samples of pus, tissue, or both to establish a histologic diagnosis of acute osteomyelitis.
The disease process involves 5 stages:
- Inflammation: This stage represents initial inflammation with vascular congestion and increased intraosseous pressure; obstruction to blood flow occurs with intravascular thrombosis.
- Suppuration: Pus within the bones forces its way through the haversian system and forms a subperiosteal abscess in 2-3 days
- Sequestrum: Increased pressure, vascular obstruction, and infective thrombus compromise the periosteal and endosteal blood supply, causing bone necrosis and sequestrum formation in approximately 7 days
- Involucrum: This is new bone formation from the stripped surface of periosteum
- Resolution or progression to complications: With antibiotics and surgical treatment early in the course of disease, osteomyelitis resolves without any complications
Although osseous changes become apparent on conventional radiographs 5-7 days into the disease process, plain radiographs are useful in ruling out other causes of bone pain, such as stress fractures. Plain radiography and radionuclide bone scanning greatly aid early diagnosis in cases of acute osteomyelitis. Plain radiography is useful for excluding other conditions; radionuclide scanning reveals evidence of inflammation at the site of bone pain.
Nuclear medicine bone scans are a highly sensitive (>90%) modality in the diagnosis of osteomyelitis. This procedure is performed in 3 stages. Technetium-99m (99m Tc) is used to create images to determine areas of infection and bone remodeling dependent on local blood flow. The sensitivity of bone scans is often helpful when the exact site and extent of the infection is not known.
CT scanning allows for 3-dimensional (3D) examination of bone and surrounding soft tissue. CT scanning is an excellent modality for depicting periosteal new-bone formation and cortical bone destruction and for determining whether any sequestration or involucrum is present. Contrast-enhanced CT may show a ring-enhancing soft tissue abscess.
MRI, if available, is another useful modality for imaging acute osteomyelitis. Findings on MRI accurately demonstrate the extent and structure of the area involved in the pathologic process. The reported sensitivity is 88-100%; the specificity is 75-100%. Fat-suppression sequences allow for better detection of bone marrow edema; however, infection and inflammation cannot be differentiated. MRI may be the imaging modality of choice for infections involving the spine, pelvis, or limbs because of its ability to provide fine details of the osseous changes and soft tissue extension in these areas.
Additional imaging may be performed with indium-111–labeled leukocytes; gallium is used as needed. Gallium seems especially valuable in monitoring the efficacy of treatment.
Urso and associates evaluated 40 pediatric patients (aged 2-16 y) with osteomyelitis to assess the roles of various imaging modalities, including conventional radiology, bone scanning with99m Tc methylene diphosphonate (MDP), scintigraphy with99m Tc hexamethylenepropyleneamineoxime (HMPAO)–labeled leukocytes, CT, and MRI. 
As for acute osteomyelitis (6 patients), conventional radiography showed a lytic lesion and periosteal new-bone formation and soft tissue swelling (4 of 6 patients). (No abnormalities were demonstrated in the other 2 patients.) Bone scanning, CT, and MRI depicted bone involvement. CT and MRI also showed involvement and the spread of an inflammatory lesion to surrounding soft tissue.99m Tc-HMPAO scintigraphy was not performed to assess acute osteomyelitis, because of technical difficulties in performing the study promptly; thus, early analysis of the nature of the lesion was made with bone biopsy.
As for subacute osteomyelitis,99m Tc-HMPAO scintigraphy was performed in 8 of 23 patients; it proved to be highly sensitive, showing cell clusters and confirming the diagnosis of an inflammatory lesion. T1-weighted MRIs showed a focal area of intermediate to low signal intensity. These MRIs also showed small, focal, intralesional areas of low intensity; a low-signal perifocal rim; and diffusely low signal intensity of the surrounding bone marrow. T2-weighted images showed high signal intensity in both the abscess lesion and the bone marrow; the latter was probably the result of edema. For 5 patients, a paramagnetic contrast agent (gadopentetate dimeglumine) was administered during MRI and resulted in inhomogeneous enhancement of both the inflammatory lesion and the surrounding bone marrow.
Regarding chronic osteomyelitis (7 patients), MRI was performed in 5 patients. In 4 patients, the lesion appeared as a hypointense area on T1-weighted images; T2-weighted images showed an inhomogeneous hyperintense area. In the same patients,99m Tc-HMPAO scintigraphy was always positive. In the fifth patient, the lesion was represented by a hypointense area on both T1- and T2-weighted images;99m Tc-HMPAO scans were negative.
Therefore, in cases of chronic osteomyelitis, both MRI and99m Tc-HMPAO were useful in detecting spinal and peripheral bone involvement, which was asymptomatic at first observation in some cases. Conventional radiography, CT findings (3 of 4 patients), and MRI findings (4 of 4 patients) of extra-axial localizations were similar to those in the subacute-chronic forms.
Limitations of techniques
Plain radiographs are often normal for at least 1 week following infection; the findings are nonspecific.
MRI is contraindicated in patients with certain implant devices and metallic clips, and it is not tolerated by all patients because of claustrophobia or morbid obesity. In addition, young children may require sedation. Use of MRI requires patient cooperation because patient motion may degrade the images.
CT is quick and inexpensive but exposes the patient to ionizing radiation. The risk of a reaction to radio-iodinated contrast material is low; the detection of bone destruction or a paraspinal mass does not require the use of contrast material.
Although radionuclide studies are sensitive, they can be time-consuming, and they have lower spatial resolution. The incidence of false-negative scans is low in neonates and in elderly patients with osteomyelitis.
The radiographic appearance of osteomyelitis may vary (see the images below). Radiographs performed early in the course of disease may show subtle swelling of the deep soft tissue or edematous subcutaneous soft tissues, but radiographs are often normal in the first 7-10 days of infection. By 10-14 days, a focal area of bone opacity develops in the metaphysis. This progresses to lytic destruction with an associated focal periosteal reaction. Radiographs typically show a well-defined, longitudinally orientated, ovoid lucency with surrounding sclerotic margin but little or no periosteal new-bone formation.
Radiographs may be normal, particularly early in the course of disease. Alternatively, they may demonstrate soft tissue swelling, periosteal reaction, subperiosteal bone resorption, and erosions and sequestra. The extension of infection through the metaphyseal cortex may lead to periosteal new-bone formation. If untreated, this may completely encircle the bone, becoming an involucrum, which can envelop the nonviable infected bone; the result is called a sequestrum.
Degree of confidence
Plain radiography is an inexpensive and noninvasive technique that is readily available worldwide. Radiography has reasonable sensitivity. Plain radiographs may help in differentiating varieties of bone lesions that may mimic osteomyelitis clinically. However, radiographs are the least sensitive method of diagnosis.
Lipman at al reported a sensitivity of 67%, a specificity of 40%, and an accuracy of 50% for radiography.  On plain images, soft tissue swelling may be seen 1-3 days after the start of infection. Destructive bone changes do not appear on plain images until 10-14 days after the start of infection. Initially, the bone may have a lucent, moth-eaten appearance.
Radiographs are often normal in the first 7-10 days after the start of infection. Radiographic mimics of osteomyelitis include septic arthritis, Ewing sarcoma, osteosarcoma, juvenile arthritis, sickle cell crisis, Gaucher disease, stress fractures, and other bone lesions that may mimic osteomyelitis clinically.
CT is the modality of choice for revealing sequestra and cortical erosions in cases of chronic osteomyelitis. 
Chandnani et al investigated acute osteomyelitis and abscesses induced in the proximal tibia and surrounding soft tissues of 67 New Zealand white rabbits and found that MRI was more sensitive than CT in the detection of osteomyelitis (94% vs 66%) and abscesses (97% vs 52%). The specificity of MRI was equal to that of CT in the exclusion of osteomyelitis (93% vs 97%), but the specificity of MRI was less than that of CT in the exclusion of abscesses (77% vs 100%). The overall accuracy of MRI was somewhat, although not significantly, greater than that of CT in the detection of both osteomyelitis (93% vs 80%) and abscesses (87% vs 75%). 
Piazza et al evaluated the role of CT in the assessment of 15 patients with acute orbital infections and emphasized that CT is the method of choice for confirming the clinical diagnosis and for demonstrating the site, extent, and complications of acute orbital cellulitis. 
In a study by Hald and Sudmann of CT for the early diagnosis of acute hematogenous osteomyelitis, the results indicated that with CT, bone marrow involvement may be detected in patients with osteomyelitis before bony changes appear on routine radiographs. 
Azouz studied 14 patients with proven septic arthritis, osteomyelitis, or spondylitis, as determined by both CT and conventional examinations,  and found that when specific problems of diagnosis were unsolved after plain radiography, standard tomography, or isotopic bone scanning, CT was of definite value for studying the entire articular surface of the bone and periarticular soft tissues. They also found that CT was able to delineate the extent of medullary and soft tissue involvement and to demonstrate cavities, serpiginous tracts, sequestra, or cloacae in osteomyelitis. In addition, CT sometimes showed soft tissue edema or bone destruction not seen on plain radiographs.
Degree of confidence
CT scanning allows for 3D examination of the bone and surrounding soft tissue. CT is excellent for depicting periosteal new-bone formation, cortical bone destruction, and sequestration or involucrum, if present. CT is the modality of choice for revealing sequestra and cortical erosions in chronic osteomyelitis. The accuracy, sensitivity, and specificity of CT in assessing chronic osteomyelitis are reportedly 96.7%, 99.1%, and 80.0%, respectively.
Hald and Sudmann found that medullary CT attenuation values were invariably increased in their study of 7 patients with acute osteomyelitis. Thus, CT may show the bone marrow involvement of osteomyelitis before bony changes appear on plain radiographs.
Technologically sophisticated procedures such as CT and MRI should be reserved for situations in which the diagnosis cannot be made by use of simpler methods, as in cases of osteomyelitis of the spine or pelvis or in cases in which the anatomic detail provided by MRI is required for surgical planning. Most children with uncomplicated osteomyelitis probably do not need to undergo CT or MRI. CT diagnosis appears particularly good when osteomyelitis secondary to sinus infection is suspected.
Theoretically, confusion may arise in cases involving other conditions associated with periosteal new-bone formation and destructive bone processes, such as primary or metastatic bone tumors.
Magnetic Resonance Imaging
On T1-weighted MRIs, osteomyelitis typically has low signal intensity; on T2-weighted and short-tau inversion recovery (STIR) images, it has high marrow signal intensity (see the image below).
MRI shows marrow edema and the extent of subperiosteal abscess collection. T1-weighted sequences with intravenous gadolinium enhancement allow the differentiation of enhancing, hyperemic inflammatory tissue from central pockets of nonenhancing pus.
Gadolinium-based contrast agents (gadopentetate dimeglumine [Magnevist], gadobenate dimeglumine [MultiHance], gadodiamide [Omniscan], gadoversetamide [OptiMARK], gadoteridol [ProHance]) have been linked to the development of nephrogenic systemic fibrosis (NSF) or nephrogenic fibrosing dermopathy (NFD). The disease has occurred in patients with moderate to end-stage renal disease after being given a gadolinium-based contrast agent to enhance MRI or MRA scans.
NSF/NFD is a debilitating and sometimes fatal disease. Characteristics include red or dark patches on the skin; burning, itching, swelling, hardening, and tightening of the skin; yellow spots on the whites of the eyes; joint stiffness with trouble moving or straightening the arms, hands, legs, or feet; pain deep in the hip bones or ribs; and muscle weakness.
The disease is generally milder in infants than in older children. In infants, subperiosteal abscess formation is more frequent and rapid, and rupture into the soft tissues is more common. In neonates, vascular channels penetrate the cartilaginous epiphysis and allow infection to spread from the metaphysis into the joint, causing adjacent septic arthritis. These changes are demonstrable on sonograms or MRIs.
Degree of confidence
MRI is considered to be the most sensitive modality for the evaluation of osteomyelitis. The sensitivity is relatively high—about 85%. MRI offers improved soft tissue resolution and multiplanar abilities, and it provides guidance for tissue sampling.
In a study by Erdman and associates, 0.35-T MRI interpretations were compared with final diagnoses in 110 patients with suspected osteomyelitis,  and the diagnostic criteria of dark marrow on T1-weighted images and bright marrow on STIR images yielded a prospective sensitivity of 98% and a prospective specificity of 75%. About 60% of uncomplicated septic joint effusions demonstrated abnormal marrow signal intensity that was mistaken for that of osteomyelitis.
Their retrospective review revealed that overall specificity could be improved to 82% without a loss of sensitivity if increased marrow signal intensity on T2-weighted images was included as an additional criterion. Specificity could be increased further by using knowledge of the morphologic patterns that distinguish various forms of osteomyelitis. Ten patients (9%) had potential pitfall diagnoses (eg, fracture, infarction, healed infection) that mimic osteomyelitis. MRI can be sensitive and specific for osteomyelitis if characteristic appearances and pitfall diagnoses are incorporated into the diagnostic criteria.
The incidence of pyogenic spondylodiscitis is rising in the elderly population, immunocompromised patients, diabetic patients, drug addicts, and patients with sickle cell disease. Leone et al reviewed the spine imaging of 118 patients with spinal infections. All patients underwent radiography, CT, and MRI. MRI greatly contributed to prompt diagnosis, allowing prompt treatment. Prompt diagnosis and treatment are essential to prevent serious bone and joint destruction and severe neurologic sequelae. 
The physical properties of bone do not usually lend themselves to ultrasonographic investigation, because of the reflection of sound waves at a soft tissue–bone interface. However, the periosteum, early new-bone formation, and soft tissue changes alongside dense bone may be imaged. Ultrasonography is useful in the acute phase and shows the presence and extent of any subperiosteal abscess collection (for which aspiration or drainage is required). Thus, it may be helpful in planning surgery. [21, 22]
In a study by Taneja and associates regarding the role of ultrasonography in the early detection of bone infectionsm,  a hypoechoic collection adjacent to bone was considered to be highly suggestive of osteomyelitis, whereas a hypoechoic collection away from the bone implied a soft tissue abscess. Cellulitis appeared as increased subcutaneous thickness. Of the 31 cases of clinically suspected osteomyelitis that were studied with sonography, 25 were proven to be osteomyelitis on the basis of surgical or subsequent radiologic findings. Four involved soft tissue abscesses, and 2 involved cellulitis. Coexistent hip joint effusion was seen in 2 patients. The investigators believed that ultrasonography is a simple and noninvasive investigation that can be used to detect bone and soft tissue infections.
Abernethy and associates found that sonographic results distinguished superficial cellulitis, soft tissue abscess, and subperiosteal abscess in 9 children with late-presenting acute osteomyelitis, in 4 children with typical superficial cellulitis, and in 4 with a soft tissue abscess.  The abscesses were confirmed at surgery. A subperiosteal abscess was also detected in a child with deep periosseous cellulitis. Ultrasonography was particularly useful in confirming a subperiosteal abscess and in precisely localizing it in children with diffuse swelling and tenderness of a limb caused by late-acute osteomyelitis. Surgical drainage of pus may be avoided in patients in the absence of sonographic evidence of an abscess. Sonography is useful in the planning of surgery for those patients who require it.
Kang and associates examined 24 children with clinically suspected acute early-stage hematogenous osteomyelitis by using sonography.  Subperiosteal abscesses were sonographically detected in all patients at 4-14 days after disease onset. The mean length and anteroposterior distance of the subperiosteal abscesses were 86.4 and 10.7 mm, respectively. Of 24 cases of subperiosteal abscesses, aspiration performed under sonographic guidance revealed purulent fluid in all patients; 23 cases were verified surgically. The results indicated that sonography may be used to diagnose acute hematogenous osteomyelitis in the early stage. The earliest case was diagnosed by means of sonography 4 days after disease onset.
Mah and associates studied the sonographic features of acute osteomyelitis in children, and deep soft tissue swelling was the earliest sign of acute osteomyelitis. In the next stage, there was periosteal elevation and a thin layer of subperiosteal fluid; in some cases, this progressed to form a subperiosteal abscess. The later stages were characterized by cortical erosion, which was common in those who had had symptoms for more than 1 week. Concurrent septic arthritis was revealed in 11 patients, most frequently in association with osteomyelitis of the proximal femur or distal humerus. Four weeks after clinical cure, sonography revealed no abnormalities. [26, 27]
According to Chao and associates, who evaluated color Doppler sonography in the diagnosis of acute osteomyelitis in 12 children with clinically suspected acute osteomyelitis,  color Doppler showed flow within or around the infected periosteum in patients who had had symptoms for 4 days or longer, whereas no such flow was observed in patients who had had symptoms for less than 4 days. During sonographic follow-up, 6 patients had increased color Doppler vascular flow within and around the affected periosteum, 2 of whom had periosteal abscess. They eventually required surgical treatment.
Persistent or increased color Doppler flow during follow-up examination was correlated with elevated serum levels of CRP. This study indicated that color Doppler vascular flow within or around the infected periosteum was correlated with advanced acute osteomyelitis; surgery usually was required in the patients. Thus, color Doppler sonography allowed the detection of advanced osteomyelitis and revealed the progression of inflammation during antibiotic therapy. The authors postulated that color Doppler ultrasonography may be valuable in determining the efficacy of antibiotic therapy and in justifying the need for an operation.
In a study of 24 infants with osteomyelitis, Riebel and associates  identified intra-articular fluid collections, subperiosteal abscess formation, or both as the most frequent early sonographic findings. Such findings preceded radiographic changes by several days in 11 cases. With positive clinical signs of inflammation, the sonographic results were usually sufficient for a correct diagnosis. In addition, sonography was also able to depict superficial cortical erosion and even an intramedullary focus in a young patient.
Degree of confidence
Ultrasonography is a noninvasive, simple, and inexpensive technique that uses nonionizing radiation; thus, it is an ideal modality for children. Ultrasonography is fairly reliable for differentiating acute hematogenous osteomyelitis from cellulites, soft tissue abscesses, acute septic arthritis, and malignant bone tumors. However, the modality remains operator dependent, and a child with acute tenderness at the site of suspected infection may not be able to tolerate the ultrasonic probe touching the surface.
Larcos and associates found that the sensitivity of ultrasound was 63% in 16 patients with acute osteomyelitis.  They prospectively examined 19 patients with high-resolution sonography for subperiosteal fluid or cortical irregularity. The diagnosis was established at surgery (3 cases) or with other tests and at clinical follow-up. Sixteen patients were identified as having osteomyelitis, and positive sonographic results were obtained in 10 of the patients. Two studies yielded false-positive results; the diagnostic accuracy was 58%. Thus, ultrasonographic results may be potentially misleading, and the authors emphasized the importance of clinical judgment and the use of other tests.
Technetium-99m diphosphonate bone scan is usually positive 24 hours after infection; a well-defined focus of tracer activity is demonstrated 1-2 hours after injection. This finding is correlated with the presence of radiotracer in the same area on dynamic scans. Bone scintigraphy may show focal uptake at the affected site; it is particularly valuable in looking for other sites of infection, because multifocal osteomyelitis may occur, especially in neonates (see the images below). [31, 32, 33]
Three-phase bone scanning
This technique uses the same radionuclide as is used in static bone scans. The 3 phases involve the following:
Phase 1: Radionuclide angiography representing the perfusion phase is performed
Phase 2: Blood-pool images are obtained
Phase 3: Static bone scans are obtained
The sensitivity of the 3-phase bone scan in the diagnosis of acute osteomyelitis is higher than that of static bone scans. The sensitivity and specificity of 3-phase bone scans has been reported to be 85-92% and 54-87%, respectively. The uptake on 3-phase bone scans is related to blood flow and osteoblastic activity. In cases of acute osteomyelitis, isotopic activity is increased in all 3 phases.
Mechanisms of gallium-67 (67Ga) citrate uptake include the following: (1) direct leukocyte and bacterial uptake, (2) lactoferrin and transferring binding, (3) increased vascularity, and (4) increased bone turnover.
Criteria for a positive result on gallium scanning include uptake exceeding that seen on the bone scan, and/or differences in distribution, as compared with distribution seen on the bone scan.
If patients with suspected acute osteomyelitis are currently untreated and if 99mTc diphosphonate and gallium scans show concordant pattern, the scans are interpreted as follows: If 99mTc diphosphonate uptake is less than gallium uptake, infection is suggested. If99m Tc diphosphonate is present, reactive bone is suggested. If 99mTc diphosphonate and gallium uptake are discordant and if uptake is truly in bone, the likely diagnosis is osteomyelitis.
Johnson et al used gallium scans to evaluate 22 diabetic patients with osteomyelitis,  and the results yielded a sensitivity of 100%, a specificity of 40%, and an accuracy of 73%.
In Schauwecker et al's review of the literature, the sensitivity was 81% and the specificity was 69%. 
Indium-111-labeled leukocyte scintigraphy
Indium-111-labeled leukocytes become localized in infectious and inflammatory lesions by means of leukotaxis. Indium-111 is regarded as the best available agent for acute infections. However, Tc-labeled leukocytes have largely replaced indium labeling, especially in studies of suspected osteomyelitis in the extremities.
Images are obtained at 2-4 hours and at 24 hours. Drawbacks include the low count rate, the cost of the radiopharmaceutical preparation, the complexity of the labeling, and the lack of bony landmarks. Johnson et al reported a sensitivity of 100%, a specificity of 70%, and an accuracy of 86%.  When combined with bone scanning, the specificity increased to 80%, and the accuracy increased to 91%.
Neuropathic joint disease and osteomyelitis in diabetes
Diabetes mellitus affects 5% of the US population. Twenty percent of adult hospitalized diabetic patients have foot disorders associated with significant disability. One third of patients have evidence of osteomyelitis. Contributing factors include angiopathy and peripheral neuropathy.
Osteomyelitis in a diabetic setting presents a difficult clinical problem, and differentiating between cellulites, osteomyelitis, and neuropathic osteoarthropathy may be problematic. These conditions most often affect the feet. Neuropathic osteoarthropathy may cause a warm and swollen extremity even without concomitant infection. Plain radiographs of osteomyelitis may show destructive changes, osteosclerosis, and periosteal new-bone formation.
Fractures may occur with minimal trauma in cases of neuropathic osteoarthropathy and may be a complicating factor. Plain radiographs of neuropathic osteoarthropathy may depict joint subluxation, fracture fragmentation, subchondral osteoporosis, sclerosis in adjacent bone, and periosteal new-bone formation. Differentiating these changes from an infection may be difficult radiographically. The results of bone scintigraphy may be positive in all 3 conditions.
Gallium-67 concentrates in inflammatory lesions and may add some diagnostic specificity. Its intense concentration is thought to indicate osteomyelitis. However, scans may show mildly increased uptake in cases of diabetic neuropathic osteoarthropathy; occasionally, accumulation is substantial, compounding the problem.
Indium-111-labeled white blood cells are specific for acute osteomyelitis and are helpful in the differentiation of nonseptic from septic osteoarthropathy in diabetic patients. Maurer et al conducted a retrospective study of 13 diabetic patients in whom111 In-labeled leukocyte studies were performed to assess possible osteomyelitis; it showed increased uptake in cases of both septic and nonseptic osteoarthropathy. The sensitivity was 75% and the specificity 56% for osteomyelitis. With leukocyte imaging, the sensitivity was the same as that of 3-phase scintigraphy; however, the specificity was 89%, making leukocyte imaging more helpful for excluding infection.
According to Schauwecker et al,  111In leukocyte imaging had a sensitivity of 100% for acute osteomyelitis and a sensitivity of 60% for chronic osteomyelitis; the specificity was 95%. Gallium-67scanning was excellent for ruling out osteomyelitis when findings were normal or for ruling it in when scans showed hyperintense uptakeas compared with bone imaging or a different distribution from the bone images. This situation occurred in 28% of the patients studied.
Splittgerber et al evaluated 6 diabetic patients with radiographic findings of osteomyelitisosteoarthropathyor both by using leukocyte and bone imagingand 3 patients were found to actually have osteomyelitis. Bone images showed increased uptake in all patientswhereas leukocyte imaging showed increased uptake in only 3 of the patients with osteomyelitis. 
In conclusion,111In-labeled leukocyte imaging may add further specificity in differentiating septic diabetic osteoarthropathy from nonseptic diabetic osteoarthropathy in patients with increased uptake on bone images, gallium images, or both.
Degree of confidence
Static 99mTc diphosphonate bone scans have a sensitivity of 83% but a 5-60% false-negative rate in neonates and children because of the masking effect from normal increase in activity in the epiphyseal plate. Another factor is the unusual spectrum of radioactivity in children with osteomyelitis. Initially, the lesion may appear photon deficient, but it eventually appears as an area of enhanced activity.
The 3-phase bone scan has a higher sensitivity and specificity, but it also has several limitations in some patients: in children; in those in posttraumatic and postoperative states; in diabetic patients with neuropathic osteoarthropathy; and in those with bone tumors, septic arthritis, healed osteomyelitis, Paget disease, or other noninflammatory bone lesions. Confusion may also arise in cases cellulitis, although activity in cellulitis decreases with time.
The sensitivity of gallium scans is 100%, and increased uptake appears a day earlier than on 99mTc diphosphonate scans. However, labeled-leukocyte scanning has largely replaced gallium scanning in cases of acute osteomyelitis because of its improved photo flux and its improved dosimetry, which allows faster scanning with better resolution.
Bone scintigraphy with 99mTc diphosphonate is a highly sensitive technique for evaluating bone pathology, but it is not specific. False-positive results may occur with fractures, primary and metastatic neoplasms, heterotopic ossification, arthritis, osteomyelitis, and neuropathic joints. False-negative bone scans may be seen if scanning is performed early in the course of infection or if scanning is performed in babies younger than 6 weeks.
On scans, cold lesions may occur in the first hours to days of the infection as a result of increased pressure in the medullary space. Therefore, the lesion may easily be missed.