Nuclear medicine imaging of posttraumatic osteomyelitis

Bone scintigraphy is one of the oldest existing nuclear medicine techniques and still one of the cornerstones in nuclear medicine practice. Radiopharmaceuticals used for bone scintigraphy are diphosphonates coupled to the radionuclide Technetium-99 m (^99mTc). These bone-seeking radiopharmaceuticals selectively accumulate on the surface on bone mineral matrix in areas of high metabolic activity and therefore depict osteoblastic activity.

When a musculoskeletal infection is suspected, a three phase bone scintigraphy can be performed as a first screening tool (Fig. 2). As revealed by its name, this bone scintigraphy consists of three phases. The first phase is the perfusion phase, or flow study, performed dynamically, over the part of interest, for the first 2 min after administration of the radiopharmaceutical. The second phase is the blood pool phase, also performed on the part of interest, directly after the first phase (2–5 min after injection). The third phase, also called the static phase, depicts the incorporation of the radiopharmaceutical into the matrix of the bone and is usually performed 3 h after administration. This late phase can be combined with a SPECT-CT to localize the area(s) of increased bone metabolism. All three phases are necessary in case of suspected bone infection, since the three phases characterize both the vascularization and the metabolic activity of a process.

The three-phase bone scan can be used as a first screening method for diagnosing PTO. Because of its good availability it can mostly be performed short (<24 h) after the request of the referring clinician and it is relatively cheap. A normal bone scan (no increased perfusion and blood pool, no uptake in the late phase) rules out almost completely an existing bone infection (high sensitivity). The role of the bone scintigraphy, however, in the acute setting is neglectable, since the specificity is rather low and uptake is visible in all sites of increased bone metabolism irrespective of the underlying cause. A positive bone scan with an increased vascularity and increased metabolic uptake may indicate PTO; yet it can also indicate healing fracture(s) or a postsurgical situation. Furthermore, in a low-grade infection even the first two phases can be negative, so the late phase is essential and when positive it may be the only indication of an infection. Literature studies trying to find out at which time point a bone scan becomes negative after fractures and/or surgery are scarce. It is known that a bone scan may be positive for at least 2 years after total hip arthroplasty (THA) and 5 years after total knee arthroplasty (TKA) due to physiological bone remodeling after implantation [5]. We do not know exactly the time frame in which the bone scan is definitely positive following trauma, fracture or after open reduction and internal fixation (ORIF) of a fracture. Probably this time period will be around 1–2 years.

In conclusion, there is no role for a bone scintigraphy for diagnosing a SSI or early PTO. There is probably a role (when negative it excludes an infection) in the long-standing PTO, but a positive bone scintigraphy must be interpreted with caution and other imaging methods are necessary to differentiate between an infection and other causes of increased osteoblastic activity.

The conventional bone scan as mentioned above is still the gold standard in bone imaging. The images are acquired on a gamma camera and most newer camera systems have also the possibility to include SPECT-CT in the imaging process. However, there is also a PET tracer for bone imaging which uses the radiopharmaceutical ¹⁸F-sodium fluoride (¹⁸F-NaF). The uptake mechanism of ¹⁸F-NaF resembles that of ^99mTc-labelled diphosphonates. The faster blood clearance and the twofold higher uptake in developing bone cells of fluoride make it possible to image faster (1 h after injection) and lead to better ratios between pathological and physiological bone uptake [10]. The advantages of using this PET tracer is the better resolution and better quantification possibilities. Limitations, however, are the higher costs and the lower availability worldwide of this techniques, and the non-possibility to perform flow and blood pool imaging. At this moment, the classical bone scan with labelled diphosphonates remains the gold standard when a bone scan is indicated; the ¹⁸F-NaF-PET could be considered for the individual patient.

Scintigraphy using labelled autologous white blood cells (WBC scintigraphy or leukocyte scintigraphy) was developed in the 1970s and is still the gold standard nuclear medicine technique for infections in the musculoskeletal system. It is a specific indicator for leukocyte infiltration into infected bone and soft tissue and is highly specific, since the WBCs accumulate by active migration to the infection. Over time, there have been major developments regarding how to correctly acquire, analyze and interpret the images, which eventually led to a high diagnostic accuracy. Also the possibility to better anatomically localize the infection due to the addition of SPECT-CT helped reach these good results mentioned in the literature.

Despite the high diagnostic accuracy, the whole procedure itself has limitations. First of all, 50–100 cc of blood has to be collected from the patient. Then, the preparation of the labelled (preferably with ^99mTc-HMPAO) white blood cells is laborious and time consuming (2–3 h) and must be performed under sterile conditions and strict regulations [11]. Subsequently, the labeled autologous leukocytes are reinjected into the patient. Finally, at least two imaging time points are necessary: 3–4 h after reinjection and 20–24 h after reinjection. This dual-time point imaging has to be performed since the accumulation of leukocytes in the infection is a dynamic process: it is the increase in size or intensity in time that indicates the presence of an infection (Fig. 3). When there is a decrease or the uptake is stable in time, then there is no infection but inflammation or physiological bone marrow uptake [8]. This change in uptake in time can be determined visually, but sometimes semi-quantitative evaluation can be a helpful tool as an addition to visual assessment. This is done by calculating ratios between the infectious focus and the contralateral side as background. Again, increase of the ratio in time points to an infection. Due to disintegration of the used radionuclide (^99mTc), the total acquisition time of the late images has to be prolonged accordingly to the half-life of the tracer to establish identical image quality. In accordance with the bone scintigraphy, a SPECT/CT can be performed to exactly localize the leukocyte uptake. The proposed correct procedure of acquisition and analysis of the scans are stated in a recent publication [12].

The role of WBC scintigraphy in peripheral osteomyelitis is extensively studied. Prandini et al. described in a meta-analysis of published papers up to December 2005, in almost 3600 cases, a diagnostic accuracy of 89 % [13]. In the included studies, however, different acquisition protocols and interpretation criteria were used. Furthermore, SPECT-CT did not exist at that time. So, probably the diagnostic accuracy is even higher when using the correct and standardized protocols and add the complementary information obtained by SPECT-CT. This was confirmed in two recent retrospective studies using these correct protocols in, respectively, 61 and 31 patients with peripheral osteomyelitis. The diagnostic accuracy in these studies was found to be very high: 97 and 100 %, respectively [12, 14].

The difficulty with interpreting the literature on the accuracy of WBC scintigraphy for diagnosing PTO is that most of these studies included patients with peripheral skeletal infections in general (including haematogenous osteomyelitis and prosthetic joint infections). Recently, we performed a systematic review of the recent literature (2000–2015) on the role of imaging modalities in patients with PTO (data not published yet). Only studies were included in which data for at least 10 patients with PTO were available, and a valid reference test (proven by histology or bacteriology, and/or clinical follow-up of more than 6 months) was described. Unfortunately, only 10 studies could be included of which 5 were performed with WBC scintigraphy (only one study reported a diagnostic accuracy, which was 98 %). So, despite the extensively available data of WBC scintigraphy in peripheral osteomyelitis in general, there is a lack of studies really focusing on PTO. Despite this disappointing finding, we believe that there is an absolute role for WBC scintigraphy in diagnosing PTO. This is based on expert opinion and best available evidence on osteomyelitis in general since it is the only existing imaging modality that is a specific indicator of an infection. Therefore, we retrospectively reviewed the diagnostic value of WBC scintigraphy ± SPECT/CT in 114 patients with suspected PTO in our hospital. Sensitivity, specificity, positive predictive value and negative predictive value were 89, 95, 86 and 97 %, respectively (data to be published).

The glucose analogue FDG is already extensively used in oncology for over a decade. It can also be used in infectious diseases, because activated leukocytes, monocytes, lymphocytes, macrophages and giant cells all use glucose as their energy source. To minimize FDG uptake in normal tissue, patients must fast for at least 4–6 h to reduce competition for glucose transporters on the cell membrane. After the injection of the labelled FDG (¹⁸F-FDG) patients must rest for an hour and limit physical activity to minimize muscle uptake and obtain a good biodistribution in the body. High contrast images of infectious lesions can be obtained with this technique. The use of FDG-PET has many advantages: no blood manipulation, high spatial resolution, one imaging time point which is already 60 min after injection, one-stop-shop possibility with diagnostic CT, etc. It is therefore an essential tool when searching for an infection or inflammation in a patient with fever of unknown origin or to establish the source of dissemination of infectious lesions in the body in a patient with an haematogeneous spread of the infection (Fig. 4).

Unfortunately, the uptake in both inflammatory and infectious cells makes this technique also very aspecific as it is often not possible to discriminate between inflammation and infection. No universal accepted interpretation criteria are available to declare a FDG-PET scan positive or negative for an infection, let alone for PTO. Furthermore, uptake of FDG in the peripheral skeleton results in a rather broad differential diagnosis. Osteomyelitis, soft tissue infection, inflammation, granulation tissue after surgery, reactive uptake around foreign body material, atherosclerosis, recent fractures and neuro-osteoarthropathy all lead to increased FDG uptake.

As is the case with WBC scintigraphy, most data are available for FDG-PET imaging in long-standing (chronic) peripheral osteomyelitis. Termaat et al. performed a systematic review and meta-analysis for the assessment of chronic osteomyelitis and found the best results for FDG-PET with a sensitivity of 96 % and a specificity of 91 % [15]. Similar results were found by Jamar et al., who pooled all available data (in total 287 cases) and found a diagnostic accuracy of 94.5 % [16]. However, most of these studies dealt with chronic osteomyelitis from a diverse etiology and the exact role of FDG-PET in the posttraumatic situation is not known. Theoretically, due to reactive inflammation, the performance may be worse with osteosynthesis in situ, as is the case in prosthetic joint infections.

In the same systematic review our group performed as mentioned earlier, only 6 FDG-PET studies (with or without combined CT scan) were identified that were published in the last 15 years and fulfilled the criteria of >10 patients with suspected PTO and a valid reference test. The reported diagnostic accuracies were between 67 and 91 % (data not published yet).

In 2013, a shared guideline for the use of FDG in inflammation and infection was published by both the European Association of Nuclear Medicine (EANM) and the Society of Nuclear Medicine and Molecular Imaging (SNMMI) in the United States. This guideline states that the evidence for the use of FDG in osteomyelitis remains low (2b at best), and that at this moment, WBC scintigraphy is the preferred imaging modality. If FDG-PET is used however it is best performed in the chronic peripheral non postoperative setting [16].

Patient A, a 37-year-old healthy male, underwent open reduction and internal fixation (ORIF) of an open fracture of his right distal tibia and fibula 22 years ago. This was complicated by posttraumatic osteomyelitis and resulted in multiple re-operations with debridements of the bone, removal of most hardware and free flap coverage of a soft tissue defect. He was referred to our hospital with a persistent clinical infection around his right distal tibia and a near wound breakdown of the scar. Medical imaging was requested (a) to confirm the diagnosis of osteomyelitis and (b) to determine the anatomical location of the suspected osteomyelitis.

First, according to the diagnostic imaging protocol in our hospital, a three-phase bone scan was performed since the fracture and surgery were >2 years ago. All three phases of the bone scan were positive, only the late phase (Fig. 5, g: anterior view, h: lateral view) is presented here. This increased osteoblastic uptake can be the result of an infection, but also due to a healing fracture or recent surgery. For further differentiation, the patient underwent a WBC scan (Fig. 5, a–d: images after 4 h, e–f: images after 24 h). This showed increased uptake in both intensity and size over time, suspect for an infection. To localize this accumulation of leukocytes a SPECT-CT was performed (Fig. 6) which revealed that the uptake was located outside the bone, in the soft tissue. Final diagnosis was a soft tissue infection.

Patient B, a 46-year-old schizophrenic but otherwise healthy male sustained an open and comminuted talar neck fracture after a fall from height. This was initially treated with multiple soft tissue debridements and an external fixation, later augmented with screw and K-wire fixation of the fracture. The soft tissue defect was closed with a local myocutaneous flap. Two months after this last procedure, the patient presented with a draining sinus in the scar on the lateral side of the ankle joint. Medical imaging was requested to (a) assess the viability of the talus and (b) to determine the anatomic location of the suspected osteomyelitis. The X-ray is shown in Fig. 7 (left image).

To answer question (a) a three phase bone scan was performed. The first phase (flow phase) is shown in Fig. 7 (upper row images): positive flow at the talar region of the left foot. Obviously also the second phase (blood pool phase, middle image lower row) and the late phase (combined with CT image, right image lower row) are positive. This means the bone is viable. However, differentiation between infection or healing fracture is not possible. Therefore, WBC scintigraphy was performed (Fig. 8, upper row: images after 4 h, lower row: images after 24 h). The uptake decreased in time, meaning that the leukocyte accumulation is the result of a healing fracture and not of an osteomyelitis. After proper wound care further healing was uneventful.

Patient C, a 33-year-old healthy male was referred to our hospital because of a suspected posttraumatic osteomyelitis in combination with a malunion of his left tibia. He sustained a gunshot wound to his left lower leg in the middle east conflict 2 years prior to this presentation which was treated with a prolonged immobilization in an external fixator combined with several wound debridements. The last operation was only a few months prior to presentation. On examination, apart from the obvious malalignment of his left lower leg, we noted a closed but unstable scar on the medial side of his left tibia. Medical imaging was requested to (a) confirm the diagnosis and (b) to determine the anatomic location of the suspected osteomyelitis.

Since his last surgery was <6 months ago, immediately a WBC scan was performed (Fig. 9, left image: anterior view after 4 h, right image: anterior view after 24 h): uptake is visible at three locations. When calculating the ratios (uptake focus-to-contralateral side) the uptake at the most proximal and most distal focus decreases in time. This means these uptakes are due to regeneration of bone marrow. However, the uptake at the middle focus increases in time, which is suspect for an infection. The SPECT-CT (Fig. 10) shows the uptake in the bone and a small fistula to the bone marrow. Indeed, surgery revealed an infection at this location.