The Diagnostic Value of Fluorine-18 Fluorodeoxyglucose Positron Emission Tomography/Computed Tomography for the Detection of Surgical Site Infection After Spine Surgery

Tomohide Segawa, MD; Hisashi Koga, MD, PhD; Masahito Oshina, MD; Masayoshi Fukushima, MD; Hirohiko Inanami, MD


Spine. 2021;46(10):E602-E610. 

In This Article

Material and Methods

Study Subjects

We retrospectively investigated 59 consecutive patients who underwent F-18 FDG PET/CT based on suspected spinal infection at Iwai Orthopedic Hospital from December 2008 to July 2018. The study protocol was approved by our ethics committee, and the patients gave informed written consent before the study. After excluding seven patients with no history of spine surgery, 52 patients were included in the study. All 52 patients were followed up by plain radiography (X-p) in addition to CT and/or magnetic resonance imaging (MRI) for potential SSI at least once a month for ≥6 months after the spine surgery. SSIs warranting F-18 FDG PET/CT were suspected when the patients exhibited fever and back pain, bacteremia and back pain, or increased C-reactive protein (CRP: >0.3 mg/dL) or white blood cell count (WBC: >9700 cells/μL) and back pain.

Final Diagnosis of SSI

The final diagnosis of SSI was based on the results of pathogen identification, plain radiography, CT and/or MRI, or response to antibiotics and/or reoperation. The diagnosis was determined when plain radiography, CT, and/or MRI imaging findings were indicative of SSI during the follow-up periods,[21–25] in addition to patient data which met at least one of the following:[26] purulent drainage; organisms isolated from an aseptically obtained culture of fluid or tissue from the incision, or in the organ/space; at least one of the following signs or symptoms of infection: fever, pain, or tenderness, localized swelling, redness, or heat; an abscess or other evidence of infection was found on direct examination, during reoperation, or by histopathologic or radiologic examination; and diagnosis of SSI by a surgeon or attending physician.

Among patients who did not meet these criteria for infection, patients treated with antibiotics for >7 days were classified as indeterminants, even if there were no imaging findings that were indicative of SSI during the follow-up period. The remaining patients were classified as noninfection.

F-18 FDG PET/CT Analysis

F-18 FDG was injected intravenously (up to 370 MBq), and scanning began 60 minutes later. No intravenous contrast agent was administered. Patients were encouraged to void before imaging and scanned with a combined FDG PET/CT scanner (GE Discovery ST Elite; GE Yokogawa Healthcare, Tokyo, Japan). Advantage Workstation 4.2 software (GE Yokogawa Healthcare) was used for image analysis. Image data were evaluated by two nuclear medical physicians blinded to the clinical and pathological results. SSI was rated as positive when focally increased uptake was seen within the bone with higher intensity compared to the uptake in the surrounding tissue (pattern-based diagnosis). The data were also quantitatively analyzed by SUVmax as an index of F-18 FDG uptake. Receiver-operating characteristic (ROC) analysis was used to determine the optimal cutoff value for SUVmax, using the area under curve with 95% confidence interval. SSI diagnosis was also performed based on the cutoff value obtained from the ROC analysis for SUVmax (SUVmax-based diagnosis).

Data Analysis and Statistics

Diagnostic yields based on F-18 FDG PET/CT data (both pattern-based and SUVmax-based diagnoses) were examined for accuracy, sensitivity, and specificity in comparison with the final diagnosis of SSIs by spine surgeons based on the overall data described above. Subgroup analyses were also performed based on the presence or absence of implants and timing after surgery (early phase, ≤12 weeks post-surgery; late phase, >12 weeks post-surgery). Statistical analysis was performed using SPSS statistics ver 24.0 (IBM Japan, Ltd., Tokyo, Japan), and a P value of <0.05 was considered statistically significant. The intergroup difference in mean SUVmax was tested for statistical significance using Student t test. The sensitivity and the specificity of both pattern-based and SUVmax-based diagnoses were tested for statistically significant difference using McNemar test. The positive predictive value and the negative predictive value were tested using the method of Moskowitz and Pepe.[27]