Is ¹⁸F-FDG Standardized Uptake Value a Predictor of Disease Extent and Therapy Duration in Tuberculosis?

Objectives: 18F-FDG-PET/CT provides information on extent and activity of disease in neoplastic and inflammatory conditions. So far, little is known whether maximum standardized uptake value (SUVmax) of a reference lesion relates to severity of disease and will allow a prognosis regarding patient specific outcome parameters in tuberculosis (TB).

Methods: We retrospectively evaluated 188 cases referred to 18F-FDG-PET/CT for possible TB between 2008 and 2014, checking correlations between extent of disease, laboratory parameters, and numerical findings (SUV). In addition, we identified 37 patients (male : female = 21 : 16; mean age 50.4 ±18.4) for the evaluation of active TB before treatment initiation. The aim was to evaluate, whether the reference lesion SUVmax differed between patients with standard- (≤9 months) and long-term (>9 months) therapy and to define a SUVmax cut-off distinguishing those groups. Additionally, we analyzed the prognostic value of SUVmax regarding the estimation of initial disease extent and lesion distribution pattern. To determine the extent of disease we grouped organ systems in lung/pleura, intrathroracic lymph nodes, extrathroracic lymph nodes, bone, parenchymal organs, other (like skin, muscles) Results: Correlating the extent of disease or SUVmax (any lesion, lung or lymph nodes) to CRP, leucocyte count and other laboratory results returned very poor figures between -0.1 and 0.2; only SUVmax (top value in each single exam) vs. number of involved systems correlated rather well at 0.7, as did the lymph node SUVmax. A top SUV of 5.5 turned out as cut-off between cases of restricted spread (only one or two organ systems involved, mostly lung plus mediastinal lymph nodes) and cases with involvement of three or more systems. In the subgroup of 37, therapy duration was 10 (interquartile range (IQR) 3.25) months and median SUVmax was 8 (IQR 8.54). SUVmax was significantly (p = 0.036) higher in patients treated longer than 9 months (10.1 vs. 5.5). In ROC analysis, a SUVmax cut-off value of 10 could differ the standard- and the long-term therapy group with a sensitivity of 50% and a specificity of 78.6%, but did not reach statistical significance (p = 0.087). Also in the therapy subgroup a higher SUVmax corresponded with a higher number of affected organ systems (p = 0.012 for >), and patients with extrathoracic involvement showed significantly higher SUVmax than patients with limited intrathoracic disease (p = 0.024) and received significantly longer anti-mycobacterial therapy (p = 0.011).

Conclusions: Extent of disease is closely related to the top SUV seen in each 18F-FDG-PET/CT, any involved organ - the higher the SUV the higher the probability of wide-spread disease. If SUV did not exceed 5.5, never more than two involved organ systems were seen. This reference lesion SUVmax from pre-therapy scans may have prognostic properties in complicated cases of TB, especially regarding therapy duration and disease extent.