High SUV uptake on FDG–PET/CT predicts for an aggressive B-cell lymphoma in a prospective study of primary FDG–PET/CT staging in lymphoma

Background: Data assessing the role of positron emission tomography (PET)/computed tomography (CT) imaging in lymphoma staging is still being accumulated and current staging is based primarily on CT. This study aims to compare the value of PET/CT over conventional CT and bone marrow biopsy (BMB) in the initial evaluation of patients with lymphoma. Methods: Data on 122 patients with PET/CT scans as part of their initial staging were prospectively collected and reviewed. All patients had complete staging, including BMB. Results: Among the 122 patients, 101 had non-Hodgkin's lymphoma (NHL) and 21 had Hodgkin's lymphoma (HL). Compared with conventional CT, PET/CT upstaged 21 (17%) cases [B-cell non-Hodgkin's lymphoma (B-NHL), 12; T-cell non-Hodgkin's lymphoma (T-NHL), 3; HL, 6]. Of significance, in 13 patients with 2-[fluorine-18]fluoro-2-deoxy-D-glucose (FDG)-avid splenic lesions, four had normal CT findings. A maximum FDG uptake of >10 standardized uptake value (SUV) seems to significantly correlate with an aggressive B-cell lineage (odds ratio 2.47, 95% confidence interval 2.23–2.70). Overall, PET scan was concordant with BMB results in 108 (89%) and discordant in 14 (11%) cases. In HL, our data show that PET scan and marrow results agreed in 19 of the cases (90%), being concordantly negative in 18 cases and concordantly positive in one, giving a negative predictive value (NPV) of 100%, sensitivity of 100% and specificity of 90%. Of note, all 13 with early-stage HL had negative PET/CT scan and BMB. In NHL, all 17 cases of T-NHL had concordant PET and BMB results. In patients with aggressive B-NHL, BMB and PET/CT agreed in 58 patients (92%) and disagreed in five (8%), while the corresponding rates in indolent B-cell lymphoma were 14 (67%) and seven patients (33%), respectively. All seven were falsely negative. Conclusions: PET/CT upstages 17% of cases and detects occult splenic involvement. This may have potential therapeutic and prognostic implications. SUV >10 may predict for an aggressive histology. Except for indolent B-NHL, our data show that PET scans have a good overall NPV in excluding lymphomatous bone marrow involvement. This is particularly true of early-stage HL, suggesting that BMB may be safely omitted in this group.