BackgroundDetermining bone marrow involvement (BMI) is a crucial element for staging of lymphoma. While the standard procedure to evaluate BMI has traditionally been bone marrow biopsy, biopsies are subject to sampling error, particularly if the involvement is focal and outside the pelvis. 18F-FDG PET/CT scans have become an increasingly popular component of the pretreatment evaluation to assess nodal and extramedullary disease. Their ability to accurately detect BMI has been suggested in Hodgkin's lymphoma, but is less well established in other histologies. This retrospective study evaluated whether 18F-FDG PET/CT scans are useful in detecting BMI in different types of lymphoma and, thus, may replace trephine biopsies as part of staging. MethodsBetween 2005 and 2013, 222 patients (pts) seen at our center underwent coinciding bone marrow biopsies and whole body 18F-FDG PET/CT scans. The most common lymphoma subtypes represented were diffuse large B-cell (DLBCL), follicular, and Hodgkin's lymphoma. Ninety-two pts were referred to our center for a new diagnosis and retrospectively enrolled on study. Unilateral bone marrow biopsy of the iliac crest was used as the standard for detecting BMI. 18F-FDG PET/CT scan was interpreted as positive for BMI when bone marrow 18F-FDG uptake was not otherwise explained by CT findings. ResultsOf the 92 newly diagnosed pts, there were 44 DLBCL, 28 follicular, and 20 Hodgkin's lymphoma. Most pts underwent 18F-FDG PET/CT scan prior to biopsy (47), as opposed to the same day (13) or after (32). The median age at diagnosis was 48 years (Range 22-89). Fifty-one of the patients were male and 41 were female.Seven of the 44 DLBCL patients had BMI documented by biopsy; 5 DLBCL, 2 follicular. When evaluating for only DLBCL marrow involvement, the sensitivity, specificity, and accuracy (concordance) of 18F-FDG PET/CT scan was 80% (CI 0.28, 0.99), 97% (CI 0.86, 1.00), and 95% (CI 0.84, 0.99) respectively. 18F-FDG PET/CT scan failed to identify 1 patient with focal DLBCL involvement. This pt already had advanced stage disease based on imaging, and would have received the same treatment regimen regardless of the extra information provided by bone marrow biopsy. When accounting for any kind of lymphomatous involvement, the sensitivity dropped to 57% (CI 0.18, 0.90) as 18F-FDG PET/CT scan failed to identify the 2 pts with follicular lymphoma of the marrow. Specificity and accuracy, however, remained high at 97% and 91%. When evaluating pts with relapsed disease, all DLBCL pts in our cohort had negative BMI by both biopsy and 18F-FDG PET/CT.In follicular lymphoma, however, the sensitivity of 18F-FDG PET/CT was 43% (CI 0.21, 0.73), specificity 93% (CI 0.64, 1.00), and accuracy 68% (CI 0.48, 0.84). Of the 9 pts with discordant results, 18F-FDG PET/CT failed to identify 8 pts with marrow involvement, 6 of whom had focal involvement. In the remaining pt, 18F-FDG PET/CT scan indicated appendicular skeletal and vertebral involvement. The bone marrow biopsy was negative in this pt, presumably due to the lack of iliac involvement. Information provided by bone marrow biopsy upstaged one of the 9 pts from limited to advanced stage disease. In the 12 pts with relapsed disease, the sensitivity was 47%, specificity 93%, and accuracy 68% based on 22 coinciding studies.In Hodgkin's lymphoma, the sensitivity, specificity, and accuracy of 18F-FDG PET/CT at diagnosis was 67% (CI 0.09, 0.99), 71% (CI 0.44, 0.90), and 70% (CI 0.46, 0.88) respectively. Of the 5 Hodgkin's pts with discordant results, 18F-FDG PET/CT scans detected marrow involvement in 4 pts with negative bone marrow biopsies. One pt had evidence of marrow involvement by biopsy but not by 18F-FDG PET/CT. As the pt was already stage III by 18F-FDG PET/CT, this information did not impact the treatment regimen. ConclusionsIn our cohort, 18F-FDG PET/CT was a highly accurate tool for detecting BMI in DLBCL and Hodgkin's lymphoma. As most pts underwent imaging first, the subsequent biopsy was unnecessary. In Hodgkin's, 18F-FDG PET/CT demonstrated the ability to detect BMI in pts who would have otherwise been considered to be negative by biopsy alone. 18F-FDG PET/CT was not as accurate in follicular lymphoma, presumably due to the low-grade nature of the disease. Further evaluation in a prospective manner is warranted, and may eliminate the need for a costly and painful procedure in many pts with lymphoma. Disclosures:No relevant conflicts of interest to declare.