Bone marrow involvement in mantle cell lymphoma could be assessed using just 18fluorodeoxyglucose (FDG)–PET/CT, according to findings from a small, retrospective study published in.
, of the University of Colorado, Aurora, and his colleagues found that, at a certain threshold of bone marrow voxels in standard uptake value (SUV), there was 100% sensitivity and 80% specificity in determining bone marrow involvement in mantle cell lymphoma (MCL).
Currently, National Comprehensive Cancer Network guidelines call for bone marrow biopsy and whole body FDG PET/CT scan to complete an initial diagnosis of MCL.
“One of the most important factors for correct staging is the identification of bone marrow involvement, occurring in approximately 55% of patients with MCL, which classifies patients as advanced stage. However, accurate analysis of bone marrow involvement can be challenging due to sampling error,” the researchers wrote. “While bone marrow biopsy remains the gold standard, it is not a perfect standard given unilateral variability.”
In previous studies, FDG PET/CT was not considered sensitive enough to detect gastrointestinal or bone marrow involvement. However, these earlier studies used SUV maximum or mean or a visual assessment of the bone marrow activity, compared with hepatic uptake. To address this issue, the researchers developed a new method of examining SUV distribution throughout the pelvic bones by analyzing thousands of bone marrow voxels within the bilateral iliacs.
During the developmental phase, an institutional dataset of 11 patients with MCL was used to define the voxel-based analysis. These patients had undergone both unilateral iliac bone marrow biopsy and FDG PET/CT at the initial diagnosis. Then, FDG PET/CT scans from another 12 patients with MCL from a different institution were used to validate the developmental phase findings. Finally, a control group of 5 people with no known malignancy were referred for FDG PET/CT pulmonary nodule evaluation.
“The hypothesis of the study was that, if the bone marrow was involved by lymphoma, then there would be a small increase in the SUV of each voxel, reflecting involvement by the lymphoma. In order to capture such changes, we analyzed the percent of total voxels in SUV ranging from 0.75 to 1.20, in increments of 0.05, as this is where the greatest divergence was visually identified,” the researchers wrote. “The goal was to identify if a percentage of voxels at a set SUV could detect lymphomatous involvement.”
The researchers identified 10 candidate thresholds in the developmental phase; 4 of these performed better than the others in the validation phase. Using those thresholds, 10 of the 12 patients in the validation cohort could be correctly staged using FDG PET/CT.
Further analysis identified a single threshold that performed best: If greater than 38% of the voxels (averaging 1,734 voxels) demonstrated an SUV of less than 0.95, the sensitivity was 100% and the specificity was 80%.
The researchers acknowledged that the findings are limited because of the study’s small sample size and said the results should be validated in a larger trial.
There was no external funding for the study and the researchers reported having no financial disclosures.