A method for quantitative total marrow imaging (QTMI) with PET/CT

Abstract

The application of PET/CT to hematological diseases and bone marrow analysis has been stunted from a lack of standardization for bone marrow analysis, preventing comparisons across studies. The purpose of this study was to establish an automated, quantitative method to analyze total bone marrow, which we term quantitative total marrow imaging (QTMI). [F-18] fluorothymidine (FLT) PET/CT imaging can be used to image the proliferation of bone marrow. In QTMI, bone marrow is isolated from FLT images and analyzed on a whole patient level as well as on the skeletal region level. PET/CT images are segmented into multiple anatomical regions, such as the pelvis. Segmented skeletal regions are registered across time points to create voxel-to-voxel response images. QTMI was applied to two cohorts of patients—patients with healthy bone marrow and patients with acute myeloid leukemia (AML)—to demonstrate its applications to assess bone marrow activity and treatment response. Two sensitivity studies of QTMI were performed; QTMI outputs were evaluated for sensitivity to different SUV thresholds and for sensitivity to misregistration. We found that using a SUV threshold between 0.25 and 0.5 g ml−1 for QTMI results in minimal changes in SUV metrics. Response values were found to be robust to misregistration up to a single voxel (0.4 × 0.4 × 0.4 cm), resulting in an average change of only 5.1 ± 0.4% (range 0%–12%). The healthy bone marrow distribution from QTMI was not significantly different from previous studies. QTMI was applied to two AML patients in a feasibility study to demonstrate treatment response analysis, which when applied to a large patient population could provide valuable insight into patient response to therapy. QTMI provides a unique tool that enables reproducible identification and analysis of total bone marrow from FLT PET/CT images, improving quantitative imaging accuracy.