Published in:
01-06-2017 | Original Research
A comparison of dosimetric parameters in PET-based active bone marrow volume and total bone volume in prediction of hematologic toxicity in cervical cancer patients treated with chemoradiation
Published in: Journal of Radiation Oncology | Issue 2/2017
Login to get accessAbstract
Objective
The purpose of this study is to evaluate whether dose to PET-defined active regions of the bone marrow is a better predictor of hematologic toxicity than dose to the total bone as a surrogate of the bone marrow.
Methods
Twenty-one patients with stage IB-IVB cervical cancer who had baseline PET imaging and underwent definitive cisplatin-based chemoradiation from 2010 to 2015 were reviewed. Total bone marrow (TBM) volume was defined as the bone from L4 to the coccyx, the pelvic bones, and proximal femoral heads. Active bone marrow (ABM) was defined by PET as the volume of the bone within the TBM greater than or equal to the mean total body standard uptake value (SUV). Dosimetric parameters evaluated were mean dose, V10, V20, and V40, which were compared by a t test. Hematologic toxicity was graded using Common Terminology Criteria for Adverse Events (CTCAEv4). Receiver operating characteristic (ROC) analysis was used to assess predictors of grade 3 or higher (grade 3+) hematologic toxicity.
Results
ABM volume (mean = 1227 mL, range 793–1671 mL) was significantly smaller than the TBM volume (mean = 1553 mL, range 1117–1920 mL) [p = 0.0001]. ROC analysis identified ABM volume (AUC = 0.800, p = 0.002) and V40 to the ABM (AUC = 0.800, p = 0.008) as the best predictors for grade 3+ hematologic toxicity. All ten patients with an ABM volume < 1201 mL had grade 3+ toxicity compared to 5/11 with an ABM volume > 1201 mL (Fisher’s p = 0.012).
Conclusion
A lower volume of ABM at a baseline (<1201 mL) was highly predictive of grade 3+ hematologic toxicity.