Dosimetric and Metabolic Parameters Predictive of Hematologic Toxicity in Cervical Cancer Patients Undergoing Definitive Chemoradiotherapy

Abstract

In cervical cancer, adding concurrent chemotherapy to definitive radiotherapy bestows a clear survival benefit, but the combination increases acute hematologic toxicity, potentially resulting in chemotherapy interruption or discontinuation. We performed a single institutional retrospective study of pelvic radiotherapy plans and FDG-PET/CT of cervical cancer patients to explore dosimetric parameters predictive of grade 3+ hematologic toxicity, with particular interest in evaluating metabolic (SUV) parameters and identifying the best predictive model. We hypothesize that metabolic metrics will be more predictive than bone metrics.Patients were included if they underwent definitive pelvic chemoradiotherapy, and the following information were available: (1) radiotherapy plan in MIM Radiation Oncology software, (2) diagnostic or treatment planning FDG-PET/CT, (3) baseline and serial CBCs. Pelvic bone and active marrow contours were retrospectively contoured by one radiation oncologist for uniformity. Active marrow was contoured as pelvic bone with SUV greater than mean pelvic bone SUV. Highly active marrow was contoured as the 10-14% volume of pelvic marrow with the highest SUV. Differences in toxicity were evaluated using logistic regression and linear mixed effects modeling.One hundred women from 2009 to 2020 were evaluable. The majority had FIGO stage (2018) IIIC1 disease (38%) and underwent IMRT (88%) with pelvic field alone (72%), prescribed to a mean dose of 48 Gy; 77% received a nodal boost to a mean dose of 58 Gy. The majority received weekly cisplatin (78%), and 82% (64/78) completed at least 5 cycles. The most common hematologic toxicity was leukopenia, grade 2 (41%, n = 41), then grade 3 (25%, n = 25). Significant predictors of grade 3+ hematologic toxicity were baseline WBC (P < 0.001), and 10 and 20 Gy dosimetric parameters to the active marrow, highly active marrow, and pelvic bone (specifically lumbosacral and lower pelvis subsites), all P-values < 0.050. Mixed model analysis demonstrated that the best predictive model of leukocyte and neutrophil trajectory were baseline WBC (P < 0.001), highly active marrow V20 (P < 0.001), and interactions between baseline WBC and time (P < 0.001) and highly active marrow V20 and baseline WBC (P < 0.001), such that patients with low baseline WBC experienced the greatest impact of highly active marrow V20.Baseline WBC was highly predictive of grade 3+ hematologic toxicity; dosimetric predictors included low dose (10 and 20 Gy) to the active marrow, highly active marrow, and pelvic bone (specifically lumbosacral and lower pelvis subsites), with potentially the greatest impact from V20 to the highly active marrow, particularly in women with low baseline WBC. Future studies should further evaluate optimal metabolic metrics incorporating baseline WBC to determine ideal radiotherapy plan parameters, and potentially better sparing of the highly active marrow in this population.E. Rahimy: None. R. Von Eyben: None. J. Lewis: None. D.H. Hristov: None. E.A. Kidd: Advisory Board; Varian Brachytherapy Advisory Board.