Prospective Validation of Bone Marrow Constraints for Hematological Toxicity in Patients Undergoing Chemoradiation for Cervical Cancer

Abstract

Dose volume constraints (DVCs) have been proposed for reducing hematological toxicity (HT) in patients undergoing chemo radiation (CRT). The present study was undertaken to investigate validity of the DVCs during three-dimensional conformal radiation (3DCRT) or intensity modulated radiotherapy (IMRT) for cervical cancer. Clinical database of patients receiving bowel sparing adjuvant IMRT or 3DCRT with weekly concurrent cisplatin (40 mg/m2) within context of an ongoing trial was reviewed. No prospective BM sparing was performed. RT planning scans were de-archived and BM was contoured using Whole Bone (WB) and free hand inner cavity (FH) technique. Segmentation into sub-regions namely whole pelvic lumbar, lumbosacral spine, sacrum, ilium, ischium, femoral head and neck, whole pelvis and lower pelvis was performed. WB and FH BM subvolumes receiving 10, 20, 30, 40 Gy were determined. Receiver operator characteristics analysis was performed to identify DVCs that predict for grade ≥ II HT. All data was dichotomized across these DVCs for univariate and multivariate analysis. Finally an independent sample t test was performed to identify treatment technique that best spares the BM sub-regions that predict for HT. Overall 128 patients (IMRT=73; 3DCRT=55) that received CRT were included. Grade ≥ II HT and ≥ III HT was observed in 33.3% and 4.3% respectively. Grade ≥ II anemia, leucopenia, neutropenia and thrombocytopenia was observed in 14.3%, 21.9%, 5% and 0.7% patients respectively. On univariate analysis none of the known predictive factors (V10 or V40) predicted for overall Grade ≥ II HT. However Grade ≥ II neutropenia was predicted by Whole pelvic lumbar V20 >75% (WB and FH, p< 0.04,0.02) and V40 FH>35% (p<0.02), Whole pelvic V10 WB>90% (p<0.03), Whole Pelvic FH V20>75% (p<0.03) and lower pelvis FH >55% (p<0.04) and femoral head and neck FH V10> 80% and V20>35% (p< 0.03 and 0.02 respectively). On multivariate analysis, whole pelvic lumbar FH V40 >35%, whole pelvic WB V10>90%, femoral head and neck V 20 >35% predicted for Grade ≥ II neutropenia. Of the identified DVCs, V40 and V20 are better spared with IMRT (p<0.0001) whereas V10 is better spared by 3DCRT (p<0.0001) BM sub-volume doses predict for Grade ≥ II neutropenia in patients receiving CRT. As both high dose and low dose regions are important, the planning aim for BM sparing IMRT should be defined accordingly in prospective studies.