Hematologic Toxicity on RTOG 0418: A Phase II Study of Post-operative IMRT for Gynecologic Cancer

Reviewer: Lara Bonner Millar, MD
The Abramson Cancer Center of the University of Pennsylvania
Last Modified: November 3, 2010

Authors: A. H. Klopp1, J. Moughan2, L. Portelance3, B. E. Miller4, M. R. Salehpour1, D. D'Souza5, L. Souhami3, W. Small6, R. Gaur7, A. Jhingran1

1 M.D. Anderson Cancer Center, Houston, TX
2 RTOG Statistical Center, Philadelphia, PA
3 McGill University Health Centre, Montreal, QC, Canada
4 Wake Forest University Baptist Medical Center, Winston-Salem, NC
5 University of Western Ontario,London Regional Cancer Center, London, ON, Canada
6 Northwestern Memorial Hospital, Chicago, IL
7 St. Luke's Cancer Institute, Kansas City, MO


  • The hypothesis that reducing low-dose radiation to bone marrow (BM) can reduce hematologic toxicity (HT) in patients undergoing pelvic RT is based on evidence from multiple in vitro and in vivo animal studies which have demonstrated high radiosensitivity of hematopoietic stem cells.
  • Up to 50% of a patient's active BM is distributed within the proximal femurs, coccyx, sacrum, and lower lumbar spine, all of which are within conventional pelvic RT treatment fields.
  • IMRT can reduce the volume of bone marrow (BM) irradiated compared to conventional 4-field treatment. Radiation- induced BM damage is dependent on both dose and volume. A retrospective study by Brixey et al. reported that acute HT was reduced with IMRT compared with four-field box techniques in gynecologic cancer patients undergoing chemotherapy. Iliac, lumbar, and sacral BM irradiation was reduced with IMRT, even though the plans were not optimized to spare the BM.
  • The addition of concurrent chemotherapy in the treatment of cervical cancer has led to improved tumor control and survival, but there is increased toxicity, particularly hematologic toxicity (HT). Whether HT is enough of a problem in cervical cancer to warrant application of highly technical treatment is unknown, but retrospective studies suggest that pelvic BM sparing produces a clinically significant reduction of HT and increases the cycles of chemotherapy delivered.
  • The authors investigated HT on RTOG 0418, a prospective study designed to test the feasibility of delivering post-operative IMRT for cervical and endometrial cancer in a multi-institutional setting. Their study asked whether V10 and V20 (volume of bone marrow receiving 10 and 20 Gy) predicted for HT.

Materials and Methods

  • RTOG 0418 required post-operative IMRT to 50.4 Gy to the pelvic lymphatics and vagina with IMRT alone for endometrial cancer, and concurrent with weekly cisplatin in cervical cancer (40 mg/m2).
    • Patients were enrolled between 2006 and 2008.
    • Simulations were done with both full and empty bladder to evaluate vaginal motion. Nodal CTV and vaginal ITVs were contoured as per RTOG protocol. A 7 mm expansion from CTV to PTV. Patients were treated with full bladders.
  • HT during concurrent chemoradiation was compared to patients treated with conventional fields and chemoradiation on RTOG 9708 in which cisplatin (50 mg/m2) was delivered on weeks 1 and 5. HT which occurred during subsequent adjuvant chemotherapy on RTOG 9708 was excluded from analysis. HT was graded using the CTCAE v 3.0 criteria.
  • Pelvic BM was defined within the treatment field using a CT-density based auto-contouring algorithm (entire bone, including cortex was contoured). V10 and V20, dosimetric parameters correlated with HT in previous studies, were determined for 39 patients (30 with endometrial and 9 with cervical cancer).


  • Eighty-three patients were eligible for analysis (43 with endometrial cancer and 40 with cervical cancer).
  • Patients with cervical cancer treated with weekly cisplatin and pelvic IMRT on RTOG 0418 had the following HT
    • 23% grade 1, 33% grade 2 , 25% grade 3, 0% grade 4, 0% grade 5
  • Ninety percent of patients with cervical cancer received at least 4 (of five) planned cycles of cisplatin. Eighty-three percent received 5 cycles.
  • While there was no difference in grade 3 or greater toxicity between the cervical arm of RTOG 0418 and RTOG 9708 (25% vs. 31%, p = NS), there was significantly less grade 4 or higher toxicity on the cervical arm of RTOG 0418 than RTOG 9708 (0% vs. 18%, p = 0.002), despite the more frequent weekly chemotherapy on 0418. Most toxicity was grade 1-2.
  • The mean V10 for evaluated patients was 88.0 (min-max: 75.8% to 96.3%) and mean V20 was 74.3 (min-max: 56.9%-90.8%). Thirty-nine percent (11/28) of patients with V10>85% had Grade 1 or worse HT compared to 27% (3/11) of patients with a V10 ?85% (p = NS).

Author's Conclusions

  • Pelvic IMRT with weekly cisplatin is associated with lower rates of HT than historical controls treated with conventional 4-field plans.
  • Although IMRT was delivered to the same anatomical target for all patients, the extent of BM sparing varied widely, suggesting that pelvic IMRT can be optimized to limit BM dose.
  • Expanded analysis of this cohort will focus on identifying the dosimetric parameter which best predicts HT.
  • The study validated the predictive value of V10. A focus on bone marrow and bowel sparing will continue to be important based on this finding.

Clinical Implications

  • Although 90% of patients on this study were able to get 4 cycles of weekly cisplatin, it is not completely clear that this is due to the IMRT technique, given that chemotherapy dosing differed between RTOG 9708 and RTOG 0418. While acute Grade 3 or greater HT is relatively low with concurrent weekly cisplatin and IMRT, bone marrow sparing-IMRT might have more of a benefit with other chemotherapy that has greater HT (ie combination cisplatin and gemcitabine concurrent with radiation, which is being studied in a phase II trial).
  • Other outstanding with regard to radiation planning warrant further investigation in the future:
    • Hematopoietically active BM is not distributed uniformly throughout the pelvis, thus it is not clear which portions of the marrow should have the highest constraints within an IMRT algorithm.
    • Furthermore, although the desire is to limit bone marrow dose, the entire bone was contoured and it remains to be seen if there is value in restricting the entire bone as opposed to only the marrow.

From the National Cancer Institute