ESTRO 38 Abstract book

S89 ESTRO 38

from Vex and Rex, was studied to demonstrate its correlation with late GI and/or GU toxicities in cervical cancer. Material and Methods 149 cervical cancer patients treated with 44-52 Gy EBRT without central shielding, parametrial or pelvic nodal boost, or re-irradiation with/without chemotherapy with Gynaecologic GEC-ESTRO II directed HDR-IGBT in 2012- 2015 were retrospectively reviewed. Late GI and/or GU toxicities were graded using CTCAE version 5.0. The iRex of each IGBT fraction was defined as ISV/Vneg . ISVs were transformed from isodose lines of fractional absorbed doses corresponded to EQD2 (α/β=3) of 60, 70, 80, 90 and 100 Gy. Vneg was the intercepted volume of ISV (60 Gy EQD2) and delineated toxicity-negligible regions including HR-CTV, uterus, and vagina. No further delineation of any OAR was performed. Numerical mean of iRex was established for each patient. The data of cumulative events of GI and/or GU toxicities was generated to determine the correlations between mean iRex and grade 2-4 late toxicities, using probit analysis.

and/or position) always occurred between CT and kV- CBCT acquisition. The mean relative percentage differences of D 2cc were: - 2.7% ± 15.6% for bladder, -1.2% ± 21.2% for rectum and 6.3% ± 24.1% for sigmoid.

Figure 1. Correlation between the planning CT and kV- CBCT doses for OARs.

Variations higher than 10% of EQD2 were found in 1/19 (5%) patients for bladder and 2/19 (10%) for rectum and sigmoid, while the (GYN) GEC-ESTRO recommendations for OARs dose limits were met in 15/19 (79%) patients for bladder and 17/19 (89%) for rectum and sigmoid. A weak yet statistically significant correlation was found between volume and D 2cc variations for bladder (R 2 =0.320, p=0.0001), but not for rectum and sigmoid. The time gap between CT and CBCT acquisitions was of 125 min ± 48 min (range: 47-204 min). No correlation was found between this time and OARs dose or volume OARs dose intra-fraction variations in HDR BT for cervical cancer are small on average, but large random variations were observed in individual patients. Differences are likely attributed to changes in volume for the bladder and to deformation and movement for the rectum and sigmoid. A kV-CBCT scan acquired before dose delivery can detect unfavorable anatomical changes, warranting adaptive strategies. Our data show that without adaptive planning, 21% of the patients will have a chance that at least one OAR will exceed the recommended limits, though a variation in EQD2 higher than 10% will occur in only 10% of them. OC-0178 Indirect Excess Dose Volume Ratio (iRex): A Novel predictor of Late Toxicity in Cervical Cancer IGBT T. Prasartseree 1 , P. Dankulchai 1 1 Faculty of Medicine- Siriraj Hospital- Mahidol University, Division of Radiation Oncology- Department of Radiology, Bangkok, Thailand Purpose or Objective Excess intermediate to high dose out of HR-CTV in IGBT associates with OAR toxicity. D2cc, D1cc, and D0.1cc demonstrate the dose-toxicity relationship with a limitation of identical locations of the most exposed volumes assumption. From observation, isodose surface volume (ISV) of intermediate dose usually covered the movable space of pelvic organs. To minimize effects of the organ motion in toxicity prediction, Excess Dose Volume (Vex) and its Excess Dose Volume Ratio (Rex) , transformed from ISV, were proposed and probably be used as IGBT dose constraints. Indirect Excess Dose Volume Ratio (iRex) , the more practical form derived variations. Conclusion

Results 50, 15, and 59 patients were affected from grade 2-4 late GI, GU, and combined GI and/or GU toxicities, respectively, with median follow-up time of 36 months. Mean iRex and toxicities correlations established the significant dose-response relationships for iRex60, iRex70 – late GI toxicities and iRex60, iRex70 – combined GI and/or GU toxicities. The effective ratio at 10 percent response values for iRex60 and iRex70 were 2.148 and 1.220 for grade 2-4 GI toxicities. For grade 2-4 combined GI and/or GU toxicities, the effective ratio at 10 percent response values for iRex60 and iRex70 were 2.039 and 1.153.