ESTRO 36 Abstract Book

S197 ESTRO 36 _______________________________________________________________________________________________

OC-0367 Dose-response curve for vaginal stenosis. Final results of a prospective study. M. Federico 1 , A. Tornero 2 , S. Torres 2 , B. Pinar 1 , M. Rey Baltar 1 , M. Lloret 1 , P. Lara 1 1 Hospital Universitario de Gran Canaria Dr. Negrín, Radiation Oncology, Las Palmas de Gran Canaria- Ca, Spain 2 Hospital Universitario de Gran Canaria Dr. Negrín, Radiación Physics, Las Palmas de Gran Canaria- Ca, Spain Purpose or Objective Vaginal stenosis as consequence of cervical cancer BT treatment severely impact quality of life. No dose constraints have been published so far. Aim of this study is to identify a threshold level for volume packing and a dose response curve for vaginal stenosis. Material and Methods 211 consecutive cervical cancer patients treated between 2008-16 (median FU time 42.4 months) with a median age at BT of 52.6 years (range 23.7- 88.5) were considered. All pts received 3DRCT (45-50 Gy with weekly concomitant CDDP 40 mg/m2 when feasible) and tandem ovoids HDR BT (or intracavitary-interstitial cylinder application when needed). Patients received simulation CT scan with radiopaque vaginal tube in place in order to delineate vagina from a plane tangential to lower border of pubic bone up to fornix. At BT vaginal packing (VP) was contoured from a plane tangential to lower border of pubic bone up above ovoids surface. Vaginal walls were delineated as a 2 mm expansion of packing subtracted of packing volume. 85 pts. (group A) received CT based BT (5 fractions of 5,5 Gy), 126 patients (group B) received MR based BT (4 fractions of 7 Gy). Group A pts had a treatment slightly optimized to OARs. Group B pts had a treatment optimized to OAR and HRCTV according GEC ESTRO recommendations. All patients entered prospective follow up. Morbidity was scored according CTCAE 4.0 vaginal volume was also measured with appropriate vaginal cylinders (diameters 10 to 45 mm).To assess the relationship between vaginal stenosis, VP and vaginal dose a median VP volume (VPm) among the 5 (group A) or 4 (group B) application each patient received was calculated. Moreover the cumulative EBRT+BT EQD2 dose to vagina was calculated. A Logistic model (LM) was used to analyze data. Results Results are summarized in Tab1. In 929 applications a double exponential fit was noticed between vaginal dose and VP volume, with a fast growing exponential part (minimal variations in VP volume corresponding to huge variations in vaginal dose), and a slow growing exponential part (variation in VP volume have modest impact on dose). VP volume cut off values dividing the two parts of the curve for all considered vagina DVH parameters were encompassed between 75 and 80 cc. LM showed good correlation (R 2 =0.97 and 0.96 respectively) between VPm and G3 or G2-3 vaginal stenosis (Fig1 A_B). Risk of vaginal stenosis G3 or G2-3 was less than 10% when a VPm volume >82 or 105cc was obtained. A dose response curve was found for G3 or G2-3 stenosis and vaginal EBRT+BT EQD2 D80 (R 2 0.99 and 0.98 respectively) with a risk of G3 or G2-3 stenosis lower of 10% when EQD2 dose parameters was lower than 63 and 44Gy EQD2 respectively (Fig1 C-D).

magnetic resonance imaging (MRI) scans reviews were carried out blind to the clinical data with focus on radiological parameters potentially correlated to the risk of VVF (necrosis, tumor height of bladder involvement, tumor volume). Times were calculated from the date of diagnosis. Survival were estimated using the Kaplan-Meier method and the Cox proportional hazards model. Results Seventy-one patients were identified. Bladder invasion was diagnosed either on imaging in 59% or endoscopically/histologically proven in 41%. All patients received pelvic external beam radiotherapy (EBRT), 45 Gy in 25 fractions ± nodal boost to macroscopically involved lymph nodes. Nineteen of the 21 patients with para-aortic nodal metastases received para-aortic EBRT. Concurrent platinum-based chemotherapy (CT) was used in 76%, neoadjuvant CT was used in 14%. After EBRT, 64 patients (90%) received uterovaginal BT (low-dose rate in 48%, pulsed-dose rate in 52%). Eight patients had VVF at diagnosis. Among the 63 patients without VVF at diagnosis, 14 patients (22.2%) developed VVF later on: four before (28.6%) and ten (71.4%) after BT (median time to onset: 3.5 months after the start of EBRT). Twelve of the 22 patients (54.5%) who presented VVF, either at diagnosis or during follow-up, needed surgery (urinary or bowel diversion ± pelvectomy). Estimated OS, PFS and LCR at 2 years were 57.3% (44.9-68.8), 45.0% (32.3-58.5) and 69.1% (54.4-80.7) respectively. Presence of para-aortic nodal metastases was significantly associated with poorer OS on multivariate analysis (HR=4, p<0.001). Only the presence of necrosis in the anterior part of the tumor on baseline MRI was strongly associated with the risk of subsequent VVF (57% vs O% at 1 year, HR=16.7, p=0.011 on a multivariate analysis taking into account the tumor volume). No correlation was found between bladder dose and risk of VVF.

Conclusion A curative intent strategy including BT as part of local treatment is feasible in patients with bladder invasion, with a rate of 22% of post-treatment VVF. MRI has a strong predictive value of VVF occurrence. This result has to be confirmed in an independent cohort. Prognosis remains poor in regard to lower-staged lesions, with a high risk of out-of-field failure. Intensification of systemic therapies should be considered.