ESTRO 2020 Abstract book

S621 ESTRO 2020

At 12 and 24 months, LFRS, MFRS, PFS and OS were respectively 67.7%/50.7 %, 67%/ 41.7%, 34.8%/14.9% and 83.2%/62.5%. Univariate analysis revealed local control improvement in low GTV (<10mL) (p=0.003) and prostate primary (p=0.02). This was however not confirmed in multivariate analysis. Of note vicinity of target with organ at risk did not significantly correlate with worse toxicity profile (p=0.14) or worse tumor coverage (GTV: p=0.8, PTV: p=0.4). Conclusion SABR pelvic re irradiation in oligometastatic patients appeared as a safe and effective procedure in this cohort. Local control was good and the pattern of relapse was more often distant relapses. Tumor size appeared to influence oncologic and toxicity outcome and should be taken account before re irradiation. PO-1085 Predicting the consensus molecular subtypes of colorectal cancer by weakly supervised deep learning S. Chen 1 , Z. Zhen 1 , J. Wang 1 , W. Hu 1 1 Fudan University Shanghai Cancer Center, Department of Radiation Oncology, Shanghai, China Purpose or Objective The consensus molecular subtypes (CMS) of colorectal cancer is the valuable reference for therapy. However, CMS is hardly applied in clinical routine since the requirement of gene sequencing which is time-consuming and expensive. In this study we manage to predict CMS using weakly supervised deep learning on whole slide images (WSI). Material and Methods The WSIs enrolled into this study contains The Cancer Genome Atlas (TCGA) COAD project and TCGA-READ project. The former project was used for training and testing while the latter project was used for validation. The training set contains 780 WSIs, the testing set contains 6 WSIs and validation set contain 302 WSIs. First, we randomly cropped 1024 patches with shape of 299 * 299 pixels from each training WSI and predicted these slices through a deep learning classification model with initial random weight. For each WSI, 4 patches with highest probability of matched CMS were selected for model training. The weight of deep learning classification model will be update after one training circle. The whole training process was run 50 times. Heatmap of each WSIs were generated by sliding window technique in validation set. A single heatmap would be spited into 25 grids and CMS score will decide on each grid. The CMS for WSI will decide base on the results of all grid. Fig.1 gives an example of CMS decision.

Pts with cT2N+ or cT3T4N0-2 were enrolled into the protocol. Elective lymph nodes and the entire mesorectum received a total dose of 45 Gy at 1.8 Gy/fraction, while gross disease and the correspondent mesorectum were boosted with 57.5 Gy at 2.3 Gy/fraction. Capecitabine (1650 mg/mq/die) was prescribed throughout the whole radiotherapy treatment. Acute and late toxicities were recorded according to CTC-AE scale. Surgery was planned at least 7-8 weeks after the chemoradiotherapy (CRT) end. Results 21 LARC pts (median age: 67 years, range 42-81 years; Male/Female ratio:14/7) entered the analysis. The vast majority of pts had a cT3 tumor (N=16, 76%) with lymph nodal involvement in 90% of cases (cN1: 15 pts, cN2: 4 pts). All patients completed the prescribed treatment; overall, most of them (16 pts, 76%) suffered from mild gastrointestinal toxicity, while 2 pts (10%) from G3 adverse events. Genitourinary and hematological toxicities were less frequent, with 5 pts (24%) referring G1 dysuria and 1 pt (5%) G1 leucopenia. Twenty patients (95%) underwent surgery, after a median interval of 12 (range 7-19) weeks from the CRT’s end. Anterior or abdominoperineal resections were performed in 18 (86%) and in 2 (10%) cases, respectively. One pt (5%), having metastatic disease at re- evaluation, was judged unfit for surgery and started to systemic therapy. Six pts (29%) obtained a pCR, while TRG 1-2 rate was 43%. Three pts (14%) suffered from perioperative complications (1 wound infection,1 anastomotic leak and 1 cardiac arrhythmia respectively). At the time of analysis, 9 pts (43%) referred mild inferior gastrointestinal discomfort (G1-G2) as late sequela. After a median follow-up of 26 months (range 5-47), all patients were alive; both the 2-year local control and the 2-year progression free-survival were 95%. Conclusion A moderately hypofractionated regimen up to a total dose of 57.5 Gy in 25 fractions with a mono-drug concomitant schedule achieves encouraging response rates and toxicity profile comparable to reported literature. With the limit inherent to the small sample size, these data can led to future scenarios on dose intensification treatments. PO-1084 Stereotactic pelvic reirradiation for locoregional cancer relapse R. Kinj 1 , J. Doyen 1 , J. Hannoun-Lévi 1 , G. Baudin 2 , J. Fererro 3 , E. Francois 3 , M. Chand 1 , D. Borchiellini 3 , L. Evesque 3 , K. Benezery 1 , P. Bondiau 1 1 Centre Antoine Lacassagne, radiation oncology, Nice, France ; 2 Centre Antoine Lacassagne, radiology, Nice, France ; 3 Centre Antoine Lacassagne, medical oncology, Nice, France Purpose or Objective Up to 40% of patients with pelvic primary will present a locoregional recurrence in a previously irradiated area. Stereotactic ablative body radiotherapy (SABR) re irradiation can be proposed for treatment of oligometastatic relapse. The aim of the study was to report the clinical outcome after SABR re irradiation in patients presenting an extra bone oligometastatic relapse from pelvic cancer. Material and Methods This is a retrospective monocentric study including patients who benefited from salvage SABR in a previously irradiated pelvic area from January 2011 to February 2018. Patients with intra-prostatic or bone relapse were not included. Results A total of 30 patients were treated with SABR, the most frequent primary tumor sites were rectum (30.8%), prostate (30.8%) and anal cancer (19.2%), and the most frequent re irradiated site was iliac node (40%). With a median follow-up of 29.4 months (1-50) there were three (10%) acute grade 3 (G3) toxicity and no late G3-5 toxicity.

Results The accuracy of CMS prediction in validation set archived to 73%. Heatmaps were generated from WSIs in testing sets and Fig.2 demonstrate an example. The heatmaps fit the clinical illustration of CMS, for example, CMS2 and CMS3 are highly related to epithelial tissues and CMS4 is highly related to stromal tissues.