ESTRO 36 Abstract Book

S117 ESTRO 36 _______________________________________________________________________________________________

B.S. Jang 1 , I.A. Kim 2 1 Seoul National University Graduate School of Medicine, Department of Radiation Oncology, Seoul, Korea Republic of 2 Seoul National University- School of Medicine, Department of Radiation Oncology, Seoul, Korea Republic of Purpose or Objective Radiosensitivity gene signature including 31 genes was identified using microarray data of NCI-60 cancer cells, however, has not been validated in independent datasets for breast cancer patients. We investigated the link between the radiosensitivity gene signature & PD-L1 and clinical outcome in order to identify a group of intensifying clinical benefit of radiotherapy (RT) combined with anti- PD1/PD-L1 therapy. Material and Methods We validated an identified gene signature alleged to radiosensitivity and analyzed PD-L1 status of invasive breast cancer in The Cancer Genome Atlas (TCGA) dataset using bioinformatic tools. First, we downloaded TCGA breast carcinoma (BRCA) gene expression data sets of 1,215 samples achieved from the Illumina HiSeq 2000 RNA Sequencing platform using UCSC Cancer Genomics Browser. To validate gene signature of our interest, 1,065 patients (or samples) were divided into two clusters using consensus clustering algorithm, then assigned radiosensitive (RS) or radioresistant (RR) group according to their prognosis. Patients were also stratified PD-L1 high or PD-L1 low group by median value of CD274 mRNA expression level as surrogates of PD-L1. Relationship between RS/RR groups and PD-L1 status was also assessed, visualized with heat maps, and their prognostic value was evaluated by Kaplan-Meier analysis and Cox proportional hazard models. Results Patents assigned to RS group had better 5-year recurrence-free survival (RFS) rate compared with RR group on univariate (89 % vs. 75 %, p-value = 0.017) only when treated with radiotherapy. RS group was independently associated with PD-L1 high group compared with RR group, as well as CD274 expression was significantly higher in RS group (p-value < 0.001). In a PD- L1 high group, RS group had better 5-year RFS rate over RR group (89 % vs. 72 %, p-value = 0.015), which was also significant on multivariate analysis. The level of PD-L1 expression could represent immunogenicity of tumors, we speculated that the PD-L1 high group had more immunogenic tumors which should be more sensitive to radiation-induced immunologic cell death. Conclusion We first validated the predictive value of radiosensitivity gene signature following adjuvant RT in TCGA data set for invasive breast cancer and also found a relationship with this radiosensitivity gene signature and PD-L1. Radiosensitivity gene signature and PD-L1 status were important factors to predict a clinical outcome of RT in patients with invasive breast cancer and could be used for selecting patients who benefit from radiation therapy combined with anti-PD1/PDL1 therapy. PV-0234 SPECT-CT visualization of axillary lymph nodes in breast cancer: the guide for radiotherapy planning S. Novikov 1 , P. Krzhivitskiy 2 , S. Kanaev 1 , P. Krivorotko 3 , A. Artemeva 4 , E. Turkevich 4 1 Prof. N.N. Petrov Research Institute of Oncology, Radiation Oncology, St. Petersburg, Russian Federation 2 Prof. N.N. Petrov Research Institute of Oncology, Nucleear Medicine, St. Petersburg, Russian Federation 3 Prof. N.N. Petrov Research Institute of Oncology, Breast Surgery, St. Petersburg, Russian Federation 4 Prof. N.N. Petrov Research Institute of Oncology, Pathology, St. Petersburg, Russian Federation

isodose overlay and gamma analysis. Pilot audits were conducted at two radiotherapy centres. Results The end-to-end audit was performed for both IMRT and VMAT treatments. Figure 1 shows the dose-area-histogram for the film measurement plane, for the target and lung OAR, for VMAT delivery at one centre, with and without phantom motion. Dose received by 95% of the GTV area was 263 cGy with the phantom static, 260 cGy with the phantom moving, confirming an appropriate ITV planning margin. Gamma analysis (3% global, 2mm, 20% cut-off) between planned and measured dose had mean passing rates of 98.3% with static phantom, 87.6% with moving phantom. There was no significant difference between IMRT and VMAT modes. Figure 2 shows an isodose comparison between planned and measured doses for a VMAT treatment. Motion blurring reduces the dose gradient around the target in the direction of motion. The 95% isodose of the TPS plan covers the ITV, while the film measured 95% isodose covers the GTV of the moving phantom.

Conclusion A novel, practical method for the dosimetric assessment of motion management strategies in radiotherapy planning and delivery has been designed and successfully piloted at two radiotherapy centres using IMRT and VMAT, enabling independent end-to-end dosimetry audit for mobile RT targets. The results showed the local 4DCT treatment planning approach was sufficient to deliver the required dose to the moving target structure at treatment delivery, and any adverse effects of MLC/gantry motion and target motion interplay were not detrimental. Initially 12 centres in the UK are being audited.

Poster Viewing : Session 5: Lung and breast

PV-0233 A Radiosensitivity Gene Signature &PD-L1 Predict Clinical Outcome of Breast Cancer in TCGA dataset