Abstract Book

S21

ESTRO 37

patients, there was no increased risk for second lung or oesophagus cancer. The risk of second thyroid cancer and sarcomas were increased overall among non-irradiated patients (RR 1.21 and 1.42), however with no remaining risk after +10 years. Comparable results are seen among women treated for DCIS, with one study showing that irradiated patients have an increased risk of in-field second cancers RR 1.37 (95% CI 1.15-1.63) compared to non-irritated women. For second solid cancers, this risk was primarily driven by second lung cancer RR 1.33 (95% CI 1.10-1.60). Second lung cancer is by far the most frequent solid cancer after breast irradiation. An evaluation of the dose-response relationship for second lung cancer risk (and other solid cancers) after high dose radiation therapy, overall indicate, that there is no evidence that the dose-response relationship departs from linearity. A recent study combined data from five individual dose-response studies on second lung cancer. All studies found an increasing risk with increasing dose, with a combined ERR/Gy of 0.11, based on both smokers and non-smokers. Data from three of the five studies however suggest, that the ERR/Gy may be higher in smokers and probably somewhat lower in non-smokers. Conclusion: Approximately 5-9% of all second cancers after breast irradiation are estimated to be attributable to radiotherapy. Although the absolute risk is relative low, the majority of breast cancer patients today are being cured of their disease, and it is these healthy women who are at risk of getting a treatment induced second cancer. As local recurrence rates after breast conserving surgery and radiotherapy today are approaching 2%, the question emerges; is there a subset of patients who have such a low risk of loco-regional failure that the absolute benefit of radiotherapy is outweighed by the potential side effects? The current challenge in radiation oncology is therefore to improve the patient selection, towards identifying those patients who have an expected benefit from radiation, from those patients with no added radiation benefit but who only gets added a risk of late side effects. OC-0049 Genomic profiling of muscle invasive bladder cancer to predict response to chemoradiation therapy D. Miyamoto 1 , E. Gibb 2 , K. Mouw 3 , Y. Liu 2 , C. Wu 1 , M. Drumm 1 , J. Lehrer 2 , H. Ashab 2 , N. Erho 2 , M. Du Plessis 2 , K. Ong 2 , W. Shipley 1 , E. Davicioni 2 , J. Efstathiou 1 1 Massachusetts General Hospital, Radiation Oncology, Boston, USA 2 GenomeDx Biosciences, Inc, Vancouver, Canada 3 Dana Farber / Brigham & Women's Hospital, Radiation Oncology, Boston, USA Purpose or Objective Bladder-sparing trimodality therapy with maximal transurethral resection of bladder tumor (TURBT) followed by chemoradiation therapy is an acceptable alternative to radical cystectomy for selected patients with muscle invasive bladder cancer (MIBC). Genomic profiling has demonstrated MIBC can be divided into molecular subtypes with differing responses to chemotherapy. We explored the utility of genomic data to select patients for bladder-sparing trimodality therapy. Material and Methods Transcriptome wide gene expression profiles were generated for 189 MIBC TURBT samples from patients treated with bladder-sparing trimodality therapy at a single institution. Of these, 103 passed microarray quality control. Molecular subtype and expression of bladder Proffered Papers: RB 1: Joint efforts in Immuno- radiobiology

cancer genes were assessed for association with overall and disease-specific survival. Transcriptome wide differential gene expression analysis was used to explore gene set enrichment in trimodality therapy response The chemoradiation cohort (n=103) had a median followup of 6.9 years for alive patients, and was classified into four subtypes: basal (n=44), basal claudin- low (n=12), infiltrated luminal (n=17) and luminal tumors (n=30). There was no significant difference in overall or disease-specific survival by subtype. However, higher expression of the luminal-associated PPARG was correlated with increased survival after adjusting for subtype and clinical factors (HR=0.52, p=0.002). In contrast, a p53 signature predicted worse survival after adjusting for clinical factors (HR=1.92, p=0.022). Elevated mRNA expression of the DNA damage repair gene MRE11 was associated with improved survival in the trimodality cohort (HR=0.69, P=0.031), consistent with its potential role as a predictive biomarker for radiation response. Gene set enrichment revealed differential regulation of immune pathways in trimodality therapy responders relative to non-responders, including enrichment of interferon gamma signaling (p=0.01) and CXCL9 (p=0.031), suggestive of an interplay between tumor immunologic microenvironment and response to chemoradiation. Conclusion Transcriptional profiling of MIBC revealed gene signatures correlated with response to chemoradiation, suggesting the potential of genomics to guide use of bladder-sparing trimodality therapy. OC-0050 Clinical significance of soluble PD-L1 level in hepatocellular carcinoma patients treated with RT H.J. Kim 1 , S. Park 1 , K.J. Kim 1 , J. Seong 1 1 Yonsei University, Radiation Oncology, SEOUL, Korea Republic of Purpose or Ojective Efforts to find an optimal combination with immune checkpoint inhibitors are underway to further increase the therapeutic effect. This study was performed to investigate the clinical implications of soluble PD-L1 (sPD-L1) level in hepatocellular carcinoma (HCC) patients treated with radiotherapy (RT). Material and Methods HCC patients treated with RT between June 2011 and March 2015 were prospectively recruited and sPD-L1 levels were analyzed. RT was performed using either stereotactic body radiotherapy (SBRT) or conventional fractionated RT. Blood samples were obtained at the each day of RT start, RT end and 1 month follow-up, respectively. sPD-L1 was measured using an enzyme- linked immunosorbent assay (ELISA) in plasma. The association between the amount of sPD-L1 and both of the clinical features and treatment outcome was analyzed. Results Fifty-three patients with HCC were included. Thirty-four patients received a conventional fractionated RT with a median dose of 45 Gy, while 19 patients received SBRT with 60 Gy in 4 fractions. Median follow-up period was 21.3 months. Initial sPD-L1 level increased as the BCLC stage increased (P = 0.047), which was also same in the mUICC stage (P = 0.015). Moreover, initial sPD-L1 level was significantly associated with portal vein tumor thrombosis (P = 0.001), vein invasion (P = 0.006), and tumor size (r = 0.279, P = 0.043). The overall-survival was significantly poor in patients with higher sPD-L1 (≥ 1.315 pg/ml). Furthermore, the higher level of sPD-L1 at 1 month follow-up (≥ 12.9 pg/ml) was significantly related to early lung metastasis within 4 months after RT (P = 0.015). sPD-L1 level was significantly increased after RT. groups. Results