Search

Immunotherapy has demonstrated its effectiveness in many cancers. In hepatocellular carcinoma (HCC), promising results shown in the first phase II studies evaluating anti-PD-1 or anti-PD-L1 monotherapies resulted in their approval in the United States. Approval was not obtained in Europe; subsequent randomized studies in first- or second-line treatment did not confirm these initial results. However, first data with immunotherapy plus antiangiogenic treatments or dual immunotherapy combinations were positive. In this context, the combination of bevacizumab and atezolizumab took the lead. The IMbrave150 trial revealed an improved objective response rate (ORR), progression-free survival, and overall survival with this combination versus the previous standard, sorafenib. Subsequent results of dual immunotherapy with the anti-CTLA-4 and anti-PD-1 monotherapies tremelimumab and durvalumab (also superior to sorafenib monotherapy) confirmed the value of using a combination in first-line treatment. These significant therapeutic advances, and the increase in ORR, raise two main questions. Whereas response was very limited with previous treatments, the ORR reported with these new combinations are between 20% and 30%. This raises the question of whether immunotherapy (ICI single agent, combination of ICI with antiangiogenic agent or other antitumoral treatment) can be used in patients beyond those in BCLC group C, the traditional candidate group for systemic therapy. We have thus seen an increasing number of patients previously treated with trans-arterial chemoembolization (BCLC group B) receiving these new treatments, and we develop the results of several studies combining loco-regional therapies and immunotherapy-based systemic treatments. The other major question is that of how and when to use these medical treatments as “adjuvants” to interventional radiology or surgery; the results of several works are discussed for this purpose. In this review, we cover all of these points in a fairly comprehensive manner.

750 Background: KRAS mutation is the most common molecular alteration (MA) in pancreatic adenocarcinoma (PDAC), and only 8-10% of patients (pts) harbor KRAS wild-type tumors (WT). This study describes clinical and molecular features of KRAS WT PDAC and compares it to mutant KRAS (m KRAS) PDAC. Methods: A retrospective chart review of clinical/molecular data was performed on all pts with PDAC who had contributive molecular profile at Gustave Roussy. KRAS status was determined on tumor molecular profiling as part of routine care using an in-house panel or Foundation Medicine NGS either on tissue (63%) or liquid biopsy (37%). Results: Overall, 41 pts with KRAS WT PDAC were compared to 197 pts with mKRAS PDAC. Median OS (mOS) from initial diagnosis was significantly higher in the KRAS WT group compared to mKRAS group (32.8 months (mts), CI95% [25.9-NR] vs 21.5 mts, CI95% [18.8-24.9] respectively), as mOS from metastatic diagnosis, 26 mts, CI95% [15.7-NR] for KRAS WT pts vs 17.7 mts CI95% [15.5-19.3] for mKRAS (all p

809 Background: Peritoneal carcinomatosis (PC) is a common manifestation in gastro-intestinal (GI) cancers.10-20 % of patients (pts) will present isolated, unresectable peritoneal disease with poor prognosis. In the past years, circulating tumor DNA (ctDNA) has emerged as a novel and less invasive biomarker for molecular diagnosis and disease monitoring. We aimed to evaluate the ability of PC to shed ctDNA into circulation in a cohort of pts with GI cancers and isolated PC treated in our institution. Methods: We reviewed retrospectively the results of liquid biopsies obtained at progression or at diagnosis. Genomic analysis was done using Foundation One liquid CDx panel. Data on demographics, tumor type, clinical outcomes and molecular alterations (MA) have been collected. Results: Out of 400 GI pts tested, 85 with exclusively PC were included in the analysis. The main characteristics were as follows: 45(53%) female, median age of 57 years (IQR: 24-79), majority have received systemic chemotherapy (all were progressive) (n=63, 74%) with median number of previous lines 1 (range: 0-3). Most common tumor types were as follows: gastric (n=24, 28%), colorectal (n=22, 26%) and primary peritoneal (n=20, 24%) cancers. The median index of peritoneal carcinomatosis was 23(IQR: 6-39). The tumor fraction was very low or undetectable in a proportion of 81% (n=69), median Tumor Mutational Burden was 1 mutation per megabase and was not correlated to tumor volume. Gene alterations were found in a proportion of 56% of pts (n=48) but most common alteration were TP 53 (n=22 pts, 26%) and KRAS mutation (n=14 pts, 16%). Targetable MA were identified only in 11 pts (13%) as follows: ATM, ARID1A, BRAFV600E, ERBB2, krasG12C. Three of them were included in a clinical trial. Germline mutations were found in 5 pts: ATM, CDH1, BRCA2, MSH6, BAP1. For colorectal pts known KRAS and BRAF V600E mutations were confirmed in plasma in a proportion of 32% (7 pts out of 22), for pancreatic cancer the KRAS in 17% (3 out of 18) and for gastric cancer, ERBB2 amplification in 5% (1 out of 22 pts). In pts with primary peritoneal disease: pseudomyxoma and mesothelioma, one patient harbored a known germline BAP 1 and kras G12C mutation without other targetable MA. The median survival in MA group was higher compared to non-MA. Conclusions: Isolated Pc and primary peritoneal cancers have distinct clinical and biological characteristics. The clinical utility of ctDNA seems to be limited in isolated PC or primary peritoneal cancers. However, it is not very clear if this to tumor location or to the ability of peritoneum to seed ctDNA.

95 Background: Oxaliplatin, a major drug in metastatic colorectal cancer (MCRC), is responsible for a cumulative and limiting peripheral neuropathy (PN). Hepatic arterial infusion (HAI) chemotherapy, which makes sense in cases of exclusive or predominant liver metastases, increases the intratumor concentration of the administered drug(s) to improve efficacy and limit systemic toxicity. Methods: We compared the frequency and severity of NP in oxaliplatin-naive patients with MCRC included in trials that evaluated treatment with oxaliplatin administered either by HAI (CHOICE, OSCAR, and PACHA-01 trials) or by intravenous (IV) route (FFCD 2000-05 trial). The primary endpoint was the occurrence of clinically significant NP (grade 2-4) according to the cumulative dose of oxaliplatin received. The secondary endpoints were the occurence of severe NP (grade 3-4) and time to onset of NP. Results: 342 patients were included (IV oxaliplatin: 300; HAI: 42). 180 patients in the IV group (60%) and 24 patients in the HAI group (57%) developed clinically significant NP, with no significant difference between the 2 groups (p = 0.85). 95 patients in the IV group (32%) and 8 patients in the HAI group (19%) developed severe NP, with no significant difference between the 2 groups (p = 0.082). NP appeared earlier in the HAI group, with more treatment discontinuations for neurotoxicity. Conclusions: The administration of oxaliplatin HAI rather than IV in the treatment of MCRC does not seem to reduce the incidence, precocity and severity of NP.

e14086 Background: PD-1 and PD-L1 inhibitors have now demonstrated their efficacy. Finding predictive or pre-emptive response factors is crucial and the role of immune-related adverse events (irAEs) is widely debated. The aim of our study was to evaluate the correlation between the occurrence of irAEs and overall survival, according to the time of onset. Methods: We retrospectively collected efficacy and safety data from patients treated with nivolumab, pembrolizumab and atezolizumab for metastatic cancer between July 2015 and January 2019 at the Pitié-Salpêtrière Hospital. Progression-free survival (PFS) and overall survival (OS) were analyzed for the global population, for patients who had an irAES, according to the time of onset (before or after 12 weeks). Results: 158 patients were treated with anti-PD1/PD-L1: 125 patients for NSCLC, 12 for melanoma, 11 for clear cell renal cancer, 5 for bladder cancer, 3 for digestive adenocarcinoma and 2 for mesothelioma. At the cut-off analysis, with a median follow-up of 8.6 months, 63 (40%) patients died, 30(19%) had a progressive disease, 31 (20%) were still receiving an immunotherapy. 25(18%) patients developed irAEs, 18(72%) before 12 weeks of treatment and 7(28%) after. 6 patients had to stop the treatment because of irAEs and 3 of them were still on immunotherapy at the cut-off analysis. Progression disease occurred in 7 patients and 9 died under treatment. Only a trend of efficacy was found between patients with irAEs and those without, with a median PFS of 13.2 vs 9.8 months (HR 1.3; p = 0,4) and a median OS of 28 vs 20 months (HR 1.4; p = 0,4). This statistically non-significant trend was found for OS between patients with an early irAEs and those without any toxicity with a median OS of 28 vs 20 months (HR 1.4; p = 0,8). No statistically difference was found between patients with early irAEs and late irAEs. Conclusions: In our study, 18% of the patients had irAEs. Patients with irAEs seemed to have better OS and PFS but no statistically difference was found. This trend is probably related to patients with late toxicity, which reflects the time of treatment and the increasing probability to develop an irAEs. In our study, early irAEs could not be considered as a reliable preemptive factor of response to immunotherapy.