Search

Although the role of epigenetic alterations in oncogenesis has been well studied, their prevalence in metastatic solid tumours is still poorly described. We therefore aimed at: (i) describing the presence of epigenetic gene alterations (EGA) - defined by an alteration in a gene encoding an epigenetic regulator; and (ii) evaluating their relationship with clinical characteristics and outcome in patients (pts) included in prospective molecular profiling trials.On-purpose tumour biopsies from pts with metastatic solid tumours enrolled in the Gustave Roussy-sponsored MOSCATO (NCT01566019) and MATCHR (NCT02517892) trials were molecularly profiled using whole exome sequencing (WES). Alterations in 176 epigenetic genes were assessed and classified as pathogenic variants (PV) or non-pathogenic variants by a molecular tumour board. Clinical characteristics and outcome were collected.Between Dec 2011 and Oct 2016, WES was successfully performed in 292 pts presenting various solid tumours. We found 496 epigenetic gene alterations in 134 patients (49%), including 237 pathogenic variants in 86 patients; 63 tumour samples (47%) presented ≥3 EGAs. The median number of previous treatment lines was 3 (1-10). The most frequently altered genes were KMT2D and KMT2C (16% each), ARID1A and SETD2 (10% each) and KMT2A (8%).; 31% of EGA co-occurred with a driver gene alteration (p amp;lt; 0.001). Outcome was not correlated with the presence of EGA.Epigenetic alterations occur frequently in metastatic solid tumours. With the current development of epigenetic modifiers, they increasingly represent actionable targets. Such genes should now be systematically analysed in molecular profiling studies.

Metastatic relapse after treatment is the leading cause of cancer mortality, and known resistance mechanisms are missing for most treatments administered to patients. To bridge this gap, we analyze a pan-cancer cohort (META-PRISM) of 1,031 refractory metastatic tumors profiled via whole-exome and transcriptome sequencing. META-PRISM tumors, particularly prostate, bladder, and pancreatic types, displayed the most transformed genomes compared with primary untreated tumors. Standard-of-care resistance biomarkers were identified only in lung and colon cancers—9.6% of META-PRISM tumors, indicating that too few resistance mechanisms have received clinical validation. In contrast, we verified the enrichment of multiple investigational and hypothetical resistance mechanisms in treated compared with nontreated patients, thereby confirming their putative role in treatment resistance. Additionally, we demonstrated that molecular markers improve 6-month survival prediction, particularly in patients with advanced breast cancer. Our analysis establishes the utility of the META-PRISM cohort for investigating resistance mechanisms and performing predictive analyses in cancer.Significance:This study highlights the paucity of standard-of-care markers that explain treatment resistance and the promise of investigational and hypothetical markers awaiting further validation. It also demonstrates the utility of molecular profiling in advanced-stage cancers, particularly breast cancer, to improve the survival prediction and assess eligibility to phase I clinical trials.This article is highlighted in the In This Issue feature, p. 1027

These figures provide complementary information for the figures presented in the main text and illustrations of the methods section.

These tables present clinical data, sample data, whole-exome and transcriptome sequencing results, literature references, and external data sources that were used for the analyses presented in the main text.

Clever-1 expression in macrophages contributes to impaired antigen presentation and suppression of anti-tumor immunity. This first-in-human trial was designed to investigate the safety and tolerability of Clever-1 blockade in patients with treatment-refractory solid tumors and to assess preliminary anti- tumor efficacy, pharmacodynamics, and immunologic correlates. Bexmarilimab was well tolerated with no observed dose limiting toxicities in part I (n=30) and no additional safety signals in part II (n=108). Disease control (DC) rates of 25-40% were observed in cutaneous melanoma, gastric, hepatocellular, estrogen receptor positive breast, and biliary tract cancers, which associated with improved survival in a landmark analysis. High pre-treatment intra-tumoral Clever-1 staining and low circulating TNFα correlated with DC. Digital spatial profiling of DC and non-DC tumors with next generation sequencing showed bexmarilimab-induced macrophage activation and robust stimulation of IFNγ and T-cell receptor signaling selectively in DC patients. These data suggest that bexmarilimab therapy is well-tolerated and show that macrophage targeting can promote tumor control in late stage cancer.HighlightsTargeting Clever-1 with bexmarilimab is well toleratedDisease control in late stage cancer can be achieved with bexmarilimab monotherapyLow baseline immune activation associates with bexmarilimab benefitBexmarilimab converts intratumoral macrophages to support adaptive immune responses

This file provides supplementary methods for genomics, including TGS, CGH array, WES, RNAseq Analysis, IHC -MET.

Individual PFS2 vs PFS1 profile plot. Each patient has been aligned at time 0 of PFS2 and sorted by the sum of PFS1 and PFS2.

Genes that contain molecular alterations matched with therapies according to tumor type.

Supplementary Figure Legends

Metastatic relapse after treatment is the leading cause of cancer mortality, and known resistance mechanisms are missing for most treatments administered to patients. To bridge this gap, we analyze a pan-cancer cohort (META-PRISM) of 1,031 refractory metastatic tumors profiled via whole-exome and transcriptome sequencing. META-PRISM tumors, particularly prostate, bladder, and pancreatic types, displayed the most transformed genomes compared with primary untreated tumors. Standard-of-care resistance biomarkers were identified only in lung and colon cancers—9.6% of META-PRISM tumors, indicating that too few resistance mechanisms have received clinical validation. In contrast, we verified the enrichment of multiple investigational and hypothetical resistance mechanisms in treated compared with nontreated patients, thereby confirming their putative role in treatment resistance. Additionally, we demonstrated that molecular markers improve 6-month survival prediction, particularly in patients with advanced breast cancer. Our analysis establishes the utility of the META-PRISM cohort for investigating resistance mechanisms and performing predictive analyses in cancer. Significance: This study highlights the paucity of standard-of-care markers that explain treatment resistance and the promise of investigational and hypothetical markers awaiting further validation. It also demonstrates the utility of molecular profiling in advanced-stage cancers, particularly breast cancer, to improve the survival prediction and assess eligibility to phase I clinical trials. This article is highlighted in the In This Issue feature, p. 1027

PURPOSE Early discontinuation affects more than one third of patients enrolled in early-phase oncology clinical trials. Early discontinuation is deleterious both for the patient and for the study, by inflating its duration and associated costs. We aimed at predicting the successful screening and dose-limiting toxicity period completion (SSD) from automatic analysis of consultation reports. MATERIALS AND METHODS We retrieved the consultation reports of patients included in phase I and/or phase II oncology trials for any tumor type at Gustave Roussy, France. We designed a preprocessing pipeline that transformed free text into numerical vectors and gathered them into semantic clusters. These document-based semantic vectors were then fed into a machine learning model that we trained to output a binary prediction of SSD status. RESULTS Between September 2012 and July 2020, 56,924 consultation reports were used to build the dictionary and 1,858 phase I or II inclusion reports were used to train (72%), validate (14%), and test (14%) a random forest model. Preprocessing could efficiently cluster words with semantic proximity. On the unseen test cohort of 264 consultation reports, the performances of the model reached: F1 score 0.80, recall 0.81, and area under the curve 0.88. Using this model, we could have reduced the screen fail rate (including dose-limiting toxicity period) from 39.8% to 12.8% (relative risk, 0.322; 95% CI, 0.209 to 0.498; P < .0001) within the test cohort. Most important semantic clusters for predictions comprised words related to hematologic malignancies, anatomopathologic features, and laboratory and imaging interpretation. CONCLUSION Machine learning with semantic conservation is a promising tool to assist physicians in selecting patients prone to achieve SSD in early-phase oncology clinical trials.

PURPOSE To facilitate implementation of precision medicine in clinical management of cancer, the European Society of Medical Oncology proposed in 2018 a new scale to harmonize and standardize the reporting and interpretation of clinically relevant genomics data (ESMO Scale of Actionability of molecular Targets [ESCAT]). This study aims to characterize the clinical impact of matching targetable genomic alterations (GAs) in patients with advanced cancer according to ESCAT. MATERIAL AND METHODS Analysis of next-generation sequencing results from 552 patients is included in two prospective precision medicine studies at Gustave Roussy. End points included objective response rates, progression-free survival, and overall survival according to ESCAT. RESULTS Molecular data from 516 patients were available and discussed within a Molecular Tumor Board. The most common tumor types were GI (n = 164; 30%), lung (n = 137; 25%), and urologic tumors (n = 68; 13%). Overall, 379 GAs were considered as actionable targets according to ESCAT in 348 (67%) patients. In 31 (6%) patients, two concomitant actionable targets were identified. On the basis of ESCAT, GAs were considered to be classified as tier I in 120 patients (29%), II in 25 patients (5%), III in 80 patients (16%), and IV in 153 patients (30%). A total of 136 patients (27%) received a matched therapy. ESCAT was significantly associated with objective response rates and clinical benefit rates. The median progression-free survival was 6.5 months (95% CI, 4.2 to 8.9), 3 months (95% CI, 1 to not available), 3 months (95% CI, 2.2 to 3.8), and 4 months (95% CI, 2.8 to 6.3) for ESCAT I, II, III, and IV, respectively ( P = .0125). CONCLUSION Implementation of ESCAT classification for clinical decision making by Molecular Tumor Board is feasible and useful to better tailor therapies in patients with cancer.

AMoNet (Artificial Molecular Networks) is a tool that aims to predict cancer patients’ survival when only targeted gene sequencing data are available. Outcome predictions from sparse data can benefit from new methods including deep learning. Our approach optimizes large recurrent directed molecular networks built from prior knowledge supported by speed-up computations and interpretations. Predictions suggested by the model simulations are available in a user-friendly interface.

Metastatic relapse after treatment is the primary cause of cancer morbidity and mortality. While genetic mechanisms of primary tumors and, to a lesser extent, metastatic cancers have been studied in large cohorts, refractory metastatic tumors are not yet sufficiently characterized. Markers of aggressiveness and resistance that molecular profiling can extract from these tumors have yet to be identified and incorporated into clinical care. In this study, we present a pan-cancer cohort of 1,031 metastatic tumors (which we refer to as META-PRISM) that are resistant to at least one systemic therapy or with no approved treatment options. We retrieved the complete clinical history of patients and performed whole-exome (n=571) and transcriptome sequencing (n=947) for this cohort. The prevalence of detected cancer biomarkers was assessed and compared to an external cohort of primary tumors. In the META-PRISM cohort, we observed an increase in (i) whole-genome duplication frequency, (ii) tumor mutational burden, (iii) germline cancer-predisposing variants, and (iv) somatic alterations in cancer genes, including KRAS, EGFR, CCND1, MYC, and TP53, as compared to the tumor type-matched primary tumors. The most extensive increase in genomic variation at metastatic stage was observed in prostate bladder, and pancreatic cancer types. We also identified enrichment of standard-of-care resistance biomarkers in most cancer types. However, only 9.3% of tumors harbored at least one such biomarker, indicating that the current understanding of resistance mechanisms remains insufficient. Our cohort demonstrated a significantly improved 6-month survival prediction from models incorporating molecular markers over models with only clinical markers for breast cancer patients and to a lesser extent for other studied tumor types. Overall, our data establish a unique resource for investigating treatment resistance mechanisms and performing predictive analyses in cancer. Citation Format: Yoann Pradat, Julien Viot, Konstantin Gunbin, Andrey Yurchenko, Luigi Cerbone, Marc Deloger, Guillaume Grisay, Loic Verlingue, Véronique Scott, Ismael Padioleau, Leonardo Panunzi, Stefan Michiels, Antoine Hollebecque, Gerome Jules-Clément, Laura Mezquita, Antoine Lainé, Yohann Loriot, Benjamin Besse, Luc Friboulet, Fabrice André, Paul-Henry Cournède, Daniel Gautheret, Sergey Nikolaev. Integrative pan-cancer genomic and transcriptomic analyses of refractory metastatic cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 6570.

The rising interest for precise characterization of the tumour immune contexture has recently brought forward the high potential of RNA sequencing (RNA-seq) in identifying molecular mechanisms engaged in the response to immunotherapy. In this review, we provide an overview of the major principles of single-cell and conventional (bulk) RNA-seq applied to onco-immunology. We describe standard preprocessing and statistical analyses of data obtained from such techniques and highlight some computational challenges relative to the sequencing of individual cells. We notably provide examples of gene expression analyses such as differential expression analysis, dimensionality reduction, clustering and enrichment analysis. Additionally, we used public data sets to exemplify how deconvolution algorithms can identify and quantify multiple immune subpopulations from either bulk or single-cell RNA-seq. We give examples of machine and deep learning models used to predict patient outcomes and treatment effect from high-dimensional data. Finally, we balance the strengths and weaknesses of single-cell and bulk RNA-seq regarding their applications in the clinic.

PD-L1 and tumor-infiltrating lymphocytes play a key role in the immune escape of cancer, although their prognostic value remains unknown in patients with refractory solid cancer compared to other known prognostic estimation methods. In this ancillary study, we assessed the prognostic value of previously-defined prognostic scores (such as the Royal Marsden Hospital (RMH) score) and of PD-L1, CD3, CD8 and FOXP3 expressions based on immunohistochemistry (IHC) and RNA sequencing (RNAseq) of tumor samples from patients included in the personalized-medicine MOSCATO-02 trial. We collected biopsies with successful IHC analysis from 266 patients treated between April 2016 and September 2017, among whom 170 (63.9%) also had a matched RNAseq. We used a Random Forest model to identify the best prognostic factor, and a Lasso-penalized Cox model to validate the findings. We found that the RMH score was the strongest prognostic factor, with high scores associated with a higher risk of death (Hazard Ratio (HR)=1.29; CI95%[1.19-1.21]). The PD-L1 expression score obtained from IHC analyses was the second-best performing predictor, with the 1+ score (low expression) linked to a lower risk of death (HR=0.564; CI95%[0.539-0.580]). Other tested variables, including primary tumor type and subsequent treatments received following biopsy, were not found significantly linked to prognosis. We found modest correlation between IHC and RNAseq expressions of immune genes, but RNAseq related better to prognosis. Overall, our study supports the use of the RMH score and the assessment of PD-L1 expression in IHC to estimate prognosis in patients with advanced cancer.

Introduction: For patients with advanced/unresectable biliary tract cancers, cisplatin-gemcitabine combination is the standard first-line treatment. Beyond the first line, the therapeutic arsenal is limited with minimal benefit. Biliary tract cancers exhibit one of the highest frequencies of targetable molecular alterations across cancer types, and several targeted therapies are emerging as treatment options.Areas covered:We discuss neurotrophic tyrosine kinase receptor gene (NTRK) fusions in biliary tract cancers and the use of NTRK inhibitors (now approved in a 'cancer-agnostic' way), mechanisms of resistance, and emerging second-generation NTRK inhibitors.Expert opinion: Despite their rarity in biliary tract cancers, NTRK fusions are promising molecular targets because i) NTRK inhibitors have proven highly effective in NTRK-rearranged cancers and are now approved in a 'cancer-agnostic' way; ii) emerging second-generation NTRK inhibitors may overcome secondary resistance; iii) NTRK rearrangements will be readily detectable with the generalization of next-generation-sequencing in biliary tract cancers, including the detection of other frequent gene rearrangements, such as those involving the fibroblast growth factor receptor 2 gene (FGFR2). However, more data are necessary regarding the prevalence and characteristics of NTRK fusions in biliary tract cancers and the efficacy of NTRK inhibitors in these patients.

498 Background: BT8009 is a Bicycle Toxin Conjugate (BTC), a novel class of chemically synthesized molecules comprising a small (~2 kDa) bicyclic peptide targeting Nectin-4 tumor antigen linked to cytotoxin MMAE via a valine-citrulline cleavable linker. BTCs represent a unique therapeutic class, distinct from ADCs and with pharmacokinetic properties analogous to small molecules. Presented are the monotherapy dose escalation results of the ongoing Phase I/II clinical trial of BT8009 (NCT04561362). Methods: In the Part A 3+3 dose escalation portion of the study, BT8009 monotherapy was administered weekly (2.5, 5, 7.5 mg/m2) or every other week (7.5, 10 mg/m2) of a 28-day cycle, or on days 1 and 8 of a 21-day cycle (7.5 mg/m2). Primary objectives were to assess safety of the monotherapy and define maximum tolerated and recommended Phase II dose(s) (RP2D). Key secondary objectives were antitumor activity per RECIST v1.1 and PK parameters. Results: At time of the 20 Sept 2022 data cut-off, 49 pts had received BT8009 (24 pts with UC): weekly (7 pts at 2.5 mg/m2, 20 pts at 5 mg/m2, 5 pts at 7.5 mg/m2); biweekly (5 pts at 7.5 mg/m2, 7 pts at 10 mg/m2); and days 1+8 of a 21-day cycle (5 pts at 7.5 mg/m2). Two RP2Ds have been determined: 5 mg/m2 given weekly on a 28-day cycle and 7.5 mg/m2 given on days 1+8 of a 21-day cycle. Median age was 66 y, 59% pts were male, and median prior lines of therapy was 3 (range 1-15). Most common treatment-related adverse events of any grade experienced by ≥15% of pts were nausea, fatigue, diarrhea, decreased appetite, alopecia, asthenia, pyrexia, and neutrophil count decreased/neutropenia; majority were mild to moderate in severity. One DLT (G3 asthenia) was reported at the 7.5 mg/m2 weekly dose level and 1 DLT (G4 sepsis) at the 10 mg/m2 biweekly dose level. Treatment-related SAEs occurred in 5 pts. No pt discontinued treatment due to an AE. No treatment-related deaths occurred. Of 8 efficacy evaluable pts with UC treated with a RP2D of 5 mg/m2, there was 1 complete response, 3 partial responses (PR), and 3 stable diseases (SD), for an ORR of 50% and clinical benefit rate of 75% (CR+PR+SD≥16 wks). Median duration of response has not yet been reached, but as of 20 Sept 2022 was at least 11 months. Both plasma BT8009 and MMAE exhibited linear pharmacokinetics, with short half-life for BT8009 (

Metastatic relapse after treatment is the primary cause of cancer morbidity and mortality. While genetic mechanisms of primary tumors and, to a lesser extent, metastatic cancers have been studied in large cohorts, refractory metastatic tumors are not yet sufficiently characterized. Markers of aggressiveness and resistance that molecular profiling can extract from these tumors have yet to be identified and incorporated into clinical care. In this study, we present a pan-cancer cohort of 1,031 metastatic tumors (which we refer to as META-PRISM) that are resistant to at least one systemic therapy or with no approved treatment options. We retrieved the complete clinical history of patients and performed whole-exome (n=571) and transcriptome sequencing (n=947) for this cohort. The prevalence of detected cancer biomarkers was assessed and compared to an external cohort of primary tumors. In the META-PRISM cohort, we observed an increase in (i) whole-genome duplication frequency, (ii) tumor mutational burden, (iii) germline cancer-predisposing variants, and (iv) somatic alterations in cancer genes, including KRAS, EGFR, CCND1, MYC, and TP53, as compared to the tumor type-matched primary tumors. The most extensive increase in genomic variation at metastatic stage was observed in prostate cancer. We also identified enrichment of standard-of-care resistance biomarkers in most cancer types. However, only 7.6% of tumors harbored at least one such biomarker, indicating that the current understanding of resistance mechanisms remains insufficient. Our cohort demonstrated a significantly improved 6-month survival prediction from models incorporating molecular markers over models with only clinical markers for breast cancer patients and to a lesser extent for other studied tumor types. Overall, our data establish a unique resource for investigating treatment resistance mechanisms and performing predictive analyses in cancer. Citation Format: Yoann Pradat, Julien Viot, Konstantin Gunbin, Andrei Iurchenko, Marc Deloger, Luigi Cerbone, Guillaume Grisay, Loic Verlingue, Veronique Scott, Stefan Michiels, Antoine Hollebecque, Gerome Jules-Clement, Antoine Laine, Luc Friboulet, Laura Mezquita, Yohann Loriot, Benjamin Besse, Fabrice Andre, Paul-Henry Cournede, Daniel Gautheret, Sergey Nikolaev. Integrative pan-cancer genomic and transcriptomic analyses of refractory metastatic cancer [abstract]. In: Proceedings of the AACR Special Conference: Cancer Metastasis; 2022 Nov 14-17; Portland, OR. Philadelphia (PA): AACR; Cancer Res 2022;83(2 Suppl_2):Abstract nr PR009.

Background Early phase clinical trials usually include patients with advanced disease who have failed standard therapies. Early palliative care for these patients has shown to improve quality of life and even survival. PALLIA 10 score (ranging from 1 to 10) is a tool developed by the French Palliative Care Society to identify the best time to introduce palliative care. Methods We assessed the PALLIA 10 score and other prognostic factors (age, ECOG, Royal Marsden Hospital (RMH) score, LDH and albumin levels, number of prior systemic treatments and metastatic sites) in patients enrolled in phase I trials at Gustave Roussy Cancer Center prospectively during 2 periods of time (cohort 1 (C1) and 2 (C2)). A double-blind assessment of the PALLIA 10 score was done during 15 days by a member of the palliative care unit in C2. A PALLIA 10 > 3 motivated a dedicated palliative care consultation. Results From July 1st 2018 to November 1st 2018 (C1) and from December 1st 2020 to April 16th 2021 (C2), a total of 86 patients were assessed in C1 and 302 in C2. No difference was observed between the two cohorts regarding prognostic factors. Median PALLIA 10 was also similar and very low (median 1, range 1-5 in C1 and 1-8 in C2). On C1 and C2, 12% and 5% of patients had a dedicated palliative consultation. Overall, 77% and 74% of patients in C1 and C2 were still alive beyond 3 months after discontinuation of the trial (p=0.78), followed by at least one subsequent treatment in 63% and 70% of pts. In C2, assessment of PALLIA 10 score was significantly different between palliative care physician (median 5, range 3-8), phase I physician (median 1, range 1 -6) and phase I nurse (median 3, range 1-8) (p

Purpose: Although the role of epigenetic alterations in oncogenesis has been well studied, their prevalence in metastatic solid tumors is still poorly described. We therefore aimed at: (i) describing the presence of epigenetic gene alterations (EGA) - defined by an alteration in a gene encoding an epigenetic regulator; and (ii) evaluating their relationship with clinical characteristics and outcome in patients (pts) included in prospective molecular profiling trials.Patients and methods: On-purpose tumor biopsies from pts with metastatic solid tumors enrolled in the Gustave Roussy-sponsored MOSCATO (NCT01566019) and MATCHR (NCT02517892) trials were molecularly profiled using Whole Exome Sequencing (WES). Alterations in 176 epigenetic genes were assessed and classified as pathogenic variants (PV) or non-pathogenic variants by a molecular tumor board. Clinical characteristics and outcome were collected.Results: Between Dec 2011 and Oct 2016, WES was successfully performed in 292 pts presenting various solid tumors. We found 496 epigenetic gene alterations in 134 patients (49%), including 237 pathogenic variants in 86 patients; 63 tumor samples (47%) presented ≥3 EGAs. The median number of previous treatment lines was 3 (1 – 10). The most frequently altered genes were KMT2D and KMT2C (16% each), ARID1A and SETD2 (10% each) and KMT2A (8%).; 31% of EGA co-occurred with a driver gene alteration (p < 0.001). Outcome was not correlated with the presence of EGA.Conclusions: Epigenetic alterations occur frequently in metastatic solid tumors. With the current development of epigenetic modifiers, they increasingly represent actionable targets. Such genes should now be systematically analyzed in molecular profiling studies.

Short summaryPersonalized biomarkers can facilitate decision making upon multiple therapeutic options in ccRCC. VEGFR2 expression denoised with 37 normal and tumor gene-expressions relates to sunitinib effect whereas raw VEGFR2 expression doesn’t relate to sunitinib effect.BackgroundSeveral studies suggested that molecular analysis of patients with advanced clear cell renal cell carcinoma (ccRCC) could indicate whether a patient is susceptible of benefiting from sunitinib in first-line systemic treatment compared to immunotherapies. However, data remain conflicting and no predictive biomarker is validated so far to decipher if sunitinib could still represent a good therapeutic option in first line setting and beyond.MethodsPREDMED® denoised the tumor RNA expression of 37 genes including KDR (encoding VEGFR2) estimated by RT-qPCR, by normalizing it on the expression of normal kidney tissue and cell types. We investigated the performance of PREDMED® VEGFR2-scoring to predict the clinical effect of sunitinib for patients affected by ccRCC.ResultsAmong the 34 ccRCC patients’ samples retrospectively retrieved from the UroCCR project (NCT03293563), high VEGFR2 scores were associated with objective clinical responses under sunitinib treatment and low scores with stable disease or progression with a sensitivity of 86%, a specificity of 67% and an AUC of 72.5% (95%CI[50.1–94.9]; p=0.04). VEGFR2 scores were significantly and positively related to progression-free survival (HR = 0.465; 95%CI[0.221–0.978]; p=0.0311) and overall survival (HR = 0.400; 95%CI[0.192–0.834]; p=0.0134) under sunitinib treatment. In our cohort, raw VEGFR2 expression (before PREDMED® processing) was not related to the above mentioned outcomes.ConclusionWe describe a gene-expression based algorithm that is accurately related to the effect of sunitinib for patients with ccRCC. We further plan a validation of PREDMED® for combinatorial strategies involving antiangiogenics and immune-checkpoint blockers.

RNA-sequencing (RNA-seq) analysis offers a tumor-centered approach of growing interest for personalizing cancer care. However, existing methods – including deep learning models – struggle to reach satisfying performances on survival prediction based upon pan-cancer RNA-seq data. Here, we present DeepOS, a novel deep learning model that predicts overall survival (OS) from pan-cancer RNA-seq with a concordance-index of 0.715 and a survival AUC of 0.752 across 33 TCGA tumor types whilst tested on an unseen test cohort. DeepOS notably uses (i) prior biological knowledge to condense inputs dimensionality, (ii) transfer learning to enlarge its training capacity through pre-training on organ prediction, and (iii) mean squared error adapted to survival loss function; all of which contributed to improve the model performances. Interpretation showed that DeepOS learned biologically-relevant prognosis biomarkers. Altogether, DeepOS achieved unprecedented and consistent performances on pan-cancer prognosis estimation from individual RNA-seq data.

ObjectivePhase I clinical trials usually include patients with advanced disease who have failed standard therapies and should benefit from early palliative care. We try to assess whether PALLIA 10, a score developed in France to help identify patients who might benefit from a palliative care referral, could be used in a phase I department trial.MethodsWe assessed PALLIA 10 score and other prognostic factors in patients enrolled in phase I trials at Gustave Roussy Cancer Center prospectively during two periods of time (cohort 1 (C1) and 2 (C2)). A double-blind assessment of the PALLIA 10 score was done in C2 by a palliative care specialist and a nurse.ResultsFrom 1 July 2018 to 1 November 2018 (C1) and from 1 December 2020 to 16 April 2021 (C2), 86 patients were assessed in C1 and 302 in C2. Median PALLIA 10 was very low in both cohorts (median 1, range 1–5 in C1 and 1–8 in C2). On C1 and C2, 12% and 5% of patients had a dedicated palliative consultation. In C2, assessment of PALLIA 10 score was significantly different between palliative care physician (median 5, range 3–8), phase I physician (median 1, range 1–6) and phase I nurse (median 3, range 1–8) (pConclusionMedian PALLIA 10 score was low when assessed by the phase I physician, which suggests the need for a better tool and appropriate clinician’s education to implement early palliative care in clinical practice and trials.

BT8009 is a Bicycle® Toxin Conjugate (BTC™), a novel class of chemically synthesized molecules, comprising a bicyclic peptide targeting Nectin-4 tumor antigen, linked to cytotoxin (monomethyl auristatin E [MMAE]) via a valine-citrulline (val-cit) cleavable linker. Due to their small size, BTCs represent a unique therapeutic class, combining the pharmacology normally associated with a biologic with the pharmacokinetic properties of a small molecule. Here, we describe the preliminary monotherapy dose escalation results of the ongoing multicenter Phase I/II clinical trial (NCT04561362) that assesses the safety and tolerability of BT8009 administration in patients with advanced solid tumors associated with Nectin-4 expression, not exposed to enfortumab vedotin. Patients were treated until disease progression or intolerable toxicity, with tumor assessments for response per RECIST taken every two months. Nectin-4 expression levels were measured in tumor tissue. Thirty-four patients were initiated on weekly BT8009 (7 patients at 2.5 mg/m2, 20 patients at 5.0 mg/m2, and 4 patients at 7.5 mg/m2) or biweekly BT8009 (3 patient at 7.5 mg/m2). Baseline demographics were 62% patients were male, 44% and 56% were ECOG 0 and ECOG 1, respectively, median age was 65.5 years. BT8009 exhibits linear pharmacokinetics. The most common treatment-related adverse events were nausea (14 patients, 41%; ≥ G3 1 patient, 3%), fatigue (13 patients, 38%; ≥ G3 2 patients, 6%), anorexia (11 patients, 32%; ≥ G3 0 patients), constipation (10 patients, 29%; ≥ G3 1 patient, 3%) and anemia (10 patients, 29%; ≥ G3 3 patients, 9%). To date, BT8009 had low incidences of ocular disorders (3%), hyperglycemia (6%), neuropathy (6%), and rash (18%). There was 1 dose-limiting toxicity (G3 asthenia) reported at the 7.5 mg/m2 weekly dose level. Six patients (18%) experienced an SAE, with only 1 related SAE (vomiting). Treatment-emergent AEs resulting in treatment discontinuation, dose interruption, or reductions were observed in 2 patients (6%), 11 patients (32%), and 5 patients (15%), respectively. Preliminary confirmed responses observed in patients with urothelial carcinoma (UC) enrolled in the 5 mg/m2 cohort show 4/8 patients with a complete response (CR) or partial response (PR), including 1/8 patients with a CR and 3/8 patients with a PR, and 2/8 patients (25%) with stable disease (SD), reflecting a 50% overall response rate and a 75% disease control rate in UC patients for this cohort. BT8009 exhibits a promising preliminary tolerability profile and preliminary antitumor activity. The molecule will continue to be explored in the current dose-escalation/dose-expansion study of BT8009 monotherapy and in combination with nivolumab. Citation Format: Meredith McKean, Capucine Baldini, Loic Verlingue, Bernard Doger, Gerald Falchook, Antoine Italiano, Julia Lostes, Oscar Reig, Roshawn Watson, Sebastien Hazard, Dominic Smethurst, Giulia Fulci, Hongmei Xu, Phil Jeffery, Kevin Lee, Irene Braña, Sophie Cousin, Elisa Fontana, Hendrik-Tobias Arkenau. BT8009-100 phase I/II study of novel bi-cyclic peptide and MMAE conjugate BT8009 in patients with advanced malignancies associated with nectin-4 expression [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr CT025.

PurposeEarly discontinuation affects over one-third of patients enrolled in early-phase oncology clinical trials. Early discontinuation is deleterious both for the patient and for the study, by inflating its duration and associated costs. We aimed at predicting the successful screening and dose-limiting toxicity period completion (SSD) from automatic analysis of consultation reports.Materials and methodsWe retrieved the consultation reports of patients included in phase I and/or phase II oncology trials for any tumor type at Gustave Roussy, France. We designed a pre-processing pipeline that transformed free-text into numerical vectors and gathered them into semantic clusters. These document-based semantic vectors were then fed into a machine learning model that we trained to output a binary prediction of SSD status.ResultsBetween September, 2012 and July, 2020, 56,924 consultation reports were used to build the dictionary, and 1,858 phase I/II inclusion reports were used to train (75%), validate (15%) and test (15%) a Random Forest model. Pre-processing could efficiently cluster words with semantic proximity. On the unseen test cohort of 264 consultation reports, the performances of the model reached: F1 score 0.80, recall 0.81 and AUC 0.88. Using this model, we could have reduced the screen fail rate (including DLT period) from 39.8% to 12.8% (RR=0.322, 95%CI[0.209-0.498], pConclusionMachine learning with semantic conservation is a promising tool to assist physicians in selecting patients prone to achieve SSD in early-phase oncology clinical trials.

PD-1 blockade represents a major therapeutic avenue in anticancer immunotherapy. Delineating mechanisms of secondary resistance to this strategy is increasingly important. Here, we identified the deleterious role of signaling via the type I interferon (IFN) receptor in tumor and antigen presenting cells, that induced the expression of nitric oxide synthase 2 (NOS2), associated with intratumor accumulation of regulatory T cells (Treg) and myeloid cells and acquired resistance to anti-PD-1 monoclonal antibody (mAb). Sustained IFNβ transcription was observed in resistant tumors, in turn inducing PD-L1 and NOS2 expression in both tumor and dendritic cells (DC). Whereas PD-L1 was not involved in secondary resistance to anti-PD-1 mAb, pharmacological or genetic inhibition of NOS2 maintained long-term control of tumors by PD-1 blockade, through reduction of Treg and DC activation. Resistance to immunotherapies, including anti-PD-1 mAb in melanoma patients, was also correlated with the induction of a type I IFN signature. Hence, the role of type I IFN in response to PD-1 blockade should be revisited as sustained type I IFN signaling may contribute to resistance to therapy. CA extern

A targeted therapy is recommended in case of ERBB2 alteration for breast and gastric carcinomas, but miscellaneous other tumor types are ERBB2-altered at low prevalence. Broadening the administration of HER2 inhibitors across tumor types and genomic alterations could benefit to patients with refractory metastatic tumors. Targeted next-generation-sequencing (tNGS) and comparative genomic hybridization array (CGH) have been performed on fresh tumor biopsies of patients included in the MOSCATO-01 and ongoing MOSCATO-02 trials to administrate HER2 inhibitors in case of ERBB2 pathogenic mutation of amplification. Between December 2011 and January 2017 a molecular analysis was performed for 934 patients (759 CGH and 912 tNGS). A novel ERBB2 alteration has been found in 4.7% (n = 44/934), including 1.5% (n = 14/912) ERBB2 mutations, and 4% (n = 30/759) ERBB2 amplifications. A matched HER2 inhibitor was administrated to 70% (31/44) of patients and consisted in trastuzumab plus chemotherapy for 90% of them (28/31). On the 31 evaluable patients, 1 complete response (CR), 10 partial response (PR) and 2 stable disease (SD) >24 weeks were observed accounting for a clinical benefit rate (CBR) of 42% (n = 13/31, 95% CI 25–61%). Besides breast and oesogastric carcinomas, 19 patients affected by 8 different tumor types had a CBR of 25% for ERBB2 mutations (n = 2/8, 95% CI 3%–65%, with 2 PR) and 64% for ERBB2 amplifications (n = 7/11, 95% CI 31%–89%; with 1 CR, 4 PR, 2 SD). ERBB2 genomic alterations were diffuse across metastatic tumor types and signs of efficacy emerged for HER2 targeted treatments, especially in case of ERBB2 amplifications or a p.S310Y ERBB2 mutation.

23 Background: Early phase clinical trials usually include patients (pts) with advanced disease who have failed to standard therapies. Early palliative care (EPC) for these pts has shown to improve quality of life and even survival. Pallia 10 score (from 1 to 10) is a tool developed by the French Palliative Care Society to identify the best time to introduce palliative care. Methods: We assessed the Pallia 10 score and other prognostic factors (age, ECOG, Royal Marsden Hospital (RMH) score, LDH and albumin levels, number (nb) of prior systemic treatments and metastatic sites) in pts enrolled in phase I trials (P1CT) prospectively during 2 periods of time (cohort 1 (C1) and 2 (C2)). A double-blind assessment of the Pallia 10 score was done during 15 days by a member of the palliative care unit in C2. A Pallia 10 > 3 motivated a dedicated palliative care consultation. Results: From 01/07/2018 to 01/11/2018 (C1) and from 01/12/2020 to 16/04/2021 (C2), a total of 85 pts were assessed in C1 and 302 in C2. Gastro-intestinal (23%), hematological (14%) and lung (11%) cancer were the most frequent tumor types. Pallia 10 score and prognostic factors were similar between both cohorts (Table). On C1 and C2, 12% and 4% of pts had a dedicated palliative consultation with median time of referral of 18 and 2 months (m) after the P1CT onset (p = 0.003), with a median Pallia 10 score of 1.5 and 2 (p = 0.65), respectively. Overall, 75% and 76% of pts in C1 and C2 were still alive beyond 3m after discontinuation of the P1CT (p = 0.91), followed by at least one subsequent treatment in 56% and 54% of pts. In C2, assessment of Pallia 10 score was significantly different between palliative care physician (median 5, range 3-8), phase I physician (median 1, range 1 -6) and phase I nurse (median 3, range 1-8) (p < 0.001). Conclusions: Only a few patients included in P1CT were referred to the palliative care unit. Median Pallia 10 score was low when assessed by the phase I physician which suggests the need for a better tool to implement EPC in clinical practice and trials.[Table: see text]

Background Recommended treatments of patients with advanced biliary tract cancer (aBTC) are limited to one chemotherapy doublet. Nevertheless, efficacy of treatment personalisation for aBTCs is supported by accumulating evidences but remains to be evaluated. Patients and methods Patients with aBTCs included in the prospective clinical trial MOSCATO-01 were treated by at least one previous systemic treatment, had an ECOG performance status of 0–1, and at least one tumour site accessible to biopsy. Multiple high-throughput molecular analysis was performed on biopsies to drive the administration of molecular targeted agents (MTAs). Results From November 2011 to March 2016, 43 patients (4%) of the 1035 adult patients included in MOSCATO-01 had aBTCs with a majority of intrahepatic localisation (67%). Successful biopsy procedures and DNA extractions led to molecular portraits for 34 patients (79%). Orientation to an appropriate early clinical trial or accessible MTA(s) was possible for 23 of these patients (68%), and 18 (53%) have received matched MTA(s). Among them, the overall response rate was 33% and the disease control rate was 88%. A PFS ≥6 months was observed in 37% and the PFS ratio was >1.3 for 50% of the patients. These patients had a lower risk for death as compared to the 20 patients not orientated to a matched MTA (HR, 0.29; 95% CI, 0.11–0.76; p = 0.008). Conclusions Within the MOSCATO-01 trial, patients with aBTCs had the highest rate of orientation to matched MTAs and derived a clear clinical benefit. A broader evaluation of these findings may improve future treatments strategies for aBTCs. Trial Registration: NCT01566019 .

Background: The COVID-19 pandemic deeply threatens the rigorous conduct of clinical trials, notably by delaying site initiation visits, patient enrolment, treatment administration, trial-associated procedures, and data monitoring. Unlike most other medical specialties, clinical trials are an integral part of patient care in oncology. Limiting access to clinical trials therefore results in a loss of chance for patients. Methods: In this retrospective single-center study, we collected clinical trial-specific items (including patient-related or trial management-related items) during the first pandemic wave (March– June 2020) and lockdown (March 17th-May 11th) at Gustave Roussy, and compared them to those of the same period in 2019. Results: In March 2020, 84 phase I (P1) and 210 phase II/III (P2/3) trials were open. During the first pandemic wave, 21 (25%) P1 and 20 (9%) P2/3 trials were temporarily halted, following a unilateral sponsor decision in virtually all cases;all but one were industry-sponsored. Despite this, all important metrics of the P1/2 trial activity remained similar to those of 2019, including the number of patients referred for inclusion (599 vs 620), inclusion consultations (215 vs 247), patients starting treatment (130 vs 130), Internal Review Board (IRB) submissions (14 vs 16), and site initiation visits (11 vs 15), all in 2020 vs 2019, respectively. The impact of the first lock-down was more marked on P2/3, with 152 patient inclusions (vs 346 in 2019), 125 randomizations (vs 278), 43 IRB submissions (vs 50) and 34 site initiation visits (vs 40). However, in parallel, 475 patients were included in three “COVID and cancer” trials. Among the 443 P1 and 2851 P2/3 patients, 198 and 628 COVID-19 PCR were performed internally, and five and 15 (2.5%) were positive, respectively. One patient with a community-based COVID-19 died after transfer in intensive care. Conclusions: Cancer clinical trials can, and must be maintained despite challenges brought by COVID-19. Sharing experiences and retrospectively evaluating the impact on patients’ safety and cancer-related outcomes will be critical to durably improve the clinical trials conduct and to anticipate at best challenges brought by future similar crises. Legal entity responsible for the study: Gustave Roussy. Funding: Has not received any funding. Disclosure: All authors have declared no conflicts of interest.

3008 Background: RET (REarranged during Transfection) gene alterations (mutations and fusions) leading to constitutive kinase activity are identified as drivers of disease progression in multiple tumor types, including non-small cell lung cancer (NSCLC) and medullary thyroid cancer (MTC). BOS172738 is an investigational, potent, selective oral RET kinase inhibitor. This next-generation inhibitor was designed with nanomolar potency against RET and >300-fold selectivity against vascular endothelial growth factor receptor 2, to maximize the potential therapeutic window. Methods: NCT03780517 is a phase 1 study consisting of a dose-escalation and dose-expansion phase. During the escalation, 67 patients with RET-altered advanced solid tumors received once-daily oral doses of BOS172738 (10-150 mg). Intra-patient dose escalation was allowed as was over-accrual to dose levels deemed to be safe. Study endpoints were safety (CTCAE v. 4.03), tolerability and confirmed overall response rate (ORR; RECIST 1.1). The data cutoff was Dec. 11, 2020. Results: BOS172738 exhibited a favorable safety profile (n=67) for long-term administration with most treatment-emergent adverse events (TEAEs) classified as grade ≤2 and deemed unrelated to drug. The most common TEAEs were creatinine phosphokinase (CPK) increase (54%), dyspnea (34%), facial edema, aspartate aminotransferase elevation, anemia (25% each), neutropenia, diarrhea (22% each), fatigue (21%), and constipation (20%). BOS172738 was not associated with hypertension or significant hepatoxicity. BOS172738 demonstrated broad anti-tumor activity with an investigator-assessed ORR of 33% (n=18/54), a NSCLC ORR of 33% (n=10/30), MTC ORR of 44% (n=7/16, including 1 complete response) and one patient with RET fusion+ pancreatic cancer reported a partial response. Responders included patients with brain metastases with one patient whose brain lesion decreased by 43%. The median duration of response has not been reached. Many patients remain on study, including the longest of 659 days, at data cutoff. BOS172738 exhibited dose-dependent exposure (AUC, Cmax), rapid absorption (median Tmax 1 to 4.5 h), and an extended half-life (approximately 65 hours) maximizing target coverage. Conclusions: BOS172738 is a highly potent and selective RET inhibitor with a differentiated safety profile and clinical activity against RET-altered tumors, including patients with brain metastases. BOS172738 continues to be evaluated in patients with NSCLC, MTC, and in patients previously treated with other selective RET inhibitors. Clinical trial information: NCT03780517.

5510 Background: Simlukafusp alfa (SIM; FAP-IL2v) comprises an interleukin-2 variant (IL-2v) moiety and an antibody against fibroblast activation protein-α (FAP). The binding of SIM to FAP, expressed on cancer-associated fibroblasts, accounts for retention and accumulation in malignant lesions. The engineered IL-2v moiety has an abolished binding to IL-2Rα while the affinity to IL-2Rβγ is retained, resulting in activation of immune effector CD8 T and NK cells, but not of regulatory T cells, and therefore may augment activity of PD-(L)1 inhibitors. Methods: The clinical activity and safety of the SIM and atezolizumab (ATZ) combination in patients with recurrent or metastatic (R/M) cervical squamous cell carcinoma (SCC) were evaluated in a phase 2 basket study (NCT03386721). Patients (pts) were treated with SIM 10 mg IV and ATZ 1200 mg IV once every 3 weeks. The primary endpoint was objective response rate (ORR) by RECIST v1.1 assessed by investigators. Secondary endpoints were: disease control rate (DCR), duration of response (DoR), progression free survival (PFS). Results: 47 Pts with ECOG ≥1 and median age of 53 years (range: 25-69) were enrolled. All pts were checkpoint inhibitor naïve and 40 (85%) had ≥ 1 previous lines of therapy in the metastatic setting. The median number of cycles was 6 (range: 1-29). The ORR was 27% (90% CI: 18, 39) and DCR was 71% (90% CI: 58, 80) in 44 response-evaluable patients: 2 (5%) had complete response, 10 (23%) partial response, and 19 (43%) stable disease. Responses were observed across PD-L1 subgroups (SP142 assay, IC/TC cut-off ≥ 1%) with 8/22 and 4/18 responders in PD-L1+ and PD-L1- patients, respectively. Responses were durable, the median DoR was 13.3 months (95% CI: 7.6, 14.7). PFS probability at 6 months was 0.4 (95% CI: 0.27, 0.59). The most common adverse events (AE) (reported in > 30% patients), irrespective of relatedness to treatment and severity, were pyrexia (74.5%), anemia (48.9%), asthenia (48.9%), AST increased (44.7%), nausea (42.6%), ALT increased (42.6%), vomiting (36.2%) and diarrhea (31.9%). Grade 3 and 4 AEs related to SIM were observed in 63.8 % and 29.8 % of pts, respectively, while serious AEs (SAE) related to SIM were reported in 40.4%. The most common SAEs (reported in > 5% pts) irrespective of relatedness to treatment were infusion related reaction (14.9%), pyrexia (6.4%) and hydronephrosis (6.4%). One Grade 5 event occurred, which was unrelated to treatment. Conclusions: SIM in combination with ATZ demonstrated an acceptable safety profile in pts with R/M cervical SCC. The anti-tumor activity compares favorably to the approved PD-1 inhibitors in this setting and supports further exploration of IL-2v and checkpoint inhibition in this patient population of high unmet medical need. Clinical trial information: NCT03386721.

Resume L’utilisation de biomarqueurs en oncologie permet de selectionner des populations pouvant beneficier de traitements anticancereux plus efficaces et moins toxiques, et suscitent ainsi l’espoir d’une medecine de precision. L’essor de la biologie moleculaire a provoque un bouleversement du developpement des therapies moleculaires ciblees, permettant d’observer des signaux d’efficacite des les essais de phase 1 et d’accelerer ainsi l’obtention d’autorisations de mise sur le marche. A ce jour, plusieurs essais academiques, dont SHIVA, SAFIR-01 et MOSCATO-01, ont evalue l’interet de fonder le traitement oncologique sur les seules anomalies moleculaires de la tumeur, avec l’ambition a terme de supplanter les classifications anatomopathologiques et cliniques historiques. L’administration de traitements cibles a egalement permis une amelioration du pronostic de nombreux cancers, et leur recours est de plus en plus envisage en situation palliative avancee. Franchir ces frontieres permettra de tendre vers une medecine de precision en oncologie, dont l’integration en pratique clinique est desormais une attente forte des praticiens et des patients.

High-throughput genomic analyses may improve outcomes in patients with advanced cancers. MOSCATO 01 is a prospective clinical trial evaluating the clinical benefit of this approach. Nucleic acids were extracted from fresh-frozen tumor biopsies and analyzed by array comparative genomic hybridization, next-generation sequencing, and RNA sequencing. The primary objective was to evaluate clinical benefit as measured by the percentage of patients presenting progression-free survival (PFS) on matched therapy (PFS2) 1.3-fold longer than the PFS on prior therapy (PFS1). A total of 1,035 adult patients were included, and a biopsy was performed in 948. An actionable molecular alteration was identified in 411 of 843 patients with a molecular portrait. A total of 199 patients were treated with a targeted therapy matched to a genomic alteration. The PFS2/PFS1 ratio was >1.3 in 33% of the patients (63/193). Objective responses were observed in 22 of 194 patients (11%; 95% CI, 7%–17%), and median overall survival was 11.9 months (95% CI, 9.5–14.3 months). Significance: This study suggests that high-throughput genomics could improve outcomes in a subset of patients with hard-to-treat cancers. Although these results are encouraging, only 7% of the successfully screened patients benefited from this approach. Randomized trials are needed to validate this hypothesis and to quantify the magnitude of benefit. Expanding drug access could increase the percentage of patients who benefit. Cancer Discov; 7(6); 586–95. ©2017 AACR. See related commentary by Schram and Hyman, p. 552. This article is highlighted in the In This Issue feature, p. 539

Background Molecular heterogeneity drives tumour evolution and resistance to therapies. Because of biopsy challenges in non-small cell lung cancer (NSCLC), few studies compared the molecular profile between primary tumour (PT) and metastasis (Met). Here, we compare somatic single-nucleotide variants (SNVs), tumour mutational burden (TMB) and expression profiles between PT and matched metachronous Met in NSCLC patients (pts). Methods DNA and RNA were extracted from 14 matched fresh-frozen samples and PBMCs from 7 pts treated at the Institut d’Oncologie Thoracique (Gustave Roussy, GR) between 2011 and 2015. GATK4 Mutect2 were used for SNV calling and TMB on whole exome sequencing. Salmon was used for expression counts from RNAseq; Z-scores were computed by comparing data to 158 other NSCLC RNAseq from GR molecular profiling trials; GSVA R package was use to assess single sample gene set enrichment score (ssGSEA). Results Median age was 67 yo (62-78); 4/7 metastases were biopsied after chemotherapy; histology was adenocarcinoma (6) or large cell carcinoma (1). The median number of SNVs unique to PT or Met was 50 (10-143) or 79 (36-324), respectively; it was 65 (0-110) for shared SNVs; 6/7 pts had known driver alterations (TP53, KRAS, PIK3CA) in both lesions; one Met acquired a pathogenic KRAS G35C mutation. Median TMB was 3.9 (0.3-4.3) in PT and 4.1 (2.2-8.4) in Met. Genes with top negative Z-scores (MRFAP1, EP400, RIOX1, DDX27) were shared between PT and Met in 6/7 pts; genes with top positive Z-score were unique. ssGSEA-enriched genesets were shared between PT and Met; top enriched hallmarks included Myc targets, Mitotic Spindle, DNA repair, oxidative phosphorylation and TGFb signaling. Conclusions Matched comparison of molecular profiles between PT and Met evidence shared and unique alterations; this informs on tumor evolution and may help guiding treatment choice. Legal entity responsible for the study Sophie Postel-Vinay. Funding Canceropole d’Ile de France. Disclosure A Kawachi: Funded by the DUERTECC Gustave Roussy program. L. Colmet-Daage: Funded by the INSERM ATIP-Avenir grant from SPV. L. Verlingue: Honoraria (self): Adaptherapy; Research grant / Funding (self): Bristol-Myers Squibb. D. Planchard: Honoraria (self): AstraZeneca; Honoraria (self): Bristol-Myers Squibb; Honoraria (self): Boehringer Ingelheim; Honoraria (self): Celgene; Honoraria (self): Eli Lilly; Honoraria (self): Merck; Honoraria (self): Novartis; Honoraria (self): Pfizer; Honoraria (self): prIME Oncology; Honoraria (self): Peer CME; Honoraria (self): Roche; Travel / Accommodation / Expenses: AstraZeneca; Travel / Accommodation / Expenses: Bristol-Myers Squibb; Travel / Accommodation / Expenses: Boehringer Ingelheim ; Travel / Accommodation / Expenses: Roche; Travel / Accommodation / Expenses: Merck; Travel / Accommodation / Expenses: Novartis; Travel / Accommodation / Expenses: prIME Oncology; Travel / Accommodation / Expenses: Pfizer. A. Hollebecque: Research grant / Funding (institution): AstraZeneca; Research grant / Funding (institution): BMS; Research grant / Funding (institution): Boehringer Ingelheim; Research grant / Funding (institution): Janssen Cilag; Research grant / Funding (institution): Merck; Research grant / Funding (institution): Novartis; Research grant / Funding (institution): Pfizer; Research grant / Funding (institution): Roche; Research grant / Funding (institution): Sanofi. C. Massard: Research grant / Funding (self): AstraZeneca; Research grant / Funding (self): BMS; Research grant / Funding (self): Boehringer Ingelheim; Research grant / Funding (self): Janssen Cilag; Research grant / Funding (self): Merck, Novartis; Research grant / Funding (self): Pfizer; Research grant / Funding (self): Roche; Research grant / Funding (self): Sanofi; Advisory / Consultancy: Amgen; Advisory / Consultancy: Astellas; Advisory / Consultancy: AstraZeneca; Advisory / Consultancy: Bayer; Advisory / Consultancy: BeiGene; Advisory / Consultancy: BMS; Advisory / Consultancy: Celgene; Advisory / Consultancy: Debiopharm; Advisory / Consultancy: Genentech; Advisory / Consultancy: Ipsen; Advisory / Consultancy: Janssen; Advisory / Consultancy: Lilly. B. Besse: Research grant / Funding (self): Abbvie; Research grant / Funding (self): Amgen; Research grant / Funding (self): AstraZeneca; Research grant / Funding (self): Biogen; Research grant / Funding (self): Blueprint Medicines; Research grant / Funding (self): BMS; Research grant / Funding (self): Celgene; Research grant / Funding (self): Eli Lilly; Research grant / Funding (self): GSK; Research grant / Funding (self): Ignyta; Research grant / Funding (self): IPSEN; Research grant / Funding (self): Merck KGaA; Research grant / Funding (self): MSD; Research grant / Funding (self): Nektar; Research grant / Funding (self): Onxeo; Research grant / Funding (self): Pfizer; Research grant / Funding (self): Pharma Mar; Research grant / Funding (self): Sanofi; Research grant / Funding (self): Spectrum Pharmaceuticals; Research grant / Funding (self): Takeda. J. Soria: Advisory / Consultancy: AstraZeneca; Advisory / Consultancy: Astex; Advisory / Consultancy: Clovis; Advisory / Consultancy: GSK; Advisory / Consultancy: GamaMabs; Advisory / Consultancy: Lilly; Advisory / Consultancy: MSD; Advisory / Consultancy: Mission Therapeutics; Advisory / Consultancy: Merus; Advisory / Consultancy: Pfizer; Advisory / Consultancy: PharmaMar; Advisory / Consultancy: Pierre Fabre; Advisory / Consultancy: Roche/Genentech; Advisory / Consultancy: Sanofi; Advisory / Consultancy: Servier; Advisory / Consultancy: Symphogen; Advisory / Consultancy: Takeda; Full / Part-time employment: AstraZeneca. S. Postel Vinay: Research grant / Funding (self): AstraZeneca; Research grant / Funding (self): BMS; Research grant / Funding (self): Boehringer Ingelheim; Research grant / Funding (self): Janssen Cilag; Research grant / Funding (self): Merck; Research grant / Funding (self): Novartis; Research grant / Funding (self): Pfizer; Research grant / Funding (self): Roche; Research grant / Funding (self): Sanofi. All other authors have declared no conflicts of interest.