Garassino, Marina Chiara

Mesothelioma is a rare cancer that originates from the mesothelial surfaces of the pleura and other sites, and is estimated to occur in approximately 3,500 people in the United States annually. Pleural mesothelioma is the most common type and represents approximately 85% of these cases. The NCCN Guidelines for Mesothelioma: Pleural provide recommendations for the diagnosis, evaluation, treatment, and follow-up for patients with pleural mesothelioma. These NCCN Guidelines Insights highlight significant updates to the NCCN Guidelines for Mesothelioma: Pleural, including revised guidance on disease classification and systemic therapy options.

8549 Background: First-line treatment of early stage non-small cell lung cancer (NSCLC) is surgical resection, but with a 5-year survival of only 54%, rates of future relapse are high. Identifying patients at high risk of relapse can help guide adjuvant treatment decisions. Deep convolutional neural network (DCNN) AI models trained on tumor histology have shown incredible flexibility as potential biomarkers, both in lung cancer and more generally across many malignancies. While DCNN models can obtain extremely accurate results when used for routine purposes such as diagnosis, subtyping, and grading of malignancies, most models trained for prognostication or treatment response prediction do not reach performance sufficient for clinical application. Uncertainty quantification (UQ) – a family of techniques that give DCNN models the ability to report confidence alongside predictions – is an underexplored avenue in cancer AI that may help further improve performance and clinical application of models designed to provide clinicians with estimations of risk. Methods: To explore the potential use of UQ in clinically-oriented DCNN models, we trained models on a single-institution, retrospective digital tumor histology slide cohort from patients with Stage I-III NSCLC who underwent surgical resection to predict risk of future relapse. Estimation of uncertainty was performed using dropout as a Bayesian approximation, and uncertainty thresholds were calculated from training sets using a novel method to identify and remove low-confidence predictions. For comparison, a separate multivariate logistic regression was trained in cross-validation using known clinical risk factors. Results: We trained DCNN models on slides from 198 patients (40 relapsed, 158 without relapse). In this cohort, 130 patients had stage I disease (65.6%), 42 had stage II (21.2%), and 26 had stage III (13.2%). The average age was 69 years, 85% were current or previous smokers, and 38 received guideline-concordant adjuvant chemotherapy. Without UQ, a DCNN model predicted risk of future relapse with an average area under receiver operator curve (AUROC) of 0.74 across three-fold cross-validation. Using UQ estimation, 63% of slides were reported with high confidence. Relapse prediction was significantly improved in the high-confidence cohort, with an average AUROC of 0.83 in cross-validation. With a specificity of 70%, this corresponds to an average sensitivity of 86.8% across the three cross-folds (79.1%, 85.5%, and 95.7%). In comparison, a clinical-only multivariate regression model predicted relapse with a cross-validated AUROC of 0.67. Conclusions: This method of uncertainty quantification appears to be a powerful tool to predict lung cancer recurrence risk from digital histopathology while simultaneously providing clinicians with a measurement of algorithm trustworthiness.

9093 Background: Treatment addressing non-small cell lung cancer (NSCLC) harboring EGFR or HER2 exon 20 insertion mutations remains an unmet need. These tumors are associated with a high incidence of CNS metastases and unfavorable survival rates. Poziotinib is a potent, irreversible, tyrosine kinase inhibitor (TKI) with a structure that can overcome the steric hindrance of the exon 20 limited binding pocket. Preclinical data suggest poziotinib CNS penetration, and here we show meaningful poziotinib CNS activity in patients with NSCLC harboring exon 20 insertion mutations in an ongoing multi-cohort, multi-center Phase 2 study (ZENITH20; NCT03318939). Methods: ZENITH20 enrolled previously treated and naïve patients with advanced/metastatic NSCLC and EGFR or HER2 exon 20 insertion mutations in several cohorts: Cohort 1 (C1) EGFR previously treated; Cohort 2 (C2) HER2 previously treated and Cohort 3 (C3) EGFR treatment-naïve. All patients with stable CNS metastases at baseline were included. Poziotinib (16 mg) was administered orally QD, with follow-up for up to 24 months. The primary endpoint was Objective Response Rate (ORR) evaluated centrally using RECIST v1.1 by an independent image review committee. Secondary endpoints included Disease Control Rate (DCR), Duration of Response (DOR), Progression-Free Survival (PFS) and safety. Primary efficacy results have been previously released. Intracranial response was determined based on the modified RECIST criteria. Results: A total of 284 patients across 3 cohorts (C1 n=115; and C2 n=90; and C3 n=79) with a median age of 60.5 years were enrolled. The median follow-up was 7.3, 8.3, and 9.2 months for all patients in C1, C2, and C3, respectively. In NSCLC patients that had baseline CNS lesions (N=36), the analysis showed a patient-based ORR of 22.2% (8/36) and a DCR of 88.9% (32/36). One patient in each cohort had a complete intracranial response and stable disease was 80.6% across 3 cohorts and 92.9% in C2. Two patients each in C1 and C3 had progressive disease (PD) and none had CNS progression in C2 (Table). Conclusions: Poziotinib exhibited clinically meaningful CNS activity in patients with EGFR or HER2 exon 20 mutations in ZENITH20 Cohorts 1-3. The majority of the patients had no CNS progression and 3/36 patients had intracranial complete responses. The preliminary data suggest that poziotinib may provide a meaningful treatment alternative for patients with NSCLC that harbor EGFR or HER2 exon 20 mutations and who present with CNS metastases that have poor prognosis. Clinical trial information: NCT03318939. [Table: see text]

PURPOSE: IMpower133 (ClinicalTrials.gov identifier: NCT02763579 ), a randomized, double-blind, phase I/III study, demonstrated that adding atezolizumab (anti-programmed death-ligand 1 [PD-L1]) to carboplatin plus etoposide (CP/ET) for first-line (1L) treatment of extensive-stage small-cell lung cancer (ES-SCLC) resulted in significant improvement in overall survival (OS) and progression-free survival (PFS) versus placebo plus CP/ET. Updated OS, disease progression patterns, safety, and exploratory biomarkers (PD-L1, blood-based tumor mutational burden [bTMB]) are reported. PATIENTS AND METHODS: Patients with untreated ES-SCLC were randomly assigned 1:1 to receive four 21-day cycles of CP (area under the curve 5 mg per mL/min intravenously [IV], day 1) plus ET (100 mg/m 2 IV, days 1-3) with atezolizumab (1,200 mg IV, day 1) or placebo, and then maintenance atezolizumab or placebo until unacceptable toxicity, disease progression, or loss of clinical benefit. Tumor specimens were collected; PD-L1 testing was not required for enrollment. The two primary end points, investigator-assessed PFS and OS, were statistically significant at the interim analysis. Updated OS and PFS and exploratory biomarker analyses were conducted. RESULTS: Patients received atezolizumab plus CP/ET (n = 201) or placebo plus CP/ET (n = 202). At the updated analysis, median follow-up for OS was 22.9 months; 302 deaths had occurred. Median OS was 12.3 and 10.3 months with atezolizumab plus CP/ET and placebo plus CP/ET, respectively (hazard ratio, 0.76; 95% CI, 0.60 to 0.95; descriptive P = .0154). At 18 months, 34.0% and 21.0% of patients were alive in atezolizumab plus CP/ET and placebo plus CP/ET arms, respectively. Patients derived benefit from the addition of atezolizumab, regardless of PD-L1 immunohistochemistry or bTMB status. CONCLUSION: Adding atezolizumab to CP/ET as 1L treatment for ES-SCLC continued to demonstrate improved OS and a tolerable safety profile at the updated analysis, confirming the regimen as a new standard of care. Exploratory analyses demonstrated treatment benefit independent of biomarker status.

Abstract Background: IMpower133 (NCT02763579), a global Phase I/III, randomized, double-blind, placebo (PBO)-controlled trial, showed that the addition of atezo (anti-PD-L1) to CE for 1L ES-SCLC led to statistically and clinically significant OS and PFS improvement vs CE alone. Here we report updated OS and exploratory analyses. Methods: Pts with untreated ES-SCLC were randomized 1:1 to receive four 21-day cycles of E (100 mg/m2 IV, days 1-3) + C (AUC 5 mg/mL/min IV, day 1) with atezo (1200 mg IV, day 1) or PBO, followed by maintenance therapy with atezo or PBO until intolerable toxicity, progression or loss of clinical benefit. PD-L1 testing was not required for enrollment. Coprimary endpoints were investigator-assessed PFS (RECIST 1.1) and OS. Interim and final OS analyses were planned for ≈240 and ≈306 events, respectively. OS was significant at the interim analysis. Updated OS, exploratory biomarkers and patterns of disease progression were analyzed. Results: 201 and 202 pts were randomized to receive atezo+CE and PBO+CE, respectively. The median follow-up was 22.9 mo and 302 deaths had occurred. Median OS for the atezo and PBO arms was 12.3 and 10.3 mo, respectively (HR, 0.76 [95% CI: 0.60, 0.95]; descriptive P = 0.0154). At the 18-mo landmark, the OS rate was 13% higher with atezo+CE than with PBO+CE (Table). Exploratory analyses showed treatment benefit with atezo+CE regardless of biomarker status. 181 (90.0%) And 194 (96.0%) pts in the atezo+CE and PBO+CE arms, respectively, had RECIST-defined disease progression. Progression at existing, new or existing and new lesions was numerically lower with atezo+CE than with PBO+CE. Common sites of new lesions included the CNS, lung, lymph node and liver, with similar incidences between arms. Conclusion: Adding atezo to CE continued to provide OS improvement for 1L ES-SCLC in an all-comer population. The updated results of IMpower133 further support this regimen for untreated ES-SCLC. Landmark OSAtezo + CE, n = 201PBO + CE, n = 20212 Mo, n (%)93 (51.9)74 (39.0)18 Mo, n (%)61 (34.0)39(21.0)Median OS in biomarker subgroupsAtezo + CEPBO + CEITT (N = 403), mo12.310.3HR (95% CI)0.76 (0.61, 0.96)aITT-BEP (n = 137), mo9.98.9HR (95% CI)0.70 (0.48, 1.02)Non-BEP (n = 266), mo14.611.2HR (95% CI)0.81 (0.61, 1.08)PD-L1 expression, 1% TC or IC< 1% (n = 65), mo10.28.3HR (95% CI)0.51 (0.30, 0.89)≥ 1% (n = 72), mo9.710.6HR (95% CI)0.87 (0.51, 1.49)PD-L1 expression, 5% TC or IC< 5% (n = 108), mo9.28.9HR (95% CI)0.77 (0.51, 1.17)≥ 5% (n = 29), mo21.69.2HR (95% CI)0.60 (0.25, 1.46)Disease progression at sites, n (%)Atezo + CEPBO + CEExisting116 (57.7)131 (64.9)New86 (42.8)99 (49.0)Existing and new42 (20.9)57 (28.2)BEP, biomarker-evaluable population; ITT, intention-to-treat population.a Stratified HR. BEP included pts evaluable by PD-L1 IHC using the VENTANA SP263 assay. Citation Format: Leora Horn, Stephen V. Liu, Aaron S. Mansfield, Tony Mok, Arnaud Scherpereel, Niels Reinmuth, Marina Chiara Garassino, Javier De Castro Carpeno, Raffaele Califano, Makoto Nishio, Francisco Orlandi, Jorge Arturo Alatorre Alexander, Ticiana Leal, Ying Cheng, Jong-Seok Lee, Sivuonthanh Lam, Mark McCleland, Yu Deng, See Phan, Martin Reck. IMpower133: Updated OS and exploratory analyses of first-line (1L) atezolizumab (atezo) + carboplatin (C) + etoposide (E) in extensive-stage SCLC (ES-SCLC) [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr CT220.