No abstract

No abstract

Background: Programmed death-ligand 1 (PD-L1) expression is the only FDA-approved biomarker for immune checkpoint inhibitors (ICIs) in patients with lung adenocarcinoma, but sensitivity is modest. Understanding the impact of molecular phenotype, clinical characteristics, and tumor features on PD-L1 expression is largely unknown and may improve prediction of response to ICI. Patients and methods: We evaluated patients with lung adenocarcinoma for whom PD-L1 testing and targeted nextgeneration sequencing (using MSK-IMPACT) was performed on the same tissue sample. Clinical and molecular features were compared across PD-L1 subgroups to examine how molecular phenotype associated with tumor PD-L1 expression. In patients treated with anti-PD-(L)1 blockade, we assessed how these interactions impacted efficacy. Results: A total of 1586 patients with lung adenocarcinoma had paired PD-L1 testing and targeted next-generation sequencing. PD-L1 negativity was more common in primary compared to metastatic samples (P < 0.001). The distribution of PD-L1 expression (lymph nodes enriched for PD-L1 high; bones predominantly PD-L1 negative) and predictiveness of PD-L1 expression on ICI response varied by organ. Mutations in KRAS, TP53, and MET significantly associated with PD-L1 high expression (each P < 0.001, Q < 0.001) and EGFR and STK11 mutations associated with PD-L1 negativity (P < 0.001, Q ¼ 0.01; P ¼ 0.001, Q < 0.001, respectively). WNT pathway alterations also associated with PD-L1 negativity (P ¼ 0.005). EGFR and STK11 mutants abrogated the predictive value of PD-L1 expression on ICI response. Conclusion: PD-L1 expression and association with ICI response vary across tissue sample sites. Specific molecular features are associated with differential expression of PD-L1 and may impact the predictive capacity of PD-L1 for response to ICIs.

Malignant pleural diseases, comprising metastatic lung and breast cancers and malignant pleural mesothelioma (MPM), are aggressive solid tumors with poor therapeutic response. We developed and conducted a first-in-human, phase I study of regionally delivered, autologous, mesothelin-targeted chimeric antigen receptor (CAR) T-cell therapy. Intrapleural administration of 0.3M to 60M CAR T cells/kg in 27 patients (25 with MPM) was safe and well tolerated. CAR T cells were detected in peripheral blood for &gt;100 days in 39% of patients. Following our demonstration that PD-1 blockade enhances CAR T-cell function in mice, 18 patients with MPM also received pembrolizumab safely. Among those patients, median overall survival from CAR T-cell infusion was 23.9 months (1-year overall survival, 83%). Stable disease was sustained for ≥6 months in 8 patients; 2 exhibited complete metabolic response on PET scan. Combination immunotherapy with CAR T cells and PD-1 blockade agents should be further evaluated in patients with solid tumors. Significance: Regional delivery of mesothelin-targeted CAR T-cell therapy followed by pembrolizumab administration is feasible, safe, and demonstrates evidence of antitumor efficacy in patients with malignant pleural diseases. Our data support the investigation of combination immunotherapy with CAR T cells and PD-1 blockade agents in solid tumors. See related commentary by Aldea et al., p. 2674. This article is highlighted in the In This Issue feature, p. 2659

BackgroundPhase contrast (PC) cardiovascular magnetic resonance (CMR) is widely employed for flow quantification, but analysis typically requires time consuming manual segmentation which can require human correction. Advances in machine learning have markedly improved automated processing, but have yet to be applied to PC-CMR. This study tested a novel machine learning model for fully automated analysis of PC-CMR aortic flow.MethodsA machine learning model was designed to track aortic valve borders based on neural network approaches. The model was trained in a derivation cohort encompassing 150 patients who underwent clinical PC-CMR then compared to manual and commercially-available automated segmentation in a prospective validation cohort. Further validation testing was performed in an external cohort acquired from a different site/CMR vendor.ResultsAmong 190 coronary artery disease patients prospectively undergoing CMR on commercial scanners (84% 1.5T, 16% 3T), machine learning segmentation was uniformly successful, requiring no human intervention: Segmentation time was < 0.01 min/case (1.2 min for entire dataset); manual segmentation required 3.96 ± 0.36 min/case (12.5 h for entire dataset). Correlations between machine learning and manual segmentation-derived flow approached unity (r = 0.99, p < 0.001). Machine learning yielded smaller absolute differences with manual segmentation than did commercial automation (1.85 ± 1.80 vs. 3.33 ± 3.18 mL, p < 0.01): Nearly all (98%) of cases differed by ≤5 mL between machine learning and manual methods. Among patients without advanced mitral regurgitation, machine learning correlated well (r = 0.63, p < 0.001) and yielded small differences with cine-CMR stroke volume (∆ 1.3 ± 17.7 mL, p = 0.36). Among advanced mitral regurgitation patients, machine learning yielded lower stroke volume than did volumetric cine-CMR (∆ 12.6 ± 20.9 mL, p = 0.005), further supporting validity of this method. Among the external validation cohort (n = 80) acquired using a different CMR vendor, the algorithm yielded equivalently small differences (∆ 1.39 ± 1.77 mL, p = 0.4) and high correlations (r = 0.99, p < 0.001) with manual segmentation, including similar results in 20 patients with bicuspid or stenotic aortic valve pathology (∆ 1.71 ± 2.25 mL, p = 0.25).ConclusionFully automated machine learning PC-CMR segmentation performs robustly for aortic flow quantification - yielding rapid segmentation, small differences with manual segmentation, and identification of differential forward/left ventricular volumetric stroke volume in context of concomitant mitral regurgitation. Findings support use of machine learning for analysis of large scale CMR datasets.Electronic supplementary materialThe online version of this article (10.1186/s12968-018-0509-0) contains supplementary material, which is available to authorized users.