P3.13-29 Patient-Derived Xenograft Models(PDX) of Lung Squamous Cell Carcinoma(SCC) for Preclinical Studies

Abstract

tumor. Because PDL1 expression in tumor is used to identify patients who might benefit from immune-modulating treatment, its detection plays a key role in clinical recommendations. Our objectives are to assess the prevalence of PDL1 expression in early stage non-small cell lung cancer (NSCLC) patients, determine its association with clinical outcomes using the Glans-Look Research (GLR) database (Calgary, AB), and validate these findings using a cohort from the Manitoba Tumor Bank (MBTB). Method: A tissue microarray (TMA) was built using pretreatment resected and biopsy tissue samples from 459 GLR database patients with early stage NSCLC, diagnosed between 2003 and 2010. Cell lines expressing varying levels of PDL1 were generated, embedded into HistoGel TM and co-mounted onto the GLR and MBTB arrays. Fluorescence immunohistochemistry was performed using anti-PDL1 E1L3N (Cell Signaling Technology), and PDL1 expression was evaluated as percent-positive and intensity scores in the cytoplasmic compartment of tumor and stromal cells using HALO TM automated image analysis software. Cell line PDL1 intensity scores served as on-slide reference standards to normalize PDL1 expression in patient specimens using R Programming software. PDL1-percent-positive tumor scores were generated to assess the cut-points of 50%("PDL1strong"), 1%-to-49%("PDL1-weak"), and <1%("PDL1-negative"), indicated by the FDA-approved companion diagnostic anti-PDL1 22C3 (Dako) for pembrolizumab. Clinicopathological outcomes were analyzed, and overall survival was assessed using the Kaplan-Meier method and compared using the log-rank test. Result: Preliminary analyses indicate PDL1-weak/negative GLR patients with adenocarcinoma experienced higher median OS (3.50yrs) compared to PDL1strong patients (1.91yrs) (p¼0.0043). This trend was not significant over all histologies, or when using mean scores. The opposite trend was found with the MBTB cohort (2.52yrs vs. 1.76yrs OS, PDL1-strong vs. PDL1-weak/negative maximum scores, p¼0.0410). Conclusion: Variations across datasets illustrate the difficulty in harmonizing PDL1 testing. Heterogeneity of protein expression, TMA sampling error, and differences between study cohorts can translate into variable correlations between PDL1-positivity and survival estimates. Increased survivorship in GLR adenocarcinoma patients with PDL1-weak/negative staining could challenge the notion of using PDL1 as a prognostic biomarker. Comparisons between the E1L3N and 22C3 anti-PDL1 assays will be performed, E1L3N percent-positive cut-points will be refined according to the lowest intensity-based statistical p-value, and further outcome findings will be presented and discussed.Background: Non-small cell lung cancer is the most prevalent type of lung cancer, and one of its subtype, squamous cell carcinoma has limited targeted therapeutic options in comparison with adenocarcinoma. For this reason, we established the patient-derived xenograft models of primary lung cancer and then focused on the SCC PDX models to be used for the disc...