Genomic and demographic landscape of non-small cell lung cancer within an ethnically-diverse population – The implications for radiation oncology and personalised medicine

Abstract

Introduction

Genomics and personalised medicine are increasingly important in managing non-small cell lung cancer (NSCLC). However, clinical trials driving practice changes are frequently lacking in ethnic diversity, with limited published data for these groups. We report real-world demographic and genomic data from an ethically-diverse Australian population.

Methods

Retrospective review of all sequential patients with NSCLC referred to a tertiary oncology service (two centres) for radiotherapy between April 2020 to February 2022. Clinicopathological data (including histopathology/genomic results) were extracted from medical records. Genomic data was only routinely available for non-squamous pathologies. We compiled and summarised this genomic data and made various correlations to clinicopathological features in the cohort, including country of birth. The Genomics Evidence Neoplasia Information Exchange (GENIE) is a publicly accessible registry of genomic and clinical data associated with 185000 patients across all tumour types. We compared the genomic features of our cohort to this population-based registry.

Results

174 patients with 189 unique malignancies were identified. Nearly 60% of patients were born overseas. 113 specimens underwent next generation sequencing (NGS). 72% of tested specimens had ≥1 mutation identified with EGFR (39%), KRAS (20.4%) and TP53 (17.7%) genes being most represented. Mutation prevalence patterns were related to ethnicity, with East Asian ancestry predicting EGFR mutation (72% vs 31%; p<0.002). Smoking exposure and Australian birth (34% vs 6%; p=0.009) predicted KRAS mutation. Genomic mutations differed compared with the GENIE cohort, with our cohort having less ATM, ERBB4, KRAS, STK11 and TP53 mutations, with a numerical trend to more EGFR mutations (39% vs 26%; p=0.19), correlating to ethnic diversity with our larger Asian representation (25% vs 7%; p<0.00001).

Conclusions

An ethnically-diverse population with NSCLC had significant genomic differences compared to major clinical databases/tissue repositories, relative to clinicopathological features. Under-representation of ethnic minorities casts doubt on the applicability of trial results due to the clinical impact of mutations in real-world populations. Further efforts to increase ethnic breadth of trial enrolment and radiotherapy-specific gene content in panel design are essential to improving personalised radiation oncology practice.