Luiza Moore, Senior Director in Global Oncology Diagnostics at AstraZeneca, shared a post on LinkedIn about a recent article she and her colleagues co-authored, adding:
“Pleased to announce our latest work is now published, highlighting AI-enabled precision oncology with a high‑accuracy EGFR mutation predictor.
Despite clear guidelines, EGFR mutation screening in lung cancer is inconsistent globally—driving under-testing and delaying access to targeted therapies.
We developed and validated a deep‑learning model that predicts EGFR mutation status in non‑small cell lung cancer (NSCLC) from routine H&E whole‑slide images. Trained on 12,000+ slides and validated across diverse, multi-site datasets, the model demonstrated strong performance. Performance was robust across specimen types (biopsy and resection), tumour subtypes, mutation subtypes and multiple slide scanners, with 90.4% concordance across scanners in a multi‑scanner cohort.
This work demonstrates the potential for computational pathology tools to aid molecular screening for patients with cancer, help prioritise testing where resources are limited, and speed time‑to‑treatment — while noting such tools are intended to complement, not replace, molecular assays.
Huge thanks and congratulations to the entire team!”
Title: Deep learning predicts EGFR mutation status from histology images in non-small cell lung cancer
Authors: Kennedy Misso, Comfort M. Takang, Kondo Chilonga, Mujaheed Suleman, Yotham Gwanika, Pius Aloyce, Glory Makupa, Augustine Musyoka, Arkadiusz Z. Dudek, Randy Hurley, Furaha Serventi
Read the Full Article on Cancer Research Communications
More posts about Lung cancer.