2023 World Conference on Lung Cancer (Posters)-P1.08. CoxMoE: A Deep Learning Model Predicts Survival and Response in T790M Positive NSCLC Patients Treated with Third-Generation EGFR-TKI

P1.08. CoxMoE: A Deep Learning Model Predicts Survival and Response in T790M Positive NSCLC Patients Treated with Third-Generation EGFR-TKI - PDF(Abstract)

Back to course

Pdf Summary

This study presents a deep learning model called CoxMoE that predicts survival and treatment response in non-small cell lung cancer (NSCLC) patients with T790M mutation, using third-generation EGFR-TKI. The model uses electronic medical records and pharmacokinetic parameters to predict treatment responses. It was trained and validated using data from 330 NSCLC patients, and its performance was measured using Area Under the Curve (AUC) and Concordance index (C-index). CoxMoE exhibited AUCs of 0.74-0.82 for predicting best objective response (BOR) and achieved a C-index of 0.67 and 0.64 for progression-free survival (PFS) prediction in the training and validating cohorts. The model outperformed traditional machine learning and linear deep learning methods. <br /><br />The study also identified several routine laboratory tests and pharmacokinetic parameters that were linked to drug resistance and tumor aggressive pathways, such as Activated Partial Thromboplastin Time (APTT), Creatinine Clearance (Ccr), Monocyte, and steady-state plasma trough concentration. APTT and Ccr showed a strong correlation with BOR and PFS, respectively. <br /><br />CoxMoE provides a non-invasive strategy for predicting treatment response and survival outcomes in NSCLC patients with T790M mutation using routine laboratory tests and pharmacokinetic parameters. This can potentially complement the current EGFR genotype detection methods.

Asset Subtitle

Ning Lou

Meta Tag

Speaker Ning Lou

Topic Tumor Biology: Translational Biology - Drug Resistance

Keywords

CoxMoE

deep learning model

NSCLC

T790M mutation

pharmacokinetic parameters

AUC

PFS

machine learning

drug resistance

routine laboratory tests