Immunotherapy efficacy predictive tool for lung adenocarcinoma based on neural network

Area of Science:

• Oncology
• Computational Biology
• Medical Informatics

Background:

• Immunotherapy has shown efficacy in lung adenocarcinoma (LUAD).
• Predicting which LUAD patients will benefit from immunotherapy remains a significant challenge.
• Accurate prediction of treatment response is crucial for optimizing patient outcomes and resource allocation.

Purpose of the Study:

• To develop and validate a predictive tool for immunotherapy efficacy in LUAD patients.
• To utilize neural networks for predicting objective response rate (ORR), disease control rate (DCR), and patient response.
• To create a tool that assists in identifying beneficiaries of immunotherapy for LUAD.

Main Methods:

• Retrospective analysis of 250 LUAD patients receiving immunotherapy.
• Development of neural network models using an 80% training dataset.
• Validation of models on a 20% test dataset to predict ORR, DCR, responder status, and overall survival (OS).

Main Results:

• The predictive tool achieved high Area Under the Curve (AUC) scores for ORR (0.9016 training, 0.8173 test) and DCR (0.8570 training, 0.8244 test).
• The tool demonstrated strong performance in predicting responders (AUC 0.8395 training, 0.8214 test).
• OS prediction showed moderate AUC scores (0.6627 training, 0.6357 test).

Conclusions:

• A neural network-based tool effectively predicts immunotherapy efficacy in LUAD patients.
• The tool demonstrates reliable prediction of ORR, DCR, and responder status.
• This predictive tool can aid clinicians in selecting appropriate patients for immunotherapy.