AI digital twin platform advances personalized cancer treatment planning

by Hong Kong Polytechnic University

edited by Swati Mestri, reviewed by Robert Egan

Swati Mestri

Scientific Editor

Meet our editorial team
Behind our editorial process

Robert Egan

Senior Editor

Meet our editorial team
Behind our editorial process Editors' notes

This article has been reviewed according to Science X's editorial process and policies. Editors have highlighted the following attributes while ensuring the content's credibility:

fact-checked

trusted source

proofread

The GIST Add as preferred source

A research team at The Hong Kong Polytechnic University (PolyU) has developed a patient-centric "Artificial Intelligence (AI) Virtual Patient Simulation System," overcoming the limitations of conventional static diagnosis.

By dynamically integrating multimodal patient data, including genomic data, medical imaging and clinical records, the system creates a continuously updated "digital twin" model. It can not only track changes in a patient's condition in real time but also predict the potential effectiveness of different cancer treatment options, helping health care teams formulate more precise and personalized medical solutions.

The work is published in the journal Medical Image Analysis.

Other medical AI tools often rely on a single CT scan, genomic report or static clinical data for analysis, making it difficult to gain a comprehensive understanding of dynamic changes in a patient's condition. Lawrence Chan, associate professor of the PolyU Department of Health Technology and Informatics, led the team that developed the "AI Virtual Patient Simulation System" based on a patient-centric digital twin platform.

Combining a platform for health care professionals with a patient-facing mobile application, the system can conduct predictive analyses in response to real-time changes in a patient's condition and simulate the effectiveness of different treatment options. It provides intelligent support for clinical diagnosis, condition monitoring and treatment assessment.

The system is particularly suitable for cancer and critical care, where disease progression can be complex, treatment options diverse and medical costs high, supporting the development of precision medicine.

The system's core strength lies in close collaboration between health care professionals and patients. The dedicated health care platform integrates multimodal data, including genomic data, medical imaging, pathology reports, laboratory test results and clinical records, helping doctors gain a comprehensive overview of a patient's condition, enhance diagnostic and treatment decision-making, and streamline multidisciplinary consultations and referral processes.

Meanwhile, the patient-facing mobile application enables patients to upload medical records, log daily symptoms and track their health status. Through an encrypted Deep Feature QR code, medical data can be securely transferred across different clinics, hospitals and devices, enhancing data-sharing efficiency while safeguarding privacy. With the system, patients can shift from passively receiving treatment to actively participating in their health management, further strengthening doctor-patient collaboration.

To advance the application of this technology in cancer care and treatment decision-making, the research team has introduced a clinical, data-driven, multiscale AI framework for predicting immunotherapy response in patients with non-small cell lung cancer.

The multimodal approach effectively integrates histopathological image features with clinical data, including gene expression profiles and cancer-type text. Named the Visual-Global Relation Fusion Network (ViGNet), the novel framework incorporates both a multiscale visual encoder and a gene-driven encoder, enabling AI to analyze image and genomic features that are closely related to cancer treatment response.

In qualitative and quantitative evaluations, ViGNet outperformed baseline approaches in response classification, achieving 82.55% discrimination performance in predicting immunotherapy response. The research enables more efficient integration of multisource data and supports the practical deployment of AI methods in clinical settings, providing insights for personalized treatment and clinical decision-making.

Chan said, "The AI Virtual Patient Simulation System is an innovative and comprehensive platform that integrates diagnosis, monitoring and treatment assessment. In addition to identifying subtle yet crucial pathological connections across multimodal data, the system can also act as a 'monitoring sentinel,' alerting health care teams when a patient's biomarkers or symptoms show abnormalities. This technology helps shorten diagnosis and assessment times, supporting health care professionals in developing more precise, effective and personalized treatment plans for patients with cancer or other critical illnesses."

More information

Luoyi Kong et al, ViGNet: A clinical data-supported deep learning approach for NSCLC immunotherapy response prediction in digital pathology, Medical Image Analysis (2026). DOI: 10.1016/j.media.2026.104154

Key medical concepts

Carcinoma, Non-Small-Cell LungExpression Profile, Gene

Clinical categories

Oncology Provided by Hong Kong Polytechnic University Who's behind this story?

Swati Mestri

Swati Mestri holds a bachelor's degree in Electronics Engineering and has worked as a content editor since 2019. She has experience editing research documents across technology, health care, and materials science, and has a particular interest in technology and space. Full profile →

Robert Egan

Bachelor's in mathematical biology, Master's in creative writing. Well-traveled with unique perspectives on science and language. Full profile →

Citation: AI digital twin platform advances personalized cancer treatment planning (2026, August 13) retrieved 14 August 2026 from https://medicalxpress.com/news/2026-08-ai-digital-twin-platform-advances.html This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no part may be reproduced without the written permission. The content is provided for information purposes only.