PO.MD01.01 · 分子诊断与数据
通过对话式人工智能赋能高危人群：一个用于患者倡导和精准肿瘤学的框架
Empowering populations at risk through conversational artificial intelligence: a framework for patient advocacy and precision oncology
作者与单位 Authors & Affiliations
摘要 Abstract
中文摘要
背景：
结直肠癌及其他癌症的高危人群仍面临着获得公平医疗的障碍，包括健康素养有限、沟通碎片化，以及在基因组研究中代表性不足。传统的患者参与方式常常忽视文化和语言的多样性，使许多患者缺乏充分参与自身诊疗所需的知识或工具。为弥补这一空白，我们开发了一个对话式 AI-患者倡导框架，旨在帮助患者以个性化、知情的方式理解、获取并受益于精准肿瘤学。
方法：
该创新平台建立在我们已验证的 AI-HOPE 生态系统之上（该系统已在多个精准肿瘤学研究应用中成功实施和测试），作为一个数字患者倡导者运行，整合了来自 AACR Project GENIE 数据库的见解——这是一个链接来自所有人群的临床和基因组数据的全球癌症登记库。该系统利用自然语言处理和可解释 AI，将复杂的癌症相关问题转化为通俗易懂的解释，帮助患者理解诊断和治疗选择，并将用户连接至倡导和教育资源。该框架强调以同理心驱动的对话、文化适应性，以及对所有人群（包括那些在精准医疗中历来代表性不足者）的可及性。
结果：
使用 GENIE 衍生数据集的初步实施使得能够获得关于癌症突变模式的群体层面见解，随后将这些见解转化为面向患者的对话式叙述。早期测试显示，患者对基因组结果的理解得到改善、对医患沟通的信心增强，以及对倡导资源和临床试验信息的参与度提高。
结论：
对话式 AI-患者倡导代表了虚拟患者倡导的一种新范式，将患者教育与源自 AACR GENIE 等大规模数据库的精准肿瘤学见解相结合。通过充当虚拟倡导者，这项技术使患者能够以清晰和信任来驾驭其诊疗历程——弥合复杂癌症数据与所有人群可付诸行动的理解之间的鸿沟。
查看英文原文 English abstract
Background:
Populations at risk for colorectal and other cancers continue to face barriers to equitable healthcare, including limited health literacy, fragmented communication, and underrepresentation in genomic research. Traditional approaches to patient engagement often overlook cultural and linguistic variety, leaving many patients without the knowledge or tools to fully participate in their care. To address this gap, we developed a Conversational AI-Patient Advocacy framework designed to help patients understand, access, and benefit from precision oncology in a personalized informed way.
Methods:
Built upon our validated AI-HOPE ecosystem, which has been successfully implemented and tested across several precision oncology research applications, this innovative platform functions as a digital patient advocate that integrates insights from the AACR Project GENIE database-a global cancer registry linking clinical and genomic data from all populations. Using natural language processing and explainable AI, the system translates complex cancer related questions into accessible, plain-language explanations, supports patient comprehension of diagnostics and treatment options, and connects users to advocacy and educational resources. The framework emphasizes empathy-driven dialogue, cultural adaptability, and accessibility for all populations, including those historically underrepresented in precision medicine.
Results:
Preliminary implementation using GENIE-derived datasets enabled population-level insights into cancer mutation patterns that were then translated into conversational narratives for patients. Early testing demonstrated improved understanding of genomic results, enhanced confidence in patient-provider communication, and increased engagement with advocacy resources and clinical trial information.
Conclusions:
Conversational AI-Patient Advocacy represents a new paradigm for virtual patient advocacy, combining patient education with precision oncology insights derived from large-scale databases like AACR GENIE. By acting as a virtual advocate, this technology empowers patients to navigate their care journey with clarity and trust-bridging the gap between complex cancer data and actionable understanding for all populations.
利益披露 Disclosure
A. Estrada, None..
B. Waldrup, None..
F. G. Carranza, None..
S. Manjarrez, None..
E. Velazquez-Villarreal, None.