📄 中文摘要
放疗技术正变得越来越精确和数据密集，当前的治疗方案生成高频率的影像和剂量数据流，非常适合用于基于人工智能的时间建模，以描述正常组织随时间的演变。在生物指导放疗（BGRT）中，每次治疗的非小细胞肺癌（NSCLC）患者都会记录新的代谢、解剖和剂量信息。然而，临床决策主要依赖于静态的、基于人群的正常组织并发症概率（NTCP）模型，这些模型忽视了顺序数据中编码的动态、独特的生物轨迹。为此，开发了COMPASS（综合个性化评估系统），作为一种安全放疗的时间数字双胞胎架构，利用患者的个体数据进行实时评估和调整。
📄 English Summary
A Patient-Specific Digital Twin for Adaptive Radiotherapy of Non-Small Cell Lung Cancer
Radiotherapy is becoming increasingly precise and data-rich, with current treatment regimens generating high-frequency imaging and dosimetry streams that are ideally suited for AI-driven temporal modeling to characterize the evolution of normal tissues over time. Each fraction of biologically guided radiotherapy (BGRT) for non-small cell lung cancer (NSCLC) patients records new metabolic, anatomical, and dose information. However, clinical decision-making is predominantly informed by static, population-based normal tissue complication probability (NTCP) models, which overlook the dynamic and unique biological trajectories encoded in sequential data. To address this, COMPASS (Comprehensive Personalized Assessment System) has been developed as a temporal digital twin architecture for safe radiotherapy, utilizing individual patient data for real-time assessment and adjustment.
Powered by Cloudflare Workers + Payload CMS + Claude 3.5
数据源: OpenAI, Google AI, DeepMind, AWS ML Blog, HuggingFace 等