高级检索
  • 摘要: 肺癌是全球癌症致死的主要原因,每年新增病例约210万,死亡人数达180万,主要由吸烟、二手烟及空气污染等因素诱发。非小细胞肺癌与小细胞肺癌为其主要类型,给患者及其家庭带来沉重负担。为降低肺癌患者负担,预防和早期筛查至关重要。早期非小细胞肺癌患者的生存率超过85%,但基层医院的医疗资源匮乏,影响该疾病的早发现和早治疗。为此,白春学教授团队结合生成式预训练转换器(generative pre-trained transformer, GPT)与物联网(Internet of Things, IoT)技术,开发了GPT赋能的肺癌筛查5A程序(GPT-enabled Lung Cancer Screening 5A program, GPT-LSapp 5A)。该程序整合患者的胸部CT、临床信息及实验室数据,通过云平台提供个性化诊治方案,缓解医疗资源不均。这一技术可提升基层医院的肺癌筛查效率,实现分级诊疗。GPT-LSapp 5A能融合多源数据,生成详细的筛查报告,缩短阅片时间,并减轻医生负担;同时,该程序支持病例信息上传、人工智能(artificial intelligence, AI)报告生成及专家远程会诊,促进资源共享。其中,“云”系统负责数据整合、分析,生成智能报告和诊断建议,优化临床路径;“端”程序则在基层医院实现数据采集、初步分析及用户交互,提升前端智能化。尽管GPT-LSapp 5A面临数据隐私、格式整合及网络性能等挑战,但可通过加密技术、标准化接口及边缘计算等措施加以解决。随着AI技术的不断进步,GPT-LSapp 5A有望在基层医院肺癌筛查中发挥关键作用,推动医疗资源下沉,改善患者预后。

     

    Abstract: As the leading cause of cancer death worldwide, lung cancer sees about 2.1 million new cases and 1.8 million deaths every year, mainly caused by smoking, secondhand smoke and air pollution. Non-small cell lung cancer (NSCLC) and small cell lung cancer (SCLC) are the main types, which put a heavy burden on patients’ family. In order to reduce the burden of lung cancer patients, prevention and early screening are crucial, and the survival rate of patients with early-stage NSCLC can exceed 85%, but the lack of primary medical resources affects early detection and early treatment. To this end, professor Bai Chunxue’ team has developed the GPT-enabled Lung Cancer Screening 5A program (GPT-LSapp 5A) by combining generative pre-trained transformer (GPT) and Internet of Things (IoT) technologies. The program integrates chest CT, clinical information and laboratory data to provide personalized diagnosis and treatment plans through the cloud platform to alleviate the imbalance of medical resources. This technology can improve the efficiency of lung cancer screening in primary hospitals and achieve hierarchical diagnosis and treatment. GPT-LSapp 5A can integrate multi-source data to generate detailed screening reports, shorten the CT reading time, reduce the burden on doctors, and help upload case information, generate artificial intelligence (AI) reports, and remote consultation with experts to achieve resource sharing. Its “cloud” system is responsible for data integration and analysis, generating intelligent reports and diagnostic recommendations, and optimizing clinical pathways. The “end” program is responsible for data collection, preliminary analysis and user interaction at primary hospitals to improve the intelligence of the front-end. Although GPT-LSapp 5A faces challenges such as data privacy, format integration, and network performance, it can be solved through measures such as encryption, standardized interfaces, and edge computing. With the advancement of AI technology, GPT-LSapp 5A is expected to play a key role in lung cancer screening at primary hospitals, channeling medical resources down to the community level and improving patient outcomes.

     

/

返回文章
返回