Large language models enable tumor-type classification and localization of cancers of unknown primary from genomic data
8.0
来源：
Cell
关键字：
neural coding
发布时间：
2025-09-05 03:31
摘要：
OncoChat is a novel AI model designed for tumor-type classification, particularly effective in identifying cancers of unknown primary (CUP). Developed using genomic data from over 158,000 tumors, it integrates various genomic alterations to enhance diagnostic accuracy. The model achieved an accuracy of 77.4% and an F1 score of 75.6%, significantly outperforming existing classification methods. Its predictions correlate with survival outcomes, indicating its potential utility in clinical decision-making and personalized treatment strategies.
原文：
查看原文
价值分投票
评分标准
新闻价值分采用0-10分制，综合考虑新闻的真实性、重要性、时效性、影响力等多个维度。
评分越高，表示该新闻的价值越大，越值得关注。
价值维度分析
domain_focus
1.0
business_impact
1.0
scientific_rigor
1.5
timeliness_innovation
1.5
investment_perspective
2.5
market_value_relevance
1.0
team_institution_background
0.5
technical_barrier_competition
1.0
关键证据
OncoChat achieved an accuracy of 0.774 and an F1 score of 0.756, outperforming baseline methods.
The model was developed on genomic data from 158,836 tumors, demonstrating its robustness.
OncoChat's predictions were prognostic for survival in cancers of unknown primary.
真实性检查
否
AI评分总结
OncoChat is a novel AI model designed for tumor-type classification, particularly effective in identifying cancers of unknown primary (CUP). Developed using genomic data from over 158,000 tumors, it integrates various genomic alterations to enhance diagnostic accuracy. The model achieved an accuracy of 77.4% and an F1 score of 75.6%, significantly outperforming existing classification methods. Its predictions correlate with survival outcomes, indicating its potential utility in clinical decision-making and personalized treatment strategies.