首页|Netherlands Cancer Institute Reports Findings in Machine Learning (Reproducing RECIST lesion selection via machine learning: Insights into intra and inter-radiologist variation)
Netherlands Cancer Institute Reports Findings in Machine Learning (Reproducing RECIST lesion selection via machine learning: Insights into intra and inter-radiologist variation)
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
By a News Reporter-Staff News Editor at Robotics & Machine Learning DailyNews Daily News - New research on Machine Learning is th e subject of a report. According to newsreporting from Amsterdam, Netherlands, by NewsRx journalists, research stated, “The Response EvaluationCriteria in Sol id Tumors (RECIST) aims to provide a standardized approach to assess treatment r esponsein solid tumors. However, discrepancies in the selection of measurable a nd target lesions among radiologistsusing these criteria pose a significant lim itation to their reproducibility and accuracy.”The news correspondents obtained a quote from the research from Netherlands Canc er Institute,“This study aimed to understand the factors contributing to this v ariability. Machine learning models wereused to replicate, in parallel, the sel ection process of measurable and target lesions by two radiologistsin a cohort of 40 patients from an internal pan-cancer dataset. The models were trained on l esioncharacteristics such as size, shape, texture, rank, and proximity to other lesions. Ablation experimentswere conducted to evaluate the impact of lesion d iameter, volume, and rank on the selection process.The models successfully repr oduced the selection of measurable lesions, relying primarily on size-relatedfe atures. Similarly, the models reproduced target lesion selection, relying mostly on lesion rank. Beyondthese features, the importance placed by different radio logists on different visual characteristics can vary,specifically when choosing target lesions. Worth noting that substantial variability was still observed between radiologists in both measurable and target lesion selection. Despite the s uccessful replication oflesion selection, our results still revealed significan t inter-radiologist disagreement.”
NETL
NSTL
