Jie Cao, Nan Chen, Lingyu Zhou, Le Yi, Zhiyu Peng, Lin Qiu, Haokun Wu, Xiyue Tan, Kunhao Wu, Huahang Lin, Zhaokang Huang, Zetao Liu, Chenglin Guo, Xiuyuan Xu, Zhang Yi, Jiandong Mei
{"title":"一项大规模诊断研究表明，具有巩固与肿瘤比（CTR）先验的贝叶斯深度学习模型彻底改变了IA期肺腺癌通过空气间隙扩散（STAS）的预测。","authors":"Jie Cao, Nan Chen, Lingyu Zhou, Le Yi, Zhiyu Peng, Lin Qiu, Haokun Wu, Xiyue Tan, Kunhao Wu, Huahang Lin, Zhaokang Huang, Zetao Liu, Chenglin Guo, Xiuyuan Xu, Zhang Yi, Jiandong Mei","doi":"10.21037/tlcr-24-890","DOIUrl":null,"url":null,"abstract":"<p><strong>Background: </strong>The preoperative prediction of spread through air spaces (STAS) in patients with early-stage lung adenocarcinoma (LUAD) is crucial for selecting the appropriate surgical approach and improving patient outcomes. Previous research has confirmed that there is a significant correlation between consolidation-to-tumor ratio (CTR) and STAS. This study aimed to develop a Bayesian deep learning (DL) model based on the CTR prior to predict STAS in patients with stage IA LUAD.</p><p><strong>Methods: </strong>This large-scale diagnostic study included patients with solitary primary invasive LUAD who underwent complete resection between November 2017 and October 2023. Enrolled patients were randomly assigned to training, validation, and test cohorts in a 7:2:1 ratio. Using a variational Bayesian inference framework, we developed a DL model based on the CTR prior (STAS-DL<sub>Prior CTR</sub>). The performance of STAS-DL<sub>Prior CTR</sub> was compared with another DL model without the CTR prior (STAS-DL<sub>Non-prior CTR</sub>) using the receiver operating characteristic (ROC) curve, calibration curve, decision curve analysis (DCA), and clinical impact curve (CIC).</p><p><strong>Results: </strong>A total of 1,374 patients were included, with 961 in the training cohort, 275 in the validation cohort, and 138 in the test cohort. The results showed that CTR in the STAS-positive group was significantly higher than that in the STAS-negative group [0.63 (interquartile range, 0.36, 0.98) <i>vs</i>. 0.35 (interquartile range, 0.19, 0.60), P<0.001]. Compared to STAS-DL<sub>Non-prior CTR</sub>, the area under the ROC curve (AUC) tends to be higher for STAS-DL<sub>Prior CTR</sub> (0.831 <i>vs</i>. 0.731, P=0.06) in the validation cohort, and STAS-DL<sub>Prior CTR</sub> demonstrated a significantly higher AUC (0.858 <i>vs</i>. 0.637, P=0.008) in the test cohort. Additionally, the calibration curve suggested better calibration for STAS-DL<sub>Prior CTR</sub>. DCA and CIC also indicated that STAS-DL<sub>Prior CTR</sub> conferred higher clinical net benefit.</p><p><strong>Conclusions: </strong>The proposed model based on the CTR prior offers significant advantages in predicting STAS in patients with stage IA LUAD, and incorporating doctors' knowledge as priors can effectively guide the development of DL models.</p>","PeriodicalId":23271,"journal":{"name":"Translational lung cancer research","volume":"14 5","pages":"1516-1530"},"PeriodicalIF":3.5000,"publicationDate":"2025-05-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12170142/pdf/","citationCount":"0","resultStr":"{\"title\":\"A Bayesian deep learning model with consolidation-to-tumor ratio (CTR) prior revolutionizes the prediction of spread through air spaces (STAS) in stage IA lung adenocarcinoma: a large-scale diagnostic study.\",\"authors\":\"Jie Cao, Nan Chen, Lingyu Zhou, Le Yi, Zhiyu Peng, Lin Qiu, Haokun Wu, Xiyue Tan, Kunhao Wu, Huahang Lin, Zhaokang Huang, Zetao Liu, Chenglin Guo, Xiuyuan Xu, Zhang Yi, Jiandong Mei\",\"doi\":\"10.21037/tlcr-24-890\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><strong>Background: </strong>The preoperative prediction of spread through air spaces (STAS) in patients with early-stage lung adenocarcinoma (LUAD) is crucial for selecting the appropriate surgical approach and improving patient outcomes. Previous research has confirmed that there is a significant correlation between consolidation-to-tumor ratio (CTR) and STAS. This study aimed to develop a Bayesian deep learning (DL) model based on the CTR prior to predict STAS in patients with stage IA LUAD.</p><p><strong>Methods: </strong>This large-scale diagnostic study included patients with solitary primary invasive LUAD who underwent complete resection between November 2017 and October 2023. Enrolled patients were randomly assigned to training, validation, and test cohorts in a 7:2:1 ratio. Using a variational Bayesian inference framework, we developed a DL model based on the CTR prior (STAS-DL<sub>Prior CTR</sub>). The performance of STAS-DL<sub>Prior CTR</sub> was compared with another DL model without the CTR prior (STAS-DL<sub>Non-prior CTR</sub>) using the receiver operating characteristic (ROC) curve, calibration curve, decision curve analysis (DCA), and clinical impact curve (CIC).</p><p><strong>Results: </strong>A total of 1,374 patients were included, with 961 in the training cohort, 275 in the validation cohort, and 138 in the test cohort. The results showed that CTR in the STAS-positive group was significantly higher than that in the STAS-negative group [0.63 (interquartile range, 0.36, 0.98) <i>vs</i>. 0.35 (interquartile range, 0.19, 0.60), P<0.001]. Compared to STAS-DL<sub>Non-prior CTR</sub>, the area under the ROC curve (AUC) tends to be higher for STAS-DL<sub>Prior CTR</sub> (0.831 <i>vs</i>. 0.731, P=0.06) in the validation cohort, and STAS-DL<sub>Prior CTR</sub> demonstrated a significantly higher AUC (0.858 <i>vs</i>. 0.637, P=0.008) in the test cohort. Additionally, the calibration curve suggested better calibration for STAS-DL<sub>Prior CTR</sub>. DCA and CIC also indicated that STAS-DL<sub>Prior CTR</sub> conferred higher clinical net benefit.</p><p><strong>Conclusions: </strong>The proposed model based on the CTR prior offers significant advantages in predicting STAS in patients with stage IA LUAD, and incorporating doctors' knowledge as priors can effectively guide the development of DL models.</p>\",\"PeriodicalId\":23271,\"journal\":{\"name\":\"Translational lung cancer research\",\"volume\":\"14 5\",\"pages\":\"1516-1530\"},\"PeriodicalIF\":3.5000,\"publicationDate\":\"2025-05-30\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12170142/pdf/\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Translational lung cancer research\",\"FirstCategoryId\":\"3\",\"ListUrlMain\":\"https://doi.org/10.21037/tlcr-24-890\",\"RegionNum\":2,\"RegionCategory\":\"医学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"2025/5/27 0:00:00\",\"PubModel\":\"Epub\",\"JCR\":\"Q2\",\"JCRName\":\"ONCOLOGY\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Translational lung cancer research","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.21037/tlcr-24-890","RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2025/5/27 0:00:00","PubModel":"Epub","JCR":"Q2","JCRName":"ONCOLOGY","Score":null,"Total":0}
A Bayesian deep learning model with consolidation-to-tumor ratio (CTR) prior revolutionizes the prediction of spread through air spaces (STAS) in stage IA lung adenocarcinoma: a large-scale diagnostic study.
Background: The preoperative prediction of spread through air spaces (STAS) in patients with early-stage lung adenocarcinoma (LUAD) is crucial for selecting the appropriate surgical approach and improving patient outcomes. Previous research has confirmed that there is a significant correlation between consolidation-to-tumor ratio (CTR) and STAS. This study aimed to develop a Bayesian deep learning (DL) model based on the CTR prior to predict STAS in patients with stage IA LUAD.
Methods: This large-scale diagnostic study included patients with solitary primary invasive LUAD who underwent complete resection between November 2017 and October 2023. Enrolled patients were randomly assigned to training, validation, and test cohorts in a 7:2:1 ratio. Using a variational Bayesian inference framework, we developed a DL model based on the CTR prior (STAS-DLPrior CTR). The performance of STAS-DLPrior CTR was compared with another DL model without the CTR prior (STAS-DLNon-prior CTR) using the receiver operating characteristic (ROC) curve, calibration curve, decision curve analysis (DCA), and clinical impact curve (CIC).
Results: A total of 1,374 patients were included, with 961 in the training cohort, 275 in the validation cohort, and 138 in the test cohort. The results showed that CTR in the STAS-positive group was significantly higher than that in the STAS-negative group [0.63 (interquartile range, 0.36, 0.98) vs. 0.35 (interquartile range, 0.19, 0.60), P<0.001]. Compared to STAS-DLNon-prior CTR, the area under the ROC curve (AUC) tends to be higher for STAS-DLPrior CTR (0.831 vs. 0.731, P=0.06) in the validation cohort, and STAS-DLPrior CTR demonstrated a significantly higher AUC (0.858 vs. 0.637, P=0.008) in the test cohort. Additionally, the calibration curve suggested better calibration for STAS-DLPrior CTR. DCA and CIC also indicated that STAS-DLPrior CTR conferred higher clinical net benefit.
Conclusions: The proposed model based on the CTR prior offers significant advantages in predicting STAS in patients with stage IA LUAD, and incorporating doctors' knowledge as priors can effectively guide the development of DL models.
期刊介绍：
Translational Lung Cancer Research(TLCR, Transl Lung Cancer Res, Print ISSN 2218-6751; Online ISSN 2226-4477) is an international, peer-reviewed, open-access journal, which was founded in March 2012. TLCR is indexed by PubMed/PubMed Central and the Chemical Abstracts Service (CAS) Databases. It is published quarterly the first year, and published bimonthly since February 2013. It provides practical up-to-date information on prevention, early detection, diagnosis, and treatment of lung cancer. Specific areas of its interest include, but not limited to, multimodality therapy, markers, imaging, tumor biology, pathology, chemoprevention, and technical advances related to lung cancer.
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号
