ヒグチ　リヨウタ   HIGUCHI Ryota   樋口　亮太    所属   医学部 医学科（附属八千代医療センター）    職種   講師
論文種別	原著
言語種別	英語
査読の有無	査読なし
表題	Risk prediction for malignant intraductal papillary mucinous neoplasm of the pancreas: logistic regression versus machine learning.
掲載誌名	正式名：Scientific reports 略　 称：Sci Rep ISSNコード：20452322/20452322
掲載区分	国外
巻・号・頁	10(1),pp.20140
著者・共著者	Kang Jae Seung, Lee Chanhee, Song Wookyeong, Choo Wonho, Lee Seungyeoun, Lee Sungyoung, Han Youngmin, Bassi Claudio, Salvia Roberto, Marchegiani Giovanni, Wolfgang Cristopher L, He Jin, Blair Alex B, Kluger Michael D, Su Gloria H, Kim Song Cheol, Song Ki-Byung, Yamamoto Masakazu, Higuchi Ryota, Hatori Takashi, Yang Ching-Yao, Yamaue Hiroki, Hirono Seiko, Satoi Sohei, Fujii Tsutomu, Hirano Satoshi, Lou Wenhui, Hashimoto Yasushi, Shimizu Yasuhiro, Del Chiaro Marco, Valente Roberto, Lohr Matthias, Choi Dong Wook, Choi Seong Ho, Heo Jin Seok, Motoi Fuyuhiko, Matsumoto Ippei, Lee Woo Jung, Kang Chang Moo, Shyr Yi-Ming, Wang Shin-E, Han Ho-Seong, Yoon Yoo-Seok, Besselink Marc G, van Huijgevoort Nadine C M, Sho Masayuki, Nagano Hiroaki, Kim Sang Geol, Honda Goro, Yang Yinmo, Yu Hee Chul, Do Yang Jae, Chung Jun Chul, Nagakawa Yuichi, Seo Hyung Il, Choi Yoo Jin, Byun Yoonhyeong, Kim Hongbeom, Kwon Wooil, Park Taesung, Jang Jin-Young
発行年月	2020/11
概要	Most models for predicting malignant pancreatic intraductal papillary mucinous neoplasms were developed based on logistic regression (LR) analysis. Our study aimed to develop risk prediction models using machine learning (ML) and LR techniques and compare their performances. This was a multinational, multi-institutional, retrospective study. Clinical variables including age, sex, main duct diameter, cyst size, mural nodule, and tumour location were factors considered for model development (MD). After the division into a MD set and a test set (2:1), the best ML and LR models were developed by training with the MD set using a tenfold cross validation. The test area under the receiver operating curves (AUCs) of the two models were calculated using an independent test set. A total of 3,708 patients were included. The stacked ensemble algorithm in the ML model and variable combinations containing all variables in the LR model were the most chosen during 200 repetitions. After 200 repetitions, the mean AUCs of the ML and LR models were comparable (0.725 vs. 0.725). The performances of the ML and LR models were comparable. The LR model was more practical than ML counterpart, because of its convenience in clinical use and simple interpretability.
DOI	10.1038/s41598-020-76974-7
PMID	33208887