AUTHOR=Gao Jiahao , Han Fang , Jin Yingying , Wang Xiaoshuang , Zhang Jiawen TITLE=A Radiomics Nomogram for the Preoperative Prediction of Lymph Node Metastasis in Pancreatic Ductal Adenocarcinoma JOURNAL=Frontiers in Oncology VOLUME=Volume 10 - 2020 YEAR=2020 URL=https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2020.01654 DOI=10.3389/fonc.2020.01654 ISSN=2234-943X ABSTRACT=Purpose:To build and verify a CT-based multi-dimensional nomogram for the evaluation of lymph node(LN) status in Pancreatic Ductal Adenocarcinoma(PDAC). Materials and methods:We retrospectively collected 172 patients with clinicopathologically confirmed PDAC who had surgical resection between February 2014 and November 2016. Patients were assigned to either a training cohort (n=121) or a validation cohort (n=51). We obtained the radiomics features from the preoperative venous phase(VP) CT scans. The maximum relevance minimum redundancy (mRMR) algorithm and the least absolute shrinkage and selection operator (LASSO) method were used to choose the optimal features. We used multivariable logistic regression to build a combined radiomics model and visualized it with the nomogram. Performance of the nomogram was evaluated by the receiver operating characteristic (ROC) curve, the calibration test and the clinical usefulness analysis. Results: The Rad-score consisted of 11 LN status-related radiomics features was associated with actual LN status significantly (P<0.01). The nomogram that consisted of Rad-score ,CT-reported parenchymal atrophy and CT-reported LN status performed good predictive effects in the training cohort (area under the curve,0.92) and validation cohort (area under the curve,0.95). The nomogram also performed well in the calibration test and decision curve analysis, demonstrating its potential clinical value. Conclusion:The multi-dimensional radiomics nomogram which consists of Rad-score, CT-reported parenchymal atrophy and CT-reported LN status may contribute to the noninvasive LN evaluation for PDAC patients.