AUTHOR=Wang Xing , Qiu Jia-Jun , Tan Chun-Lu , Chen Yong-Hua , Tan Qing-Quan , Ren Shu-Jie , Yang Fan , Yao Wen-Qing , Cao Dan , Ke Neng-Wen , Liu Xu-Bao 

TITLE=Development and Validation of a Novel Radiomics-Based Nomogram With Machine Learning to Preoperatively Predict Histologic Grade in Pancreatic Neuroendocrine Tumors

JOURNAL=Frontiers in Oncology

VOLUME=Volume 12 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2022.843376

DOI=10.3389/fonc.2022.843376

ISSN=2234-943X

ABSTRACT=Backgroud: Tumor grade is the determinant of biological aggressiveness of pancreatic neuroendocrine tumors (PNETs) and the best current tool to help establish individualized therapeutic strategies. A non-invasive way to accurately predict the histology grade of PNETs preoperatively is urgently needed and extremely limited. 

Methods: The models training and construction of the radiomic signature were carried out separately in three phases (plain, arterial, and venous) CT. The Mann-Whitney U test and LASSO (Least absolute shrinkage and selection operator) were applied for feature preselection and radiomic signature construction. SVM-linear models were trained by incorporating the radiomics signature with clinical characteristics. Then, an optimal model was chosen to build a nomogram. 

Results: 
A total of 139 PNETs (including 83 in the training set and 56 in the independent validation set) were included at present study. We build a model basing on an eight-feature radiomics signature (Group 1) to stratify PNETs patients into grade 1 and grade 2/3 groups with an AUC of 0.911 (95% confidence intervals (CI), 0.908-0.914) and 0.837 (95% CI, 0.827-0.847) in the training and validation cohorts respectively. The nomogram combining the radiomics signature of plain phase CT with T stage and Dilated main pancreatic duct (MPD) / bile duct (BD) (Group 2) showed the best performance (training set: AUC = 0.919, 95% CI = 0.916-0.922; validation set: AUC = 0.875, 95% CI = 0.867-0.883). 

Conclusions: Our developed nomogram that integrates radiomics signature with clinical characteristics could be useful in predicting grade 1 and grade 2/3 PNETs preoperatively with powerful capability.