Preoperative prediction of pancreatic neuroendocrine tumor grade based on 68Ga-DOTATATE PET/CT

• DOI: 10.1007/s12020-023-03515-3
• Publication Date: 2023-09-16T10:01:42Z
• Authors: Jiao Ma; Xiaoyong Wang; Mingsong Tang; Chunyin Zhang
• Keywords: Nomogram; Grading (engineering); PET-CT

Abstract Objective To establish a prediction model for preoperatively predicting grade 1 and 2/3 tumors in patients with pancreatic neuroendocrine (PNETs) based on 68 Ga-DOTATATE PET/CT. Methods Clinical data of 41 PNETs were included this study. According to the pathological results, they divided into 2/3. PET/CT images collected within one month before surgery. The clinical risk factors significant radiological features filtered, predictive these was established. 3D slicer used extracted 107 radiomic from region interest (ROI) Ga-dotata images. Pearson correlation coefficient (PCC), recursive feature elimination (REF) five-fold cross validation adopted selection, score computed subsequently. comprehensive combining rad-score established as well nomogram. performance above evaluated compared. Results Adjacent organ invasion, N staging, M staging PNET grading ( p < 0.05). 12 optimal (3 PET features, 9 CT features) screen out. achieved an area under curve (AUC) 0.785. has better (AUC = 0.953). Conclusion We proposed nomogram predict assist personalized diagnosis treatment plans PNETs.