MENU

Click here to

×

Are you sure ?

Yes, do it No, cancel
  1. Home
  2. Meeting Program
  3. All Sessions
  4. All Imaging Scientific Session
  5. A Radiomics Nomogram Model to Preoperatively Predict Histologic Grade in Pancreatic Neuroendocrine Tumors

A Radiomics Nomogram Model to Preoperatively Predict Histologic Grade in Pancreatic Neuroendocrine Tumors

P Yang*, L Xu , R Huang , Y Jiang , C Yang , Y Xue , T Niu , Zhejiang University, Hangzhou, Zhejiang

Presentations

(Tuesday, 7/31/2018) 4:30 PM - 6:00 PM

Room: Room 202

Purpose: Tumor grade is an important factor in the staging system of pancreatic neuroendocrine tumors (pNETs). The accuracy of preoperative fine-needle aspiration biopsy to determine tumor grade is still challenging in clinic. Existing tumor grade classification is assessed on post-operative specimens and does not help on preoperatively individualized therapeutic decisions. We develop and validate a radiomics nomogram model for preoperatively differentiate Grade 1 and Grade 2/3 tumors in patients with pNETs.

Methods: A total of 137 patients who underwent contrast enhanced CT from two hospitals were included in this study. The patients from the second hospital (n=51) were selected as the independent validation set for the model developed. The arterial phase was used for radiomics feature extraction. The Mann-Whitney U test and LASSO regression were used for feature selection and establishing radiomics signature. A radiomics nomogram model was constructed by combing the radiomics signature with clinical factors using multivariable logistic regression method. The receiver operating characteristic (ROC), area under ROC curve (AUC), calibration curve and decision curve analysis (DCA) were employed to assess the diagnostic value of the model developed. The Kaplan-Meier analysis was used for survival analysis.

Results: A total of 467 features including histogram-based, texture-based and wavelet-based features were extracted, from which an eight-feature-combined radiomics signature was finally constructed. The radiomics signature companied by clinical stage showed the best performance among all the clinical factors (AUC=0.907, CI: 0.826-0.959 (training), AUC=0.891, CI：0.772-0.961(validation)). The calibration curve and DCA showed good agreement and demonstrated the clinical usefulness of radiomics nomogram. The Kaplan-Meier analysis showed significant difference between the survival of predicted G1 group and G2/3 group (P = 0.017<0.05, 95%CI) using developed model.

Conclusion: The nomogram incorporating radiomics signature and clinical stage can be used in differentiating G1 and G2/3 tumor in patients with pNETs.

Keywords

Quantitative Imaging

Taxonomy

IM- CT: Quantitative imaging/analysis

Contact Email

Additional Links