Prediction of solid pseudopapillary tumor invasiveness of the pancreas based on multiphase contrast-enhanced CT radiomics nomogram

Dabin Ren ¹, Liqiu Liu ¹, Aiyun Sun ²

⁰Department of Radiology, Taizhou Central Hospital (Taizhou University Hospital), Taizhou, China.

Frontiers in Oncology +

|

April 22, 2025

View abstract on PubMed

Summary

This study developed a radiomics nomogram using multiphase contrast-enhanced CT scans to predict pancreatic solid pseudopapillary neoplasm (PSPN) invasiveness. The nomogram, combining radiomics and traditional CT features, demonstrated superior accuracy in identifying invasive PSPN compared to models using only CT or radiomics alone.

Area Of Science

Radiology
Oncology
Medical Imaging Analysis

Background

Pancreatic solid pseudopapillary neoplasms (PSPN) require accurate assessment of invasiveness for appropriate clinical management.
Traditional CT features have limitations in predicting PSPN invasiveness.
Radiomics offers a potential non-invasive method to extract quantitative features from medical images.

Purpose Of The Study

To develop and validate a multiphase contrast-enhanced CT-based radiomics nomogram for predicting PSPN invasiveness.
To combine traditional CT features with radiomics signatures for improved predictive performance.
To assess the clinical utility of the developed nomogram.

Main Methods

Retrospective analysis of 114 patients with pathologically diagnosed PSPN.
Development of a traditional CT model using univariate and multivariate analyses to identify independent predictors of invasiveness.
Extraction of radiomics features from multiphase contrast-enhanced CT images and development of radiomics models (U, A, V, U+A+V).
Construction of a radiomics nomogram integrating CT predictors and radiomics signatures.
Performance evaluation using ROC curve analysis, Delong's test, and Decision Curve Analysis (DCA).

Main Results

Solid tumors and ill-defined tumor margins were identified as independent predictors of PSPN invasiveness.
The combined radiomics model (U+A+V) showed high diagnostic performance (AUCs: 0.857 training, 0.839 validation).
The radiomics nomogram achieved superior AUCs (0.87 training, 0.867 validation) compared to individual models.
DCA indicated that the radiomics nomogram provides increased net benefit for clinical decision-making.

Conclusions

Multiphase contrast-enhanced CT radiomics can effectively and non-invasively predict the invasiveness of PSPN.
The developed radiomics nomogram, integrating radiomics and traditional CT features, significantly enhances classification ability for PSPN invasiveness.
This nomogram holds potential as a valuable tool for clinical decision support in managing PSPN.

Keywords:

contrast-enhanced computed tomography (CECT) invasiveness nomogram pancreatic solid pseudopapillary neoplasm (PSPN) radiomics