Coupling Habitat Radiomic Analysis with the Diversification of the Tumor ecosystem: Illuminating New Strategy in the Assessment of Postoperative Recurrence of Non-Muscle Invasive Bladder Cancer

Hong Li ¹, Yiqun Sui ², Yongli Tao ¹

⁰Department of Radiology, The Second Affiliated Hospital of Soochow University, Suzhou, China (H.L., Y.T.).

Academic Radiology +

|

October 25, 2024

View abstract on PubMed

Summary

Novel habitat-based radiomics effectively predicts non-muscle-invasive bladder cancer (NMIBC) recurrence. This approach improves risk stratification and may enhance clinical management of NMIBC patients.

Area Of Science

Radiology and Imaging
Oncology
Artificial Intelligence in Medicine

Background

Non-muscle-invasive bladder cancer (NMIBC) is characterized by high recurrence rates, posing a significant challenge to patient prognosis.
Intratumoral heterogeneity is a key factor influencing NMIBC recurrence.
Accurate risk stratification is essential for effective NMIBC management.

Purpose Of The Study

To investigate a novel habitat-based radiomic analysis for stratifying NMIBC recurrence risk.
To develop and validate a machine learning model for predicting two-year recurrence of NMIBC.
To compare the diagnostic efficacy of the habitat-based model (HBM) with clinical and multiphase radiomic models.

Main Methods

Retrospective analysis of CT images from 382 NMIBC patients.
Identification of intratumoral habitats using K-means clustering on texture features.
Extraction of radiomic features and development of a habitat-based model (HBM) using machine learning.

Main Results

Three distinct intratumoral habitats were identified in NMIBC.
The HBM demonstrated high predictive performance (AUC 0.932 training, 0.782 validation) for two-year recurrence.
The HBM showed superior diagnostic efficacy compared to the clinical model (p < 0.001) and correlated significantly with tumor-stroma ratio.

Conclusions

Habitat-based radiomics combined with machine learning offers an efficient method for predicting NMIBC recurrence.
This approach has the potential to improve the clinical management of NMIBC.
Further research into habitat-based radiomics could lead to enhanced patient outcomes.

Keywords:

Bladder cancer Habitat Intratumoral Heterogeneity Machine Learning Radiomics