Related Experiment Video
Updated: Jun 16, 2025
08:43
An Orthotopic Bladder Tumor Model and the Evaluation of Intravesical saRNA Treatment
Published on: July 28, 2012
14.7K
Non-Invasive Tumor Budding Evaluation and Correlation with Treatment Response in Bladder Cancer: A Multi-Center
Xiaoyang Li1,2, Chen Zou1,2, Chunhui Wang3
1Department of Urology, Third Affiliated Hospital of Sun Yat-sen University, Sun Yat-sen University, 600th Tianhe Road, Guangzhou, 510630, P. R. China.
Summary
Neoadjuvant chemoimmunotherapy (NACI) shows promise for bladder cancer (BCa), but many patients do not respond. This study developed a deep learning model using CT scans to predict tumor budding (TB) status, identifying patients likely to have poor outcomes and lower response rates to NACI.
Area of Science:
- Oncology
- Medical Imaging
- Artificial Intelligence
Background:
- Neoadjuvant chemoimmunotherapy (NACI) offers clinical benefits for bladder cancer (BCa), yet over half of patients do not achieve pathological complete response (pCR).
- Tumor budding (TB) is a prognostic factor in BCa, but its assessment is typically invasive.
- Predicting NACI response and prognosis noninvasively is crucial for optimizing treatment strategies.
Purpose of the Study:
- To investigate the correlation between tumor budding (TB) status and response to neoadjuvant chemoimmunotherapy (NACI) in bladder cancer (BCa).
- To develop and validate a deep learning model for noninvasive prediction of TB status using CT images.
- To assess the prognostic value of the deep learning-based TB status prediction in BCa patients undergoing NACI.
Main Methods:
- Utilized multi-center cohorts of 2322 pathologically diagnosed BCa patients.
- Developed a deep learning model to evaluate TB status from CT images.
- Validated the model across training, internal validation, external validation (2 cohorts), and NACI-specific validation cohorts.
Main Results:
- The deep learning model demonstrated high accuracy in predicting TB status (AUCs ranging from 0.854 to 0.944 across cohorts).
- Patients predicted with high TB status had significantly worse prognosis (p < 0.05).
- High TB status was associated with a substantially lower pCR rate (25.9%) compared to low TB status (73.9%, p < 0.001).
Conclusions:
- A deep learning model can reliably and noninvasively predict tumor budding (TB) status in bladder cancer (BCa) patients.
- This AI tool aids in assessing prognosis and formulating personalized NACI strategies.
- Noninvasive TB status prediction may improve patient selection and treatment outcomes for BCa.
