Integrating the continuous-time random-walk diffusion model and the vesical imaging-reporting and data system to predict muscle invasion of bladder cancer
10.3760/cma.j.cn112149-20230911-00182
- VernacularTitle:连续时间随机行走扩散模型联合膀胱影像报告与数据系统诊断膀胱癌肌层浸润状态的研究
- Author:
Wei WANG
1
;
Wei LI
;
Junzhe YANG
;
Jingyun WU
;
Jianxing QIU
Author Information
1. 北京大学第一医院医学影像科，北京 100034
- Keywords:
Urinary bladder neoplasms;
Magnetic resonance imaging;
Diffusion model;
Muscle invasion;
Vesical imaging-reporting and data system
- From:
Chinese Journal of Radiology
2024;58(4):394-400
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To investigate the diagnostic performance of continuous-time random-walk (CTRW) diffusion model combined with vesical imaging-reporting and data system (VI-RADS) in the diagnosis of muscle invasion of bladder cancer.Methods:In this case-control study, 64 patients with pathologically confirmed bladder urothelial carcinoma in Peking University First Hospital were retrospectively enrolled from August 2022 to March 2023. The patients were divided into the muscle invasive bladder cancer (MIBC) group and the nonmuscle invasive bladder cancer (NMIBC) group (29 cases and 35 cases, respectively) according to the pathological results. All patients underwent bladder MRI within 4 weeks before surgery, including T 2WI, conventional diffusion weighted imaging (DWI), and multi-b-value DWI. The CTRW model was used to obtain three quantitative diffusion parameters, including D m (an anomalous diffusion coefficient), α (related to temporal diffusion heterogeneity), and β (related to spatial diffusion heterogeneity). The apparent diffusion coefficient (ADC) was calculated using a mono-exponential model. The VI-RADS scores were evaluated based on T 2WI and conventional DWI. The Mann-Whitney U test was used to compare the diffusion parameters between the MIBC group and the NMIBC group. The combination of the parameters was investigated with logistic regression analysis. The diagnostic performance for muscle invasion of bladder cancer was evaluated by receiver operating characteristic analysis and the area under the curve (AUC). The difference between AUC was compared using the DeLong test. Results:There were statistically significant differences in ADC, D m, and α between the MIBC group and the NMIBC group ( Z=-2.31, -2.91, -3.97, P=0.021, 0.004,<0.001). No significant difference was found in β between the two groups ( Z=1.69, P=0.091). The AUC (95% CI) of D m and α for diagnosing MIBC were 0.712 (0.587-0.838) and 0.790 (0.676-0.904) respectively, both of which were higher than that of ADC (AUC 0.669, 95% CI 0.537-0.801) with statistically significant differences ( Z=2.86, 2.27, P=0.004, 0.023). The AUC (95% CI) of CTRW (D m+α) was 0.782 (0.661-0.876), which was significantly higher than that of ADC ( Z=2.35, P=0.019). The AUC (95% CI) of VI-RADS score and VI-RADS combined with CTRW parameter (VI-RADS+D m+α) were 0.823 (0.716-0.930) and 0.900 (0.799-0.961) respectively, with a statistically significant difference between them ( Z=2.16, P=0.031). Conclusion:The D m and α parameters in the CTRW diffusion model show better performance than the ADC in the mono-exponential model for muscle-invasive evaluation of bladder cancer, and the CTRW diffusion model can enhance the diagnostic performance of VI-RADS.
