Research Presentation Session: Genitourinary
RPS 407 - Bladder cancer and ureteral cancer imaging
March 4, 13:00 - 14:30 CET
6 min
An artificial intelligence framework based on contrast-enhanced CT for predicting postoperative recurrence risk in non-muscle-invasive bladder cancer: a multicenter study
Jialu Wu, Inner Mongolia Autonomous Region / China
Author Block: J. Wu, J. Li, J. Han, R. Liu; Inner Mongolia Autonomous Region/CN
Purpose: To determine whether the deep learning model based on automatic segmentation of CT images can provide a robust prediction of postoperative recurrence for non-muscle-invasive bladder cancer (NMIBC).
Methods or Background: Preoperative CT scans were retrospectively acquired in patients with surgically proven NMIBC at multiple centers from January 2015 to May 2023. Follow-up was extended until May 2025, and cases were categorized as recurrent or nonrecurrent based on
postoperative outcomes. The proposed framework comprised four modules, including a 3D tumor segmentation model by 3D-UNet, a deep learning feature extraction module by Swin Transformer, a clinical-radiological feature screening module, and a fully-connected classification module that combines features from different sources. The predictive performance of the model was evaluated based on discrimination, calibration, and clinical usefulness.
Results or Findings: The training data set was comprised of 310 patients from three centers, and 115 and 72 patients were included in the two external test data sets. The average Dice coefficient of the 3D-UNet automatic segmentation network in the test sets was 0.88 and 0.83. Subsequently, the model's ability to predict the risk of postoperative recurrence was evaluated. The integrated model exhibited excellent performance with AUCs of 0.92 in the training set, 0.85 and 0.88 in the external validation set, along with satisfactory calibration across all cohorts (P>0.05). Furthermore, the integrated model significantly outperformed both the clinical-radiological model and deep learning model (P<0.05).
Conclusion: The integrated model based on clinical features and deep learning features focusing on NMIBC patients undergoing the transurethral resection of bladder tumor(TURBT), accurately identifies cases at high-risk of postoperative recurrence, which outperformed the clinical-radiological feature model.
Limitations: This is a retrospective study and we only included NMIBC with a pathological diagnosis after TURBT, with some selection bias.
Funding for this study: This work was supported in part by the Key Scientific Research and Development Achievement Transformation Programme of the Department of Science and Technology of the Inner Mongolia Autonomous Region (2025YFSH0050); and partially supported by the Public Hospital Research Joint Fund Programme of Inner Mongolia Medical University (2023GLLH0086, 2024GLLH0013).
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: This study was approved by the Institutional Review Board, and written informed consent was obtained.
Purpose: To determine whether the deep learning model based on automatic segmentation of CT images can provide a robust prediction of postoperative recurrence for non-muscle-invasive bladder cancer (NMIBC).
Methods or Background: Preoperative CT scans were retrospectively acquired in patients with surgically proven NMIBC at multiple centers from January 2015 to May 2023. Follow-up was extended until May 2025, and cases were categorized as recurrent or nonrecurrent based on
postoperative outcomes. The proposed framework comprised four modules, including a 3D tumor segmentation model by 3D-UNet, a deep learning feature extraction module by Swin Transformer, a clinical-radiological feature screening module, and a fully-connected classification module that combines features from different sources. The predictive performance of the model was evaluated based on discrimination, calibration, and clinical usefulness.
Results or Findings: The training data set was comprised of 310 patients from three centers, and 115 and 72 patients were included in the two external test data sets. The average Dice coefficient of the 3D-UNet automatic segmentation network in the test sets was 0.88 and 0.83. Subsequently, the model's ability to predict the risk of postoperative recurrence was evaluated. The integrated model exhibited excellent performance with AUCs of 0.92 in the training set, 0.85 and 0.88 in the external validation set, along with satisfactory calibration across all cohorts (P>0.05). Furthermore, the integrated model significantly outperformed both the clinical-radiological model and deep learning model (P<0.05).
Conclusion: The integrated model based on clinical features and deep learning features focusing on NMIBC patients undergoing the transurethral resection of bladder tumor(TURBT), accurately identifies cases at high-risk of postoperative recurrence, which outperformed the clinical-radiological feature model.
Limitations: This is a retrospective study and we only included NMIBC with a pathological diagnosis after TURBT, with some selection bias.
Funding for this study: This work was supported in part by the Key Scientific Research and Development Achievement Transformation Programme of the Department of Science and Technology of the Inner Mongolia Autonomous Region (2025YFSH0050); and partially supported by the Public Hospital Research Joint Fund Programme of Inner Mongolia Medical University (2023GLLH0086, 2024GLLH0013).
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: This study was approved by the Institutional Review Board, and written informed consent was obtained.
6 min
Utility of nacVI-RADS for Predicting Pathological Complete Response After Chemoradiotherapy in Muscle-Invasive Bladder Cancer
Kenichi Fukui, Adachi-ku Tokyo-to / Japan
Author Block: K. Fukui1, M. Kobayashi1, S. Yoshida1, Y. Arita2, K. Yamagiwa1, J. Tsuchiya1, Y. Fujii1, U. Tateishi1; 1Tokyo/JP, 2San Francisco, CA/US
Purpose: Accurate post-treatment imaging evaluation remains a key challenge in muscle-invasive bladder cancer (MIBC), particularly for assessing the effect of chemoradiotherapy (CRT) in the context of bladder-sparing strategies. To pathologically validate CRT response, analysis of a patient cohort who underwent partial or radical cystectomy after CRT is required. This study evaluated the effectiveness of nacVI-RADS on multiparametric MRI for predicting pathological complete response (pCR) after CRT and explored its potential role in standardized post-treatment assessment to guide bladder-preserving management.
Methods or Background: In this single-center retrospective study, 73 patients with non-metastatic MIBC who underwent CRT (40 Gy) followed by cystectomy and had multiparametric MRI both pre-TURBT and post-CRT were included. Two board-certified radiologists with genitourinary subspecialty experience (9 and >10 years), blinded to pathology, independently assigned nacVI-RADS on post-CRT MRI. Diagnostic performance for predicting pCR was assessed using a prespecified threshold of nacVI-RADS <3; inter-reader agreement was quantified with Cohen’s κ.
Results or Findings: Median age was 70 years; 30% were female. Clinical stages were T2 55%, T3 43%, and T4 3%. Radical and partial cystectomy were performed in 37% and 63%, respectively. pCR occurred in 67% (49/73). Patients with nacVI-RADS <3 had significantly higher pCR rates than those with ≥3 (reader 1: 81% vs 42%, p=0.002; reader 2: 77% vs 46%, p=0.021). For predicting pCR, sensitivity/specificity of nacVI-RADS <3 were 78%/63% (reader 1) and 80%/50% (reader 2). Inter-reader agreement was almost perfect (κ=0.84).
Conclusion: To our knowledge, this is the first study to validate nacVI-RADS as a reproducible imaging biomarker for pathological response assessment after CRT. These results support its potential role in standardized post-treatment MRI evaluation and in guiding bladder-preserving strategies in MIBC.
Limitations: Retrospective single-center design with a modest cohort and a fixed CRT dose regimen.
Funding for this study: No funding was received for this study.
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:
Purpose: Accurate post-treatment imaging evaluation remains a key challenge in muscle-invasive bladder cancer (MIBC), particularly for assessing the effect of chemoradiotherapy (CRT) in the context of bladder-sparing strategies. To pathologically validate CRT response, analysis of a patient cohort who underwent partial or radical cystectomy after CRT is required. This study evaluated the effectiveness of nacVI-RADS on multiparametric MRI for predicting pathological complete response (pCR) after CRT and explored its potential role in standardized post-treatment assessment to guide bladder-preserving management.
Methods or Background: In this single-center retrospective study, 73 patients with non-metastatic MIBC who underwent CRT (40 Gy) followed by cystectomy and had multiparametric MRI both pre-TURBT and post-CRT were included. Two board-certified radiologists with genitourinary subspecialty experience (9 and >10 years), blinded to pathology, independently assigned nacVI-RADS on post-CRT MRI. Diagnostic performance for predicting pCR was assessed using a prespecified threshold of nacVI-RADS <3; inter-reader agreement was quantified with Cohen’s κ.
Results or Findings: Median age was 70 years; 30% were female. Clinical stages were T2 55%, T3 43%, and T4 3%. Radical and partial cystectomy were performed in 37% and 63%, respectively. pCR occurred in 67% (49/73). Patients with nacVI-RADS <3 had significantly higher pCR rates than those with ≥3 (reader 1: 81% vs 42%, p=0.002; reader 2: 77% vs 46%, p=0.021). For predicting pCR, sensitivity/specificity of nacVI-RADS <3 were 78%/63% (reader 1) and 80%/50% (reader 2). Inter-reader agreement was almost perfect (κ=0.84).
Conclusion: To our knowledge, this is the first study to validate nacVI-RADS as a reproducible imaging biomarker for pathological response assessment after CRT. These results support its potential role in standardized post-treatment MRI evaluation and in guiding bladder-preserving strategies in MIBC.
Limitations: Retrospective single-center design with a modest cohort and a fixed CRT dose regimen.
Funding for this study: No funding was received for this study.
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:
6 min
Optimizing Multi-b-Value DWI Quantitative Parameters for Assessing Detrusor Muscle Invasion and Histological Grade in Bladder Tumors
Yunyun Shang, Kunming / China
Author Block: Y. Wu, X. Zhu, Y. Shang, C. Liu, Y. Tong, L. Ai, Z. Xiaolong; Kunming/CN
Purpose: Given bladder tumors' high incidence, malignancy, and invasiveness, traditional VI-RADS scoring, based on morphology, inadequately captures tumor microstructure. This study innovatively evaluates MRI multi-b-value DWI parameters' predictive power for muscular invasion and histological grading in bladder urothelial carcinoma, exploring their synergy with VI-RADS.
Methods or Background: Prospectively, 160 patients with suspected bladder masses underwent conventional MRI and multi-b-value DWI (11 b-values, 0-2000 s/mm²). Among them, 106 pathologically confirmed cases (age 68±11 years, 84.9% male) were analyzed. Using iCareSpinx software, entire lesion volumes were delineated to compute 15 diffusion parameters from six non-Gaussian models (CTRW, DKI, FROC, IVIM, SEM, SE). Statistical tests compared parameter differences between groups with varying muscular invasion and pathological grades. Binary logistic regression constructed combined diagnostic models, and ROC curves with AUC values evaluated diagnostic performance.
Results or Findings: Significant differences were observed in CTRW-ADC, DKI-ADC, IVIM-D slow, SEM-ADC, SM-ADC, SM-Theta, and monoADC for discriminating muscular invasion and pathological grading (all P < 0.05). Individually, these parameters showed moderate AUCs (0.704-0.779 for invasion, 0.629-0.712 for grading). VI-RADS alone had AUCs of 0.774 (95% CI: 0.681-0.867) for invasion and 0.720 (95% CI: 0.621-0.818) for grading. Combining parameters with VI-RADS improved AUCs (e.g., 0.870 for invasion with four parameters, outperforming traditional monoADC-VI-RADS combination's 0.845; 0.849 for grading, outperforming 0.765).
Conclusion: MRI multi-b-value DWI quantitative parameters offer clinical value in assessing bladder urothelial tumor invasiveness and grading. While their individual efficacy is limited, their combination with VI-RADS significantly enhances diagnostic performance, providing a more reliable tool.
Limitations: Firstly, the single-center study has a relatively small sample size. Secondly, not delineating smaller or unclear-boundary lesions may omit some. However, volume-based delineation minimizes this bias. Additionally, post-processing times for some models are lengthy, necessitating further optimization to enhance clinical practicality.
Funding for this study: Yunnan Provincial Education Department Project: Research on Predicting Pathological Grading of Bladder Cancer Based on MRI Radiomics Model (2024J03480)
General Project of Yunnan Provincial Science and Technology Department: Comprehensive Study on Assessing Bladder Cancer Aggressiveness, Prognosis, and Therapeutic Efficacy Using MRI-Based Deep Learning Models (202401AY070001-337)
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: None
Purpose: Given bladder tumors' high incidence, malignancy, and invasiveness, traditional VI-RADS scoring, based on morphology, inadequately captures tumor microstructure. This study innovatively evaluates MRI multi-b-value DWI parameters' predictive power for muscular invasion and histological grading in bladder urothelial carcinoma, exploring their synergy with VI-RADS.
Methods or Background: Prospectively, 160 patients with suspected bladder masses underwent conventional MRI and multi-b-value DWI (11 b-values, 0-2000 s/mm²). Among them, 106 pathologically confirmed cases (age 68±11 years, 84.9% male) were analyzed. Using iCareSpinx software, entire lesion volumes were delineated to compute 15 diffusion parameters from six non-Gaussian models (CTRW, DKI, FROC, IVIM, SEM, SE). Statistical tests compared parameter differences between groups with varying muscular invasion and pathological grades. Binary logistic regression constructed combined diagnostic models, and ROC curves with AUC values evaluated diagnostic performance.
Results or Findings: Significant differences were observed in CTRW-ADC, DKI-ADC, IVIM-D slow, SEM-ADC, SM-ADC, SM-Theta, and monoADC for discriminating muscular invasion and pathological grading (all P < 0.05). Individually, these parameters showed moderate AUCs (0.704-0.779 for invasion, 0.629-0.712 for grading). VI-RADS alone had AUCs of 0.774 (95% CI: 0.681-0.867) for invasion and 0.720 (95% CI: 0.621-0.818) for grading. Combining parameters with VI-RADS improved AUCs (e.g., 0.870 for invasion with four parameters, outperforming traditional monoADC-VI-RADS combination's 0.845; 0.849 for grading, outperforming 0.765).
Conclusion: MRI multi-b-value DWI quantitative parameters offer clinical value in assessing bladder urothelial tumor invasiveness and grading. While their individual efficacy is limited, their combination with VI-RADS significantly enhances diagnostic performance, providing a more reliable tool.
Limitations: Firstly, the single-center study has a relatively small sample size. Secondly, not delineating smaller or unclear-boundary lesions may omit some. However, volume-based delineation minimizes this bias. Additionally, post-processing times for some models are lengthy, necessitating further optimization to enhance clinical practicality.
Funding for this study: Yunnan Provincial Education Department Project: Research on Predicting Pathological Grading of Bladder Cancer Based on MRI Radiomics Model (2024J03480)
General Project of Yunnan Provincial Science and Technology Department: Comprehensive Study on Assessing Bladder Cancer Aggressiveness, Prognosis, and Therapeutic Efficacy Using MRI-Based Deep Learning Models (202401AY070001-337)
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: None
6 min
VI-RADS quality score: reproducibility validation study
Luca De Giacomo, Rome / Italy
Author Block: L. De Giacomo, A. Dehghanpour, A. Borrelli, G. Bincoletto, M. Pecoraro, V. Panebianco; Rome/IT
Purpose: To retrospectively validate the reproducibility of the recently proposed Vesical Imaging Quality Score (VI-RADS QS) for multiparametric magnetic resonance imaging (mpMRI) of the bladder, assessing inter-reader agreement across radiologists with different levels of expertise.
Methods or Background: One hundred consecutive bladder mpMRI examinations acquired at 1.5T and 3T scanners were retrospectively retrieved and anonymized. Three radiologists with varying experience in genitourinary imaging (expert, intermediate, beginner) independently scored each examination according to the VI-RADS QS, which stratifies image quality as inadequate, adequate, or optimal per sequence (T2-weighted, diffusion-weighted, and dynamic contrast-enhanced) and overall. Inter- and intra-reader agreement were measured using Gwet’s AC1 statistics.
Results or Findings: Preliminary analyses are expected to show moderate to substantial inter-reader agreement for the overall VI-RADS QS (anticipated Gwet’s AC1 ≈ 0.50–0.70), with higher reproducibility among expert readers compared to beginners. Per-sequence agreement is projected to be higher for DWI and DCE compared to T2-weighted imaging, reflecting the key role of these sequences in bladder cancer staging.
Conclusion: This retrospective study on 150 bladder mpMRI examinations will provide the first validation of the VI-RADS Quality Score, demonstrating its reproducibility across readers of varying expertise. Adoption of the VI-RADS QS may facilitate protocol optimization, harmonize reporting across institutions, and strengthen the reliability of bladder MRI in multidisciplinary decision-making.
Limitations: Small sample size; single center retrospective study
Funding for this study: None
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:
Purpose: To retrospectively validate the reproducibility of the recently proposed Vesical Imaging Quality Score (VI-RADS QS) for multiparametric magnetic resonance imaging (mpMRI) of the bladder, assessing inter-reader agreement across radiologists with different levels of expertise.
Methods or Background: One hundred consecutive bladder mpMRI examinations acquired at 1.5T and 3T scanners were retrospectively retrieved and anonymized. Three radiologists with varying experience in genitourinary imaging (expert, intermediate, beginner) independently scored each examination according to the VI-RADS QS, which stratifies image quality as inadequate, adequate, or optimal per sequence (T2-weighted, diffusion-weighted, and dynamic contrast-enhanced) and overall. Inter- and intra-reader agreement were measured using Gwet’s AC1 statistics.
Results or Findings: Preliminary analyses are expected to show moderate to substantial inter-reader agreement for the overall VI-RADS QS (anticipated Gwet’s AC1 ≈ 0.50–0.70), with higher reproducibility among expert readers compared to beginners. Per-sequence agreement is projected to be higher for DWI and DCE compared to T2-weighted imaging, reflecting the key role of these sequences in bladder cancer staging.
Conclusion: This retrospective study on 150 bladder mpMRI examinations will provide the first validation of the VI-RADS Quality Score, demonstrating its reproducibility across readers of varying expertise. Adoption of the VI-RADS QS may facilitate protocol optimization, harmonize reporting across institutions, and strengthen the reliability of bladder MRI in multidisciplinary decision-making.
Limitations: Small sample size; single center retrospective study
Funding for this study: None
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:
6 min
Added value of synthetic MRI in refining VI-RADS for predicting muscle-invasive bladder cancer
Zihui Zhao, Nanjing / China
Author Block: Z. Zhao, X-G. Peng; Nanjing/CN
Purpose: To evaluate the diagnostic value of synthetic MRI (SyMRI)-derived T1 and T2 values in predicting muscle-invasive bladder cancer (MIBC) and assess the feasibility of integrating them into biparametric MRI VI-RADS (bpVI-RADS).
Methods or Background: This prospective, dual-center study included 36 patients with histologically confirmed urothelial carcinoma who underwent standardized bladder MRI. Two radiologists independently assigned bpVI-RADS scores based on T2WI and DWI, blinded to pathological. Discrepancies were resolved by another senior radiologist. Tumor volumes were manually delineated on SyMRI-derived maps to extract T1 and T2 values. Two modified models (bpVI-RADS+T1 and bpVI-RADS+T2) were developed by adjusting bpVI-RADS scores based on predefined T1 or T2 thresholds. Diagnostic performance was compared among original and modified scoring systems using ROC analysis.
Results or Findings: Thirty-six patients (mean age: 71.25 ± 10.27 years, 22.2% MIBC) were enrolled in this study. MIBC lesions showed significantly lower T1 and T2 values than non–muscle-invasive bladder cancer (NMIBC). Optimal cutoff values for differentiating MIBC from NMIBC were 1370 msec for T1 (sensitivity: 87.5%, specificity: 89.3%) and 113 msec for T2 (sensitivity: 87.5%, specificity: 100%). Among all scoring methods, bpVI-RADS+T2 achieved the highest diagnostic performance at the ≥3 cutoff, with higher specificity (89.3% vs. 67.9%, p=0.031), accuracy (91.7% vs. 75.0%, p=0.031), and AUC (0.993 vs. 0.900, p=0.048) compared to bpVI-RADS. The bpVI-RADS+T1 showed non-significant AUC improvement (0.973 vs. 0.900, P = 0.144).
Conclusion: Incorporating SyMRI-derived T2 values into the bpVI-RADS framework significantly enhances diagnostic performance for predicting MIBC.
Limitations: First, the modest sample size and limited MIBC cases may reduce statistical power and generalizability. Second, inter-center variations in MRI equipment and software could have introduced bias despite high reader consistency. Finally, the modified thresholds were derived from this cohort without external validation, warranting cautious clinical application.
Funding for this study: This study has received funding by National Natural Science Foundation of China (82272064), Jiangsu Provincial Science and Technique Program (BK20221461), Zhongda Hospital Affiliated to Southeast University, Jiangsu Province High-Level Hospital Pairing Assistance Construction (zdlyg08), Postgraduate Research & Practice Innovation Program of Jiangsu Province (KYCX23_0323 and KYCX22_0297).
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Approval No. 2024ZDSYLL061-P01
Purpose: To evaluate the diagnostic value of synthetic MRI (SyMRI)-derived T1 and T2 values in predicting muscle-invasive bladder cancer (MIBC) and assess the feasibility of integrating them into biparametric MRI VI-RADS (bpVI-RADS).
Methods or Background: This prospective, dual-center study included 36 patients with histologically confirmed urothelial carcinoma who underwent standardized bladder MRI. Two radiologists independently assigned bpVI-RADS scores based on T2WI and DWI, blinded to pathological. Discrepancies were resolved by another senior radiologist. Tumor volumes were manually delineated on SyMRI-derived maps to extract T1 and T2 values. Two modified models (bpVI-RADS+T1 and bpVI-RADS+T2) were developed by adjusting bpVI-RADS scores based on predefined T1 or T2 thresholds. Diagnostic performance was compared among original and modified scoring systems using ROC analysis.
Results or Findings: Thirty-six patients (mean age: 71.25 ± 10.27 years, 22.2% MIBC) were enrolled in this study. MIBC lesions showed significantly lower T1 and T2 values than non–muscle-invasive bladder cancer (NMIBC). Optimal cutoff values for differentiating MIBC from NMIBC were 1370 msec for T1 (sensitivity: 87.5%, specificity: 89.3%) and 113 msec for T2 (sensitivity: 87.5%, specificity: 100%). Among all scoring methods, bpVI-RADS+T2 achieved the highest diagnostic performance at the ≥3 cutoff, with higher specificity (89.3% vs. 67.9%, p=0.031), accuracy (91.7% vs. 75.0%, p=0.031), and AUC (0.993 vs. 0.900, p=0.048) compared to bpVI-RADS. The bpVI-RADS+T1 showed non-significant AUC improvement (0.973 vs. 0.900, P = 0.144).
Conclusion: Incorporating SyMRI-derived T2 values into the bpVI-RADS framework significantly enhances diagnostic performance for predicting MIBC.
Limitations: First, the modest sample size and limited MIBC cases may reduce statistical power and generalizability. Second, inter-center variations in MRI equipment and software could have introduced bias despite high reader consistency. Finally, the modified thresholds were derived from this cohort without external validation, warranting cautious clinical application.
Funding for this study: This study has received funding by National Natural Science Foundation of China (82272064), Jiangsu Provincial Science and Technique Program (BK20221461), Zhongda Hospital Affiliated to Southeast University, Jiangsu Province High-Level Hospital Pairing Assistance Construction (zdlyg08), Postgraduate Research & Practice Innovation Program of Jiangsu Province (KYCX23_0323 and KYCX22_0297).
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Approval No. 2024ZDSYLL061-P01
6 min
Optimizing VI-RADS: Tumor Contact Length for Precise Bladder Cancer Staging
Divyansh Gupta, Jaipur / India
Author Block: D. Gupta, A. Bhandari, N. Gupta; Jaipur/IN
Purpose: To investigate VI-RADS diagnostic performance on multiparametric MRI for predicting detrusor muscle invasion in bladder cancer, with histopathology as gold standard, and evaluate tumor contact length (TCL) as a quantitative biomarker for equivocal cases.
Methods or Background: This prospective cross-sectional study enrolled 70 patients with suspected bladder masses undergoing 3.0T Mp-MRI prior to transurethral resection (June 2023-July 2024). Inclusion criteria: gross hematuria and untreated bladder cancer. Exclusion criteria: prior treatment, MRI contraindications, or elevated creatinine. Imaging protocol included T2-weighted imaging, diffusion-weighted imaging with ADC mapping, and dynamic contrast-enhanced sequences. VI-RADS scoring used standardized algorithms with DWI as dominant sequence. TCL was measured as maximum curvilinear tumor-bladder wall interface on T2-weighted images. Statistical analysis included ROC analysis, Youden index optimization, and multivariate logistic regression.
Results or Findings: The cohort comprised 44 patients (62.9%) with non-muscle-invasive bladder cancer (NMIBC) and 26 patients (37.1%) with muscle-invasive bladder cancer (MIBC). VI-RADS (≥3 threshold) achieved 96.2% sensitivity, 84.7% specificity, 89.7% accuracy, and 95.5% negative predictive value. Diffusion-weighted imaging showed superior individual sequence performance (92.5% sensitivity, 88.1% specificity). TCL (≥3.2cm threshold) demonstrated 96.2% sensitivity, 94.6% specificity, and 98.6% accuracy. Within challenging VI-RADS 2-3 intermediate cases (n=36), TCL integration achieved 89% sensitivity, 95% specificity, and 94% accuracy. Multivariate analysis confirmed VI-RADS ≥3 (OR 4.2, p<0.05) and TCL (OR 8.7, p<0.01) as independent predictors.
Conclusion: VI-RADS provides highly accurate bladder cancer staging. TCL integration transforms equivocal intermediate-risk lesions into definitive clinical decisions, fundamentally improving patient management using standard imaging protocols.
Limitations: single-center design, moderate sample size, predominance of urothelial carcinoma, and dependence on TURBT reference standard with inherent sampling variability.
Funding for this study: The authors received no specific funding for this work
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Ethics approval obtained from SMS Medical college Ethics Committee. Written informed consent obtained from all participants prior to enrollment."
Purpose: To investigate VI-RADS diagnostic performance on multiparametric MRI for predicting detrusor muscle invasion in bladder cancer, with histopathology as gold standard, and evaluate tumor contact length (TCL) as a quantitative biomarker for equivocal cases.
Methods or Background: This prospective cross-sectional study enrolled 70 patients with suspected bladder masses undergoing 3.0T Mp-MRI prior to transurethral resection (June 2023-July 2024). Inclusion criteria: gross hematuria and untreated bladder cancer. Exclusion criteria: prior treatment, MRI contraindications, or elevated creatinine. Imaging protocol included T2-weighted imaging, diffusion-weighted imaging with ADC mapping, and dynamic contrast-enhanced sequences. VI-RADS scoring used standardized algorithms with DWI as dominant sequence. TCL was measured as maximum curvilinear tumor-bladder wall interface on T2-weighted images. Statistical analysis included ROC analysis, Youden index optimization, and multivariate logistic regression.
Results or Findings: The cohort comprised 44 patients (62.9%) with non-muscle-invasive bladder cancer (NMIBC) and 26 patients (37.1%) with muscle-invasive bladder cancer (MIBC). VI-RADS (≥3 threshold) achieved 96.2% sensitivity, 84.7% specificity, 89.7% accuracy, and 95.5% negative predictive value. Diffusion-weighted imaging showed superior individual sequence performance (92.5% sensitivity, 88.1% specificity). TCL (≥3.2cm threshold) demonstrated 96.2% sensitivity, 94.6% specificity, and 98.6% accuracy. Within challenging VI-RADS 2-3 intermediate cases (n=36), TCL integration achieved 89% sensitivity, 95% specificity, and 94% accuracy. Multivariate analysis confirmed VI-RADS ≥3 (OR 4.2, p<0.05) and TCL (OR 8.7, p<0.01) as independent predictors.
Conclusion: VI-RADS provides highly accurate bladder cancer staging. TCL integration transforms equivocal intermediate-risk lesions into definitive clinical decisions, fundamentally improving patient management using standard imaging protocols.
Limitations: single-center design, moderate sample size, predominance of urothelial carcinoma, and dependence on TURBT reference standard with inherent sampling variability.
Funding for this study: The authors received no specific funding for this work
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Ethics approval obtained from SMS Medical college Ethics Committee. Written informed consent obtained from all participants prior to enrollment."
6 min
Dual-layer detector spectral CT parameters for differentiating low-grade and high-grade bladder urothelial cancer and predicting muscular invasion
Jian Lv, Guilin / China
Author Block: J. Lv1, R. X. Mu1, X. Li1, W. Zheng1, P. Yang1, X. Liu2, X. Qin1, X. Zhu3; 1Guilin/CN, 2Guangzhou/CN, 3Baise/CN
Purpose: To evaluate the diagnostic efficacy of dual-layer detector spectral CT (DLCT) in distinguishing low-grade (LG) from high-grade (HG) bladder urothelial cancer and in predicting muscular invasion (MI).
Methods or Background: A total of 116 patients (74 HG, 42 LG) who underwent preoperative DLCT scan were retrospectively included. Quantitative parameters included iodine density (ID) and extracellular volume (ECV) fractions in the arterial (AECV), venous (VECV), and equilibrium (EECV) phases were calculated. ECV was calculated as (1–hematocrit) × (lesion iodine concentration / blood pool iodine concentration). Univariate and multivariate regression analysis were performed to construct the combined diagnostic mode for predicting bladder cancer grading, as well as MI versus non muscular invasion (NMI) within HG and LG tumors, respectively. ROC analysis was used to assess diagnostic performance.
Results or Findings: V-ID, AECV, VECV, and EECV significantly differentiated HG from LG tumors, with AUCs of 0.801, 0.727, 0.770, and 0.833, respectively. Tumor calcification was identified as a risk factor, achieving an AUC of 0.638. The combined diagnostic model significantly improved the discriminative performance for HG versus LG bladder cancer, yielding an AUC of 0.891 (sensitivity 0.905, specificity 0.833). For predicting MI within HG tumors, heterogeneous muscular enhancement (HME) at VMI 40 keV, VECV and EECV served as predictors, with AUCs of 0.733, 0.692, and 0.714, respectively. The combined model for MI prediction achieved an AUC of 0.788 (sensitivity=0.800, specificity=0.708) within HG tumors. No spectral parameters demonstrated statistically significant difference in predicting MI within the LG subgroup.
Conclusion: DLCT-derived parameters, particularly ECV, allow reliable differentiation between low- and high-grade bladder urothelial carcinoma and provide valuable prediction of muscular invasion in HG tumors.
Limitations: None
Funding for this study: None
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: None
Purpose: To evaluate the diagnostic efficacy of dual-layer detector spectral CT (DLCT) in distinguishing low-grade (LG) from high-grade (HG) bladder urothelial cancer and in predicting muscular invasion (MI).
Methods or Background: A total of 116 patients (74 HG, 42 LG) who underwent preoperative DLCT scan were retrospectively included. Quantitative parameters included iodine density (ID) and extracellular volume (ECV) fractions in the arterial (AECV), venous (VECV), and equilibrium (EECV) phases were calculated. ECV was calculated as (1–hematocrit) × (lesion iodine concentration / blood pool iodine concentration). Univariate and multivariate regression analysis were performed to construct the combined diagnostic mode for predicting bladder cancer grading, as well as MI versus non muscular invasion (NMI) within HG and LG tumors, respectively. ROC analysis was used to assess diagnostic performance.
Results or Findings: V-ID, AECV, VECV, and EECV significantly differentiated HG from LG tumors, with AUCs of 0.801, 0.727, 0.770, and 0.833, respectively. Tumor calcification was identified as a risk factor, achieving an AUC of 0.638. The combined diagnostic model significantly improved the discriminative performance for HG versus LG bladder cancer, yielding an AUC of 0.891 (sensitivity 0.905, specificity 0.833). For predicting MI within HG tumors, heterogeneous muscular enhancement (HME) at VMI 40 keV, VECV and EECV served as predictors, with AUCs of 0.733, 0.692, and 0.714, respectively. The combined model for MI prediction achieved an AUC of 0.788 (sensitivity=0.800, specificity=0.708) within HG tumors. No spectral parameters demonstrated statistically significant difference in predicting MI within the LG subgroup.
Conclusion: DLCT-derived parameters, particularly ECV, allow reliable differentiation between low- and high-grade bladder urothelial carcinoma and provide valuable prediction of muscular invasion in HG tumors.
Limitations: None
Funding for this study: None
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: None
6 min
Diagnostic performance of Intravoxel Incoherent Motion MRI In Diagnosis and Assessment of Invasiveness of Urothelial Bladder Carcinoma
Mona Shawky Korkor, Mansoura / Egypt
Author Block: M. S. Korkor; Mansoura/EG
Purpose: The aim of the current study is to evaluate the role of intravoxel incoherent motion (IVIM) MRI and apparent diffusion coefficient (ADC) map in differentiating muscle invasive from non-muscle invasive bladder carcinomas.
Methods or Background: Urinary bladder cancer is a common malignancy. Preoperative accurate judgement of the degree of tissue invasion is crucial to determine the plan of therapy. The current study included 83 patients with bladder cancer. We included patients with clinical and ultrasonographical suspicion of bladder cancer. Patients with contraindications to MRI study as those with cardiac pacemaker, metallic implants or those received neoadjuvant chemotherapy were excluded. All patients were prepared and examined in the supine position using 3-Tesla MRI machine. The protocol included 3 main sequences: high resolution T2WI and DWI with post processing ADC map (b value = 1000). Additional IVIM sequences were performed with axial multiple b values DWI (b values = 0, 10, 25, 50, 75, 100, 150, 200, 400, 600, 800, 1000, 1400 s/mm2).
Regarding IVIM parameters: the D, D* and F were measured and assessed for the mass and wall beside mass. In addition, the ADC values of mass and wall beside mass were analyzed. All values were correlated to the histopathological findings.
Results or Findings: Values of both D (true diffusion coefficient) and ADC were lower in muscle invasive carcinomas than non-invasive masses with p values < 0.001 and 0.001, respectively.
Conclusion: IVIM MRI parameters and ADC map can be utilized to detect the degree of tissue invasion by bladder cancer.
Limitations: • Being a single center study.
• The analyzed parameters are obtained from single region of interest (ROI) that may not fully reflect the information of the whole tumor.
Funding for this study: No funding.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Study protocol was submitted for approval by Institution Research Board (IRB) of Mansoura Medical College (MS.23.01. 2275.R1).
Purpose: The aim of the current study is to evaluate the role of intravoxel incoherent motion (IVIM) MRI and apparent diffusion coefficient (ADC) map in differentiating muscle invasive from non-muscle invasive bladder carcinomas.
Methods or Background: Urinary bladder cancer is a common malignancy. Preoperative accurate judgement of the degree of tissue invasion is crucial to determine the plan of therapy. The current study included 83 patients with bladder cancer. We included patients with clinical and ultrasonographical suspicion of bladder cancer. Patients with contraindications to MRI study as those with cardiac pacemaker, metallic implants or those received neoadjuvant chemotherapy were excluded. All patients were prepared and examined in the supine position using 3-Tesla MRI machine. The protocol included 3 main sequences: high resolution T2WI and DWI with post processing ADC map (b value = 1000). Additional IVIM sequences were performed with axial multiple b values DWI (b values = 0, 10, 25, 50, 75, 100, 150, 200, 400, 600, 800, 1000, 1400 s/mm2).
Regarding IVIM parameters: the D, D* and F were measured and assessed for the mass and wall beside mass. In addition, the ADC values of mass and wall beside mass were analyzed. All values were correlated to the histopathological findings.
Results or Findings: Values of both D (true diffusion coefficient) and ADC were lower in muscle invasive carcinomas than non-invasive masses with p values < 0.001 and 0.001, respectively.
Conclusion: IVIM MRI parameters and ADC map can be utilized to detect the degree of tissue invasion by bladder cancer.
Limitations: • Being a single center study.
• The analyzed parameters are obtained from single region of interest (ROI) that may not fully reflect the information of the whole tumor.
Funding for this study: No funding.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Study protocol was submitted for approval by Institution Research Board (IRB) of Mansoura Medical College (MS.23.01. 2275.R1).
6 min
INcidence of malignant ureteric obStruction In patienTs with non-organ confined abdominopElvic malignancy (INSITE MUO) a multi-centre observational study
Oliver Llewellyn, Edinburgh / United Kingdom
Author Block: O. Llewellyn1, J. Blackmur1, J. Aning2, N. Shaida3, M. Bagkeris4, T. Barrett3, A. Laird1, .. BAUS Oncology1, .. UNITE Collaborative1; 1Edinburgh/UK, 2Bristol/UK, 3Cambridge/UK, 4London/UK
Purpose: Interventional radiologists and Urologists commonly encounter malignant ureteric obstruction (MUO). Capturing this patient population is challenging due to the pan-cancer aetiology. The aim of this study was to estimate the incidence of MUO in advanced abdominopelvic cancer and provide groundwork for further prospective work.
Methods or Background: National multicentre observational research study using national Interventional Radiology and Urology stakeholder networks (UNITE & BAUS). 24 sites contributed HRA approved, RCR funded study. Study teams contained at least one post-FRCR Trainee/Consultant.
Cases of advanced cancer involving the abdomen and cases of MUO were identified by trainees filtering reports. To do this, a standardised list of CT examination protocols involving the abdomen/pelvis over a 4 week study period was acquired at each site. Pseudonymised data was collected centrally from sites using the REDCap platform.
Results or Findings: 6781 abdominopelvic malignancy cases were identified. 2972 cancers were classified as advanced. 1387 cases missing TNM data were excluded from this interim analysis. The incidence of MUO in advanced cancer was 5.2% and the prevalence of was 9.8%. Survival analysis (Kaplan-Meier) indicated a protective influence of intervention (nephrostomy or stent) on overall survival in MUO in advanced cancer, p = 0.049. Intra-abdominal involvement from cancers originating outside the abdomen represented the largest proportion of the advanced intra-abdominal cancers (22.7%). The most common cause for MUO was prostate cancer (18.1%).
Conclusion: The incidence of MUO in patients with advanced abdominopelvic cancer was 5.2% and the prevalence of was 9.8%. Intervention (nephrostomy or ureteric stent) in MUO in advanced cancer appears to have a protective effect on overall survival.
Limitations: Observational study methodology
Funding for this study: Royal college of radiologists Kodak fellowship grant
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:
Purpose: Interventional radiologists and Urologists commonly encounter malignant ureteric obstruction (MUO). Capturing this patient population is challenging due to the pan-cancer aetiology. The aim of this study was to estimate the incidence of MUO in advanced abdominopelvic cancer and provide groundwork for further prospective work.
Methods or Background: National multicentre observational research study using national Interventional Radiology and Urology stakeholder networks (UNITE & BAUS). 24 sites contributed HRA approved, RCR funded study. Study teams contained at least one post-FRCR Trainee/Consultant.
Cases of advanced cancer involving the abdomen and cases of MUO were identified by trainees filtering reports. To do this, a standardised list of CT examination protocols involving the abdomen/pelvis over a 4 week study period was acquired at each site. Pseudonymised data was collected centrally from sites using the REDCap platform.
Results or Findings: 6781 abdominopelvic malignancy cases were identified. 2972 cancers were classified as advanced. 1387 cases missing TNM data were excluded from this interim analysis. The incidence of MUO in advanced cancer was 5.2% and the prevalence of was 9.8%. Survival analysis (Kaplan-Meier) indicated a protective influence of intervention (nephrostomy or stent) on overall survival in MUO in advanced cancer, p = 0.049. Intra-abdominal involvement from cancers originating outside the abdomen represented the largest proportion of the advanced intra-abdominal cancers (22.7%). The most common cause for MUO was prostate cancer (18.1%).
Conclusion: The incidence of MUO in patients with advanced abdominopelvic cancer was 5.2% and the prevalence of was 9.8%. Intervention (nephrostomy or ureteric stent) in MUO in advanced cancer appears to have a protective effect on overall survival.
Limitations: Observational study methodology
Funding for this study: Royal college of radiologists Kodak fellowship grant
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:
6 min
Preliminary evaluation of 5T MRI for bladder imaging: feasibility and image quality
Zihui Zhao, Nanjing / China
Author Block: Z. Zhao, X-G. Peng; Nanjing/CN
Purpose: To evaluate the feasibility and image quality of ultra-high-field 5T MRI for bladder imaging.
Methods or Background: This prospective study included 6 healthy volunteers undergoing 5T MRI and 7 bladder cancer patients undergoing 3T MRI between June and July 2025, with one patient scanned at both field strengths, yielding 14 bladder MRI datasets. Imaging sequences included T1-weighted, T2-weighted, and diffusion-weighted imaging (DWI). All participants drank 300–500 mL of water 40 minutes before scanning or voided 1–2 hours beforehand, with intravenous spasmolytic administered 15 minutes prior. Three experienced radiologists independently evaluated T2WI and DWI sequences using a 5-point Likert scale for bladder distension, artifacts, tumor–bladder wall interface clarity (or bladder wall clarity in volunteers), and overall image quality (a weighted composite score of the above metrics, on a 5-point scale). 5T images were compared with 3T images.
Results or Findings: Among 5T scans, 6/7 (85.7%) achieved optimal bladder distension, 1/7 (14.3%) showed notable artifacts, and all 7 (100%) exhibited clear bladder wall or tumor–bladder wall interface details. Notably, two volunteers prepared via waiting 1–2 hours after voiding showed almost no artifacts. Compared with 3T MRI, 5T MRI demonstrated marginally superior overall image quality (4.4±0.6 vs. 3.7±0.7, p=0.065), with better bladder distension, clearer bladder wall details, and fewer artifacts, although differences did not reach statistical significance (p > 0.05). Visual assessment in one patient scanned on both field strengths confirmed better image quality at 5T.
Conclusion: Ultra-high-field 5T MRI shows excellent feasibility for bladder imaging, with trends toward improved anatomic delineation and artifact reduction over 3T.
Limitations: The study is limited by a small, unevenly distributed sample, and most results were not statistically significant. Larger prospective studies are needed to confirm the clinical value of 5T bladder MRI.
Funding for this study: This study has received funding by National Natural Science Foundation of China (82272064), Jiangsu Provincial Science and Technique Program (BK20221461), Zhongda Hospital Affiliated to Southeast University, Jiangsu Province High-Level Hospital Pairing Assistance Construction (zdlyg08), Postgraduate Research & Practice Innovation Program of Jiangsu Province (KYCX23_0323 and KYCX22_0297).
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Approval No. 2024ZDSYLL061-P01
Purpose: To evaluate the feasibility and image quality of ultra-high-field 5T MRI for bladder imaging.
Methods or Background: This prospective study included 6 healthy volunteers undergoing 5T MRI and 7 bladder cancer patients undergoing 3T MRI between June and July 2025, with one patient scanned at both field strengths, yielding 14 bladder MRI datasets. Imaging sequences included T1-weighted, T2-weighted, and diffusion-weighted imaging (DWI). All participants drank 300–500 mL of water 40 minutes before scanning or voided 1–2 hours beforehand, with intravenous spasmolytic administered 15 minutes prior. Three experienced radiologists independently evaluated T2WI and DWI sequences using a 5-point Likert scale for bladder distension, artifacts, tumor–bladder wall interface clarity (or bladder wall clarity in volunteers), and overall image quality (a weighted composite score of the above metrics, on a 5-point scale). 5T images were compared with 3T images.
Results or Findings: Among 5T scans, 6/7 (85.7%) achieved optimal bladder distension, 1/7 (14.3%) showed notable artifacts, and all 7 (100%) exhibited clear bladder wall or tumor–bladder wall interface details. Notably, two volunteers prepared via waiting 1–2 hours after voiding showed almost no artifacts. Compared with 3T MRI, 5T MRI demonstrated marginally superior overall image quality (4.4±0.6 vs. 3.7±0.7, p=0.065), with better bladder distension, clearer bladder wall details, and fewer artifacts, although differences did not reach statistical significance (p > 0.05). Visual assessment in one patient scanned on both field strengths confirmed better image quality at 5T.
Conclusion: Ultra-high-field 5T MRI shows excellent feasibility for bladder imaging, with trends toward improved anatomic delineation and artifact reduction over 3T.
Limitations: The study is limited by a small, unevenly distributed sample, and most results were not statistically significant. Larger prospective studies are needed to confirm the clinical value of 5T bladder MRI.
Funding for this study: This study has received funding by National Natural Science Foundation of China (82272064), Jiangsu Provincial Science and Technique Program (BK20221461), Zhongda Hospital Affiliated to Southeast University, Jiangsu Province High-Level Hospital Pairing Assistance Construction (zdlyg08), Postgraduate Research & Practice Innovation Program of Jiangsu Province (KYCX23_0323 and KYCX22_0297).
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Approval No. 2024ZDSYLL061-P01
6 min
MRI-based Intratumoral Heterogeneity Quantification Accessing Early Neoadjuvant Treatment Response in Bladder Cancer
Lingmin Kong, Guangzhou / China
Author Block: L. Kong, Y. Qin, H. Wang; Guangzhou/CN
Purpose: To develop a quantitative measure of Intratumoral Heterogeneity (ITH) on pretreatment MRI scans and test its performance for predicting pathologic complete response (pCR) after neoadjuvant treatment (NAT) in bladder cancer (BCa)
Methods or Background: Pretreatment MRI scans, including contrast-enhanced (CE) and non-contrast-enhanced (NE) T1-weighted images (WI), were retrospectively acquired in patients with BCa who received NAT followed by surgery from April 2020 and August 2024. Tumor regions on MRI scans were clustered and integrated with global pixel distribution patterns to calculate ITH scores. Multivariable logistic regression analysis was used to identify variables associated with pCR, and significant variables, including clinicopathologic variables and ITH index, were combined into a predictive model for which performance was assessed using the area under the receiver operating characteristic curve (AUC).
Results or Findings: Among the 74 patients, the median age was 64 years (IQR: 57, 71); with 61 (82%) being male. Forty patients (54%) achieved pCR. The combined model showed good performance for predicting pCR to NAT in the training data set (AUC, 0.976 ) and test data sets (AUC, 0.875).
Conclusion: ITH analysis on pre-therapeutic CE- and NE-MRI demonstrated promising predictive ability for NAT responsiveness prior to treatment initiation.
Limitations: All patients in this study were recruited from single center, which need a further verification of the generalizability of the findings. Second, in this study, the use of manual tumor delineation was time consuming and also could affect the stability and repeatability of radiomic features. In future, it is imperative to employ an automated tumor segmentation method.
Funding for this study: This study received funding from the following sources: National Natural Science Foundation of China (Nos. 82372075, 82371911), Natural Science Outstanding Youth Fund Project of Guangdong province, China (No. 2024B1515020061)
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: the Institutional Review Board of the First Affiliated Hospital
Purpose: To develop a quantitative measure of Intratumoral Heterogeneity (ITH) on pretreatment MRI scans and test its performance for predicting pathologic complete response (pCR) after neoadjuvant treatment (NAT) in bladder cancer (BCa)
Methods or Background: Pretreatment MRI scans, including contrast-enhanced (CE) and non-contrast-enhanced (NE) T1-weighted images (WI), were retrospectively acquired in patients with BCa who received NAT followed by surgery from April 2020 and August 2024. Tumor regions on MRI scans were clustered and integrated with global pixel distribution patterns to calculate ITH scores. Multivariable logistic regression analysis was used to identify variables associated with pCR, and significant variables, including clinicopathologic variables and ITH index, were combined into a predictive model for which performance was assessed using the area under the receiver operating characteristic curve (AUC).
Results or Findings: Among the 74 patients, the median age was 64 years (IQR: 57, 71); with 61 (82%) being male. Forty patients (54%) achieved pCR. The combined model showed good performance for predicting pCR to NAT in the training data set (AUC, 0.976 ) and test data sets (AUC, 0.875).
Conclusion: ITH analysis on pre-therapeutic CE- and NE-MRI demonstrated promising predictive ability for NAT responsiveness prior to treatment initiation.
Limitations: All patients in this study were recruited from single center, which need a further verification of the generalizability of the findings. Second, in this study, the use of manual tumor delineation was time consuming and also could affect the stability and repeatability of radiomic features. In future, it is imperative to employ an automated tumor segmentation method.
Funding for this study: This study received funding from the following sources: National Natural Science Foundation of China (Nos. 82372075, 82371911), Natural Science Outstanding Youth Fund Project of Guangdong province, China (No. 2024B1515020061)
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: the Institutional Review Board of the First Affiliated Hospital
6 min
The role of multiparametric MRI and VIRADS in the preoperative differentiation of Ta and T1 bladder cancer
Merve Nur Tasdemir, Giresun / Turkey
Author Block: M. N. Tasdemir, U. Eryürük, S. Aslan; Giresun/TR
Purpose: To assess multiparametric MRI's utility, using VI-RADS and quantitative metrics – ADC, Tumor Contact Lenght (TCL) – , for distinguishing Ta from T1 stage non-muscle-invasive bladder cancer (NMIBC).
Methods or Background: This retrospective study enrolled patients with histopathologically confirmed NMIBC (2021–2025) who underwent mpMRI. Index tumors were evaluated using mp-VIRADS, and ADC values and TCL were measured. ROC analysis determined optimal cut-offs for ADC and TCL. The Mann-Whitney U and chi-square tests compared continuous variables and VI-RADS associations, respectively.
Results or Findings: Among 157 patients (83 Ta, 74 T1), T1 tumors exhibited significantly lower mean tumor ADC (1136 vs. 1326 x10⁻³ mm²/s) and mean normalized ADC (0.47 vs. 0.55), and higher mean TCL (30.4 vs. 14.0 mm, all p<0.001). ROC analysis for Ta-T1 discrimination yielded AUCs of 0.72 for tumor ADC (cut-off: 1216, sensitivity 79%, specificity 74%), 0.72 for normalized ADC (cut-off: 0.51, sensitivity 79%, specificity 68%), and 0.75 for TCL (cut-off: 17 mm, sensitivity 66%, specificity 79%), identifying TCL as the strongest discriminator. T1 stage tumors were significantly more frequent with higher mpVIRADS scores, with a prevalence of 22.2% (6/27) for score 1, 46.7% (42/90) for score 2, 52.4% (11/21) for score 3, 66.7% (6/9) for score 4, and 90.0% (9/10) for score 5 (p < 0.001).
Conclusion: MpMRI effectively differentiates Ta from T1 NMIBC. Both VI-RADS system and quantitative assessments are valuable, with TCL being the most potent discriminator. These findings underscore mpMRI's pivotal role in preoperative planning.
Limitations: This was a retrospective study.
Funding for this study: None.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: The study was approved by the local ethics commitee.
Purpose: To assess multiparametric MRI's utility, using VI-RADS and quantitative metrics – ADC, Tumor Contact Lenght (TCL) – , for distinguishing Ta from T1 stage non-muscle-invasive bladder cancer (NMIBC).
Methods or Background: This retrospective study enrolled patients with histopathologically confirmed NMIBC (2021–2025) who underwent mpMRI. Index tumors were evaluated using mp-VIRADS, and ADC values and TCL were measured. ROC analysis determined optimal cut-offs for ADC and TCL. The Mann-Whitney U and chi-square tests compared continuous variables and VI-RADS associations, respectively.
Results or Findings: Among 157 patients (83 Ta, 74 T1), T1 tumors exhibited significantly lower mean tumor ADC (1136 vs. 1326 x10⁻³ mm²/s) and mean normalized ADC (0.47 vs. 0.55), and higher mean TCL (30.4 vs. 14.0 mm, all p<0.001). ROC analysis for Ta-T1 discrimination yielded AUCs of 0.72 for tumor ADC (cut-off: 1216, sensitivity 79%, specificity 74%), 0.72 for normalized ADC (cut-off: 0.51, sensitivity 79%, specificity 68%), and 0.75 for TCL (cut-off: 17 mm, sensitivity 66%, specificity 79%), identifying TCL as the strongest discriminator. T1 stage tumors were significantly more frequent with higher mpVIRADS scores, with a prevalence of 22.2% (6/27) for score 1, 46.7% (42/90) for score 2, 52.4% (11/21) for score 3, 66.7% (6/9) for score 4, and 90.0% (9/10) for score 5 (p < 0.001).
Conclusion: MpMRI effectively differentiates Ta from T1 NMIBC. Both VI-RADS system and quantitative assessments are valuable, with TCL being the most potent discriminator. These findings underscore mpMRI's pivotal role in preoperative planning.
Limitations: This was a retrospective study.
Funding for this study: None.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: The study was approved by the local ethics commitee.