The role of VI-RADS scoring criteria for predicting oncological outcomes in bladder cancer.

PURPOSE: Our purpose was to evaluate the prognostic value of Vesical Imaging Reporting and Data System (VI-RADS) in bladder cancer (BCa) staging and predicting recurrence or progression. METHODS: We retrospectively analyzed the prospectively collected data from 96 patients with bladder tumors who underwent VI-RADS-based multiparametric magnetic resonance imaging (mpMRI) before endourological treatment from April 2021 to December 2022. Diagnostic performance was evaluated by comparing mpMRI reports with final pathology, using logistic regression for muscle-invasive bladder cancer (MIBC) predictors. Follow-up until May 2023 included Kaplan-Meier and Cox regression analysis to assess VI-RADS predictive roles for recurrence-free survival (RFS) and progression-free survival (PFS). RESULTS: A total of 96 patients (19.8% women, 80.2% men; median age 68.0 years) were included, with 71% having primary tumors and 29% recurrent BCa. Multiparametric MRI exhibited high sensitivity (92%) and specificity (79%) in predicting MIBC, showing no significant differences between primary and recurrent cancers (AUC: 0.96 vs. 0.92, P = .565). VI-RADS emerged as a key predictor for MIBC in both univariate (OR: 40.3, P < .001) and multivariate (OR: 54.6, P < .001) analyses. Primary tumors with VI-RADS ≥ 3 demonstrated significantly shorter RFS (P = .02) and PFS (P = .04). CONCLUSIONS: In conclusion, mpMRI with VI-RADS has a high diagnostic value in predicting MIBC in both primary and recurrent BCa. A VI-RADS threshold ≥ 3 is a strong predictor for MIBC, and in primary tumors predicts early recurrence and progression.

PENet: Prior evidence deep neural network for bladder cancer staging.

Bladder cancer is a heterogeneous, complicated, and widespread illness with high rates of morbidity, death, and expense if not treated adequately. The accurate and exact stage of bladder cancer is fundamental for treatment choices and prognostic forecasts, as indicated by convincing evidence from randomized trials. The extraordinary capability of Deep Convolutional Neural Networks (DCNNs) to extract features is one of the primary advantages offered by these types of networks. DCNNs work well in numerous real clinical medical applications as it demands costly large-scale data annotation. However, a lack of background information hinders its effectiveness and interpretability. Clinicians identify the stage of a tumor by evaluating whether the tumor is muscle-invasive, as shown in images by the tumor's infiltration of the bladder wall. Incorporating this clinical knowledge in DCNN has the ability to enhance the performance of bladder cancer staging and bring the prediction into accordance with medical principles. Therefore, we introduce PENet, an innovative prior evidence deep neural network, for classifying MR images of bladder cancer staging in line with clinical knowledge. To do this, first, the degree to which the tumor has penetrated the bladder wall is measured to get prior distribution parameters of class probability called prior evidence. Second, we formulate the posterior distribution of class probability according to Bayesian Theorem. Last, we modify the loss function based on posterior distribution of class probability which parameters include both prior evidence and prediction evidence in the learning procedure. Our investigation reveals that the prediction error and the variance of PENet may be reduced by giving the network prior evidence that is consistent with the ground truth. Using MR image datasets, experiments show that PENet performs better than image-based DCNN algorithms for bladder cancer staging.

The accuracy of Vesical Imaging-Reporting and Data System (VI-RADS): an updated comprehensive multi-institutional, multi-readers systematic review and meta-analysis from diagnostic evidence into future clinical recommendations.

PURPOSE: To determine through a comprehensive systematic review and meta-analysis the cumulative diagnostic performance of vesical imaging-reporting and data system (VIRADS) to predict preoperative muscle-invasiveness among different institutions, readers, and optimal scoring accuracy thresholds. METHODS: PubMed, Cochrane and Embase were searched from inception up to May 2021. Sensitivity (Sn), Specificity (Sp) were first estimated and subsequently pooled using hierarchical summary receiver operating characteristics (HSROC) modeling for both cut-off ≥ 3 and ≥ 4 to predict muscle-invasive bladder cancer (MIBC). Further sensitivity analysis, subgroup analysis and meta-regression were conducted to investigate contribution of moderators to heterogeneity. RESULTS: In total, n = 20 studies from 2019 to 2021 with n = 2477 patients by n = 53 genitourinary radiologists met the inclusion criteria. Pooled weighted Sn and Sp were 0.87 (95% CI 0.82-0.91) and 0.86 (95% CI 0.80-0.90) for cut-off ≥ 3 while 0.78 (95% CI 0.74-0.81) and 0.94 (95% CI 0.91-0.96) for cut-off ≥ 4. The area under the HSROC curve was 0.93 (95% CI 0.90-0.95) and 0.91 (95% CI 0.88-0.93) for cut-off ≥ 3 and ≥ 4, respectively. Meta-regression analyses showed no influence of clinical characteristics nor cumulative reader's experience while study design and radiological characteristics were found to influence the estimated outcome. CONCLUSION: We demonstrated excellent worldwide diagnostic performance of VI-RADS to determine pre-trans urethral resection of bladder tumor (TURBT) staging. Our findings corroborate wide reliability of VI-RADS accuracy also between different centers with varying experience underling the importance that standardization and reproducibility of VI-RADS may confer to multiparametric magnetic resonance imaging (mpMRI) for preoperative BCa discrimination.

Preoperative detection of Vesical Imaging-Reporting and Data System (VI-RADS) score 5 reliably identifies extravesical extension of urothelial carcinoma of the urinary bladder and predicts significant delayed time to cystectomy: time to reconsider the need for primary deep transurethral resection of bladder tumour in cases of locally advanced disease?

OBJECTIVES: (I) To determine Vesical Imaging-Reporting and Data System (VI-RADS) score 5 accuracy in predicting locally advanced bladder cancer (BCa), so as to potentially identify those patients who could avoid the morbidity of deep transurethral resection of bladder tumour (TURBT) in favour of histological sampling-TUR prior to radical cystectomy (RC). (II) To explore the predictive value of VI-RADS score 5 on time-to-cystectomy (TTC) outcomes. PATIENTS AND METHODS: We retrospectively reviewed patients' ineligible or refusing cisplatin-based combination neoadjuvant chemotherapy who underwent multiparametric magnetic resonance imaging (mpMRI) of the bladder prior to staging TURBT followed by RC for muscle-invasive BCa. Sensitivity, specificity, positive and negative predictive values (PPV, NPV) were calculated for VI-RADS score 5 vs. score 2-4 cases to assess the accuracy of mpMRI for extravesical BCa detection (≥pT3). VI-RADS score performance was assessed by receiver operating characteristics curve analysis. A Κ statistic was calculated to estimate mpMRI and pathological diagnostic agreement. The risk of delayed TTC (i.e. time from initial BCa diagnosis of >3 months) was assessed using multivariable logistic regression model. RESULTS: A total of 149 T2-T4a, cN0-M0 patients (VI-RADS score 5, n = 39 vs VI-RADS score 2-4, n = 110) were examined. VI-RADS score 5 demonstrated sensitivity, specificity, PPV and NPV, in detecting extravesical disease of 90.2% (95% confidence interval [CI] 84-94.3), 98.1% (95% CI 94-99.6), 94.9% (95% CI 89.6-97.6) and 96.4% (95% CI 91.6-98.6), respectively. The area under the curve was 94.2% (95% CI 88.7-99.7) and inter-reader agreement was excellent (Κinter 0.89). The mean (SD) TTC was 4.2 (2.3) and 2.8 (1.1) months for score 5 vs 2-4, respectively (P < 0.001). VI-RADS score 5 was found to independently increase risk of delayed TTC (odds ratio 2.81, 95% CI 1.20-6.62). CONCLUSION: The VI-RADS is valid and reliable in differentiating patients with extravesical disease from those with muscle-confined BCa before TURBT. Detection of VI-RADS score 5 was found to predict significant delay in TTC independently from other clinicopathological features. In the future, higher VI-RADS scores could potentially avoid the morbidity of extensive primary resections in favour of sampling-TUR for histology. Further prospective, larger, and multi-institutional trials are required to validate clinical applicability of our findings.

Photoacoustic imaging of the bladder: a pilot study.

Photoacoustic imaging is a promising new technology that combines tissue optical characteristics with ultrasound transmission and can potentially visualize tumor depth in bladder cancer. We imaged simulated tumors in 5 fresh porcine bladders with conventional pulse-echo sonography and photoacoustic imaging. Isoechoic biomaterials of different optical qualities were used. In all 5 of the bladder specimens, photoacoustic imaging showed injected biomaterials, containing varying degrees of pigment, better than control pulse-echo sonography. Photoacoustic imaging may be complementary to diagnostic information obtained by cystoscopy and urine cytologic analysis and could potentially obviate the need for biopsy in some tumors before definitive treatment.

Performance of Node-RADS Scoring System for a Standardized Assessment of Regional Lymph Nodes in Bladder Cancer Patients.

BACKGROUND: Current cross-sectional imaging modalities exhibit heterogenous diagnostic performances for the detection of a lymph node invasion (LNI) in bladder cancer (BCa) patients. Recently, the Node-RADS score was introduced to provide a standardized comprehensive evaluation of LNI, based on a five-item Likert scale accounting for both size and configuration criteria. In the current study, we hypothesized that the Node-RADS score accurately predicts the LNI and tested its diagnostic performance. METHODS: We retrospectively reviewed BCa patients treated with radical cystectomy (RC) and bilateral extended pelvic lymph node dissection, from January 2019 to June 2022. Patients receiving preoperative systemic chemotherapy were excluded. A logistic regression analysis tested the correlation between the Node-RADS score and LNI both at patient and lymph-node level. The ROC curves and the AUC depicted the overall diagnostic performance. In addition, the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated for different cut-off values (>1, >2, >3, >4). RESULTS: Overall, data from 49 patients were collected. Node-RADS assigned on CT scans images, was found to independently predict the LNI after an adjusted multivariable regression analysis, both at the patient (OR 3.36, 95%CI 1.68-9.40, p = 0.004) and lymph node (OR 5.18, 95%CI 3.39-8.64, p < 0.001) levels. Node-RADS exhibited an AUC of 0.87 and 0.91 at the patient and lymph node levels, respectively. With increasing Node-RADS cut-off values, the specificity and PPV increased from 57.1 to 97.1% and from 48.3 to 83.3%, respectively. Conversely, the sensitivity and NPV decreased from 100 to 35.7% and from 100 to 79.1%, respectively. Similar trends were recorded at the lymph node level. Potentially, Node-RADS > 2 could be considered as the best cut-off value due to balanced values at both the patient (77.1 and 78.6%, respectively) and lymph node levels (82.4 and 93.4%, respectively). CONCLUSIONS: The current study lays the foundation for the introduction of Node-RADS for the regional lymph-node evaluation in BCa patients. Interestingly, the Node-RADS score exhibited a moderate-to-high overall accuracy for the identification of LNI, with the possibility of setting different cut-off values according to specific clinical scenarios. However, these results need to be validated on larger cohorts before drawing definitive conclusions.

A multiparametric MRI-based CAD system for accurate diagnosis of bladder cancer staging.

Appropriate treatment of bladder cancer (BC) is widely based on accurate and early BC staging. In this paper, a multiparametric computer-aided diagnostic (MP-CAD) system is developed to differentiate between BC staging, especially T1 and T2 stages, using T2-weighted (T2W) magnetic resonance imaging (MRI) and diffusion-weighted (DW) MRI. Our framework starts with the segmentation of the bladder wall (BW) and localization of the whole BC volume (Vt) and its extent inside the wall (Vw). Our segmentation framework is based on a fully connected convolution neural network (CNN) and utilized an adaptive shape model followed by estimating a set of functional, texture, and morphological features. The functional features are derived from the cumulative distribution function (CDF) of the apparent diffusion coefficient. Texture features are radiomic features estimated from T2W-MRI, and morphological features are used to describe the tumors' geometric. Due to the significant texture difference between the wall and bladder lumen cells, Vt is parcelled into a set of nested equidistance surfaces (i.e., iso-surfaces). Finally, features are estimated for individual iso-surfaces, which are then augmented and used to train and test machine learning (ML) classifier based on neural networks. The system has been evaluated using 42 data sets, and a leave-one-subject-out approach is employed. The overall accuracy, sensitivity, specificity, and area under the receiver operating characteristics (ROC) curve (AUC) are 95.24%, 95.24%, 95.24%, and 0.9864, respectively. The advantage of fusion multiparametric iso-features is highlighted by comparing the diagnostic accuracy of individual MRI modality, which is confirmed by the ROC analysis. Moreover, the accuracy of our pipeline is compared against other statistical ML classifiers (i.e., random forest (RF) and support vector machine (SVM)). Our CAD system is also compared with other techniques (e.g., end-to-end convolution neural networks (i.e., ResNet50).

Systematic Review and Meta-Analysis of Vesical Imaging-Reporting and Data System (VI-RADS) Inter-Observer Reliability: An Added Value for Muscle Invasive Bladder Cancer Detection.

The Vesical Imaging-Reporting and Data System (VI-RADS) has been introduced to provide preoperative bladder cancer staging and has proved to be reliable in assessing the presence of muscle invasion in the pre-TURBT (trans-urethral resection of bladder tumor). We aimed to assess through a systematic review and meta-analysis the inter-reader variability of VI-RADS criteria for discriminating non-muscle vs. muscle invasive bladder cancer (NMIBC, MIBC). PubMed, Web of Science, Cochrane, and Embase were searched up until 30 July 2020. The Quality Appraisal of Diagnostic Reliability (QAREL) checklist was utilized to assess the quality of included studies and a pooled measure of inter-rater reliability (Cohen's Kappa [κ] and/or Intraclass correlation coefficients (ICCs)) was calculated. Further sensitivity analysis, subgroup analysis, and meta-regression were conducted to investigate the contribution of moderators to heterogeneity. In total, eight studies between 2018 and 2020, which evaluated a total of 1016 patients via 21 interpreting genitourinary (GU) radiologists, met inclusion criteria and were critically examined. No study was considered to be significantly flawed with publication bias. The pooled weighted mean κ estimate was 0.83 (95%CI: 0.78-0.88). Heterogeneity was present among the studies (Q = 185.92, d.f. = 7, p < 0.001; I2 = 92.7%). Meta-regression analyses showed that the relative % of MIBC diagnosis and cumulative reader's experience to influence the estimated outcome (Coeff: 0.019, SE: 0.007; p= 0.003 and 0.036, SE: 0.009; p = 0.001). In the present study, we confirm excellent pooled inter-reader agreement of VI-RADS to discriminate NMIBC from MIBC underlying the importance that standardization and reproducibility of VI-RADS may confer to multiparametric magnetic resonance (mpMRI) for preoperative BCa staging.

Preoperative imaging for locoregional staging of bladder cancer.

Bladder cancer is the ninth most common cancer, expected to lead to an estimated 17,670 deaths in the United States in 2019. Clinical management and prognosis of bladder cancer mainly depend on the extent of locoregional disease, particularly whether bladder muscle is involved. Therefore, bladder cancer is often divided into superficial, non-muscle-invasive bladder cancer and muscle-invasive bladder cancer; the latter often prompts consideration for cystectomy. While precise staging prior to cystectomy is crucial, the optimal preoperative imaging modality used to stage the disease remains controversial. Transurethral resection of bladder tumor (TURBT) followed by computed tomography (CT) urography is the current recommended approach for staging bladder cancer but suffers from a high rate of understaging. We review the recent literature and compare different imaging modalities for assessing the presence of muscle invasion and lymph node involvement prior to cystectomy and highlight the advantages of each modality.

Urinary bladder cancer staging in CT urography using machine learning.

PURPOSE: To evaluate the feasibility of using an objective computer-aided system to assess bladder cancer stage in CT Urography (CTU). MATERIALS AND METHODS: A dataset consisting of 84 bladder cancer lesions from 76 CTU cases was used to develop the computerized system for bladder cancer staging based on machine learning approaches. The cases were grouped into two classes based on pathological stage ≥ T2 or below T2, which is the decision threshold for neoadjuvant chemotherapy treatment clinically. There were 43 cancers below stage T2 and 41 cancers at stage T2 or above. All 84 lesions were automatically segmented using our previously developed auto-initialized cascaded level sets (AI-CALS) method. Morphological and texture features were extracted. The features were divided into subspaces of morphological features only, texture features only, and a combined set of both morphological and texture features. The dataset was split into Set 1 and Set 2 for two-fold cross-validation. Stepwise feature selection was used to select the most effective features. A linear discriminant analysis (LDA), a neural network (NN), a support vector machine (SVM), and a random forest (RAF) classifier were used to combine the features into a single score. The classification accuracy of the four classifiers was compared using the area under the receiver operating characteristic (ROC) curve (Az ). RESULTS: Based on the texture features only, the LDA classifier achieved a test Az of 0.91 on Set 1 and a test Az of 0.88 on Set 2. The test Az of the NN classifier for Set 1 and Set 2 were 0.89 and 0.92, respectively. The SVM classifier achieved test Az of 0.91 on Set 1 and test Az of 0.89 on Set 2. The test Az of the RAF classifier for Set 1 and Set 2 was 0.89 and 0.97, respectively. The morphological features alone, the texture features alone, and the combined feature set achieved comparable classification performance. CONCLUSION: The predictive model developed in this study shows promise as a classification tool for stratifying bladder cancer into two staging categories: greater than or equal to stage T2 and below stage T2.