Bladder Urothelial Carcinoma: Machine Learning-based Computed Tomography Radiomics for Prediction of Histological Variant

Published:March 26, 2022DOI:

      Rationale and Objectives

      Histological variant (HV) of bladder urothelial carcinoma (UC) is a significant factor for therapy management. We aim to assess the predictive performance of machine learning (ML)-based Computed Tomography radiomics of UC for HV.

      Materials and Methods

      Volume of interest of 37 bladder UC tumors, of which 21 were pure and 16 were HV, were manually segmented. The extracted first- and second-order texture features (n = 117) using 3-D Slicer radiomics were compared to the radical cystectomy histopathological results. ML algorithms were performed to determine the significant models using Python 2.3, Pycaret library. The sample size was increased to 74 by synthetic data generation, and three outliers from the training set were removed (training dataset; n = 52, test dataset; n = 19). The predictive performances of 15 ML algorithms were compared. Then, the best two models were evaluated on the test set and ensembled by Voting Classifier.


      The ML algorithms demonstrated area under curve (AUC) and accuracy ranging 0.79-0.97 and 50%-90%, respectively on the train set. The best models were Gradient Boosting Classifier (AUC: 0.95, accuracy: 90%) and CatBoost Classifier (AUC: 0.97, accuracy: 85%). On the test set; the Voting Classifier of these two models demonstrated AUC, accuracy, recall, precision, and F1 scores as follows; 0.93, 79%, 86%, 67%, and 75%, respectively.


      ML-based Computed Tomography radiomics of UC can predict HV, a prognostic factor that is indeterminable by qualitative radiological evaluation and can be missed in the preoperative histopathological specimens.

      Key Words


      AUC (Area under curve), ChT (Chemotherapy), CT (Computed Tomography), catboost (CatBoost classifier), gbc (Gradient Boosting classifier), GLCM (Gray-level co-occurrence matrix), GLDM (Gray-level difference method), GLRLM (Gray-level run-length matrix), GLSZM (Gray-level size zone), HV (Histological variant), ML (Machine learning), NGTDM (Neighbouring gray tone difference matrix), ROC (Receiver operating characteristic), RT (Radiotherapy), TUR-BT (Transurethral resection of bladder tumor), UC (Urothelial carcinoma)
      To read this article in full you will need to make a payment

      Purchase one-time access:

      Academic & Personal: 24 hour online accessCorporate R&D Professionals: 24 hour online access
      One-time access price info
      • For academic or personal research use, select 'Academic and Personal'
      • For corporate R&D use, select 'Corporate R&D Professionals'


      Subscribe to Academic Radiology
      Already a print subscriber? Claim online access
      Already an online subscriber? Sign in
      Institutional Access: Sign in to ScienceDirect


        • Moch H
        • Cubilla AL
        • Humphrey PA
        • et al.
        The 2016 WHO classification of tumours of the urinary system and male genital organs—part A: renal, penile, and testicular tumours.
        Eur Urol. 2016; 70: 93-105
        • Black PC
        • Brown GA
        • Dinney CP.
        The impact of variant histology on the outcome of bladder cancer treated with curative intent.
        Urol Oncol: Semin Orig Investig. 2009; 27: 3-7
        • Ehdaie B
        • Maschino A
        • Shariat SF
        • et al.
        Comparative outcomes of pure squamous cell carcinoma and urothelial carcinoma with squamous differentiation in patients treated with radical cystectomy.
        The J Urol. 2012; 187: 74-79
        • Koguchi D
        • Matsumoto K
        • Ikeda M
        • et al.
        Histologic variants associated with biological aggressiveness and poor prognosis in patients treated with radical cystectomy.
        Jpn J Clin. 2019; 49: 373-378
        • Mori K
        • Abufaraj M
        • Mostafaei H
        • et al.
        A systematic review and meta-analysis of variant histology in urothelial carcinoma of the bladder treated with radical cystectomy.
        J Urol. 2020; 204: 1129-1140
        • Takemoto K
        • Teishima J
        • Kohada Y
        • et al.
        The impact of histological variant on oncological outcomes in patients with urothelial carcinoma of the bladder treated with radical cystectomy.
        Anticancer Res. 2020; 40: 4787-4793
        • Lopez-Beltran A
        • Henriques V
        • Montironi R
        • et al.
        Variants and new entities of bladder cancer.
        Histopathology. 2019; 74: 77-96
        • Abufaraj M
        • Shariat SF
        • Foerster B
        • et al.
        Accuracy and prognostic value of variant histology and lymphovascular invasion at transurethral resection of bladder.
        World J Urol. 2018; 36: 231-240
        • Cai T
        • Tiscione D
        • Verze P
        • et al.
        Concordance and clinical significance of uncommon variants of bladder urothelial carcinoma in transurethral resection and radical cystectomy specimens.
        Urology. 2014; 84: 1141-1146
        • Shah RB
        • Montgomery JS
        • Montie JE
        • et al.
        Variant (divergent) histologic differentiation in urothelial carcinoma is under-recognized in community practice: impact of mandatory central pathology review at a large referral hospital.
        Urol Oncol: Semin Orig Investig. 2013; 31: 1650-1655
        • Amin MB.
        Histological variants of urothelial carcinoma: diagnostic, therapeutic and prognostic implications.
        Mod Pathol. 2009; 22: S96-S118
        • Fan T-W
        • Malhi H
        • Varghese B
        • et al.
        Computed tomography-based texture analysis of bladder cancer: differentiating urothelial carcinoma from micropapillary carcinoma.
        Abdom Radiol. 2019; 44: 201-208
        • Wang J
        • Wu C-J
        • Bao M-L
        • et al.
        Machine learning-based analysis of MR radiomics can help to improve the diagnostic performance of PI-RADS v2 in clinically relevant prostate cancer.
        Eur Radiol. 2017; 27: 4082-4090
        • Van Griethuysen JJ
        • Fedorov A
        • Parmar C
        • et al.
        Computational radiomics system to decode the radiographic phenotype.
        Cancer Res. 2017; 77: e104-ee07
        • Rankin D
        • Black M
        • Bond R
        • et al.
        Reliability of supervised machine learning using synthetic data in health care: Model to preserve privacy for data sharing.
        JMIR Med Inform. 2020; 8: e18910
        • Patro S
        • Sahu KK.
        Normalization: a preprocessing stage.
        arXiv preprint arXiv, India2015 (150306462)
        • Chawla NV
        • Bowyer KW
        • Hall LO
        • et al.
        SMOTE: synthetic minority over-sampling technique.
        J Artif Intell Res. 2002; 16: 321-357
        • Duda RO
        • Hart PE
        • Stork DG
        Pattern classification.
        John Willey and Sons, New York2001
        • Varma S
        • Simon R.
        Bias in error estimation when using cross-validation for model selection.
        BMC Bioinform. 2006; 7: 1-8
        • Mwangi B
        • Tian TS
        • Soares JC.
        A review of feature reduction techniques in neuroimaging.
        Neuroinformatics. 2014; 12: 229-244
        • Dormann CF
        • Elith J
        • Bacher S
        • et al.
        Collinearity: a review of methods to deal with it and a simulation study evaluating their performance.
        Ecography. 2013; 36: 27-46
        • Monn MF
        • Kaimakliotis HZ
        • Pedrosa JA
        • et al.
        Contemporary bladder cancer: variant histology may be a significant driver of disease.
        Urol Oncol: Semin Orig Investig. 2015; 33: 18.e15-18.e20
        • Moschini M
        • Dell’Oglio P
        • Gandaglia G
        • et al.
        Incidence and effect of variant histology on oncological outcomes in patients with bladder cancer treated with radical cystectomy.
        Urol Oncol: Semin Orig Investig. 2017; 35: 335-341
        • Soave A
        • Schmidt S
        • Dahlem R
        • et al.
        Does the extent of variant histology affect oncological outcomes in patients with urothelial carcinoma of the bladder treated with radical cystectomy?.
        Urol Oncol: Semin Orig Investig. 2015; 33: 21.e1-21.e9
        • Vetterlein MW
        • Wankowicz SA
        • Seisen T
        • et al.
        Neoadjuvant chemotherapy prior to radical cystectomy for muscle-invasive bladder cancer with variant histology.
        Cancer. 2017; 123: 4346-4355
        • Moschini M
        • D'andrea D
        • Korn S
        • et al.
        Characteristics and clinical significance of histological variants of bladder cancer.
        Nat Rev Urol. 2017; 14: 651-668
        • Xu X
        • Zhang X
        • Tian Q
        • et al.
        Three-dimensional texture features from intensity and high-order derivative maps for the discrimination between bladder tumors and wall tissues via MRI.
        Int J Comput Assist Radiol Surg. 2017; 12: 645-656
        • Wang Z
        • He Y
        • Wang N
        • et al.
        Clinical value of texture analysis in differentiation of urothelial carcinoma based on multiphase computed tomography images.
        Medicine. 2020; 99: e20093-e20098
        • Eminaga O
        • Eminaga N
        • Semjonow A
        • et al.
        Diagnostic classification of cystoscopic images using deep convolutional neural networks.
        J Urol Surg. 2019; 6: 264-265
        • Wang H
        • Xu X
        • Zhang X
        • et al.
        Elaboration of a multisequence MRI-based radiomics signature for the preoperative prediction of the muscle-invasive status of bladder cancer: a double-center study.
        Eur Radiol. 2020; 30: 4816-4827
        • Xu S
        • Yao Q
        • Liu G
        • et al.
        Combining DWI radiomics features with transurethral resection promotes the differentiation between muscle-invasive bladder cancer and non-muscle-invasive bladder cancer.
        Eur Radiol. 2020; 30: 1804-1812
        • Zhang X
        • Xu X
        • Tian Q
        • et al.
        Radiomics assessment of bladder cancer grade using texture features from diffusion-weighted imaging.
        J Magn Reson Imaging. 2017; 46: 1281-1288
        • Zhang G
        • Xu L
        • Zhao L
        • et al.
        CT-based radiomics to predict the pathological grade of bladder cancer.
        Eur Radiol. 2020; 30: 6749-6756