Machine Learning in Prediction of Bladder Cancer on Clinical Laboratory Data

Bladder cancer has been increasing globally. Urinary cytology is considered a major screening method for bladder cancer, but it has poor sensitivity. This study aimed to utilize clinical laboratory data and machine learning methods to build predictive models of bladder cancer. A total of 1336 patients with cystitis, bladder cancer, kidney cancer, uterus cancer, and prostate cancer were enrolled in this study. Two-step feature selection combined with WEKA and forward selection was performed. Furthermore, five machine learning models, including decision tree, random forest, support vector machine, extreme gradient boosting (XGBoost), and light gradient boosting machine (GBM) were applied. Features, including calcium, alkaline phosphatase (ALP), albumin, urine ketone, urine occult blood, creatinine, alanine aminotransferase (ALT), and diabetes were selected. The lightGBM model obtained an accuracy of 84.8% to 86.9%, a sensitivity 84% to 87.8%, a specificity of 82.9% to 86.7%, and an area under the curve (AUC) of 0.88 to 0.92 in discriminating bladder cancer from cystitis and other cancers. Our study provides a demonstration of utilizing clinical laboratory data to predict bladder cancer.

Diagnostic accuracy of NMP-52, NMP-22 and urine cytology in the diagnosis of bladder cancer

Bladder cancer (BC) is the most important tumor problem of urologic cancer. Therefore, noninvasive urinary biomarkers were used for diagnosis of BC. However, the new biomarkers failed to reach higher accuracy. The aim of this study was to assess the diagnostic efficacy of nuclear matrix protein-22 (NMP- 22), nuclear matrix protein-52 (NMP-52), urinary cytology and to investigate combinations of urine NMP-52 with urinary cytology as noninvasive biomarkers to increase diagnostic performance of bladder cancer at different grades and stages. Overall, there were 156 subjects (62 BC, 54 cystitis patients and 40 healthy volunteers). The NMP-22 and NMP-52 were quantified in urine samples by ELISA. The urinary cytology is used by some physicians routinely for diagnosis of BC. The sensitivity and specificity for NMP-52 were 94% and 82%, for NMP-22 69% and 80.8% and for cytology 56% and 94.6% respectively and also, both urinary NMP-22 and NMP-52 have extremely significant relation (p<0.0001) to BC vs. healthy individuals and cystitis patients. Moreover, the combination of NMP- 52 with urinary cytology could predict all BC stages and grade with 95.6% sensitivity and 94.3% specificity. In conclusion, NMP-52 and urinary cytology in combination improve diagnostic performance for BC detection in different pathological types.

Significance of 5-Aminolevulinic Acid-mediated Photodynamic Diagnosis Following Standard Transurethral Resection in Non-muscle Invasive Bladder Cancer

Background/Aim: Transurethral resection of bladder tumors (TURBT) guided by photodynamic diagnosis (PDD) with 5-aminolevulinic acid (5-ALA) has been shown to be superior to conventional white light (WL) cystoscopy with regard to diagnostic accuracy, for non-muscle invasive bladder cancer (NMIBC); however, PDD is usually performed prior to WL TURBT. The objective of this study was to investigate the diagnostic significance of 5-ALA-mediated PDD following TUR of NMIBC. Patients and Methods: In 83 NMIBC patients, all visible tumors were completely resected under the conventional WL guidance followed by additional resection if fluorescent lesions were observed under the blue light (BL) guidance, and random biopsy was further conducted. Results: With the BL source, 39 (47.0%) patients were judged to have fluorescent lesions. Twenty (51.3%) of the 39 patients with positive findings by PDD were diagnosed with residual cancer by additional resection, while random biopsy detected cancers in 2 (4.5%) of the 44 patients with negative findings by PDD. Accordingly, the sensitivity, specificity, positive predictive value and negative predictive value in diagnosing residual cancer by our PDD system were 90.9, 68.9, 51.3 and 95.5%, respectively. Univariate analysis showed that the presence of residual cancers shown by WL TURBT was significantly associated with the preoperative findings of urinary cytology, tumor multiplicity, pathological T stage and tumor grade, of which only the preoperative finding on urinary cytology was shown to have an independent impact on the diagnosis of residual cancer. Conclusion: 5-ALA-mediated PDD following standard WL TURBT could improve the diagnostic accuracy in NMIBC patients, particularly those who are positive for preoperative urinary cytology.