Stratifying risk of disease in haematuria patients using machine learning techniques to improve diagnostics.

Background: Detailed and invasive clinical investigations are required to identify the causes of haematuria. Highly unbalanced patient population (predominantly male) and a wide range of potential causes make the ability to correctly classify patients and identify patient-specific biomarkers a major challenge. Studies have shown that it is possible to improve the diagnosis using multi-marker analysis, even in unbalanced datasets, by applying advanced analytical methods. Here, we applied several machine learning algorithms to classify patients from the haematuria patient cohort (HaBio) by analysing multiple biomarkers and to identify the most relevant ones. Materials and methods: We applied several classification and feature selection methods (k-means clustering, decision trees, random forest with LIME explainer and CACTUS algorithm) to stratify patients into two groups: healthy (with no clear cause of haematuria) or sick (with an identified cause of haematuria e.g., bladder cancer, or infection). The classification performance of the models was compared. Biomarkers identified as important by the algorithms were also analysed in relation to their involvement in the pathological processes. Results: Results showed that a high unbalance in the datasets significantly affected the classification by random forest and decision trees, leading to the overestimation of the sick class and low model performance. CACTUS algorithm was more robust to the unbalance in the dataset. CACTUS obtained a balanced accuracy of 0.747 for both genders, 0.718 for females and 0.803 for males. The analysis showed that in the classification process for the whole dataset: microalbumin, male gender, and tPSA emerged as the most informative biomarkers. For males: age, microalbumin, tPSA, cystatin C, BTA, HAD and S100A4 were the most significant biomarkers while for females microalbumin, IL-8, pERK, and CXCL16. Conclusions: CACTUS algorithm demonstrated improved performance compared with other methods such as decision trees and random forest. Additionally, we identified the most relevant biomarkers for the specific patient group, which could be considered in the future as novel biomarkers for diagnosis. Our results have the potential to inform future research and provide new personalised diagnostic approaches tailored directly to the needs of the individuals.

Biomarkers to assess the risk of bladder cancer in patients presenting with haematuria are gender-specific.

Introduction: Haematuria is a common red flag symptom of urinary tract cancer. Bladder cancer (BC) is the most common cancer to present with haematuria. Women presenting with haematuria are often underdiagnosed. Currently, no gender-specific tests are utilized in clinical practice. Considerable healthcare resources are needed to investigate causes of haematuria and this study was set up to help identify markers of BC. The aim of the study was to define biomarker algorithms in haematuria patients using an expanded panel of biomarkers to diagnose BC and investigate if the algorithms are gender-specific. Materials and Methods: A total of n=675 patients with a history of haematuria were recruited from Northern Ireland hospitals. Patients were collected on a 2:1 ratio, non-BC (control) n=474: BC n=201. A detailed clinical history, urine and blood samples were collected. Biomarkers, known to be involved in the pathobiology underlying bladder carcinogenesis were investigated. Biomarkers differentially expressed between groups were investigated using Wilcoxon rank sum and linear regression. Results: Biomarkers were gender specific. Two biomarker-algorithms were identified to triage haematuria patients; male - u_NSE, s_PAI-1/tPA, u_midkine, u_NGAL, u_MMP-9/TIMP-1 and s_prolactin (u=urine; s=serum); sensitivity 71.8%, specificity 72.8%; AUROC 0.795; and female urine biomarkers - IL-12p70, IL-13, midkine and clusterin; sensitivity 83.7%, specificity 79.7%; AUROC 0.865. Addition of the clinical variable infection to both algorithms increased both AUROC to 0.822 (DeLong p=0.014) and to 0.923 (DeLong p=0.004) for males and females, respectively. Combining clinical risk factors with biomarker algorithms would enable application of the algorithms to triage haematuria patients. Conclusion: Using gender-specific biomarker algorithms in combination with clinical risks that are associated with BC would allow clinicians to better manage haematuria patients and potentially reduce underdiagnosis in females. In this study, we demonstrate, for the first time, that blood and urine biomarkers are gender-specific when assessing risk of BC in patients who present with blood in their urine. Combining biomarker data with clinical factors could improve triage when referring patients for further investigations.

Clinical features and predictive biomarkers for bladder cancer in patients with type 2 diabetes presenting with haematuria.

AIMS: To identify clinical features and protein biomarkers associated with bladder cancer (BC) in individuals with type 2 diabetes mellitus presenting with haematuria. MATERIALS AND METHODS: Data collected from the Haematuria Biomarker (HaBio) study was used in this analysis. A matched sub-cohort of patients with type 2 diabetes and patients without diabetes was created based on age, sex, and BC diagnosis, using approximately a 1:2 fixed ratio. Randox Biochip Array Technology and ELISA were applied for measurement of 66 candidate serum and urine protein biomarkers. Hazard ratios and 95% confidence intervals were estimated by chi-squared and Wilcoxon rank sum test for clinical features and candidate protein biomarkers. Diagnostic protein biomarker models were identified using Lasso-based binominal regression analysis. RESULTS: There was no difference in BC grade, stage, and severity between individuals with type 2 diabetes and matched controls. Incidence of chronic kidney disease (CKD) was significantly higher in patients with type 2 diabetes (p = 0.008), and CKD was significantly associated with BC in patients with type 2 diabetes (p = 0.032). A biomarker model, incorporating two serum (monocyte chemoattractant protein 1 and vascular endothelial growth factor) and three urine (interleukin 6, cytokeratin 18, and cytokeratin 8) proteins, predicted incidence of BC with an Area Under the Curve (AUC) of 0.84 in individuals with type 2 diabetes. In people without diabetes, the AUC was 0.66. CONCLUSIONS: We demonstrate the potential clinical utility of a biomarker panel, which includes proteins related to BC pathogenesis and type 2 diabetes, for monitoring risk of BC in patients with type 2 diabetes. Earlier urology referral of patients with type 2 diabetes will improve outcomes for these patients. TRIAL REGISTRATION: http://www.isrctn.com/ISRCTN25823942.

Non-alcoholic fatty liver disease-A pilot study investigating early inflammatory and fibrotic biomarkers of NAFLD with alcoholic liver disease.

Introduction: Non-alcoholic fatty liver disease (NAFLD) is a condition where excess fat accumulates in the liver (hepatic steatosis) and there is no history of alcohol abuse or other secondary causes of chronic liver disease. NAFLD is a very common disorder, occurring in 25% of the global population. NAFLD is now the most common chronic liver disorder in Western countries. Liver biopsy is the gold standard for NAFLD diagnosis and staging; however, this is invasive, costly and not without risk. Biomarkers that could diagnose and stage disease would reduce the need for biopsy and allow stratification of patients at risk of progression to non-alcoholic steatohepatitis (NASH). Methods: One hundred and thirty-five patients were involved in the study [N = 135: n = 34 controls; n = 26 simple steatosis; n = 61 NAFLD/NASH, and n = 14 alcoholic liver disease (ALD)]. Clinically diagnosed (ICD-10) patient serum samples were obtained from Discovery Life Sciences (US) along with clinical history. Samples were run in duplicate using high-sensitivity cytokine array I, immunoassays and ELISAs. In total, n = 20 individual biomarkers were investigated in this pilot study. Results: Thirteen/20 (65%) biomarkers were identified as significantly different between groups; IFNγ, EGF, IL-1ß, IL-6, IL-8, IL-10, TNFα, FABP-1, PIIINP, ST2/IL-33R, albumin, AST and ALT. Five/20 (25%) biomarker candidates were identified for further investigation; namely, three biomarkers of inflammation, IL-6, IL-8, and TNFα, and two biomarkers of fibrosis, PIIINP and ST2/IL-33R. Discussion: Single biomarkers are unlikely to be diagnostic or predictive at staging NAFLD due to the complex heterogeneity of the disease. However, biomarker combinations may help stratify risk and stage disease where patients are averse to biopsy. Further studies comparing the 5 biomarkers identified in this study with current diagnostic tests and fibrotic deposition in liver tissue are warranted.

Thrombomodulin Expression in Bladder Cancer Tissue and Its Association with Prognosis and Patient Survival.

BACKGROUND: Decreased expression of thrombomodulin (TM) in bladder cancer tissue has been shown to be associated with cell proliferation, increased malignancy and a poor prognosis. The aim of this study was to investigate the immunoexpression of TM in bladder tissue cores by immunohistochemistry (IHC) and the relationship between TM score and patient survival for the following pathologies: transitional cell papillary carcinoma (TCPC), transitional cell carcinoma (non-papillary) (TCC), squamous cell carcinoma (SCC), adenocarcinoma, and sarcoma. TM immunoexpression was also evaluated in normal adjacent bladder tissue cores. METHODS: TM immunoexpression was assessed in n=185 formalin-fixed paraffin-embedded (FFPE) bladder tissue cores from n=98 patients by IHC. Tissue cores included TCPC (n=29), TCC (n=85), SCC (n=21), adenocarcinoma (n=12), sarcoma (n=4), and normal tissue cores (n=34). RESULTS: TM immunoexpression scores are stronger in TCPC, TCC and SCC bladder cancer tissue cores with respect to adenocarcinoma and sarcoma (mean TM immunoexpression scores: 3.04, 2.57, 2.55, 1.55 and 1.19, respectively) (Kruskal-Wallis p<0.001). TM immunoexpression scores significantly decreased in bladder cancer tissue cores across both stage (p<0.001) and grade (p<0.001) (Kruskal-Wallis). Survival data were available for n=45 bladder cancer patients (mean follow-up of 34 months). Applying a TM immunoexpression cut-off score of 3.0 demonstrated that patients with bladder cancer who had a TM immunoexpression score <3.0 had lower survival rates (median survival 23.5 months). In contrast, patients with TM immunoexpression scores ≥3.0 had longer survival rates (median survival 40 months) (log-rank; p=0.045). CONCLUSION: TM immunoexpression in bladder cancer tissue may be a clinically relevant predictor of tumor progression and survival. Low expression of TM in bladder cancer biopsies or in recurrent bladder cancer may be indicative of a poor prognosis. TM immunoexpression could be used to guide clinical decision making.

An early analysis of the cost-effectiveness of a diagnostic classifier for risk stratification of haematuria patients (DCRSHP) compared to flexible cystoscopy in the diagnosis of bladder cancer.

BACKGROUND: Urothelial bladder cancer (UBC) is the 5th most common cancer in Western societies. The most common symptom of UBC is haematuria. Cystoscopy the gold standard for UBC detection, allows direct observation of the bladder, but is expensive, invasive, and uncomfortable. This study examines whether an alternative new urine-based diagnostic test, the DCRSHP, is cost-effective as a triage diagnostic tool compared to flexible cystoscopy in the diagnosis of UBC in haematuria patients. METHODS: A model-based cost-utility analysis using cost per quality adjusted life year and life year gained, parameterised with secondary data sources. RESULTS: If the DCRSHP is targeted at haematuria patients at lower risk of having bladder cancer e.g. younger patients, non-smokers, then it can be priced as high as £620, and be both effective and cost-effective. Sensitivity analysis found that DCRSHP is approximately 80% likely to be cost-effective across all willingness to pay values (for a QALY) and prevalence estimates. CONCLUSION: This analysis shows the potential for a non-invasive test to be added to the diagnostic pathway for haematuria patients suspected of having UBC. If the DCRSHP is applied targeting haematuria patients at low risk of UBC, then it has the potential to be both effective and cost-effective.

Collectives of diagnostic biomarkers identify high-risk subpopulations of hematuria patients: exploiting heterogeneity in large-scale biomarker data.

BACKGROUND: Ineffective risk stratification can delay diagnosis of serious disease in patients with hematuria. We applied a systems biology approach to analyze clinical, demographic and biomarker measurements (n = 29) collected from 157 hematuric patients: 80 urothelial cancer (UC) and 77 controls with confounding pathologies. METHODS: On the basis of biomarkers, we conducted agglomerative hierarchical clustering to identify patient and biomarker clusters. We then explored the relationship between the patient clusters and clinical characteristics using Chi-square analyses. We determined classification errors and areas under the receiver operating curve of Random Forest Classifiers (RFC) for patient subpopulations using the biomarker clusters to reduce the dimensionality of the data. RESULTS: Agglomerative clustering identified five patient clusters and seven biomarker clusters. Final diagnoses categories were non-randomly distributed across the five patient clusters. In addition, two of the patient clusters were enriched with patients with 'low cancer-risk' characteristics. The biomarkers which contributed to the diagnostic classifiers for these two patient clusters were similar. In contrast, three of the patient clusters were significantly enriched with patients harboring 'high cancer-risk" characteristics including proteinuria, aggressive pathological stage and grade, and malignant cytology. Patients in these three clusters included controls, that is, patients with other serious disease and patients with cancers other than UC. Biomarkers which contributed to the diagnostic classifiers for the largest 'high cancer- risk' cluster were different than those contributing to the classifiers for the 'low cancer-risk' clusters. Biomarkers which contributed to subpopulations that were split according to smoking status, gender and medication were different. CONCLUSIONS: The systems biology approach applied in this study allowed the hematuric patients to cluster naturally on the basis of the heterogeneity within their biomarker data, into five distinct risk subpopulations. Our findings highlight an approach with the promise to unlock the potential of biomarkers. This will be especially valuable in the field of diagnostic bladder cancer where biomarkers are urgently required. Clinicians could interpret risk classification scores in the context of clinical parameters at the time of triage. This could reduce cystoscopies and enable priority diagnosis of aggressive diseases, leading to improved patient outcomes at reduced costs.

Standardization of diagnostic biomarker concentrations in urine: the hematuria caveat.

Sensitive and specific urinary biomarkers can improve patient outcomes in many diseases through informing early diagnosis. Unfortunately, to date, the accuracy and translation of diagnostic urinary biomarkers into clinical practice has been disappointing. We believe this may be due to inappropriate standardization of diagnostic urinary biomarkers. Our objective was therefore to characterize the effects of standardizing urinary levels of IL-6, IL-8, and VEGF using the commonly applied standards namely urinary creatinine, osmolarity and protein. First, we report results based on the biomarker levels measured in 120 hematuric patients, 80 with pathologically confirmed bladder cancer, 27 with confounding pathologies and 13 in whom no underlying cause for their hematuria was identified, designated "no diagnosis". Protein levels were related to final diagnostic categories (p = 0.022, ANOVA). Osmolarity (mean = 529 mOsm; median = 528 mOsm) was normally distributed, while creatinine (mean = 10163 µmol/l, median = 9350 µmol/l) and protein (0.3297, 0.1155 mg/ml) distributions were not. When we compared AUROCs for IL-6, IL-8 and VEGF levels, we found that protein standardized levels consistently resulted in the lowest AUROCs. The latter suggests that protein standardization attenuates the "true" differences in biomarker levels across controls and bladder cancer samples. Second, in 72 hematuric patients; 48 bladder cancer and 24 controls, in whom urine samples had been collected on recruitment and at follow-up (median = 11 (1 to 20 months)), we demonstrate that protein levels were approximately 24% lower at follow-up (Bland Altman plots). There was an association between differences in individual biomarkers and differences in protein levels over time, particularly in control patients. Collectively, our findings identify caveats intrinsic to the common practice of protein standardization in biomarker discovery studies conducted on urine, particularly in patients with hematuria.

The impact of biomarkers in multivariate algorithms for bladder cancer diagnosis in patients with hematuria.

BACKGROUND: We appraised 23 biomarkers previously associated with urothelial cancer in a case-control study. Our aim was to determine whether single biomarkers and/or multivariate algorithms significantly improved on the predictive power of an algorithm based on demographics for prediction of urothelial cancer in patients presenting with hematuria. METHODS: Twenty-two biomarkers in urine and carcinoembryonic antigen (CEA) in serum were evaluated using enzyme-linked immunosorbent assays (ELISAs) and biochip array technology in 2 patient cohorts: 80 patients with urothelial cancer, and 77 controls with confounding pathologies. We used Forward Wald binary logistic regression analyses to create algorithms based on demographic variables designated prior predicted probability (PPP) and multivariate algorithms, which included PPP as a single variable. Areas under the curve (AUC) were determined after receiver-operator characteristic (ROC) analysis for single biomarkers and algorithms. RESULTS: After univariate analysis, 9 biomarkers were differentially expressed (t test; P < .05). CEA AUC 0.74; bladder tumor antigen (BTA) AUC 0.74; and nuclear matrix protein (NMP22) 0.79. PPP included age and smoking years; AUC 0.76. An algorithm including PPP, NMP22, and epidermal growth factor (EGF) significantly improved AUC to 0.90 when compared with PPP. The algorithm including PPP, BTA, CEA, and thrombomodulin (TM) increased AUC to 0.86. Sensitivities = 91%, 91%; and specificities = 80%, 71%, respectively, for the algorithms. CONCLUSIONS: Addition of biomarkers representing diverse carcinogenic pathways can significantly impact on the ROC statistic based on demographics. Benign prostate hyperplasia was a significant confounding pathology and identification of nonmuscle invasive urothelial cancer remains a challenge.

Optimising ultrasound-mediated gene transfer (sonoporation) in vitro and prolonged expression of a transgene in vivo: potential applications for gene therapy of cancer.

Therapeutic approaches using gene-based medicines promise alternatives or adjuncts to conventional cancer treatment. Because of its non-invasive nature, ultrasound, as a membrane-permeabilising stimulus has the potential to be highly competitive with viral gene delivery and existing non-viral alternatives. In optimising ultrasound-mediated, microbubble-assisted (MB101) gene tranfection in vitro, we demonstrate efficiencies of up to 18% using ultrasound at 1 MHz at a duty cycle of 25% at intensities ranging from 1 to 4 W cm(-2). Using ultrasound-mediated transfection together with an episomal plasmid-based gene expression system, we demonstrate prolonged functional gene expression of luciferase in mouse hind leg muscle and in tumours in vivo.

Enhancing ultrasound-mediated cell membrane permeabilisation (sonoporation) using a high frequency pulse regime and implications for ultrasound-aided cancer chemotherapy.

Delivering ultrasound to HeLa cells at 1MHz using a high frequency pulse regime (40kHz) and at a maximum energy density of 270Jcm(-2) resulted in significant cell membrane permeabilisation. Using FITC-dextran as a fluorogenic marker, optimally up to 64% of treated populations were permeabilised with cell viability remaining above 80%. Although cell membrane permeabilisation was observed in the presence of the microbubble-based ultrasound contrast agent, SonoVue, cell viability was severely compromised. Using the high frequency pulse regime in the absence of microbubbles, the LD50 of the cancer chemotherapeutic agent, camptothecin, was reduced from 58 to 18nM.

Kassinakinin S: a novel histamine-releasing heptadecapeptide from frog (Kassina senegalensis) skin secretion.

Amphibian defensive skin secretions remain a largely untapped resource for the peptide biochemist with an interest in the identification, structural characterization, and precursor cDNA cloning of novel bioactive peptides. Here we report the isolation, structural characterization, functional profiling, and nucleotide sequence of precursor cDNA of a novel histamine-releasing heptadecapeptide, FIPVTLLALHKIKEKLN-amide, from the defensive skin secretion of the African running frog, Kassina senegalensis. This peptide was found to be a potent histamine secretagogue (EC(50) = 6 microM; maximal release = 25 microM) in a rat peritoneal mast cell model system and was accordingly named kassinakinin S. The open-reading frame of the cDNA encoding prepro-kassinakinin S was found to consist of 71 amino acid residues containing a single copy of kassinakinin S and its glycyl residue amide donor at the C-terminus. Kassinakinin S can thus be added to the growing list of amphibian skin bioactive peptide prototypes.