Collectives of diagnostic biomarkers identify high-risk subpopulations of hematuria patients, "Beyond the Abstract," by Kate Williamson, PhD, et al

BERKELEY, CA ( - Natural clustering of hematuria patients according to their biomarker profiles
This paper describes how hematuria patients cluster naturally into clinical groupings on the basis of their biomarker profiles. This is very promising because it suggests the feasibility of a "biomarker assessment" as a means to categorise patients into low- and high-risk groupings. 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. This approach allowed us to analyse collectives of biomarkers in the context of the comprehensive clinical and demographic data that had been collected for each of the 157 hematuria patients; 80 were diagnosed with urothelial cancer and 77 had confounding pathologies. The profile for each patient comprised the measurements of the 29 biomarkers in their samples. The biomarkers were assessed by scientists at Randox Laboratories Ltd, Crumlin Northern Ireland using Biochip Array Technology and commercially available ELISAs.

Final diagnoses of the hematuria patients recruited to the study
For 36 patients we were not able to identify the underlying cause of their hematuria. We defined the diagnosis for these patients as “no diagnosis.” For the remaining 41 controls, 6 patients had benign pathologies, 17 stones or inflammatory conditions, 12 benign prostate enlargement, and 6 were diagnosed with cancer other than urothelial cancers. Forty-seven of the urothelial cancer patients had pTa stage disease, 13 had pT1, 10 pT2a, 2 pT2b, 4 had pT3 and one pT4; two had carcinoma in situ; and the remaining patient had cancer of the kidney ureter.

Low-risk clusters
Following agglomerative clustering, we noticed that the 157 patients had clustered naturally into five distinct risk groupings. Two of the identified clusters had characteristics associated with benign or less serious disease. The patients in these groups were more likely to have a final diagnosis of no diagnosis or a benign pathology; and if they were diagnosed with urothelial cancer it was more likely to be low stage and grade disease. Further, these patients were less likely to have proteinuria or malignant cytology.

High-risk clusters
In contrast to the low-risk clusters, the 3 remaining clusters were exemplified by patients who had more serious diseases, including other cancers. High proportions of patients in these high-risk clusters had proteinuria and when they were diagnosed with urothelial cancer, it was usually of high stage and grade 3 disease.

Six of the patients who fell into the high-risk clusters had negative cystoscopy and had been designated as controls in our study. However, 3 of these patients had renal cell carcinoma and 3 had prostate cancer. Patients with aggressive urothelial cancer also fell into the 2 high-risk clusters.

HaBio study
We wish to build on our findings, reported in BMC Medicine, in our ongoing Haematuria Biomarker study, (HaBio). We have obtained funding from Invest Northern Ireland for a project entitled “Development of protein based algorithms for the detection of bladder cancer in haematuria populations.” HaBio is a collaborative project between scientists and academics in Queen’s University Belfast, Scientists in Randox Laboratories Ltd. and urologists and pathologists in the Belfast City and Dundonald hospitals in Northern Ireland. HaBio will recruit 999 patients who have a history of hematuria. Recruitment began in late 2012. We are collecting extensive clinical, demographic, and pathological data for each patient. Approximately 35 protein biomarkers will be assessed in samples from each patient. The aims of HaBio are to create diagnostic classifiers for patients with hematuria and to gain understanding of the underlying pathobiology of bladder carcinogenesis. We will employ both classical statistics and systems biology approaches to analyse the data.


Written by:
Kate Williamson,a Brian Duggan,b Frank Emmert-Streib,c and Mark Ruddockd as part of Beyond the Abstract on This initiative offers a method of publishing for the professional urology community. Authors are given an opportunity to expand on the circumstances, limitations etc... of their research by referencing the published abstract.

aChief Investigator, HaBio / Senior Lecturer, Centre for Cancer Research and Cell Biology, Queen’s University Belfast
bClinical Lead / Consultant Urologist, Belfast Health and Social Care Trust, Belfast
cComputational Biology and Machine Learning Laboratory, Centre for Cancer Research and Cell Biology, Queen’s University Belfast
dTeam Leader in Life Sciences, Randox Laboratories, Crumlin, Northern Ireland

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

More Information about Beyond the Abstract