Victor Tso

A Machine-Learned Predictor of Colonic Polyps Based on Urinary Metabolomics

We report an automated diagnostic test that uses the NMR spectrum of a single spot urine sample to accurately distinguish patients who require a colonoscopy from those who do not. Moreover, our approach can be adjusted to tradeoff between sensitivity and specificity. We developed our system using a group of 988 patients (633 normal and 355 who required colonoscopy) who were all at average or above-average risk for developing colorectal cancer. We obtained a metabolic profile of each subject, based on the urine samples collected from these subjects, analyzed via 1H-NMR and quantified using targeted profiling. Each subject then underwent a colonoscopy, the gold standard to determine whether he/she actually had an adenomatous polyp, a precursor to colorectal cancer. The metabolic profiles, colonoscopy outcomes, and medical histories were then analysed using machine learning to create a classifier that could predict whether a future patient requires a colonoscopy. Our empirical studies show that this classifier has a sensitivity of 64% and a specificity of 65% and, unlike the current fecal tests, allows the administrators of the test to adjust the tradeoff between the two.

Development and Validation of a Highly Sensitive Urine-Based Test to Identify Patients with Colonic Adenomatous Polyps

Objectives:Adenomatous polyps are precursors of colorectal cancer; their detection and removal is the goal of colon cancer screening programs. However, fecal-based methods identify patients with adenomatous polyps with low levels of sensitivity. The aim or this study was to develop a highly accurate, prototypic, proof-of-concept, spot urine-based diagnostic test using metabolomic technology to distinguish persons with adenomatous polyps from those without polyps.Methods:Prospective urine and stool samples were collected from 876 participants undergoing colonoscopy examination in a colon cancer screening program, from April 2008 to October 2009 at the University of Alberta. Colonoscopy reference standard identified 633 participants with no colonic polyps and 243 with colonic adenomatous polyps. One-dimensional nuclear magnetic resonance spectra of urine metabolites were analyzed to define a diagnostic metabolomic profile for colonic adenomas. A urine metabolomic diagnostic test for colonic adenomatous polyps was established using 67% of the samples (un-blinded training set) and validated using the other 33% of the samples (blinded testing set). The urine metabolomic diagnostic test’s specificity and sensitivity were compared with those of fecal-based tests.Results:Using a two-component, orthogonal, partial least-squares model of the metabolomic profile, the un-blinded training set identified patients with colonic adenomatous polyps with 88.9% sensitivity and 50.2% specificity. Validation using the blinded testing set confirmed sensitivity and specificity values of 82.7% and 51.2%, respectively. Sensitivities of fecal-based tests to identify colonic adenomas ranged from 2.5 to 11.9%.Conclusions:We describe a proof-of-concept spot urine-based metabolomic diagnostic test that identifies patients with colonic adenomatous polyps with a greater level of sensitivity (83%) than fecal-based tests.

Metabolomic Profiles Are Gender, Disease and Time Specific in the Interleukin-10 Gene-Deficient Mouse Model of Inflammatory Bowel Disease

Metabolomic profiling can be used to study disease-induced changes in inflammatory bowel diseases (IBD). The aim of this study was to investigate the difference in the metabolomic profile of males and females as they developed IBD. Using the IL-10 gene-deficient mouse model of IBD and wild-type mice, urine at age 4, 6, 8, 12, 16, and 20 weeks was collected and analyzed by nuclear magnetic resonance (NMR) spectroscopy. Multivariate data analysis was employed to assess differences in metabolomic profiles that occurred as a consequence of IBD development and severity (at week 20). These changes were contrasted to those that occurred as a consequence of gender. Our results demonstrate that both IL-10 gene-deficient and wild-type mice exhibit gender-related changes in urinary metabolomic profile over time. Some male-female separating metabolites are common to both IL-10 gene-deficient and control wild-type mice and, therefore, appear to be related predominantly to gender maturation. In addition, we were able to identify gender-separating metabolites that are unique for IL-10 gene-deficient and wild-type mice and, therefore, may be indicative of a gender-specific involvement in the development and severity of the intestinal inflammation. The comparison of the gender-separating metabolomic profile from IL-10 gene-deficient mice and wild-type mice during the development of IBD allowed us to identify changes in profile patterns that appear to be imperative in the development of intestinal inflammation, but yet central to gender-related differences in IBD development. The knowledge of metabolomic profile differences by gender and by disease severity has potential clinical implications in the design of both biomarkers of disease as well as the development of optimal therapies.

Development and Validation of a High-Throughput Mass Spectrometry Based Urine Metabolomic Test for the Detection of Colonic Adenomatous Polyps

Background: Colorectal cancer is one of the leading causes of cancer deaths worldwide. The detection and removal of the precursors to colorectal cancer, adenomatous polyps, is the key for screening. The aim of this study was to develop a clinically scalable (high throughput, low cost, and high sensitivity) mass spectrometry (MS)-based urine metabolomic test for the detection of adenomatous polyps. Methods: Prospective urine and stool samples were collected from 685 participants enrolled in a colorectal cancer screening program to undergo colonoscopy examination. Statistical analysis was performed on 69 urine metabolites measured by one-dimensional nuclear magnetic resonance spectroscopy to identify key metabolites. A targeted MS assay was then developed to quantify the key metabolites in urine. A MS-based urine metabolomic diagnostic test for adenomatous polyps was established using 67% samples (un-blinded training set) and validated using the remaining 33% samples (blinded testing set). Results: The MS-based urine metabolomic test identifies patients with colonic adenomatous polyps with an AUC of 0.692, outperforming the NMR based predictor with an AUC of 0.670. Conclusion: Here we describe a clinically scalable MS-based urine metabolomic test that identifies patients with adenomatous polyps at a higher level of sensitivity (86%) over current fecal-based tests (<18%).

Clinical validation of a novel urine-based metabolomic test for the detection of colonic polyps on Chinese population

PurposeColorectal cancer is the fifth leading cause of cancer-related deaths in China. When detected early, with the removal of adenomatous polyps, precursors of colorectal cancer, it is preventable. The aim of this study was to evaluate a novel urine-based metabolomic diagnostic test for the detection of adenomatous polyps, PolypDx™, that was originally developed and validated using 1000 samples from Canadian Cohort, on Chinese population.MethodsProspective urine samples were collected from 1000 participants undergoing colonoscopy examination, from March 2013 to July 2014 at Minhang District, Shanghai Centre for Disease Control and Prevention. One-dimensional nuclear magnetic resonance spectra of urine metabolites were analyzed to determine the concentrations of three key metabolites used in PolypDx™. The predicted results were then compared to the gold standard for colorectal cancer diagnostic, colonoscopy. Area under curve (AUC) was calculated specifically for the Chinese population and compared with the Canadian dataset. Sensitivity and specificity of this urine-based metabolomic diagnostic test were also compared with three commercially available fecal-based tests.ResultsAn AUC of 0.717 for PolypDx™ was calculated on Chinese dataset which is slightly lower than the AUC on the Canadian dataset. A sensitivity of 82.6% and a specificity of 42.4% were achieved on Chinese dataset.ConclusionsHere, we validated a novel urine-based metabolomic diagnostic test for the detection of adenomatous polyps, PolypDx™, on Chinese population through a sample size of 1000 participants with a greater level of sensitivity than fecal-based tests.

Metabolomic profiling to characterize acute intestinal ischemia/reperfusion injury

Sepsis and septic shock are the leading causes of death in critically ill patients. Acute intestinal ischemia/reperfusion (AII/R) is an adaptive response to shock. The high mortality rate from AII/R is due to the severity of the disease and, more importantly, the failure of timely diagnosis. The objective of this investigation is to use nuclear magnetic resonance (NMR) analysis to characterize urine metabolomic profile of AII/R injury in a mouse model. Animals were exposed to sham, early (30 min) or late (60 min) acute intestinal ischemia by complete occlusion of the superior mesenteric artery, followed by 2 hrs of reperfusion. Urine was collected and analyzed by NMR spectroscopy. Urinary metabolite concentrations demonstrated that different profiles could be delineated based on the duration of the intestinal ischemia. Metabolites such as allantoin, creatinine, proline, and methylamine could be predictive of AII/R injury. Lactate, currently used for clinical diagnosis, was found not to significantly contribute to the classification model for either early or late ischemia. This study demonstrates that patterns of changes in urinary metabolites are effective at distinguishing AII/R progression in an animal model. This is a proof-of-concept study to further support examination of metabolites in the clinical diagnosis of intestinal ischemia reperfusion injury in patients. The discovery of a fingerprint metabolite profile of AII/R will be a major advancement in the diagnosis, treatment, and prevention of systemic injury in critically ill patients.

Urine metabolomics as a predictor of patient tolerance and response to adjuvant chemotherapy in colorectal cancer

Colorectal cancer is the third leading cause of cancer-associated mortality in the western world. The ability to predict a patients response to chemotherapy may be of great value for clinicians and patients when planning cancer treatment. The aim of the current study was to develop a urine metabolomics-based biomarker panel to predict adverse events and response to chemotherapy in patients with colorectal cancer. A retrospective chart review of patients diagnosed with stage III or IV colorectal cancer between 2008 and 2012 was performed. The exclusion criteria included chemotherapy for palliation and patients living outside of Alberta. Data was collected concerning the chemotherapy regimen, adverse events associated with chemotherapy, disease progression and recurrence and 5-year survival. Adverse events were subdivided as follows: Delays in treatment, dose reductions, hospitalizations and chemotherapy regime changes. Patients provided urine samples for analysis prior to any intervention. Nuclear magnetic resonance (NMR) spectra of urine samples were acquired. The 1H NMR spectrum of each urine sample was analyzed using Chenomx NMRSuite v7.0. Using machine learning, predictors were generated and evaluated using 10-fold cross-validation. Urine spectra were obtained for 62 patients. The best predictors resulted in area under the receiver operating characteristic curve values of: 0.542 for chemotherapy dose reduction, 0.612 for 5-year survival, 0.650 for cancer recurrence and 0.750 for treatment delay. Therefore, predictors were developed for response to and adverse events from chemotherapy for patients with colorectal cancer patients. The predictor for treatment delay has the most promise, and further studies will aid its refinement and improvement of its accuracy.

Tu1763 Nuclear Magnetic Resonance (NMR)-Base Urine Metabolomics Analysis of Patients With Clostridium difficile Infection

Background: Clostridium difficile infection (CDI) is a common hospital acquired infections with significant morbidity and mortality. The risk of recurrent CDI is approximately 20%, 40% and up to 80% after the first, second, and third episode, respectively. Furthermore, recurrent CDI is associated with intestinal dysbiosis, characterized by reduced biodiversity and expansion of potentially pathogenic bacteria. The novel science of metabolomics provides a powerful tool to characterize vast numbers of metabolites in biological fluids with complex compositions and has recently yielded several clinically important diagnostic tests in cases where the human intestinal microbiota is altered. Aim: the aim of this pilot project was to examine the urine metabolomics profile in patients with CDI to determine if a unique metabolic profile was associated with recurrent infections. Method: Spot urine samples were prospectively collected from 31 patients (male 19, mean age 55 years) with CDI: 12 during 1st episode, 5 during 2nd episode, 5 during 3rd episode, and 9 with > 4th episode. Onedimensional NMR spectra were acquired using an Oxford 600MHz NMR instrument with a VNMRJ software. The 1H NMR spectrum of each urine sample was analyzed using Chenomx NMRSuite v7.6 (Chenomx Inc., Edmonton, AB). Multivariate statistical analysis was performed by using Simca P+ v 12.0.1 from Umetrics (Uema, Sweden). Results: Using 69 metabolites, a two-component orthogonal partial least squares (OPLS) model for 1, 2, 3 or at least 4 episodes of CIDwas built, with R2Y= 0.471 andQ2= 0.12 (models predictability of data in 7-fold cross validation). Based on OPLS model, clustering was observed among patients with 1 episode, 2 episodes and 3 or more episodes of CDI, respectively as shown in figure 1. Conclusion: This is the first study to demonstrate that NMR urine metabolomics has the ability to distinguish patients with various frequencies of recurrent CDI. Urine metabolomics has the potential to become an accurate, noninvasive and inexpensive diagnostic test for predicting recurrent CDI, and can help clinicians in risk stratification and selection of the most appropriate therapy for each patient.

A New and Highly Sensitive Screening Tool for Colorectal Adenomatous Polyps Using a Spot Urine Metabolomics Test

• Colorectal cancer (CRC) is a major public health concern • Adenomatous polyps are precursors of CRC and their identification is the basis for population-based CRC screening programs • Current world-wide population-based screening uses guaiac-based fecal occult blood tests (FOBT) which has a sensitivity of 5-30% for adenomatous polyps • Newer fecal immunochemical tests (FITs) use antibodies directed against human globin and have a sensitivity of 20-67% for colonic adenomas • Urine metabolomics has the potential to be a more accurate and patient-friendly screening test for colorectal cancer and precancerous lesions.