Urine samples were collected from 102 RCC patients and 178 control patients. After this, a subset of age matched patients made up the study’s cohort, with 32 RCC patients (14 females, 18 males) and 34 control patients (14 females, 20 males). Urine samples were diluted with MeOH (1:5) and then analyzed with both positive and negative ion modes HILIC-UPLC-MS in a Q Extractive HF mass spectrometer. First, a sample blank was utilized to analyze background signals. Then, spectral parameters such as retention time and m/z pairs were measured for each ionization mode using Compound Discoverer 2.1 and then retained if the parameters’ signal had 5 times the signal in the sample blank. With a standard student t-test, the parameters were filtered based on a p-value of less than 0.05 and with a two median fold change.
As a result, 569 metabolomic parameters were found for the 66 patients. With the PLS toolbox 8.1.1 with MATLAB R2012b, a PCA model was built to show the class separation between the control and RCC. When differentiating the RCC from the controls, Ms. Bonvillain and colleagues found an accuracy, sensitivity, and specificity that were 96.3, 93.9, and 98.8 respectively.
With these results, Ms. Bonvillain concluded that there is a potential for clinical translation to non-invasively screen, detect, and diagnose renal masses using urine samples. A future direction their research team proposed was to identify if there are metabolomics differences between different types of renal cancers.
Presented by: Annelise Bonvillain, MD
Written by: Cyrus Lin, Department of Urology, University of California-Irvine at the 2018 AUA Annual Meeting - May 18 - 21, 2018 – San Francisco, CA USA