br Extraction of VOCs From Urine Samples by Stir
Extraction of VOCs From Urine Samples by Stir Bar Sorptive Extraction
Urine samples were thawed in ice, transferred to centrifuge tubes, and centrifuged for 10 minutes at 300 g prior to extraction. To extract the organic metabolites (ie, VOCs), 1.0 mL of urine sample supernatant, 19.0 mL of DI water, 300 mL of 100-ppm Mirex solution (internal standard), and 600 mL of 2-M HCl were added into a 20-mL amber vial. A commercially available stir bar coated with polydimethylsiloxane (Twister, 10 mm 1 mm, Gerstel, Mülheim an der Ruhr, Germany) was then placed into the vial, and the solution was stirred for 2 hours at 1000 rpm. At the end of the stirring, the stir bar was removed from the solution, rinsed with DI water, dried with lint-free paper, and placed into a thermal desorption tube for chemical analysis.
Materials and Methods
Chemicals and Materials
All chemicals were of analytical grade. Mirex (99.0%, Dr. Ehrenstorfer GmbH, Germany), used as the internal standard, was purchased from the Laboratories of Dr. Ehrenstorfer, Germany. Mirex solution of 100 mg/L was prepared in methanol (liquid chromatography/mass spectrometry grade, Burdick & Jackson, Muskegon, MI). Hydrochloric 172922-91-7 (HCl) (37%) was purchased from Sigma-Aldrich (St. Louis, MO). Ultra-pure deionized (DI) water from Milli-Q system (Millipore, Bedford, MA) was used in the preparation of HCl solution and dilution of urine samples.
Patient Recruitment and Sample Collection
Internal Review Board approvals and written informed consents were obtained for the multi-institutional study. Patients presenting to the Geisinger Medical Center urologic clinic for evaluation of elevated PSA > 2.5 ng/mL or abnormal digital rectal exam in the urology clinic were selected for the study. Patients who did not wish to participate in this study or whose urinalysis was suspicious for infection were excluded. All patients provided urine specimens for dipstick urinalysis GC-MS Analysis
Urinary VOCs were analyzed in a thermal desorption unit (TDU, Gerstel), coupled with a 6890-GC system and a 5973-N Mass Selective Detector (Agilent Technologies, Wilmington DE). The thermal desorption process was programmed as follows. The initial temperature was set at 45 C holding for 0.5 minutes; the temperature was increased to 300 C at 60 C/minute and held for 5 minutes. Desorption gas flow was set at 1.0 mL/minute. During desorption, all the desorbed compounds were concentrated in a cold injection system, CIS4 (Gerstel), at 40 C prior to GC injection. Once the desorption process was completed, the CIS4 was heated to 300 C at 12 C/second and held for 5 minutes in a solvent vent mode. The VOCs were separated and analyzed by GC-MS using splitless mode through a ZB-5ms capillary column (30 m 0.25 mm 0.25 mm; Phenomenex, Torrence, CA). The oven temperature was programmed as follows: held at 35 C for 5 mi-nutes; heated to 300 C at 10 C/minute, and held for 10 minutes. The VOCs were detected by mass selective detector in scan mode (20-500 m/z). The National Institute of Standards and Technology Library was used for the identification of VOC profiles in urine samples.
Qin Gao et al
Table 1 Demographic Information of Patients With Prostate Cancer and Controls in the VOC Prostate Cancer Diagnosis Model Development
Training Cohort (Model Development)
Testing Cohort (Model Evaluation)
P Valuea Patients With
Prostate Cancer Controls Prostate Cancer Controls
Data are presented as median (interquartile range) for continuous variables and n (%) for categorical variables. Abbreviations: PSA ¼ prostate-specific antigen; VOC ¼ volatile organic compound. aThe P value from the t test of the PSA numbers between prostate cancer and control groups.
Data Processing and Statistical Analysis
We used Mirex as the internal standard because of its nonexis-tence in human urine. The relative intensity of each VOC peak could then be normalized against that of Mirex, allowing semi-quantitative statistical analysis of VOCs.
Over 9000 different VOC types were found in the urine samples, resulting in a high-dimensional modeling problem. To streamline the analysis, we first removed the VOCs that could be observed in less than 3% of the entire population. The remaining variables were screened by testing the difference in each VOC between the PCa-positive and control groups. The Wilcoxon rank-sum test was used because it can accommodate the zero inflation among many VOCs. Heat maps were generated to visualize those significant VOCs (P < .05) in the PCa-positive and control groups. Applying a liberal cutoff of 0.2 to the P values, over 800 VOCs remained for the model development. To deal with this p>>n scenario (ie, numbers of VOCs are much greater than the number of samples), we fit regularized logistic regression models18 with LASSO19 penalty, and the 10-fold cross-validation was used to select the optimal tuning parameter. The final logistic model was then evaluated via the receiver operating characteristic (ROC) curve and other performance measures on the basis of jackknife predic-tion,20 which helps alleviate the over-optimism induced by variable selection. Furthermore, the Firth approach was taken to fit the final logistic model in order to achieve bias reduction for the small sample scenario and deal with the nearly complete separation seen in the data.21 Another R package, known as OptimalCutpoints, was used to determine the optimal cut point for the diagnostic model cor-responding to the maximum Youden Index.22 All statistical analyses