154P - Highly sensitive serum volatolomic biomarkers for pancreatic cancer diagnosis and prognosis

Background

Pancreatic adenocarcinoma (PAC) is one of the most lethal kinds of cancer. PAC is usually diagnosed very late, thus limiting treatment options and leading to low survival rates. Volatile organic compounds (VOCs) are final metabolism products that can be detected in a number of fluids. Our goal was to identify and validate diagnostic and prognostic PAC biomarkers in serum samples.

Methods

Twenty PAC patients and a sex/age matched control group (n=19) were recruited. A second group of 5 patients and 5 healthy volunteers was used for validation. IRB approval was provided by “Comité de Ética de Investigación con Medicamentos de La Rioja (CEImLar)” and all subjects signed the informed consent. A Varian CP-3800 gas chromatographer coupled to a Saturn 2200 ion trap mass spectrometer was used for the non-targeted analysis of VOCs. Surrogate standards were added to all serum samples, that were heated at 50°C and VOCs were collected into a 50/30 μm CAR/PDMS/DVB fiber (Supelco). Desorption was done at 250°C for 30 seconds in the injection port. MassHunter software (Agilent) was used for chromatogram processing including deconvolution of the signals.

Results

The volatolomic profile identified 433 VOCs in serum. Of these, 40 were able to distinguish cases from controls (diagnostic biomarkers). Unsupervised PCA and heatmaps, as well as supervised PLS-DA, showed a clear separation of cases vs controls. ROC curves indicated that 19 VOCs showed excellent predictive potential, with AUC > 0.90. The validation cohort confirmed the predictive value of 11 VOCs. Prognostic VOCs were identified by comparing patients that survived more than one year after diagnosis with those who did not. Sixteen VOCs showed prognostic discrimination and 3 of them had ROC with AUC > 0.86. Furthermore, the chemical structure of 4 of the main VOCs was confirmed by comparison with standards.

Conclusions

We have identified a panel of VOCs that correctly classify patients from controls and predict patient survival with high specificity and sensitivity. Reliable biomarkers may open the door to new technologies and the early detection of PAC.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

ADER (Project Hequipa, 2018-I-IDD-00059).

Disclosure

All authors have declared no conflicts of interest.