Abstract
Background: We investigated the possibility of predicting tumor burden with salivary extracellular RNA (exRNA) biomarkers in gastric cancer patients. Methods: Saliva samples were prospectively collected from 50 gastric cancer patients who underwent gastrectomy with curative intent. Approximately 5 mL of saliva was collected before surgery and on the 5th to 7th days after surgery. The expression of three mRNAs (SPINK7, PPL, and SEMA4B) and two miRNAs (miR140-5p and miR301a) that were previously validated was determined by reverse transcription quantitative real-time PCR. Results: There were significant differences in the pre-operative expression of PPL (p = 0.025), SEMA4B (p = 0.012), and miR140-5p (p = 0.036) between pathologic stage I/II and III/IV groups. The area under the curve (AUC) of each respective multivariable model in predicting stage III/IV, which was adjusted for age and sex, was 75.4% (PPL), 82.5% (SEMA4B), and 75.5% (miR140-5p). In the multivariable model, including all three biomarkers, the AUC was 89.2%. On the other hand, none of the conventional tumor markers (CEA, CA19-9, and CA72-4) could predict tumor burden before surgery. The AUC of the multivariable model, including CEA, CA19-9, and CA72-4, was 67.2%, 66.2%, and 67.4%, respectively. When all three tumor markers were included in the multivariable model, the AUC was 70.5%. Conclusions: Noninvasively detected salivary biomarkers have been shown to have higher diagnostic accuracy than conventional tumor markers detected by invasive blood tests for estimating pre-operative tumor burden. This study demonstrates the potential utility of these biomarkers in pre-operative risk assessment and monitoring surgical treatment response to gastric cancer.