Prediction for Gastro-Oesophageal Adenocarcinoma, Precision at Biomarkers

Delays in GOA presentation worsen prognosis, despite better surgical and chemotherapeutic therapy. Due to GOA's molecular variability, biomarkers for prognosis and early diagnosis are lacking. This study uses Ectonucleotide Pyrophosphatase/Phosphodiesterase 2 (ENPP2), or Autotaxin (ATX), its metabolite LPA, and Carcinoembryonic Antigen (CEA) to mimic GOA development. We used bioinformatics and advanced molecular biology to examine ENPP2 and the diurnal variation of circulating ATX and LPA concentrations in no-treatment and treated GOA patients. In patients surgically treated with or without neoadjuvant chemotherapy, biomarker alterations were measured serially to establish their significance in disease progression, tumor burden, and treatment response. Each patient's genetic, biochemical, and clinical biomarker data was combined with machine learning actionable biomarker signals to create a risk and probable therapy response model. We found that ENPP2 is mutated in a small percentage of PDAC patients, but it is associated with more aggressive disease stage and shorter OS using TCGA data. Interestingly, diffuse GOA had greater baseline circulating ATX and LPA levels, which independently predicted poor survival. After surgery, these biomarkers are dramatically decreased independent of treatment, demonstrating that they closely correspond with tumors. CEA, ATX, and LPA levels boosted predictive power and enabled biomarker-based customized treatment planning.