Abstract
Tumor-educated platelets (TEPs) are potential biomarkers for cancer diagnostics. We employ TEP-derived RNA panels, determined by swarm intelligence, to detect and monitor glioblastoma. We assessed specificity by comparing the spliced RNA profile of TEPs from glioblastoma patients with multiple sclerosis and brain metastasis patients (validation series, n = 157; accuracy, 80%; AUC, 0.81 [95% CI, 0.74-0.89; p < 0.001]). Second, analysis of patients with glioblastoma versus asymptomatic healthy controls in an independent validation series (n = 347) provided a detection accuracy of 95% and AUC of 0.97 (95% CI, 0.95-0.99; p < 0.001). Finally, we developed the digitalSWARM algorithm to improve monitoring of glioblastoma progression and demonstrate that the TEP tumor scores of individual glioblastoma patients represent tumor behavior and could be used to distinguish false positive progression from true progression (validation series, n = 20; accuracy, 85%; AUC, 0.86 [95% CI, 0.70-1.00; p < 0.012]). In conclusion, TEPs have potential as a minimally invasive biosource for blood-based diagnostics and monitoring of glioblastoma patients.
Keywords:
blood platelets; glioblastoma; liquid biopsies; machine learning; swarm intelligence; tumor-educated platelets.