Cancer diagnostics using mRNA sequencing of tumor-educated platelets

In Chapter 1, we explore literature that describes the relationship between cancer and platelets, focusing on tumor-educated platelets (TEPs) as a novel biosource for cancer diagnostics. TEPs, which are influenced by tumor cells, show altered RNA profiles in cancer patients, making them promising candidates for liquid biopsies—a minimally invasive method for cancer detection. Historically, a link between platelets and cancer was noted, and modern research shows TEPs process RNA in response to tumor signals, possibly sequestering tumor-derived RNA. TEPs are abundant, easy to isolate, and provide high-quality RNA, making them viable for use alongside other biosources like circulating tumor DNA for enhanced diagnostic accuracy. Chapter 2 highlights an algorithm developed for detecting non-small cell lung cancer (NSCLC) using TEPs, yielding high accuracy in identifying both early and late-stage cancers. Using particle swarm optimization (PSO) to enhance RNA biomarker selection from platelet RNA libraries, the algorithm achieved 88% accuracy (AUC 0.94) in late-stage and 81% (AUC 0.89) in early-stage validation. This approach shows that TEP RNA could reliably detect various stages of NSCLC. Chapter 3 outlines the thromboSeq protocol, detailing both wet- and dry-lab methods for generating platelet RNA sequencing libraries and developing machine learning algorithms. The wet-lab protocol includes isolating and amplifying platelet RNA, preparing it for sequencing, while the dry-lab protocol involves quality control, data normalization, and PSO-enhanced support vector machine development. This protocol allows the scientific community to leverage TEP RNA for diagnostic development within a week. In Chapter 4, we examine the potential of TEP RNA as a sarcoma screening tool. Sarcomas are rare and recurrent malignancies with no current blood-based biomarkers. This study identified a sarcoma-specific TEP RNA signature in 57 sarcoma patients and a validation cohort (AUC 0.93) with 87% accuracy, indicating that TEP RNA may be effective for sarcoma diagnostics. Chapter 5 discusses using TEPs to distinguish glioblastoma from other brain conditions like multiple sclerosis (MS) and brain metastasis. TEP-derived RNA panels accurately identified glioblastoma with 80% accuracy (AUC 0.81) and 95% accuracy (AUC 0.97) in validation studies. Additionally, a digitalSWARM algorithm improved glioblastoma monitoring by distinguishing true progression from false-positive results. Chapters 6 and 7 show the methodology’s ability to differentiate malignant from non-malignant cases, as well as non-malignant diseases from controls. Chapter 6 presents a proof-of-concept study on detecting MS through TEP RNA, achieving 80% accuracy in validation. Chapter 7 explores pulmonary hypertension (PH) detection, where TEP RNA panels distinguished PH patients with an accuracy of 77% (AUC 0.89). Chapter 8 assesses TEPs for pan-cancer detection and tissue of origin (TOO) testing. In a study of 1,096 samples, TEP RNA accurately detected 18 cancer types, with a specificity of 99% in controls and correctly identified location of tumor origin in over 80% of the cancer cases. This study underscores TEP RNA's potential in blood-based cancer screening. Chapter 9 demonstrates the translation of a 58-gene signature to a companion diagnostic for BRAF-mutation-like subtypes in colorectal cancer using FFPE tissue samples. This diagnostic tool provides a consistent gene expression profile for selecting patients likely to respond to targeted therapy, validated across multiple experiments with high agreement. Finally, Chapter 10 draws overarching conclusions and discusses technical, biological, and clinical implications of thromboSeq. While TEP-based algorithms show promise, challenges in validation remain due to RNA contamination, platelet activation, and variability between patient groups. Further research into platelet-specific bioinformatics and clinical applications is essential for advancing thromboSeq toward routine clinical use.