Article of the Month – October 2022
October 1, 2022OrCam MyEye©
October 1, 2022Freenome
Freenome's vision is to use computational biology and ML techniques to reinvent disease management through early detection and precision intervention.
Overview
Since its beginning in 2014, Freenome has had a transparent vision – building a multi-disciplinary team expertly in computational biology and machine learning (ML) techniques to reinvent disease management through early detection and precision intervention. They seek to radically transform how patients with cancer are managed, equipping people everywhere with the knowledge and tools they need for a healthier life, and ultimately preventing cancer altogether.
Freenome AI Aspects
At Freenome, they associate people with next-generation blood tests for early cancer detection powered by their multiomics platform.
Freenome‘s blood tests look for mutations to detect the body’s early-warning signs for cancer and integrate a multidimensional view of both tumors and immune-derived signatures that facilitate the primary detection of cancer. By combining deep expertise in bioscience and advanced computational techniques to acknowledge disease-associated patterns among billions of circulating, cell-free biomarkers, they are developing simple and accurate blood tests for early cancer detection and integrating actionable insights into health systems to operationalize a circuit between care and science. (1)
At Freenome, they associate people with next-generation blood tests for early cancer detection powered by their multiomics platform.
Freenome‘s blood tests look for mutations to detect the body’s early-warning signs for cancer and integrate a multidimensional view of both tumors and immune-derived signatures that facilitate the primary detection of cancer. By combining deep expertise in bioscience and advanced computational techniques to acknowledge disease-associated patterns among billions of circulating, cell-free biomarkers, they are developing simple and accurate blood tests for early cancer detection and integrating actionable insights into health systems to operationalize a circuit between care and science. (1)
How does the multiomics platform work?
By training on thousands of cancer-positive blood samples, the platform learns which biomarkers’ patterns signify a cancer stage, type, and most effective treatment pathways. Also, training on healthy samples helps to establish what a normal composition of cell-free biomarkers should look like.
What does the platform detect?
Freenome‘s multiomics platform detects key biological signals from a routine blood draw. The platform integrates assays for cell-free DNA (cfDNA), methylation, and proteins with advanced computational biology and ML technology to understand additive signals for early cancer detection. The strategy involves a multidimensional view of both tumor and immune-derived signatures that enable the early detection of cancer instead of depending only on tumor-derived markers, which may miss early signs of cancer.
Mapping the multiomics of blood
- cfDNA: provides insights into immune response and tumor heterogeneity.
- cfRNA: reveals gene expression changes in the body related to tumor formation.
- Proteins: provides a higher-order dimension to immunogenic and oncogenic signals.
Freenome’s Financials Aspects
Freenome announced a $270 million Series C financing, bringing the company’s total financing to over $500 million since the company’s launch. These funds will be used to conduct the PREEMPT CRC study for Freenome‘s blood screening for early detection of colorectal cancer and early detection and intervention for other cancers. A part will also be devoted to detecting precancerous lesions, advancing in a line of blood detection tests, and thus continuing to build the company’s patented multi-omics platform. PREEMPT CRC is a study launched in May 2020 to support FDA approval of the blood test to reach the 45 million people who are currently not up-to-date with colorectal cancer screening in the United States. Bain Capital Life Sciences and Perceptive Advisors are financing the Series C, along with a group of other new investors, such as Fidelity Management and Research Company, LLC, Janus Handerson Investors, and Colorectal Cancer Alliance, among others. Existing Freenome investors, including RA Capital Management and Novartis, also participated in the financing.(3)
Early-Stage Colorectal Cancer Detected by Machine Learning
Circulating cell-free DNA (cfDNA) and DNA derived from non-tumor cells have been proposed to replace existing cancer screening methods. Machine learning (ML) can learn disease-related patterns from genome signals directly from patients with more aggressive or benign cancer.
ML could approximate the fetal fraction in the cfDNA of pregnant women without detecting single-nucleotide polymorphism differences between mother and fetus. A significant advantage is the use of low-depth sequencing, which decreases the cost of testing. (4)
Circulating cell-free DNA (cfDNA) and DNA derived from non-tumor cells have been proposed to replace existing cancer screening methods. Machine learning (ML) can learn disease-related patterns from genome signals directly from patients with more aggressive or benign cancer.
ML could approximate the fetal fraction in the cfDNA of pregnant women without detecting single-nucleotide polymorphism differences between mother and fetus. A significant advantage is the use of low-depth sequencing, which decreases the cost of testing. (4)
Plasma-Derived cfDNA Reveals Potential Biomarkers for Response Prediction and Monitoring in NSCLC Patients on Immunotherapy
Immune checkpoint inhibitors (ICI) improve clinical response to the treatment of non-small cell lung cancer (NSCLC) compared to chemotherapy. Biomarkers like elevated tumor-infiltrating cytotoxic T cells and natural killer cells at baseline are associated with ICI response. Accurate biomarkers that predict ICI response are needed. In this study, the objective was to evaluate the potential of cfDNA biomarkers to predict the response of nivolumab (PD-1 ICI) in patients with refractory metastatic NSCLC. Plasma from 30 patients with stage IV NSCLC was collected before and at week 8 of nivolumab therapy. Whole-genome sequencing was performed, and estimated cell proportions were determined by deconvolution of cfDNA co-fragmentation patterns. Results showed that high levels of monocytes at week eight were associated with more favorable overall survival, suggesting that changes in the immune system may provide insights for biomarkers discovery. (5)
Immune checkpoint inhibitors (ICI) improve clinical response to the treatment of non-small cell lung cancer (NSCLC) compared to chemotherapy. Biomarkers like elevated tumor-infiltrating cytotoxic T cells and natural killer cells at baseline are associated with ICI response. Accurate biomarkers that predict ICI response are needed. In this study, the objective was to evaluate the potential of cfDNA biomarkers to predict the response of nivolumab (PD-1 ICI) in patients with refractory metastatic NSCLC. Plasma from 30 patients with stage IV NSCLC was collected before and at week 8 of nivolumab therapy. Whole-genome sequencing was performed, and estimated cell proportions were determined by deconvolution of cfDNA co-fragmentation patterns. Results showed that high levels of monocytes at week eight were associated with more favorable overall survival, suggesting that changes in the immune system may provide insights for biomarkers discovery. (5)
Blood-Based Detection of Early-Stage Colorectal Cancer Using Multiomics and ML
Up to a third of adults within the appropriate age are not updated with recommended colorectal cancer (CRC) guideline screening. In this context, a blood-based test in early-stage CRC could improve adherence and eventually reduce mortality; however, biomarker tests alone have limited sensitivity, especially in the early stages.
The study’s objective was to evaluate their multiomics blood test in prospectively collected CRC samples and colonoscopy-confirmed negative controls by combining tumor- and non-tumor-derived signals from cfDNA, epigenetic, and protein biomarkers. They analyzed 43 subjects with CRC and 548 colonoscopy-confirmed negative controls across assays.
Results showed that the multi-center study achieved high sensitivity (94%) and high specificity (94%) for early-stage (I/II) colorectal adenocarcinoma. Also, the multiomics blood test demonstrated higher sensitivity for CRC when compared to stool-based FIT, plasma ctDNA, and plasma CEA. (6)
Plasma Analysis in Disease Progression in Early-Stage NSCLC
Lung cancer remains the leading cause of cancer-associated mortality in the United States, accounting for approximately 25% of all cancer deaths.A study investigated the potential of pre-surgery plasma-derived cfDNA and circulating proteins as biomarkers of progression in NSCLC patients after surgery.
Results showed that both circulating proteins and cfDNA are needed to identify prognostic biomarkers of disease progression and potential drug targets in early-stage NSCLC. Moreover, IL-1 was found to play a role in cancer progression. SOX-9 was observed in patients with advanced stages of cancer and ultimately had poorer clinical outcomes. IL-1RN and IL-1RA can be studied as potential prognostic biomarkers for NSCLC progression. (7)
Lung cancer remains the leading cause of cancer-associated mortality in the United States, accounting for approximately 25% of all cancer deaths.A study investigated the potential of pre-surgery plasma-derived cfDNA and circulating proteins as biomarkers of progression in NSCLC patients after surgery.
Results showed that both circulating proteins and cfDNA are needed to identify prognostic biomarkers of disease progression and potential drug targets in early-stage NSCLC. Moreover, IL-1 was found to play a role in cancer progression. SOX-9 was observed in patients with advanced stages of cancer and ultimately had poorer clinical outcomes. IL-1RN and IL-1RA can be studied as potential prognostic biomarkers for NSCLC progression. (7)