Biodesix’s Diagnostic Cortex™ Platform Used in Three Studies Presented at SITC
Blood-Based Assays to Predict Patient Response to Combination Checkpoint Blockade, High-dose interleukin 2 (HDIL-2) in Melanoma November 15, Boulder, CO…Studies presented by Biodesix® at the 31st Annual Meeting of the Society for Immunotherapy of Cancer (SITC) demonstrated the potential for the utility of the company’s proprietary Diagnostic Cortex™ platform to design blood-based tests that provide prognostic information and support immunotherapy selection. This proprietary platform has been successfully applied to a wide range of clinical questions related to therapeutic vaccines and immunotherapies, including prognosis and therapy guidance in advanced melanoma. “Biodesix’s approach is based on the unbiased analysis of a very large number of proteins found in serum from patients with various malignancies. The study results indicate that the tests have the potential to offer clinically relevant information that could help physicians selecting immunotherapies for their patients make the best decision to optimize benefit,” said Dr. Jeffrey Weber, Deputy Director of the Laura and Isaac Perlmutter Cancer Center at NYU-Langone Medical Center in New York City. Oral Poster Presentation: Abstract Title: A test identifying advanced melanoma patients with long survival outcomes on nivolumab shows potential for selection for benefit from combination checkpoint blockade Authors: J. Weber, H. Kluger, R. Halaban, M. Sznol, H. Roder, J. Roder, S. Asmellash, A. Steingrimsson, S. Blackmon, R. Sullivan The Study: Using the Diagnostic Cortex™ platform, a test was designed to identify two groups of patients with metastatic melanoma who have high vs. low likelihood of good outcome when treated with anti-PD-1 monotherapy (Group 1 and Group 2 classifications, respectively). The performance of the test was validated in two independent patient cohorts. The test was then applied to an investigational cohort of patients treated with nivolumab/ipilimumab combination therapy. Comparison of the outcomes of patients treated with the combination versus mono-therapy indicated that while patients classified as Group 2 have better OS when treated with nivolumab/ipilimumab than with a single agent, Group 1 patients have similar outcomes on both treatments. Adaptation of Gene Set Enrichment Analysis (GSEA) to protein data suggests that complement system, acute phase and wound healing processes are correlated with the Group 1 vs. Group 2 classification. While further validation in randomized studies is needed, these preliminary results suggest that a blood-based test may provide a basis for identifying patients who require only monotherapy and those who may benefit from combination therapy. Poster Presentation: Title: High Dose Interleukin 2 (HD IL‑2) Select Trial in Melanoma: A tissue and blood collection protocol to identify predictive biomarkers of response to HD IL‑2 in patients with advanced melanoma Authors: Ryan J. Sullivan, Yujin Hoshida, Theodore Logan, Nikhil I. Khushalani, Anita Giobbie-Hurder, Kim Margolin, Johanna Roder, Rupal Bhatt, Henry Koon, Thomas Olencki, Thomas Hutson, Brendan Curti, Shauna Blackmon, James Mier, Heinrich Roder, John Stewart, Asim Amin, Marc Ernstoff, Joseph Clark, Michael B. Atkins, Howard Kaufman, Jeffrey Sosman, Sabina Signoretti, David F. McDermott The Study: The primary objective of this study was to validate the Class 2 (C2) gene signature as a predictor of response to HD IL‑2; a secondary objective was to evaluate serum markers as predictors of response to HD IL‑2. Because this therapy requires ICU-level inpatient care and has significant side effects, biomarkers of response to HD IL‑2 could be clinically relevant. The study’s conclusions included the finding that the serum protein assay developed with the Diagnostic Cortex™ platform discriminated between Early and Late progression and that further validation is needed. Poster Presentation: Title: Application of a test developed for prediction of response to high-dose interleukin-2 (HD IL‑2) and the BDX008 test for prediction of outcomes following checkpoint inhibitors to cohorts of patients treated with HD IL‑2 or nivolumab Authors: Ryan J. Sullivan, Theodore Logan, Nikhil I. Khushalani, Kim Margolin, Henry Koon, Thomas Olencki, Thomas Hutson, Brendan Curti, Joanna Roder, Shauna Blackmon, Heinrich Roder, John Stewart, Asim Amin, Marc Ernstoff, Joseph Clark, Michael B. Atkins, Howard Kaufman, Jeffrey Sosman, Jeffrey Weber, David F. McDermott The Study: The Diagnostic Cortex™platform has been used to design two tests relevant for immunotherapy treatment of patients with advanced melanoma: BDX008, a test identifying patients with better (BDX008+) or worse (BDX008-) clinical outcomes when treated with anti‑PD‑1 therapy, and the IL‑2 Test for identifying patients who are more (Group A) or less (Group B) likely to have durable benefit from HD IL‑2 treatment. Both tests were investigated in each of two cohorts of advanced melanoma patients, one treated with nivolumab and a second treated with HD IL‑2. All Group A patients were found to be BDX008+. In the cohort of patients treated with nivolumab, overall survival (OS) and progression-free survival (PFS) were found to depend on BDX008 status but not IL‑2 Test status. Within the HD IL-2 treated cohort, patients in Group A had better PFS than patients in Group B, and PFS did not depend on BDX008 status. While further validation of the IL-2 test is necessary, the results demonstrate the differential performance of the two tests in these two immunotherapeutic regimens. About the Diagnostic Cortex™ Platform The Diagnostic Cortex enables scientific research services for the development of clinically meaningful tests. It incorporates standard machine learning and techniques adapted from deep learning to design generalizable and performance-tuned tests from clinically defined data sets. This proprietary platform enables the design of blood-based tests from relatively small sample sets that:
- Minimize the risk of overfitting
- Are generated together with reliable estimators of performance
- Can deal with sub-optimal development sets (e.g., those with confounding factors)
To date, the platform has been used to design more than ten tests in a variety of clinical applications. About Biodesix Biodesix® is a molecular diagnostics company advancing the development of innovative blood tests in oncology to enable precision medicine. Biodesix discovers, develops and commercializes multivariate protein and genomic diagnostic blood tests, including the GeneStrat® and VeriStrat® tests, that deliver results within 72 hours. The company is changing the standard of care by providing physicians with diagnostic tests for better therapeutic guidance, more accurate prognosis and enhanced disease monitoring to improve patient outcomes. At the forefront of precision medicine, Biodesix is developing new blood tests to identify patients who may benefit from immunotherapies. In addition to developing novel diagnostics independently, the company partners with biotechnology and pharmaceutical companies to develop companion diagnostics for use with therapeutic agents.
Topics: cancer immunotherapy, choosing the right immunotherapy, deep learning, Dr. Jeffrey Weber, Press Releases, HD-IL2 therapy, immunotherapy, machine learning, melanoma, overfitting, SITC