Immunotherapy—a type of treatment that harnesses a patient’s own immune system to fight cancer—has been a “game changer” for many patients. It can be difficult, however, to predict whether or not a given person would benefit from immunotherapy, especially when it comes to drugs that target the immune system molecules PD-1 and PD-L1. Our Curious Dr. George asks Scott Tomlins, MD, PhD, Co-founder and Chief Medical Officer of Strata Oncology, about his company’s new tool to help ensure that anyone who could benefit from these drugs is identified.
Curious Dr. George: One predictor of response to immunotherapy, tumor mutation burden (TMB), has been very helpful for patients whose tumors have a high TMB score. But the tumors of many other patients may also respond to immunotherapy, providing substantial benefits in both progression-free and overall survival. What is Strata Oncology’s Immunotherapy Response ScoreTM (IRSTM), how effective is it, and how can patients with solid tumors access it?
Scott Tomlins, MD, PhD: Current biomarkers for immunotherapy are not enough. Immune checkpoint inhibition with monoclonal antibodies targeting PD-1 and PD-L1 has revolutionized the care of patients with advanced cancer. TMB is an important pan-tumor biomarker for anti-PD-1/PD-L1 therapy, but it does not identify most patients who benefit.
For example, let’s look at the KEYNOTE-158 study of 10 tumor types that led to pan-solid tumor approval by the U.S. Food and Drug Administration (FDA) of second-line pembrolizumab monotherapy in patients who are TMB-high (TMB-H). While a higher objective response rate (ORR) was observed in TMB-H patients versus TMB-low (TMB-L), numerically more objective responses were observed in TMB-L patients (43 out of 688 in TMB-L versus 30 out of 102 in TMB-H). This study demonstrates that TMB alone is not sufficient to identify all responders.
Likewise, in an analysis of TMB in 1,772 patients treated with pembrolizumab monotherapy across 24 tumor types, although ORR was significantly higher in TMB-H versus TMB-L patients, numerically similar numbers of responding patients were again observed (136 out of 433 in TMB-H versus 127 out of 1339 in TMB-L).
Clearly, there is more to the story when it comes to who will benefit from anti-PD-1/PD-L1 therapy. An improved predictive biomarker is needed to ensure that all patients who can benefit from immunotherapy are given the opportunity.
In our recent publication in Communications Medicine, we describe the development, validation, and clinical utility of a new biomarker that fulfills this unmet medical need—the Immunotherapy Response Score or “IRS”.
Immunotherapy Response Score (IRS) is an improved predictive biomarker
IRS combines TMB with four target gene expression measurements (PD-L1, PD-1, TOP2A, and ADAM12) from simultaneous, analytically valid, comprehensive genomic profiling (CGP) plus quantitative transcriptional profiling (qTP) of formalin-fixed paraffin-embedded (FFPE) tumor specimens.
We developed IRS using molecular profiling data combined with treatment and outcome data from the Strata Trial® (NCT03061305), an ongoing observational clinical trial evaluating the impact of molecular profiling for patients with advanced solid tumors.
IRS was trained to predict real-world progression-free survival (rwPFS, by time-to-next-therapy) in a discovery cohort (26 tumor types) of patients treated with pembrolizumab systemic therapy and was then validated in an independent cohort (24 tumor types) of patients treated with non-pembrolizumab anti-PD-1/PD-L1 monotherapy.
In the validation cohort, IRS-high (IRS-H) versus IRS-low (IRS-L) patients had significantly longer anti-PD-1/PD-L1 monotherapy rwPFS (median rwPFS 23.1 months versus 10.2 months, adjusted hazard ratio = 0.52) and overall survival (OS; median OS 40.4 months versus 21.4 months, adjusted hazard ratio = 0.49).
The predictive nature of IRS was confirmed in a case-crossover analysis of 146 patients from the pembrolizumab cohort who had received a previous line of systemic therapy prior to pembrolizumab monotherapy, and by a lack of association with rwPFS in 3,184 patients treated with a non-anti-PD-1/PD-1 or anti-CTLA4 containing first line systemic therapy.
IRS-H/TMB-L patients had similar outcomes as IRS-H/TMB-H patients, demonstrating the clinical utility of IRS beyond TMB alone. Across all Strata Trial patients, IRS-H identified a population nearly twice the size of the TMB-H population (n=24,463; 20.9% versus 10.8%). Critically, 7.6% of patients with tumor types not approved for PD-1/PD-L1 monotherapy were IRS-H/TMB-L, representing a sizable population predicted to have benefit but not currently eligible for treatment.
Bringing new options to people with cancer
We are excited to put this novel biomarker into the hands of physicians to help them ensure every patient gets their best-possible therapy. We are currently providing early access to IRS to a select group of key opinion leaders. IRS will be more broadly available later in 2023.
IRS is just the beginning. By leveraging data from tens of thousands of patients collected under the Strata Trial protocol we are discovering, validating, and demonstrating clinical utility for multiple new, highly quantitative, DNA- and RNA-based multivariate predictive treatment selection biomarker algorithms that span therapeutic modalities.
Dr. Tomlins can be reached at email@example.com.
- Demystifying Cancer Biomarkers: Webinar Recap
- Taking Your Cancer Management to the Next Level: Webinar Recap
- A New Technology for Rapid Therapy Guidance: Webinar Recording
A message from Curious Dr. George:
The goal of Cancer Commons is to help patients identify and access their best possible treatments, one patient at a time, while moving the field forward. If you have advanced cancer, let our molecular oncology Scientists provide personalized information about your options.
Copyright: This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.