Combination feasibility prediction for checkpoint inhibitors for a biologic

Overview

A European biotech company had a large molecule (biologic) in its development pipeline for cancer indications and sought a combination feasibility prediction study. They were interested in pairing their proprietary molecule with already approved immune checkpoint inhibitors (such as anti-PD-1/PDL-1) to significantly improve therapeutic efficacy and patient outcomes

Our client

Our client is a Biotech company based in Europe, specializing in the oncology sector. Their strategic goal involves advancing biologic drug development by leveraging advanced computational methods. They faced the need for high-confidence data to support a crucial combination strategy, requiring predictive analytical support to guide their high-stakes clinical decisions.

Client’s challenge

The main challenge was to systematically prioritize specific cancer indications most likely to exhibit sensitivity towards the combination of their proprietary biologic with a checkpoint inhibitor. Building this strategy required establishing robust predictive models from heterogeneous public data on successful and failed drug combinations, necessitating advanced statistical and biological interpretation.

Client’s goals

The core objective was to utilize advanced predictive modeling to prioritize cancer indications and patient cohorts based on their predicted sensitivity to the combination therapy. This analysis was intended to:

• Validate the synergistic role of the combination partners.
• Widen the scope of indications where the proprietary drug could be developed.
• Identify patients where monotherapy resistance could be overcome by the combination.

Our approach

Excelra supported the program using its Computational Biology Services and expertise in Scientific Application Development. This work leveraged predictive analytical workflows, guided by our insights on Selecting and Preparing Data for AI/ML Predictive Modeling, to ensure the models were robust and reliable.

The methodology involved several key steps:

Data preparation

Robust predictive models were built by combining curated public datasets using our dedicated Data Curation services, ensuring adherence to FAIR data principles for reuse across downstream analytic pipelines.

Predictive modeling

Machine learning models were built to assess the sensitivity of various cancer indications and patient types to the drug combination. This included cancer cohort analytics and patient stratification.

Hypothesis generation

Custom pathways were generated to understand the precise signaling and crosstalk between the drug-induced signaling and checkpoint inhibitor pathways, establishing a biological hypothesis for synergy.

Our solution

Based on this rigorous analysis, Excelra’s contribution provided immediately actionable insights:

• We performed a high-confidence combination feasibility prediction for the two-drug regimen.
• We prioritized the specific cancer indications where combination therapy with a PD-1 blocker would be most effective.
• We successfully predicted that certain indications resistant or only partially sensitive to monotherapy would become sensitive toward the combination with the checkpoint inhibitor.
• Key biomarker signals were cross-referenced against Excelra’s Custom Biomarker Knowledgebase, and results were presented using our Visualization services, helping the client translate predictive insights into hypothesis-driven follow-up studies and move closer to Precision Medicine goals.