We’re excited to share an update to ImpriMed’s Personalized Prediction Profile (PPP) and Drug Response Predictions (DRP) — specifically for CHOP protocol response predictions.
Thanks to your support, we’ve added more real-world data to our machine-learning models, training them with outcome information from 494 unique canine lymphoma patients.
C-index (concordance index): The c-index is a score that tells us how well the model ranks patients by risk. For example, if we predict that Patient A is at higher risk than Patient B and Patient A does indeed experience progression sooner, that’s a correct prediction. The c-index focuses on the order of predicted risk across patients rather than relying on a binary outcome at a specific time point.
While sensitivity and specificity are widely used — and still valuable in many contexts — they’re not always ideal for evaluating outcomes like time to progression.
For example, if a T-cell lymphoma patient lives 61 days without progression and our model labels them as high risk, sensitivity/specificity may treat that as a misclassification. However, if another patient predicted to be low risk lives 74 days, the c-index considers that a correct prediction because the higher-risk patient had a shorter PFS.
By moving to metrics like c-index and hazard ratio, we aim to provide clinically meaningful insights for time-based outcomes like progression-free survival.
Hazard Ratio (HR): HR compares the relative risk of progression over time between different groups.
This gives insight into how much more or less likely an event is to occur in one group compared to another, at any given point during the study period.
We’re committed to continuously improving our models and providing clinically meaningful insights to support your treatment decisions. By working closely with the veterinary oncology community, we aim to make our tools even more helpful and relevant in your day-to-day practice.
Thank you for being part of this journey with us. We look forward to continuing to grow and innovate together.
Interested in learning more? Please email lunch@imprimedicine.com or fill out the form below to book an Oncology Team Lunch & Learn with the ImpriMed Team.