In modern oncology, targeted treatments show tremendous promise, but during clinical trials, it is necessary to identify if a subpopulation is responding to the treatment, and after approval, those patients who will benefit from the treatment.
Predict survival for a patient based on a treatment and determine if the treatment increases the survival time.
The performance of oncology treatments is being revolutionized with personalized medicine and our ability to target treatments to the individual. This revolution requires a change in the way survival analysis is conducted. Instead of trying to understand the survival rate for the population or subset of the population, survival can be predicted at the individual level taking into account their genes, demographics and patient history. This new approach is challenging because the necessary data is stored in multiple different locations and at different levels. First, patient information from the Electronic Medical Record (EMR) is needed to understand the patients demographics, diagnosis, pathology reports, and treatment history. To this, the patients genetic information needs to be added from either the EMR or the pharmaceutical company’s tests. All this data needs to be cleaned, joined and transformed into valuable ML features before going into model training. This pre-modeling prep process can be frustrating and time consuming. We are here to help.
Build machine learning models using standard techniques (from linear regression to more advanced machine learning algorithms) to predict survival. From this model, determine if a treatment improves survival.