Kaiku has continued its pioneering research and efforts to find potential ways to improve safety and quality of life of cancer patients treated with immune checkpoint inhibitor (ICI) therapies. ICIs can be associated with severe adverse events which may occur even after treatment discontinuation. Research conducted in collaboration between Kaiku Health and the Oulu University Hospital in Finland indicates that by combining real-world clinical data with patient-reported symptom data, it is possible to predict development of immune-related adverse events from 0 up to 21 days prior the onset of these novel toxicities.
The trained machine learning based prediction model combined data from laboratory measurements of immune checkpoint inhibitor (ICI) treated cancer patients, clinician-assessed adverse events and patient-reported symptom data on the toxicities during treatment. Prediction model had an excellent performance in predicting onset of adverse events when previously unseen data was presented to the model. The performance was evaluated with commonly used machine learning performance metrics.
For the patient, earlier detection of severe adverse events may mean better chances of coping successfully through the treatment, increased safety, better quality of life, and even better treatment results.
Detecting and managing upcoming adverse events early on is an important, but not a straight-forward task. That is why Kaiku has leveraged machine learning, a branch of artificial intelligence, to tackle this challenge. The findings of this study hold the potential to improve toxicity management of cancer patients but more research is needed in the future to further validate the results. Thus, Kaiku is determined to continue its research efforts within this area.
Results of this study are currently being presented in the ESMO (European Society for Medical Oncology) conference 2020. Interested to learn more? Contact us for more information here.