Speaker
Description
Computed tomography scans remain a common modality for measuring tumor burden in cancer patients. However, diagnostic imaging can be burdensome or invasive for patients due to enclosed environments and clinical discomfort. Tumor volume data is also collected infrequently. Patient-Reported Outcomes (PROs) offer a less intrusive avenue for gauging cancer progression through measurements regarding the quality of life and symptomology. Past studies have shown that certain PROs, such as insomnia, are highly correlated with changes in tumor volume in non-small cell lung cancer (NSCLC).
We analyzed PROs and tumor volumes collected from 80 NSCLC patients undergoing immunotherapy to determine how PRO dynamics could inform when volumetric treatment progression would occur. We calibrated the tumor growth inhibition (TGI) model to patient-specific tumor volume dynamics and evaluated its predictive performance. We then analyzed changes in PROs to assess how they correlated with model parameters and could be used to improve predictive ability.
The TGI model can accurately describe patient-specific volume dynamics. Predictions using only the TGI model yielded a true positive rate of 62.5% and an overall accuracy of 64.7%. We found that incorporating changes in insomnia into the model increased the true positive rate to 72.2%, with an accuracy of 71.3%. Additionally, using this innovative framework, we can predict precisely when progression occurs, which may aid clinical decision-making.
