Predicting Patient Reported Outcomes of Cognitive Function Using Connectome-Based Predictive Modeling in Breast...

Area of Science:

• Neuroscience
• Oncology
• Medical Imaging

Background:

• Cancer-related cognitive impairment (CRCI) affects patient care and incurs significant costs.
• Predicting CRCI is crucial for timely intervention and improved patient outcomes.
• Previous research has not fully explored predictive neuroimaging markers for CRCI.

Purpose of the Study:

• To determine if baseline resting-state functional magnetic resonance imaging (fMRI) can predict patient-reported CRCI in breast cancer patients.
• To assess cognitive function changes over time in breast cancer patients undergoing chemotherapy versus those not receiving chemotherapy.
• To identify neuroimaging predictors of executive dysfunction and memory decline.

Main Methods:

• Recruited 76 newly diagnosed breast cancer patients and 50 healthy controls.
• Collected self-reported cognitive function (executive dysfunction, memory) and psychological distress data at three time points.
• Acquired resting-state fMRI data, converted to connectome matrices, and applied connectome-based predictive modeling (CPM) and support vector regression (SVR).

Main Results:

• Chemotherapy patients showed increased executive dysfunction over time; memory function declined in both patient groups compared to controls.
• CPM models successfully predicted executive dysfunction and memory function scores (r > 0.31, p < 0.002).
• SVR RBF demonstrated high performance in predicting executive dysfunction (r=0.68) and memory function (r=0.44).

Conclusions:

• Baseline resting-state fMRI shows potential for predicting patient-reported cognitive outcomes in breast cancer survivors.
• Neuroimaging markers may help identify patients requiring surveillance or early intervention for treatment-related cognitive effects.
• This approach could personalize care and mitigate the impact of CRCI.