Prognostication in Advanced Cancer by Combining Actigraphy-Derived Rest-Activity and Sleep Parameters with Routine...

Area of Science:

• Oncology
• Palliative Care
• Biomedical Engineering

Background:

• Accurate survival prediction is crucial in oncology and palliative care.
• Existing prognostic models often lack robustness.
• Integrating objective and subjective patient data may improve prognostic accuracy.

Purpose of the Study:

• To assess the feasibility of combining actigraphy, sleep diary, and clinical data for cancer survival prognostication.
• To develop and validate a predictive model for survival in advanced cancer patients.

Main Methods:

• Fifty adult outpatients with advanced cancer and <1 year prognosis were recruited.
• Patients wore a wrist actigraph for 8 days and completed a sleep diary.
• Univariate and regularised multivariate regression (Lasso) were used to identify predictors from 66 variables.

Main Results:

• Forty-nine patients completed the study; 34 died within 1 year.
• A Lasso-derived algorithm successfully differentiated shorter/longer survival (log rank p < 0.0001).
• Predictors of longer survival included sleep efficiency, subjective sleep quality, clinician estimate, global health status, and hemoglobin. Shorter survival was linked to sleep disturbance, neutrophil count, urea, creatinine, and C-reactive protein.

Conclusions:

• Machine learning applied to actigraphy, sleep, and clinical data shows promise for developing novel prognostic tools in oncology.
• Objective sleep and activity data, alongside subjective reports and clinical markers, can contribute to more accurate survival predictions.
• This integrated approach may enhance personalized palliative care and treatment planning for advanced cancer patients.