Associations among anesthesia-induced brain biomarkers during surgery, preoperative exposure to cancer therapeutics, and clinical outcomes.

Abstract

12095 Background: Cognitive impairment represents a key concern in oncology, as issues with memory and concentration can be a debilitating side effect of cancer and its treatment. Prior work demonstrates that the power of the alpha wave, a pattern of electroencephalographic (EEG) signals in the brain in patients under general anesthesia, correlates with cognitive function. In this study, we sought to investigate the associations among anesthesia-induced alpha power, preoperative exposure to cancer therapeutics, and clinical outcomes in a large cohort of patients undergoing surgery. Methods: We analyzed intra-operative EEG recordings of a retrospective cohort of patients with and without cancer who underwent surgery with general anesthesia at a single institution between 4/2016 - 12/2019. We used multivariable regression models (adjusted for age, sex, history of dementia) to determine associations of cancer-related covariates with patients’ alpha power (measured in decibels [dB]), and to determine associations of alpha power with clinical outcomes, such as hospital length of stay, discharge location, hospital readmission, and survival. Results: Among 1,084 total patients included in this study, 377 (35%) had a preoperative diagnosis of cancer. The most common cancer diagnoses were breast (87, 23% of those with cancer), gastrointestinal (85, 23%), genitourinary (73, 19%), and lung (47, 12%). We found that a cancer diagnosis alone was not associated with alpha power (-0.015 dB, p = 0.95). However, among patients with cancer, receiving chemotherapy within 3 months before surgery was associated with a 1.6dB drop in alpha power (p < 0.001). By comparison, having a preoperative diagnosis of dementia was associated with a drop of 2.86dB (p < 0.001). Decreased alpha power was associated with longer hospital stays (1.03 times longer per dB drop, p < 0.001), higher rates of discharge to a rehab facility (odds ratio [OR] = 1.22 per dB drop, p < 0.001), and higher mortality (hazard ratio = 1.085 per dB drop, p < 0.001). We found no association between alpha power and 30-day readmission (OR = 1.04, p = 0.24). Conclusions: In this large cohort of patients undergoing surgery, we found that alpha power correlated with several important clinical outcomes, and that recent receipt of chemotherapy was associated with a decrease in alpha power. Of note, the observed drop of 1.6 dB among those with recent chemotherapy is roughly equivalent to aging 10-15 years, while a diagnosis of dementia was associated with a nearly 3dB drop in alpha power. As evidence suggests that alpha power corresponds with patients’ cognitive function, our findings highlight the need to prospectively monitor and track longitudinal alpha power in patients receiving cancer treatment to determine if early detection of changes in alpha power may predict for future treatment-related cognitive decline.