Trajectories of Cognitive Complaints in Patients With Breast Cancer and Their Association With Psychosocial and Neurobiological Factors

Abstract

Introduction: Cancer-related cognitive impairment (CRCI) is common among patients with breast cancer, with significant variability in both severity and duration of cognitive complaints. This study aimed to elucidate this variability by identifying distinct trajectories of cognitive complaints over time in patients with breast cancer. Additionally, we explored the relationships between these trajectories and various clinical, demographic, psychosocial, neuropsychological, neuroimaging (resting-state functional MRI, rs-fMRI), and serum markers.

Methods: In this prospective study, 67 patients with non-metastatic breast cancer underwent psychosocial questionnaires, neuropsychological testing, rs-fMRI, and serum markers at diagnosis (T0), 8 months after diagnosis (T1), and 16 months after diagnosis (T2). Using the partition around medoids (PAM) algorithm on the difference scores of cognitive complaints, patients were clustered into distinct groups. Differences between the groups were assessed using linear mixed effects models. Rs-fMRI was analyzed using whole-brain graph theory and connectivity in four cognition-related networks.

Results: Four different trajectories of cognitive complaints were identified: stable with no changes in complaints (n = 24), improving with a decrease in complaints at T1 (n = 15), a short-term affected with an increase in complaints at T1 and recovery at T2 (n = 13), and a long-term affected with an increase in complaints at T1 and T2 (n = 15). While the groups strongly correlated with changes in the psychosocial measures, only subtle associations were found with neuropsychological tests, serum markers, or rs-fMRI analyses.

Conclusions: This research affirmed the presence of distinct trajectories of cognitive complaints in patients with breast cancer, which were associated with differences in anxiety, fatigue, stress, and depression.

Keywords: CRCI; Rs‐fMRI breast cancer; clustering; cognitive complaints; longitudinal trajectories; psychosocial.

FIGURE 1

Flowchart of study participant inclusion. Non‐inclusions refer to recruited participants that did not start the study. Dropouts and non‐inclusions occurred due to withdrawal, Covid‐19 restrictions, discomfort in the scanner, and age‐matching of patients with healthy controls. The exclusions within this study were due to changes in treatment and/or discovery of neurological conditions during scanning (very high number of white matter lesions). Some patients had missing modalities at one or two time‐points due to time constraints, fear of the MRI, or forgetting to complete a questionnaire. BDI, Beck's depression inventory; NfL, neurofilament light chain; PSS, perceived stress scale; rs‐fMRI, resting‐state functional MRI; TMT, trail making test.

FIGURE 2

Changes in cognitive complaints between the 3 time‐points for each solution of the PAM algorithm. Individual and mean lines are shown in each plot. The 4‐cluster solution was selected as the optimal clustering. HCs are not shown in the figure. (A) The 2‐cluster solution. (B) The 3‐cluster solution. (C) The 4‐cluster solution. (D) 5‐cluster solution. CFQ, cognitive failure questionnaire; HC, healthy control.

FIGURE 3

Trajectories of self‐reported measures between the 3 time‐points for each group. Individual and mean lines are shown in each plot. (A) Trajectories of depression, measured by BDI. (B) Trajectories of fatigue, measured by FAS. (C) Trajectories of stress, measured by PSS. (D) Trajectories of anxiety, measured by STAI. BDI, Beck's depression inventory; FAS; fatigue assessment scale; PSS, perceived stress scale; STAI, state–trait anxiety inventory.

FIGURE 4

Heat maps of the change in functional connectivity patterns between T0 and T1 for each group. The change in correlation coefficients of each pair of regions is shown. Connections that were not significant after FDR correction are made transparent. All groups, except the improving group, showed significant changes in functional connectivity between T0 and T1, but not between T1 and T2. A list of regions and their abbreviations is provided in Table S2. DMN, default mode network; DA, dorsal attention network; FDR, false discovery rate; FP, frontoparietal network; SN, salience network.