Trajectories of cognitive performance over five years in a prospective cohort of patients with breast cancer (NEON-BC)

Abstract

Purpose: To identify trajectories of cognitive performance up to five years since diagnosis and their predictors, in a cohort of patients with breast cancer (BCa).

Methods: A total of 464 women with BCa admitted to the Portuguese Institute of Oncology, Porto, during 2012, were evaluated with the Montreal Cognitive Assessment (MoCA) before any treatment, and after one, three and five years. Probable cognitive impairment (PCI) at baseline was defined based on normative age- and education-specific reference values. Mclust was used to define MoCA trajectories. Receiver Operating Characteristic curves were used to assess the predictive accuracy for cognitive trajectories.

Results: Two trajectories were identified, one with higher scores and increasing overtime, and the other, including 25.9% of the participants, showing a continuous decline. To further characterize each trajectory, participants were also classified as scoring above or below the median baseline MoCA scores. This resulted in four groups: 1) highest baseline scores, stable overtime (0.0% with PCI); 2) lowest baseline scores (29.5% with PCI); 3) mid-range scores at baseline, increasing overtime (10.5% with PCI); 4) mid-range scores at baseline, decreasing overtime (0.0% with PCI). Adding the change in MoCA during the first year to baseline variables significantly increased the accuracy to predict the downward trajectory (area under the curve [AUC] = 0.732 vs. AUC = 0.841, P < 0.001).

Conclusion: Four groups of patients with BCa with different cognitive performance trends were identified. The assessment of cognitive performance before treatments and after one year allows for the identification of patients more likely to have cognitive decline in the long term.

Keywords: Breast neoplasms; Cognitive trajectory; Longitudinal studies; Neurocognitive disorders.

Fig. 1

Cognitive trajectories since before treatment to five years after diagnosis, represented with the raw score of the Montreal Cognitive Assessment (MoCA) and with its age- and education-adjusted value (aMoCA score). Graphs A and B: the two model-based trajectories, Upward and Downward; Graphs C and D: patterns of cognitive performance in the groups Consistently high - women of the Upward trajectory with a baseline MoCA score > median; Mid-upward - women of the Upward trajectory with a baseline MoCA score ≤ median; Mid-downward - women of the Downward trajectory with a baseline MoCA score > median; Consistently low - women of the Downward trajectory with a baseline MoCA score ≤ median. ∗P < 0.01 and ∗∗P < 0.001, for the change between consecutive evaluations, within each trajectory.

Fig. 2

Association of cognitive performance at baseline and its variation after one year, socio-demographic characteristics of the patients, lifestyle, co-morbidities, clinical characteristics of the tumor, treatments, neurological complications and patient-reported outcomes (PRO) with cognitive trajectories - Downward vs.Upward. CIPN, chemotherapy-induced peripheral neuropathy; PRO, patient-reported outcomes; Tx, treatments (a) Categories of age and education as they are used in the classification for cognitive impairment based on normative data. (b) When menopausal status was not specified, all women with at least 60 years of age, women who underwent a bilateral oophorectomy and those with an intact uterus and being amenorrheic for 12 or more consecutive months prior to the diagnosis in the absence of alternative pathological or physiological cause and follicle stimulating hormone and serum estradiol levels within the laboratory's reference ranges were classified as postmenopausal, or otherwise as premenopausal. (c) According to drug classification of the WHO Collaborating Centre for Drug Statistics Methodology (https://www.whocc.no/atc_ddd_index). (d) One patient only performed axillary surgery. (e) Patients who had both lymph node dissection and sentinel lymph node biopsy are reported as lymph node dissection. https://www.whocc.no/atc_ddd_index (f) Depression and anxiety were defined as presenting the respective sub-score equal to or higher than 11 in the Hospital Anxiety and Depression Scale. (g) Poor quality of sleep was defined as presenting a total score equal to or higher than five in the Pittsburg Sleep Quality Index. (h) Adjusted for age. (i) Adjusted for age, education. (j) Adjusted for age, education and baseline MoCA score. (k) Adjusted for age, education and cancer stage.

Fig. 3

Receiver operating characteristic curves of predictive models of the downward trajectory in women with breast cancer. AUC, Area Under the Curve; MoCA, Montreal Cognitive assessment; PCI, Probable cognitive impairment at the baseline evaluation defined as scoring below two standard deviations of the age- and education-specific distribution from normative data; ROC, Receiver Operating Characteristic. Age in years, education in four categories (≤4, 5–9, 10–12, >12 years); Baseline predictors: age, education and baseline MoCA score. AUC(model with age)≠AUC(model with education), P = 0.378. AUC(model with age)≠AUC(model with PCI), P = 0.319. AUC(model with education)≠AUC(model with baseline MoCA score), P < 0.001. AUC (model with baseline MoCA score)≠AUC(model with baseline predictors), P = 0.295. AUC (model with baseline predictors)≠AUC(model with baseline predictors + change in MoCA score during the first year), P < 0.001. AUC(model with MoCA score at the one-year evaluation)≠AUC(model with baseline predictors + change in MoCA score during the first year), P = 0.102.

Fig. 4

Distribution of the probabilities of belonging to the downward trajectory estimated by the model based on the baseline predictors age, education (≤4, 5–9, 10–12, >12 years) and Montreal Cognitive Assessment (MoCA) score, and by the same model plus the variation in the MoCA score during the first year (score at the one-year evaluation - baseline score).