Longitudinal patterns of patient-reported cognitive impairment in patients with non-Hodgkin lymphoma (NHL) treated with chemotherapy: A group-based trajectory analysis of a nationwide community-based cohort.

Abstract

7071 Background: Cognitive problems are an important clinical concern for cancer patients. We previously published that patients with lymphoma experienced worsening in cognitive functions and perceived function from pre-to post-treatment. Limited data exists in NHL, importantly previously reported group means of cognitive function may not capture interindividual heterogeneity. Therefore, we identified distinct subgroups of NHL patients with similar trajectories of cognitive function over time and investigated potential predictors. Methods: We included 181 NHL patients receiving chemotherapy (mean age [SD]: 59.0 [11.8]; 63.5% male; 54% diffuse large B-cell, 22% follicular lymphoma) enrolled in the National Cancer Institute Community Oncology Research Program. Cognitive function was assessed by the patient-reported Functional Assessment of Cancer-Therapy-Cognitive Function (total score and perceived cognitive impairment [PCI]) prior to chemotherapy (T1), after chemotherapy (T2), and 6 months following T2 (T3). Patient and cancer characteristics were collected through questionnaires or medical records at T1. We utilized group-based trajectory modeling (GBTM) to identify latent clusters of individuals with similar cognitive function patterns over time. Model selection and assessment were based on the Bayes factor, group size, model diagnostics, and clinical interpretability. Multinomial logistic regression models were identified characteristics associated with group membership. Results: Four groups were identified by GBTM: Group A had high perceived impairment, which decreased at T2 and remained high at T3 (n=17, 9.4%); Group B had moderate perceived impairment, with a continuous decline at T2 and T3 (n=21, 11.6%); Group C had low perceived impairment and remained stable at T2 and T3 (n=95, 52.5%); Group D had moderate perceived impairment, without changes at T2 and T3 (n=48, 26.5%). Similar trajectory groups were identified for the PCI subscale. Compared to ≥ some college education, ≤ high school was associated with an increased odds of membership in Group D relative to Group C (OR 3.45, 95% CI 1.44, 8.30). Patients with a higher fatigue score (MFSI) were more likely to report worse cognitive impairment (Group A vs. C: OR 1.10, 95% CI 1.07, 1.14; Group B vs. C: OR 1.05, 95% CI 1.02, 1.08; Group D vs. C: 1.05, 95% CI 1.03, 1.07). Stage, lymphoma subtype, and treatment regimen were not associated with group membership. Conclusions: Our analysis revealed that patients with perceived impairment experienced stable impairment or further decline over time, which is associated with education and fatigue; a subgroup of patients did not experience cognitive worsening. These data help identify patients with different patterns of cognitive impairment to inform personalized interventions.