AUTHORS
Jowon Laura Kim, MD, Bri Negaard2, Melissa Larson, MS3, Annalynn Williams, PHD4, Kathleen Yost5, Mazie Tsang, MD, MAS, MS6, Jonathon B. Cohen, MD, MS7, Andrew L. Feldman, MD8, Christopher R. Flowers Jr, MD, MS9, Chijioke C Nze, MD, MPH10, Jonathan W. Friedberg, MD11, Brad S. Kahl, MD12, Izidore S. S. Lossos, M.D.13, Eric Mou, MD14, Jia Ruan, MD, PhD15, Matthew E. Maurer, ScD16, James R. Cerhan, MD, PhD17, Carrie Thompson, MD18, Thomas M. Habermann, MD19, Urshila Durani, MD, MPH20
AFFILIATIONS
1 Mayo Clinic, Rochester, Minnesota, United States
2 Mayo Clinic, Rochester, Minnesota, United States
3 Mayo Clinic, Rochester, Minnesota, United States
4 University of Rochester Medical Center, Rochester, New York, United States
5 Mayo Clinic, Rochester, Minnesota, United States
6 Mayo Clinic, Phoenix, Arizona, United States
7 Emory University, Atlanta, Georgia, United States
8 Mayo Clinic, Rochester, Minnesota, United States
9 The University of Texas MD Anderson Cancer Center, Houston, Texas, United States
10 The University of Texas MD Anderson Cancer Center, Houston, Texas, United States
11 University of Rochester, Rochester, New York, United States
12 Washington University School of Medicine, St. Louis, Missouri, United States
13 Sylvester Comprehensive Cancer Center, University of Miami, Miami, Florida, United States
14 University of Iowa, Iowa City, Iowa, United States
15 Meyer Cancer Center, Weill Cornell Medicine and New York-Presbyterian Hospital, New York, New York, United States
16 Mayo Clinic, Rochester, Minnesota, United States
17 Mayo Clinic, Rochester, Minnesota, United States
18 Mayo Clinic, Rochester, Minnesota, United States
19 Mayo Clinic, Rochester, Minnesota, United States
20 Mayo Clinic, Rochester, Minnesota, United States
ABSTRACT
Introduction: Cognitive dysfunction has been reported in lymphoma survivors treated with systemic chemotherapy, but the prevalence, longitudinal course, and risk factors for patient-reported cognitive impairment are poorly characterized. We evaluated perceived cognitive impairment and associated risk factors in the prospective Lymphoma Epidemiology of Outcomes (LEO) cohort.
Methods: LEO is a longitudinal prospective cohort of patients with non-Hodgkin lymphoma enrolled within 6 months of diagnosis at 8 academic centers in the US from 2015-2020, with baseline clinical annotation and longitudinal patient-reported outcome assessments. The Patient-Reported Outcomes Measurement Information System (PROMIS) cognitive survey was administered at 3- and 5-year (y) follow-up, and patients with an available cognitive score at either timepoint were eligible for analysis. Raw cognitive scores were converted to standardized T-scores (population mean=50, standard deviation=10). Severe cognitive impairment was defined as T-score <29, moderate 30-39, mild 40-44, and normal ≥45. A clinically meaningful between-group difference in PROMIS cognitive T score (ΔT) was defined as an absolute difference of ≥2 points. Clinically meaningful within-person change between 3y and 5y was defined as ≥5 points.
Baseline characteristics (at enrollment) assessed for association with cognitive scores included age, gender, race/ethnicity, education, disease stage, lymphoma subtype, B symptoms, bulky disease, extranodal disease, treatment, opioid use, ECOG performance status (PS), International Prognostic Index (IPI), Lymphoma Co-morbidity Index (LCI), ENRICHD Social Support Inventory (ESSI), history of depression or anxiety, and self-reported memory problems. The following variables were assessed at both 3y and 5y and evaluated for association with cognitive scores at the corresponding timepoints: marital status, employment status, ESSI, physical activity (GODIN survey), ECOG PS, history of depression or anxiety. Variables meeting both clinical (ΔT score ≥2) and statistical significance (p<0.05) in univariable analyses were entered into separate multivariable linear regression models for the 3y and 5y cognitive assessments, to identify independent predictors of cognitive function.
Results: Among 7718 patients enrolled, 2117 had an available 3y (n=1319) and/or 5y (n=1362) PROMIS cognitive score. Median age at diagnosis was 63y (range 18-95), 45% were female, and lymphoma subtypes included large B cell (35.5%), follicular (25%), mantle cell (11%), marginal zone (10.5%), other B-cell (10%), T-cell (6%), and composite (2.5%). Among 305 patients with CNS lymphoma, only 69 patients (23%) had a 3/5y PROMIS cognitive score; therefore, subgroup analysis by CNS involvement was not conducted.
At 3y, median cognitive function T score was 52 (range 25-61); 77% normal cognition, 13% mild, 8% moderate, and 2% severe impairment. At 5y, median cognitive function T score was 55 (range 25-61); 80% normal cognition, 12% mild, 6% moderate, and 1% severe impairment. Among 564 patients with both 3/5y cognitive scores, most (57%) had no clinically meaningful change, while 21% improved, and 22% worsened.
In multivariable analysis, disability related unemployment at 3y (vs full-time employment, ΔT -2.8, p=0.025), worse baseline ECOG (vs ECOG 0, associated with lower T scores at both 3y [ΔT range -1 to -5.3 for ECOG 1-4] and 5y [ΔT range -1.3 to -3.6 for ECOG 1-4], all p<0.001), worse ECOG PS at 3/5y (ΔT range -4.6 to -13.0 for ECOG 1-4 at 3y, ΔT range -4.4 to -8.5 for ECOG 1-4 at 5y, all p<0.001), lower ESSI scores at 3/5y (p<0.001); depression at 3/5y (vs no depression, ΔT at 3y -3.0, ΔT -2.1 at 5y, all p<0.001), anxiety at 3/5y (vs no anxiety, ΔT at 3y -2.3, ΔT at 5y -2.8, all p<0.001), and baseline memory concerns (vs no memory concerns, associated with lower T scores at both 3y [ΔT -8.2] and 5y [ΔT -11.7], all p<0.001) were independently associated with worse cognitive function.
Conclusion: Approximately one in five lymphoma survivors reported persistent perceived cognitive impairment 3-5 years after diagnosis. Functional status, social support, depression, anxiety, and baseline memory concerns were more strongly associated with long-term cognitive impairment than lymphoma-related disease characteristics or treatment class, suggesting survivorship interventions should prioritize these potentially modifiable risk factors.
DISCLOSURES
Jowon Laura Kim, MD: No relevant conflicts of interest to declare.