Physical activity patterns track with thinking and memory problems in newly diagnosed lung cancer patients

Within six months of a lung cancer diagnosis, roughly one in four patients shows measurable impairment on an objective test of thinking and memory, and the way their physical activity changes over that period is closely linked to whether those problems appear. That is the central finding of a prospective cohort study of 327 patients newly diagnosed with non-small cell lung cancer (NSCLC) in Hong Kong, published open access in Supportive Care in Cancer. The research, led by Mu-Hsing Ho with collaborators including Janette Vardy and Chia-Chin Lin, is among the first to follow cognitive function monthly from the point of diagnosis in this population, and to model physical activity not as a single snapshot but as a trajectory over time. Lung cancer is the most common cancer worldwide and the leading cause of cancer-related death, and NSCLC accounts for 80 to 85 percent of cases, making the size of the affected population substantial.

Cancer-related cognitive impairment, often abbreviated CRCI, describes the difficulties with short-term memory, attention, executive function, language, orientation, and processing speed that many patients experience as a result of the cancer itself or its treatments. Patients report trouble forming thoughts, finding words, multitasking, organizing, and completing familiar tasks because of mental slowing or reduced concentration. The problem is well documented in breast cancer, where more than half of patients treated with chemotherapy experience some degree of CRCI, but lung cancer has been comparatively understudied. A proportional meta-analysis in lung cancer patients reported a pooled prevalence of 26 percent, and the International Cancer and Cognition Task Force highlights the likely contributions of cancer treatments and underlying pathophysiological mechanisms. Beyond the clinic, CRCI can erode self-confidence, strain social relationships, and make work and daily tasks markedly harder.

The Hong Kong team recruited adults diagnosed with NSCLC within the preceding four weeks from the oncology departments of Queen Mary Hospital and Gleneagles Hospital, supplemented by community recruitment through District Health Centres, Neighbourhood Elderly Centres, and District Elderly Community Centres. Patients with pre-existing cognitive impairment or dementia were excluded. Each participant completed a baseline assessment followed by monthly face-to-face follow-ups for six months, seven measurement points in total, a window chosen because six-month prognosis is commonly adopted in cancer care research for newly diagnosed lung cancer. Objective cognition was tested with the Hong Kong version of the Montreal Cognitive Assessment 5-Minute Protocol, which examines memory, language, orientation, and delayed recall and was scored against age- and education-adjusted norms. Subjective cognition was measured with the 37-item Functional Assessment of Cancer Therapy-Cognitive Function scale, a tool recommended by the Cancer Neuroscience Initiative Working Group. Physical activity was quantified with the International Physical Activity Questionnaire Short Form, converted into MET-minutes per week across walking, moderate, and vigorous activity. Symptom burden was tracked with the MD Anderson Symptom Inventory, and exercise confidence with the Self-Efficacy for Exercise scale.

The cohort of 327 patients was enrolled between October 2023 and August 2025, with 54 patients withdrawing over the study period, an attrition rate of 16.5 percent. The withdrawals included 32 deaths, 14 patients unwilling to continue follow-up, and 8 lost to contact. Men made up 58.1 percent of the sample, the mean age was 63.7 years with a standard deviation of 8.8, and most patients had stage III (26 percent) or stage IV (34.6 percent) disease. Hypertension was present in 31.2 percent and diabetes in 9.5 percent. At baseline, 25.7 percent of patients did not engage in moderate or high levels of physical activity, and the mean MET total was 12.8 hours per week.

The prevalence of objective CRCI, defined as a protocol score at or below the seventh percentile of normative data, fluctuated between 17.5 percent and 29.1 percent across the observation period. At the six-month follow-up, 24.1 percent of participants were affected, with a 95 percent confidence interval of 19.2 to 29.5 percent. The authors note that this figure mirrors previous estimates in lung cancer and broader oncology populations, confirming CRCI as a common survivorship challenge. The temporal variability suggests that cognitive function may rise and fall during the cancer trajectory, possibly influenced by treatment phases, symptom burden, and psychosocial stressors. That fluctuation carries a practical warning: screening at a single time point may either miss patients who later deteriorate or overlook perceived impairment that develops between visits, which is why the authors argue for longitudinal monitoring rather than one-off assessment.

To capture how activity actually changes after diagnosis, the researchers applied group-based trajectory modelling, a statistical technique that assigns individuals to clusters with comparable repeated-measures paths modelled as polynomial functions of time, with model fit evaluated by the Bayesian Information Criterion. Three distinct trajectories emerged. The largest group, 45 percent of patients, started at high activity and declined to moderate levels over the study period. Another 43.1 percent remained consistently low in physical activity throughout, while 11.9 percent maintained consistently high activity. The groups began the study at considerably different baseline activity levels, and by the end of the six months the declining group was engaging in less activity than the consistently low group, whose members gradually increased theirs, though they never reached a moderate level.

After adjustment for demographic and clinical covariates in multivariate logistic regression, both non-low trajectories were associated with substantially lower odds of CRCI at six months. Patients with consistently high physical activity had 93.8 percent lower odds of objective impairment compared with the consistently low group, corresponding to an odds ratio of 0.062 with a 95 percent confidence interval of 0.008 to 0.494. The declining group had 85.9 percent lower odds, with an odds ratio of 0.141 and a confidence interval of 0.062 to 0.324. The consistently high group also outperformed the others on objective subdomains including memory, language, and delayed recall, and on every subjective subdomain of the cognitive scale, from perceived impairments to their impact on quality of life. Notably, the declining group reported the lowest exercise self-efficacy and the highest cancer symptom severity and interference, suggesting that fatigue, pain, breathlessness, and mood symptoms may erode both motivation and capability over time.

Physical activity trajectory was not the only factor associated with cognitive outcomes. Older age, male sex, shorter duration of education, advanced NSCLC stage, and lower baseline scores for perceived cognitive function were all linked to CRCI, and lower self-efficacy for exercise emerged as an additional predictor. The authors interpret the education finding through the lens of cognitive reserve, the idea that greater reserve buffers the impact of neurotoxic injury, while the sex difference warrants further study because evidence on gender and CRCI remains mixed, potentially shaped by hormonal, genetic, or psychosocial factors. The role of exercise confidence underscores the behavioural and psychological components of cognitive health, hinting that building patients’ belief in their ability to be active may indirectly support cognition. The team also observed that subjective and objective measures were poorly correlated, with the consistently low activity group reporting better subjective scores over time than the declining group even where objective differences between groups were not obvious.

The findings demand careful interpretation, and the authors are explicit about the limits. This was an observational study, so the associations should not be read as proof that increasing physical activity independently reduces CRCI. Physical activity trajectories may partly reflect underlying health status, including disease severity, symptom burden, and functional capacity, rather than an independent protective effect. Because activity and cognition were assessed over overlapping periods, reverse causation cannot be excluded. Actual treatment exposure during follow-up was not captured, so treatment-related confounding remains possible. Depression, anxiety, and genetic predisposition were not included as confounders. Mortality-related attrition, with 32 of the 54 withdrawals caused by death, may have introduced survivor selection, potentially underestimating impairment in the baseline cohort. Convenience sampling limits generalisability, and because the cognitive assessment is essentially a screening test for mild impairment, it may underestimate the subtler deficits typical of cancer patients.

Even with those caveats, the study offers a template for clinical practice. The authors argue that routine cognitive screening with brief validated instruments at multiple post-diagnosis time points should be integrated into lung cancer care where comprehensive neuropsychological testing is not feasible, and that efforts to promote physical activity must be tailored to each patient’s symptom severity, functional status, motivation, and psychosocial context. Multidisciplinary collaboration among oncologists, rehabilitation and exercise specialists, psychologists, and nursing staff will be vital for combining exercise prescription, symptom management, and self-efficacy enhancement. Mechanistically, physical activity is hypothesised to attenuate CRCI by enhancing neuroplasticity, promoting cerebral perfusion, reducing systemic inflammation, and mitigating the neurotoxic effects of cancer and its treatments, but randomized controlled trials are needed to establish causality and determine the optimal timing, intensity, and type of activity. For now, the message for the growing population of newly diagnosed lung cancer patients is that cognitive health after diagnosis is dynamic, monitorable, and tightly interwoven with how much patients move.

Subject of Research: Prevalence, predictors, and physical activity trajectories of cancer-related cognitive impairment in newly diagnosed non-small cell lung cancer patients

Article Title: Prevalence, predictive factors, and the influential role of physical activity trajectories of cancer-related cognitive impairment among patients newly diagnosed with non-small cell lung cancer

Article References: Ho, M.-H., Vardy, J. L., Lin, Y.-K., Chan, W. L., Takemura, N., Wang, L., Cheung, D. S. T., & Lin, C.-C. (2026). Prevalence, predictive factors, and the influential role of physical activity trajectories of cancer-related cognitive impairment among patients newly diagnosed with non-small cell lung cancer. Supportive Care in Cancer, 34(11), 1152. https://doi.org/10.1007/s00520-026-11368-5

Image Credits: AI Generated

DOI: 10.1007/s00520-026-11368-5

Keywords: cancer-related cognitive impairment, non-small cell lung cancer, physical activity trajectories, cognitive function, exercise self-efficacy, prospective cohort study, Hong Kong, cancer survivorship, group-based trajectory modelling, symptom burden, Montreal Cognitive Assessment, FACT-Cog