Cognitive flexibility, memory, and attention in colorectal cancer: findings from a comprehensive neuropsychological assessment.

Introduction
Colorectal cancer is among the most prevalent malignancies worldwide, and its management often requires surgical intervention plus adjuvant chemotherapy [1]. Postoperative cognitive dysfunction is particularly concerning in older patients, who already face age-related cognitive challenges [2, 3].
Chemotherapy, while critical for cancer treatment, is associated with a range of side effects that can significantly affect patients’ quality of life. Among these, neurocognitive dysfunction has emerged as a growing concern. Characterized by memory, attention, and executive function impairments, such cognitive changes can adversely impact daily functioning and long-term recovery [4].
Cancer and its treatments have increasingly been linked to neurocognitive impairments, with contributing factors including the malignancy itself, aging, genetic predispositions, psychological stress, and neurotoxic effects of chemotherapy and radiotherapy [5, 6]. These impairments, often referred to as “chemobrain” commonly affect memory, attention, executive function, and overall quality of life [7, 8]. Although studies suggest that more than 50% of cancer survivors may experience some degree of cognitive decline, true prevalence remains uncertain due to heterogeneous methodologies and assessment criteria [5, 9, 10]. In colorectal cancer specifically, recent reviews report a broad range of post-surgical cognitive decline from 0 to 64%, reflecting variability in evaluation tools [3]. Furthermore, the long-term functional impact of these deficits is still inadequately understood.
More studies, such as Linder et al. [7], noted memory and attention issues in chemotherapy patients, with some recovery over time. Muzzatti et al. [11] found significant impacts on verbal fluency and memory in long-term survivors. Cruzado et al. conducted a longitudinal study on colon cancer patients receiving FOLFOX4 chemotherapy, finding that verbal memory decline was a significant cognitive impairment, especially in older patients and those with lower education. Assessments before, after, and six months post-treatment showed mostly mild and transient cognitive changes, measured using tests like the Trail Making Test [12]. Iranzo et al. studied breast and colon cancer patients and found that age ≥ 55, lower education, and comorbidities significantly increased the risk of cognitive decline with some partial recovery over time, especially after chemotherapy [13]. However, Morin et al. [14], found no significant correlation between chemotherapy and long-term neurocognitive scores.
Given the heterogeneity in CRCI prevalence and methodology across studies, country-specific investigations are essential for enabling future meta-analyses that can more accurately define the scope and clinical significance of cognitive dysfunction in cancer survivors. Our study is the first in Türkiye to comprehensively evaluate cognitive function in CRC patients using a broader range of neuropsychological assessments than most previous research. By comparing post-chemotherapy patients with healthy controls, we aim to contribute valuable data and promote awareness of cognitive monitoring in cancer care.

Material method
This prospective study included 58 patients diagnosed with colorectal cancer, all of whom were evaluated in a multidisciplinary tumor board and staged according to the 8th edition of the American Joint Committee on Cancer (AJCC) TNM Staging System for Colon Cancer. Stage 2 or 3 patients underwent appropriate surgical intervention and then adjuvant treatment. Stage 4 patients with nonsurgical candidates underwent palliative therapy with respect to their ras, raf, and MSI status [15].
Inclusion criteria were as follows: The ability to read and write, supported by a basic educational background. Completion of either adjuvant-modified FOLFOX chemotherapy for 6 months or palliative chemotherapy consisting of FOLFOX combined with bevacizumab or cetuximab. Cognitive assessments were conducted within four weeks post-chemotherapy to capture the acute phase of impairment, during which subjective complaints are most prevalent and have been shown to correlate with reduced quality of life [16].
Patients with cognitive-affecting comorbidities were excluded from the study. Additional exclusion criteria included pre-existing cognitive impairments or documented memory deficits prior to treatment, brain metastases, and any visual or auditory impairments that could disrupt test performance. Participants who declined to provide informed consent were also excluded. Furthermore, individuals using psychotropic medications (such as SSRIs, antipsychotics, and benzodiazepines) were excluded to prevent potential confounding effects on cognitive assessments, given that these medications could influence neuropsychological test results.
The control group comprised 32 individuals matched for age and educational level with the cancer patients. These controls had no history of cancer or other medical conditions that could affect cognitive function.
Cognitive assessments were carefully selected in consultation with a psychiatrist and a psychologist, ensuring their suitability for the study population. All tests were conducted under their supervision to maintain methodological rigor and ensure the validity of the results.

Cognitive and neuropsychological assessments in study groups
Sensory Perception Test (Duyumsama):
Assesses an individual’s ability to process and respond to sensory stimuli, contributing to understanding perceptual processing differences between groups.
Total Learning Score (Toplam Öğrenme Puanı):

Measures the cumulative information retained over multiple trials, evaluating learning capacity and efficiency.

Long-term Memory Score (Uzun Süreli Belek Skoru):

Examines the ability to recall information after a delay, reflecting the integrity of long-term memory storage and retrieval.

Stroop Test:
This test measures cognitive flexibility, selective attention, and processing speed. It includes five conditions:

Stroop 1: Time to read color words.

Stroop 2: Time to name the color of rectangles.

Stroop 3: Time to name the color of words printed in incongruent colors.

Stroop 4: Time to switch between reading color words and naming colors alternately.

Stroop 5: Error rates and corrections for Stroop 4 and 5 are also recorded to assess inhibitory control and error monitoring.

Trail-Making Test (İz Sürme):
Evaluates visual attention and task-switching capabilities.
Trail A (İz Sürme A): Participants connect numbered circles sequentially.
Trail B (İz Sürme B): Involves alternating between numbers and letters, requiring higher cognitive flexibility.
Completion time, error frequency, and corrections are recorded for both tasks.

Statistical analysis
All statistical analyses were conducted using IBM SPSS Statistics for Windows, version 24.0 (IBM Corp., Armonk, NY, USA). The normality of continuous variables was assessed using the Shapiro-Wilk test. As the data did not meet the assumptions of normality, non-parametric tests were applied throughout the analysis. Descriptive statistics were presented as medians and ranges for continuous variables and as frequencies and percentages for categorical variables. Between-group comparisons of demographic and clinical characteristics were performed using the Chi-square test or Fisher’s exact test for categorical variables and the Mann-Whitney U test for continuous variables. A p-value < 0.05 was considered statistically significant.

Results
The median age of the cancer group was 56.5 years (range: 29–70), while the median age for the control group was 52 years (range: 28–66). There was no significant difference regarding age between the two groups (p = 0.08).
The gender distribution was similar in both groups, with 58.6% females and 41.4% males in the cancer group, compared to 46.9% females and 53.1% males in the control group (p = 0.2). Regarding civil status, 86.2% of cancer patients and 81.3% of controls were married, which also showed no significant difference (p = 0.3). Regarding reading habits, 34.5% of cancer patients and 43.8% of controls reported reading books or newspapers, with no significant difference observed (p = 0.3). Family support was reported by 87.9% of cancer patients and 93.8% of controls, with no significant difference (p = 0.4).
Table 1 summarizes the patient’s characteristics across groups.

Regarding education levels, 74.1% of the cancer group had primary or secondary education, while 19.0% held a high school diploma, and 6.9% earned a college degree. In comparison, the control group included 56.2% with primary or secondary education, 31.3% with a high school diploma, and 12.5% with a college degree, indicating no significant difference (p = 0.1). Despite the lack of significance between groups, the cancer group’s Kruskal-Wallis analysis revealed a significant difference across education levels for Total Learning Score (χ² = 8.953, p = 0.011), Stroop Test 1 st (χ² = 8.153, p = 0.017), 3rd (χ² = 9.332, p = 0.009), 4th (χ² = 12.354, p = 0.002), and 5th seconds (χ² = 10.136, p = 0.006). No significant difference was found for Sensory Perception (p = 0.080) or Long-term Memory Score (p = 0.134). Conversely, in the control group, a significant difference across education levels was noted for Stroop Test 3rd (χ² = 11.669, p = 0.003), 4th (χ² = 14.088, p = 0.001), and 5th seconds (χ² = 8.987, p = 0.011). No significant differences were identified for Sensory Perception (p = 0.433), Total Learning Score (p = 0.099), Long-term Memory Score (p = 0.794), or Stroop 1 st and 2nd seconds (p = 0.061 and p = 0.187, respectively).
The Total Learning Score was significantly lower in the cancer group (mean rank = 40.23) compared to the control group (mean rank = 55.05). This finding was supported by a Mann-Whitney U value of 1233.500 and a p-value of 0.01, indicating that cancer patients performed worse overall.
A Mann-Whitney U test compared the cognitive tests in cancer and control groups (Table 2, Figs. 1 and 2).
The sensory Perception test did not differ between groups (p > 0.05).
The Total Learning Score was significantly lower in the cancer group (mean rank = 40.23) compared to the control group (mean rank = 55.05). This finding was supported by a Mann-Whitney U value of 1233.500 and a p-value of 0.01, indicating that cancer patients performed worse overall.
However, there was no significant difference between the groups in terms of long-term memory performance. The mean rank for the cancer group was 43.92, while the control group’s mean rank was 48.36, with a U value of 1019.500 and a p-value of 0.4.
In the Trail-Making Test A, cancer patients (mean rank = 51.37) also performed significantly worse than the controls (mean rank = 36.86), with a U value of 587.500 and a p-value of 0.004. Similarly, in the Trail-Making Test B, the cancer group (mean rank = 51.43) exhibited poorer performance than the control group (mean rank = 34.75), with a significant difference noted (U = 584.000, p = 0.004).
The Stroop Test results also revealed significant differences. In the 1 st second of the test, the cancer group (mean rank = 51.88) had worse performance than the control group (mean rank = 33.94), with U = 558.000 and p = 0.002. This trend continued in the 2nd second (cancer: mean rank = 52.34, control: mean rank = 33.11, U = 531.500, p = 0.001), the 3rd second (cancer: mean rank = 51.65, control: mean rank = 34.36, U = 571.500, p = 0.003), and the 4th second (cancer: mean rank = 50.28, control: mean rank = 36.83, U = 650.500, p = 0.02). Additionally, the cancer group performed significantly worse in the 5th second (mean rank = 50.59) than the control group (mean rank = 36.28), with U = 633.000 and p = 0.01.
The control group was also compared with the subgroup of cancer patients, which revealed no significant difference in Total Learning Score (U = 478.000, \(p = 0.6\)), Long-Term Memory Score (U = 32.62, \(p = 0.3\)), and Stroop tests across all time intervals (\(p > 0.05\)). However, a trend indicating slower performance in cancer patients was noted in the Trail-Making Test A (U = 321.500, Z = -1.87, \(p = 0.06\)) and in the Stroop test during the 5th second (\(p = 0.06\)). A significant difference was found in the Trail-Making Test B (U = 291.000, Z = -2.32, \(p = 0.02\)), suggesting a potential impairment in executive function among cancer patients at these stages (Table 3).

Compared to control group the cancer patients with stage 4 exhibited notably lower Total Learning Scores (U = 411.000, Z = 3.39, \(p = 0.001\)) and demonstrated slower completion times in both Trail-Making Test A (U = 606.500, Z = -3.88, \(p < 0.001\)) and Trail-Making Test B (U = 62.500, Z = -4.23, \(p < 0.001\)). Results from the Stroop test indicated significant delays in response times across all measured time intervals: the 1 st second (U = 120.000, Z = -2.99, \(p = 0.003\)), the 3rd second (U = 93.500, Z = -3.57, \(p < 0.001\)), and the 5th second (U = 118.000, Z = -3.02, \(p = 0.003\)) (Table 4).

Discussion
Cognitive impairment in cancer patients can present as subacute, chronic, or progressive, influenced by both cancer itself and its treatments. Despite this, cognitive function is not routinely assessed in clinical practice due to a lack of awareness, time constraints, and the misconception that cognitive issues stem solely from cancer-related stress or yield normal results on objective tests [15, 17]. However, preclinical studies have demonstrated that various cancer treatments induce glial and neuronal cell damage, leading to structural brain changes and cognitive deficits [18]. Additionally, neuroimaging studies have revealed widespread structural and functional brain alterations associated with cancer treatments [19, 20], while biomarker studies link chemotherapy to accelerated cellular and brain aging [21–23]. While many studies have been conducted on breast cancer survivors, cognitive impairment might be called a disorder that significantly reduces the patient’s quality of life and well-being in every cancer patient [24].
Lange et al. [25] have comprehensively reviewed CRCI, identifying memory, attention, executive function, and processing speed as the most commonly affected cognitive domains in patients treated with chemotherapy and other systemic therapies. Our results, particularly the poorer performance observed in the Trail-Making Test B and the Stroop Test among CRC patients, align with these findings and add to the evidence that cognitive flexibility, task-switching, and inhibitory control are frequently compromised following chemotherapy. But it is worth mentioning that despite no statistically significant difference in educational background between groups, Kruskal-Wallis analyses revealed that test performance, particularly in tasks involving executive function, varied across education levels. This suggests that residual confounding due to unequal distribution of educational attainment may have influenced some of the cognitive outcomes.
In terms of prevalence, Whitaker et al. [26] systematically reviewed 52 studies, reporting that approximately one in three breast cancer survivors may experience clinically significant cognitive impairment. They highlighted variability in prevalence estimates depending on the methods used: self-reported rates were as high as 44%, while objective assessments using neuropsychological batteries showed impairment rates ranging from 21 to 34%. Although our study focused on CRC rather than breast cancer, Whitaker’s findings underscore the challenge of accurately quantifying cognitive dysfunction across different cancer populations. Our results, based on objective testing, fall within this prevalence range, particularly regarding impairments in attention and executive functioning.
Yang et al. [27] conducted a study specifically examining cognitive function in CRC patients at different stages of their treatment course. They reported that attention and processing speed were significantly worse in patients who had completed chemotherapy within the previous two years, compared to those newly diagnosed. However, they found no significant differences in memory performance or overall cognitive function between groups. Our findings are consistent with these observations, as we also identified significant deficits in attention and processing speed in our cohort, particularly in more advanced-stage patients. While Yang’s study suggests that these deficits may improve over time, our cross-sectional design captured data within four weeks post-treatment, potentially reflecting a more acute phase of cognitive impairment.
Conversely, Hwang et al. [28] performed a meta-analysis of 11 longitudinal prospective studies involving CRC and rectal cancer patients treated with chemotherapy or chemoradiotherapy. They found no significant pooled effect of chemotherapy on cognitive function, either in subjective reports or objective neuropsychological testing (SMD for objective assessments, 0.000; 95% CI, − 0.093 to 0.093). However, their analysis also reported moderate to high heterogeneity across studies (I² = 60%), which complicates interpretation. Several factors may account for the differences between our findings and theirs, including variations in study design, patient populations, timing of assessments, and cognitive measures employed. Notably, our study excluded patients with pre-existing cognitive impairments and those with comorbidities known to affect cognition, potentially yielding a more homogeneous sample in which chemotherapy-related effects were more readily detectable.
In a recent prospective cohort study, Dwek et al. [29] examined the effects of chemotherapy on cognitive performance in colorectal cancer patients following surgery. Their findings showed that although cognitive impairment was common in both groups, those who received chemotherapy and those who had surgery alone, there was no significant difference in overall cognitive outcomes between them. However, they noted that patients who underwent chemotherapy experienced a slower cognitive recovery over time compared to those who had surgery only. In contrast, our study identified more pronounced and persistent deficits in attention, processing speed, and executive function in patients treated with chemotherapy, particularly in those with advanced-stage disease.
One of the key issues highlighted both in Lange et al.‘s review and in Whitaker’s systematic analysis is the frequent discordance between subjective cognitive complaints and objective test results. In many studies, patients report cognitive difficulties despite performing within normal limits on standardized neuropsychological assessments. Our study did not assess subjective cognitive complaints, which limits our ability to comment on this relationship. It is also important to note that the heterogeneity in cognitive assessment methods across studies, as noted by Lange et al. and Hwang et al., poses a significant challenge to drawing firm conclusions about the prevalence and severity of cognitive impairment. Our study used validated tools such as the Stroop Test and Trail-Making Test, which are recommended by the International Cancer and Cognition Task Force (ICCTF) [30] for evaluating attention and executive function. Nonetheless, we did not include other commonly recommended tests, such as the Hopkins Verbal Learning Test or Controlled Oral Word Association Test, which limits comparability with some other studies.
Recent longitudinal data by Vardy et al. [31] demonstrated that although cognitive impairment is common in colorectal cancer patients shortly after diagnosis and treatment, no significant long-term differences were found between survivors and healthy controls 6–12 years later. The discrepancy with our result might be explained by a major limitation, which was the challenge in recruiting education- and age-matched controls. Low literacy levels among patients limited participation, while the length and complexity of cognitive tests led many healthy volunteers to withdraw prematurely. These factors may have contributed to minor group differences despite statistical non-significance.

Our research contributed also to the current literature by stratifying cognitive outcomes according to disease stage. We found that patients with stage IV disease exhibited the most pronounced deficits, particularly in executive function and processing speed.

Future research should adopt longitudinal designs with pre- and post-treatment assessments, incorporate patient-reported outcomes, and explore potential biomarkers and neuroimaging correlates. Interventions such as cognitive rehabilitation, physical exercise, and pharmacologic strategies also warrant further study, as they may help mitigate cognitive deficits and improve the quality of life for cancer survivors.

Conclusion

Cancer patients showed reduced cognitive function, particularly in learning, attention, and task-switching abilities. Cancer patients with stage 4 colorectal cancer demonstrated more pronounced cognitive decline, suggesting that advanced disease stages may worsen cognitive impairment.

Our results highlighted the importance of routine cognitive screening in CRC patients, particularly those receiving chemotherapy. This is especially crucial for advanced-stage patients, who may benefit from cognitive support strategies in addition to their medical care.