Relation between perceived cognitive function and neuropsychological performance in survivors of breast and colorectal cancer

Reviewer: Christine Hill, MD
Abramson Cancer Center of the University of Pennsylvania
Last Modified: June 3, 2008

Presenter: J.L. Vardy
Presenter's Affiliation: Sydney Cancer Centre, Sydney, Australia
Type of Session: Scientific


  • “Chemobrain” or “chemofog” was originally described by the female breast cancer population, and refers to symptoms of decreased cognitive acuity described by patients. Often, these symptoms include problems with linear thought and multi-tasking.
    • 50 – 70% of breast cancer patients self-report cognitive impairment after treatment.
    • Neuropsychologic impairment on formal testing is identified in 15-50% (Weineke, 1995).
  • Several retrospective studies have found that chemotherapy in breast cancer is associated with persistent cognitive effects that may be present even years after completion of therapy (Brezden, 2000; Schagen, 1999; Ahles, 2002).
  • Other groups have demonstrated dissociation between self-reported cognitive impairment and formal testing of cognitive function in cancer survivors, and have described a “disorder of insight” in some patients (Vardy, 2008).
  • The type of cognitive impairment occurring in cancer patients remains somewhat unclear, and its etiology largely unknown.
    • Brain parenchymal loss has been described in patients receiving adjuvant chemotherapy for breast cancer (Swayampakula, 2007), although the contribution of this finding to cognitive changes is not fully understood.
    • Other symptoms, such as depression and fatigue, are likely confounding factors.
    • Cortical and subcortical toxicities of antineoplastic medications routinely used in cancer treatment are also not clearly understood.
  • This study was carried out in order to evaluate perceived cognitive function (PCF) and formal neuropsychological (NP) performance in patients with breast and colorectal cancer, and to assess the association of perceived impairment with impairment found on NP testing.

Materials and Methods

  • This study sought to evaluate PCF and NP performance in patients with colorectal and breast cancers.
    • Colorectal cancer patients were eligible for this study if they were 75 years old or less. They were considered in three groups: those receiving chemotherapy with curative intent; those receiving chemotherapy with palliative intent; and those not receiving chemotherapy.
      • Colorectal patients underwent NP evaluation at baseline (post-surgery but before chemotherapy), and at 6 months, 12 months, and 24 months after having begun chemotherapy.
    • Breast cancer patients were eligible for this study if they were 60 years old or less, and were considered in two groups: those receiving chemotherapy and those not receiving chemotherapy.
      • Breast cancer patients underwent neuropsychologic evaluation once within five years of breast cancer diagnosis.
    • Neuropsychologic assessment was performed using classical NP testing as well as computerized NP testing (CANTAB), and quantified by global deficit scores (GDS). The latter tests made use of a computer with a touch screen.
    • At the time of assessment, patients completed the FACT-COG to evaluate perceived cognitive function.
    • Patients completed concurrent questionnaires assessing fatigue, quality of life, and anxiety/ depression.
    • Mixed model regression adjusted for repeated measures, group, and disease diagnosis was used to assess association between PCF and NP.


  • Four-hundred and twenty patients were assessed during this study, with a total of 843 assessments performed.
    • Median age among colorectal patients was 58 years, and among breast cancer patients was 49 years. Overall median age was 56 years.
    • Of colorectal cancer patients, 38% were female, and 58% of the entire study group was female.
  • Overall, 15% of patients reported major perceived cognitive impairment (12% of colorectal patients, and 33% of breast cancer patients). Over the four assessments performed for colorectal cancer patients, perceived cognitive impairment was approximately constant at baseline, 6 months, and 12 months, but appeared to decrease somewhat at 24 months.
  • Neuropsychological deficit was identified in 22% of patients on classical NP testing (22% of colorectal patients, and 14% of breast cancer patients).
    • Interestingly, 31% of colorectal patients had deficits on NP testing at baseline. This improved to 10% at 24 months after treatment.
  • Neuropsychological deficit was identified in 23% of patients on CANTAB testing, and this number was the same for patients with breast and colorectal cancer.
  • Significant fatigue was reported by 26% of patients, 25% of colorectal cancer patients, and 28% of breast cancer patients.
  • A weak association was identified between PCF and NP deficit, with association ratios of 0.15 in breast cancer patients, and 0.17 in colorectal cancer patients.
  • Perceived cognitive functional deficits were correlated more strongly with fatigue (association ratio 0.48), quality of life (association ratio 0.51), and depression (association ratio 0.43). 
  • Patients who received chemotherapy perceived greater impairment (p < 0.0001), worse quality of life (p = 0.07), and worse fatigue (p = 0.01), but demonstrated no significant difference on NP testing from those who did not receive chemotherapy.
  • Among colorectal cancer patients, females were more likely to report perceived cognitive deficits (p = 0.06), although male patients were more likely to demonstrate deficits on NP testing (p = 0.02).
  • No difference was detected on classical testing or in terms of perceived cognitive function according to age, although patients over 60 appeared to have more NP deficits detected by CANTAB, a computer based assessment. Neuropsychologic deficits were detected in 31% of patients over 60 years old, and in 15% of patients less than 60 years old based only on CANTAB.
  • Correlation between perceived cognitive function and actual NP varied according to diagnosis and gender:
    • Among colorectal cancer patients, 71% reported normal perceived function and performed normally on NP testing. This was the case for 51% of breast cancer patients (p < 0.001).
    • Among colorectal cancer patients, 7% reported impaired function, but performed normally on NP testing. This was the case for 35% of breast cancer patients (p < 0.001).
    • Overall, female patients were more likely to report impairment in the setting of normal NP testing, while male patients were more likely to report normal function in the setting of NP deficits on testing.

Author's Conclusions

  • The authors conclude that, at best, a weak association exists between self-reported cognitive impairment and deficits on NP testing.
  • They note that fatigue, anxiety, depression, and quality of life appear to impact self-reported cognitive impairment.
  • They note that breast cancer patients are more likely to report cognitive impairment in the setting of normal function on NP testing than are colorectal cancer patients.
  • They conclude that self-reported cognitive impairment and NP deficits are both important, but that they are not surrogates and measure different things.

Clinical/Scientific Implications

  • This study represents an investigation of an interesting question – namely, the correlation between perceived cognitive deficit after cancer treatment and objective deficits detected on NP testing.
  • The authors have attempted to perform this evaluation in multiple subsets of patients in order to compare this issue across cancer diagnosis, gender, age, and presence or absence of chemotherapy. They have also used multiple testing techniques and have evaluated concurrent depression, fatigue, and quality of life issues. All of these increase the comprehensive nature of this study.
  • Several aspects of this study, however, weaken its conclusions:
    • First, one of the major issues in terms of evaluating cognitive deficits after cancer treatment is the lack of techniques which have been validated for doing so. Classical NP testing was not designed to assess this specific population, although it is frequently used to do so. Furthermore, classical NP testing is known to test issues of “multitasking” poorly; given that difficulty with multitasking is a common complaint of breast cancer patients, the validity of these tests in accurately assessing the cognitive issues encountered by breast cancer patients is questionable.
    • In a recent meta-analysis, the nature of cognitive evaluation (prospective versus retrospective) was found to significantly impact results in terms of detection of cognitive impact after breast cancer treatment (Falletti, 2005). In this study, colorectal cancer patients were evaluated in a prospective nature, but breast cancer patients were evaluated only once. The authors offered no explanation for this discrepancy; however, it serves to decrease the impact of the results presented, especially in light of data that study design may significantly alter results.
  • Within the oncologic community, the contributions of depression and fatigue are recognized as important in cognition. In this study, 30% of colorectal cancer patients were noted to have deficits on NP testing at baseline (after surgical treatment, but before beginning chemotherapy). These deficits improved over time. The impact of anxiety at the time of cancer diagnosis should not be underestimated, and may have contributed to the high rate of NP deficit at baseline in this group; additionally, aspects of surgery and the presence of neoplasm itself may contribute.
  • This study demonstrates significant cognitive impairment reported by a high percentage of patients undergoing cancer treatment. The authors demonstrate that this impairment is often not detectable on classical or computer based NP tests. This finding does not decrease the importance or validity of perceived cognitive impairment, however. As the authors point out, both perceived cognitive deficit and objective deficit on NP testing are extremely important, and should be considered as separate but equally important contributions to patient quality of life after cancer.

View Patient Summary