Cognitive Function After Systemic Therapy for Breast Cancer

May 1, 2001

Anecdotal reports of cognitive compromise among patients treated with chemotherapy are relatively common among breast cancer survivors and may play an important role in adversely affecting functioning in multiple domains. As noted by Dr. Olin,

Anecdotal reports of cognitivecompromise among patientstreated with chemotherapy are relatively common among breast cancer survivorsand may play an important role in adversely affecting functioning in multipledomains. As noted by Dr. Olin, the empiric exploration of the neurocognitiveeffects of chemohormonal treatment among breast cancer survivors has beeninitiated only in the past few years. In her article, she cogently reviews therelevant literature with a focus on the relationship between neurocognitivestatus and quality of life, and concludes by offering several suggestions fordirecting future research endeavors in this field. We agree that this topic is atimely one, especially with the increasing indications for adjuvant therapy inpatients with earlier-stage disease, who, with therapy, can anticipate nearlynormal life expectancies. If neurocognitive changes are prevalent and/ordisabling, then this late effect must be considered in the process of informeddecision-making.

Noteworthy Findings

Dr. Olin’s review of the literature suggests that asignificant minority of studied breast cancer survivors demonstrated some degreeof cognitive deficit. Exact rates of impairment are bound to vary, depending inpart upon how neurocognitive compromise is operationalized. Some studies havefailed to use adequate control groups and have relied instead on comparison withextant normative data. As Dr. Olin points out, this approach has seriouslimitations, particularly if premorbid differences exist between the individualscomprising normative samples and patients with breast cancer; also, the qualityof the normative data for women in this age group varies tremendously from taskto task.

Comparison to normative data often leads to an inflatedestimation of neurocognitive compromise when the comparison group represents aclinical population that may suffer from some combination of medical,psychiatric, and/or neurologic condition(s). Unfortunately, the two reviewedDutch studies used the same control group and, therefore, their results shouldbe interpreted with caution. However, their findings strongly suggest adose-response relationship between chemotherapy and cognitive function, withmore intensive systemic treatment leading to higher rates of impairment andlower group mean scores on various neuropsychological instruments. Theimplications of such findings are particularly important in light of recentfindings that fail to suggest that high-dose chemotherapy is superior toconventional-dose chemotherapy for adjuvant treatment.[1]

Another noteworthy finding that Dr. Olin briefly discusses isthe fact that some change in cognitive function is present during, shortlyafter, and up to (on average) 2 years posttreatment. In fact, a recent study byAhles and Saykin,[2] along with pilot data from our own lab, suggest that somecognitive compromise was noted in a subset of breast cancer survivors more than5 years posttreatment and was more likely to be present in patientsreceiving more intense therapy. Data from both our own lab and Ahles and Saykinsuggest that tasks that place demands on working memory (eg, simultaneouslystoring and manipulating information), verbal memory, and acceleratedinformation processing are particularly affected.

Correlates of Neurocognitive Compromise

Dr. Olin notes that there are undoubtedly multiple pathways tocognitive compromise in survivors of breast cancer. Systemic treatment effectsare clearly one but certainly not the only etiologic agent responsible fordecreased cognitive efficiency. Research to date has suggested that constructs,such as depression, fatigue, and pain, which are often associated with eachother among breast cancer survivors,[3] are unlikely to explain the cognitivedeficits that were observed in the reviewed studies. This finding is notsurprising given the presence of psychiatric exclusionary criteria and/or thegenerally low levels of depression reported in the studies published to date.

While depression, fatigue, and pain are all known to adverselyaffect cognitive function, it is typically when these symptoms or syndromes arepresent to a pronounced and persistent degree that a consistent relationship tocognition is noted.[4] Also, in some patients, depression and fatigue may beconsequences of central nervous system involvement, in which case their presencemay have more prognostic significance (ie, be more likely to be related tocognitive compromise) than when these constructs are instead a response orreaction to loss or increased stress.

Individual differences in age and premorbid intellectual levelare likely to be associated with neurocognitive performance followingchemotherapy as each of these constructs has shown a clear and consistentrelationship with various cognitive domains (including verbal learning andmemory, information-processing speed, and complex attentional function).Finally, there does seem to be a relationship between menopausal status andneurocognitive function,[5] meaning that this important potential confound (ormoderator variable) should be carefully accounted for in future studies.

Consequences of Cognitive Compromise

The impact of cognitive compromise can vary tremendously, fromsubtle changes in information-processing speed and efficiency to frank dementiasyndromes. There is currently insufficient evidence to conclude with any suretythe clinical significance of lower performance on neuropsychological tasks amonga subgroup of breast cancer survivors. Characteristics of the injury or insult(eg, structural vs neurochemical, location, severity), characteristics of theindividual (eg, premorbid cognitive functioning, age, presence of comorbidconditions), and characteristics of the environment (eg, demands of job, degreeof social support, presence of rehabilitation opportunities) each influence howdisruptive the cognitive dysfunction will be.

Dr. Olin notes that quality of life may be dependent upon manyof these factors and does not always share a linear relationship with the degreeof cognitive deficit. An important "next step" in this research willbe the elucidation of mediators of cognitive dysfunction within this populationof breast cancer survivors. Assessment tools borrowed not only from psychologyand neuropsychology, but also from psychiatry and neuroradiology, will beimportant in determining which individual and/or environmental differences aremost relevant in predicting cognitive function.

Cautions and Future Directions

Neuropsychological findings drawn from research in otherpopulations suggest that we should avoid overinterpreting poor performance onany one task until more data accumulate. Many neurocognitive tasks aremultifactorial in nature, meaning that they measure cognitive processes inaddition to that process or domain that is the main target of assessment. Forexample, a task purporting to index "verbal memory" is also likely tomake demands on other cognitive domains such as language comprehension,sustained attention, information-processing speed, and/or auditory perception.The tendency to interpret low scores on such a task as evidence of verbal memorydeficits runs the risk of mislabeling the deficit if, for example, it is indeeda problem with sustaining attention that is driving poor performance on thistask. It is not uncommon to find that one or two more basic cognitive processes(eg, processing speed) mediate performance on higher-order tasks such as memoryor executive functioning (ie, planning, organization, initiation).

The role of neuroimaging techniques in helping to betterelucidate the neuroanatomic and neurophysiologic substrates associated withneurocognitive compromise among affected breast cancer survivors may beparticularly informative. More specifically, using positron-emission tomographyand magnetic resonance imaging, the brain function of breast cancer survivorscan be examined using activation paradigms in which brain glucose metabolism ismeasured and compared to controls. Finally, the role of cognitive rehabilitationin increasing quality of life among affected survivors has received littleempiric attention, but is clearly a topic with important real-worldramifications.


1. McCarthy NJ, Swain SM: Update on adjuvant chemotherapy forearly breast cancer. Oncology 14:1267-1280, 2000.

2. Ahles TA, Saykin AJ: Cognitive effects of standard-dosechemotherapy in patients with cancer. Cancer Invest. In press.

3. Bower JE, Ganz PA, Desmond KA, et al: Fatigue in breastcancer survivors: Occurrence, correlates, and impact on quality of life. J ClinOncol 18:743-753, 2000.

4. Cassens G, Wolfe L, Zola M: The neuropsychology ofdepressions. J Neuropsychiatry Clin Neurosci 2:202-213, 1990.

5. Sherwin BB: Estrogen and cognitive function in women. ProcSoc Exp Biol Med 217:17-22, 1998.