Breast Cancer Research and Treatment

, Volume 116, Issue 1, pp 125–127

Neuropsychological studies in breast cancer: in search of chemobrain


  • Steven Castellon
    • Department of Psychiatry and Biobehavioral SciencesDavid Geffen School of Medicine at UCLA
    • Department of Mental Health and PsychiatryWest Los Angeles Veterans Affairs Medical Center
    • Division of Cancer Prevention & Control Research, Jonsson Comprehensive Cancer CenterUCLA Schools of Medicine & Public Health
Invited Commentary

DOI: 10.1007/s10549-008-0211-2

Cite this article as:
Castellon, S. & Ganz, P.A. Breast Cancer Res Treat (2009) 116: 125. doi:10.1007/s10549-008-0211-2

Earlier detection and diagnosis of breast cancer has resulted in a growing number of patients with non-invasive tumors or localized invasive breast cancers. Among these early-stage patients, survival rates are high and most women can anticipate a normal life expectancy—making understanding and attending to post-treatment quality of life issues particularly pertinent. One area that has received increasing attention over the last several years is cognitive functioning [13], which is known to be strongly related to quality of life. The focus on post-treatment neurocognitive function has grown out of the recognition that a sizeable minority of patients undergoing adjuvant chemotherapy report and/or demonstrate some degree of cognitive compromise.

In this issue of ‘Breast Cancer Research and Treatment’, Quesnel and colleagues [4], report findings from a longitudinal study of early-stage breast cancer patients, (half of whom received adjuvant chemotherapy in addition to radiotherapy, half of whom received radiotherapy only), assessed before starting treatment, immediately after stopping, and 3 months post-treatment. They also compare pre-treatment neuropsychological performance of the two breast cancer groups to that of matched, healthy controls, to assess the possibility that rates of cognitive compromise may be elevated in breast cancer patients even prior to starting adjuvant treatment [5, 6]. There are several aspects of this study that we feel are important to address, as they illustrate issues that must be attended to as we move beyond the early wave of cross-sectional studies in this area of research [7].

First, these authors undertook the challenging task of assessing cognitive function in a fairly comprehensive fashion at three different time points along the adjuvant treatment continuum. As the authors suggest, several previous cross-sectional neuropsychological studies of breast cancer survivors have shown higher rates of cognitive compromise among patients at different stages during and after their chemotherapy treatment—but, while provocative, such studies are ultimately limited by their design. Prospective longitudinal studies such as this one can provide the important pre-treatment data that are needed to fully understand whether “change” in cognitive functioning has occurred in a given patient. That said, the analytic strategy that was employed by Quesnel and colleagues did not take maximal advantage of the study design. While their mixed-model ANCOVA offers the ability to look at both Group and Group × Time effects, it does not address individual change in performance over time, which is especially relevant when many patients in each clinical group are likely to show minimal or no treatment effect. This has been established in the chemobrain literature—that only a minority of all the patient’s exposed to chemotherapy will show detectable cognitive compromise [8]. We would have preferred to see them employ an analytic technique that addressed the known clinical reality that people do better, often much better, when assessed with the same task repeatedly (commonly called a “practice effect”), and that this improvement may be more robust on some cognitive tasks than others. Individual growth curve modeling is a relatively recent technique that allows one to measure change over time at both the group and individual level and might be considered in longitudinal studies such as this one [9]. Also, implementation of a Reliable Change Index that accounts for practice effects [10, 11] can provide an assessment of whether significant change is occurring over time—and whether this occurs in a differential proportion of patients in the groups that are being assessed. Studies assessing neurocognitive functioning at different time points would be wise to at least consider an analytic approach that can address whether an individual patient is showing significant deterioration (or improvement) at post-baseline assessment, above and beyond what would be expected from practice effect alone.

Second, the Quesnel et al. [4] study does not fully address an issue we feel to be increasingly important-namely the hormonal milieu of the breast cancer patients being assessed. All study groups differed significantly in menopausal status-between the two cancer groups and also between each cancer group and the matched healthy controls. Given the hypothesized relationship between estrogen levels and various cognitive functions [12], including especially verbal memory, it is quite possible that menopause status and timing of exposure to cancer directed endocrine therapy (or premature menopause, induced by chemotherapy) could confound group comparisons. Recall that in this study, verbal memory was the only measure—of many administered—on which both cancer groups showed decline in function. In fact, in a recently published prospective study, Jenkins and her colleagues [8] found that women with chemo-induced menopause were particularly vulnerable to cognitive compromise. Another confound is that the majority of patients (more than controls) had been on menopausal endocrine therapy prior to cancer treatment, and withdrawal of this therapy at the time of diagnosis could confound results. Also, according to the authors, nearly 75% of the breast cancer patients in both the chemotherapy and the radiotherapy only groups were on an adjuvant endocrine therapy at some point during treatment (similar proportions on Tamoxifen or Arimidex in each group), meaning that this study really may be assessing adjuvant endocrine therapy effects as well, about which little is currently known [13]. The study provides no information on the exact timing of initiation of endocrine therapy for the two cancer groups, which could be a confounding factor if a different proportion of patients in the two cancer groups were on endocrine therapy at the second and third assessments. Our research group found in a cross-sectional study of cancer survivors at least 5 years post-treatment, that endocrine therapy appeared to contribute to poorer cognitive function when added to chemotherapy, warranting further investigation for its potential impact on cognition [14]. We showed that women who had received both adjuvant chemotherapy and adjuvant hormonal therapy were particularly likely to show both neuropsychological compromise [14] and altered brain metabolism, as assessed by brain positron emission tomography [15].

Finally, as Quesnel and her colleagues point out, there are scant data on the ecological validity of neuropsychological testing in cancer patients at the current time. Given important differences in the laboratory testing setting (controlled environment, distraction free, behavior guided by examiner etc.), it may be that real-world functioning problems—that require multi-tasking and focus when distractions abound—are not well predicted by neuropsychological testing performance. Many teams, including ours at UCLA, have shown that self-reported cognitive functioning is only weakly correlated (or not correlated at all!) with actual neurocognitive testing performance, suggesting the possibility that neuropsychological tests do not accurately assess how well (or how poorly) a woman may perform in her everyday life. While some suggest that self-reported cognitive complaints are merely a by-product of emotional distress—often correlations are seen between self-reported depression/anxiety and self-report of cognitive compromise—there is some intriguing data suggesting that self-reported cognitive complaints do bear a relationship to neuroimaging parameters [16]. To better understand the frequent mis-match between objective and subjective cognitive functioning, obviously both methods of assessment should be employed (as was seen in the Quesnel et al. study). However, perhaps analyses of associations between pre- and post-treatment change scores for each “type” of functioning may be more fruitful than simply looking at correlations between objective and subjective functioning at each different time point.

It will be some time before there is agreement on whether or not cancer treatments are responsible for cognitive complaints in breast cancer survivors, and whether chemotherapy is the principal culprit. Hormonal changes in the woman (e.g., induced by chemotherapy or oophorectomy, influenced by endocrine therapy), are also likely to play an important role, especially in relationship to the interaction with normal aging [17]. Comprehensive neuropsychological testing studies are laborious to conduct, are challenged by the identification of suitable controls, and are limited by the fact that even when neuropsychological changes are noted, the effect sizes are modest and not necessarily associated with impairment in a specific domain of neurocognitive function or the woman’s own cognitive complaints. Certainly, the judicious use of adjuvant therapies (especially chemotherapy, which has many other side effects and may induce premature menopause) is advised, particularly for patients with small tumors with a low risk for recurrence. In addition, it is time for us to focus on strategies to assist those women who do complain of cognitive difficulties after treatment—whatever the cause—just as we would try to palliate pain, fatigue or other post-treatment symptoms in breast cancer survivors. To date, few efforts have been made to study ways to provide cognitive rehabilitation for those who are suffering with this problem [18]. First, it is important to address untreated depression, anxiety, and insomnia, in breast cancer survivors with cognitive complaints, as these may all contribute to poorer performance. If these are under control, then efforts to help women with organizing, planning, dividing their attention (i.e. multi-tasking) and improving their ability to concentrate—as these are the most commonly reported deficits—should be the next step forward on the research agenda. At UCLA, we are beginning such work in our research program to enhance survivorship and quality of life in women reporting cognitive problems after breast cancer treatments.


Supported in part by the Breast Cancer Research Foundation.

Copyright information

© Springer Science+Business Media, LLC. 2008