Cognitive Impairment in Patients With Breast Cancer: Understanding the Impact of Chemotherapy and Endocrine Therapy

During the past 2 decades, greater awareness of the impact of cancer treatments on cognition has become apparent.¹ Although initial patient reports of post-treatment difficulties with memory and attention were often met with skepticism by clinicians, increasing numbers of well-conducted observational studies have assessed cognitive function with standard neuropsychological testing and identified impairments.^2-4 Additional studies, using questionnaires sensitive to cancer-related cognitive complaints, demonstrate that patients treated for cancer differ from noncancer controls.^5,6 These observations are further supported by sophisticated neuroimaging studies that document brain structure and function changes that emerge after cancer treatment exposures.⁷ Although cancer-related cognitive impairment (CRCI) is potentially a problem for any patient group treated with cancer therapy (chemotherapy, radiation, high-dose chemotherapy with stem cell transplantation, endocrine therapy [ET]), patients with breast cancer have been the most extensively studied.

Most CRCI research has focused on exposure to chemotherapy because its acute toxicities are readily apparent. The association of cognitive complaints in patients with breast cancer emerged in the 1990s with the more widespread use of adjuvant chemotherapy and high-dose chemotherapy with stem cell transplantation.^8-11 Subsequently, ET emerged as a more widely used form of breast cancer adjuvant therapy, either with chemotherapy or more recently alone. ET is now the primary adjuvant therapy for the majority of women with early-stage breast cancer, yet less is known about its effects on cognitive function.^12-16 This change came about because of the shift toward smaller, favorable-histology tumors, as well as the more widespread use of genomic evaluation of primary tumors to define the risk of recurrence and/or need for chemotherapy.^17-19 Outcomes from the TAILORx study further enlarged the group of patients with hormone-sensitive early-stage breast cancer who no longer require chemotherapy in addition to ET.²⁰ However, the question remains: does omission of adjuvant chemotherapy for breast cancer reduce the risk of CRCI?

In the article that accompanies this editorial, Wagner et al²¹ report on the opportunistic inclusion of a study of patient-reported outcomes (PROs) that was added on to the TAILORx study near the time it was reaching its final target accrual. The beauty of the parent study design was the random assignment of women with intermediate values (11-25) of the 21-gene recurrence score to receive ET with or without chemotherapy.²⁰ Although observational studies have addressed this comparison in part,^11,22 the lack of bias associated with randomization allows for control of factors that might be influencing the choice of therapy. Furthermore, capturing PROs before the initiation of any adjuvant therapy provides a common entry point to cleanly assess subsequent changes in cognitive complaints that are associated with the specific treatment exposures. The main focus of the report by Wagner et al²¹ was cognitive impairment assessed by the Functional Assessment of Cancer Therapy-Cognitive Function (FACT-Cog), although other PRO measurements were collected and will be reported later (fatigue, endocrine symptoms, fear of recurrence, and health-related quality of life).

The prespecified primary endpoint for this trial was the score on the 20-item FACT-Cog Perceived Cognitive Impairment (PCI) scale at 3 months after randomization. The PCI is scored from 0-80, with a higher score being better. The sample size and power calculation for the study were based on a specific effect size (0.3 standard deviation [SD]) between the 2 treatment groups at 3 months. This was initially estimated using data from patients enrolled early in the study and then finally determined based on the final SD for the PCI at study entry for the entire sample. The absolute difference score on the PCI required to claim a statistically significant difference for the comparison of the 2 treatment groups at 3 months was −3.75, which is quite small in terms of cognitive symptom burden. However, the primary study endpoint was met, with the observed difference between the 2 groups being −3.82, demonstrating statistically significantly worse PCI scores among those women who received chemotherapy and ET compared with those who received ET alone. Of note, at the 3-month assessment, most women assigned to chemotherapy would not yet have started their ET; therefore, this is a direct comparison of the chemotherapy versus ET rather than the combined exposure.

In prespecified subgroup analyses by menopausal status, an interesting and divergent pattern occurred for the premenopausal and postmenopausal groups. In the postmenopausal women, there was a significant difference between the ET-alone and chemotherapy-plus-ET groups at both 3 and 6 months, with a sustained pattern of lowered PCI scores in the combined treatment group out to 36 months. In contrast, for the premenopausal women, the only significant difference between the 2 groups was at 3 months, with subsequent overlap in PCI decline of the 2 treatment groups, primarily due to a continued decline in PCI for the ET-only group, which was greater in magnitude than what was observed in the postmenopausal women. Interestingly, the modest decline in PCI observed in the postmenopausal ET-alone group is comparable to what was observed in the NSABP B-35 trial that compared anastrozole and tamoxifen in postmenopausal women with ductal carcinoma in situ.²³

From an analytic perspective, one cannot criticize the primary outcome conclusion or the interpretation of the comparisons made for the 2 treatment groups in this study. But how severe or serious are the cognitive changes observed between the 2 treatment groups, and what do these declines in scores actually mean? The pretreatment PCI score for both groups was approximately 69 and declined to approximately 63 (chemotherapy-plus-ET group) and 65.5 (ET-only group) at 3 months, without much additional decline thereafter, out to 36 months (Appendix Table 2).²¹ Until recently, there have been no normative data available to interpret PCI scores for “clinical caseness” of cognitive impairment; however, using a sample of healthy women for comparison, we recently developed a cut-point score for the PCI and found that a score < 60 was the equivalent of 2 SDs below the mean among healthy, middle-aged women on another reference self-report measure.^24,25 Thus, despite the observed treatment differences in the PCI in this trial, the actual scores, on average, do not reach the severity threshold for cognitive impairment at any time point for either treatment group. It is therefore difficult to interpret how meaningful these differences and declines might be. Of course, the mean may easily hide severe cognitive difficulties in women with much lower scores than average, reflected in the large SD of 17.3 at 3 months in the chemotherapy group. Furthermore, the impact of PCI decline, albeit subtle and subthreshold, was shown to have a statistically significant impact on health-related quality of life at 12 months of follow-up in this study. For better clinical interpretation, these investigators could apply this PCI cut-point to the study sample to better understand the subgroup of patients in both arms of the trial who have substantial impairment.

There were several challenges associated with the conduct of this study. First, adding on a PRO study to a multisite clinical trial, especially as an amendment several years after the study initiation, led to poor participation and missing or unevaluable data. In addition, there were some trial participants who were randomly assigned as part of the parent treatment study who did not receive the per-protocol treatment. The investigators anticipated this situation and knew that they would not be able to conduct an intention-to-treat analysis as the main outcome for the PRO study. Overall, there were few medical or demographic differences between the randomized and per-protocol treatment groups, and the authors did an excellent job describing the flow of study participants and differences between the final study sample and the enrolled trial population. It is unfortunate that the authors were not able to include the PRO data on fatigue and depressive symptoms as covariates in their evaluation of changes in the PCI over time, because these common treatment-related symptoms may affect cognitive function.

Overall, the results of this study make an important contribution toward understanding the potential adverse cognitive effects of chemotherapy in combination with ET and ET alone. The finding of different responses to chemotherapy plus ET and ET alone in the 2 menopausal groups suggests that we need to be considering menopausal status as a covariate or stratification variable when examining cognitive changes after breast cancer treatment. Estrogen receptors are widespread throughout the brain, and hormonal effects are important for brain function.^16,26,27 The influence of the woman’s pretreatment hormonal milieu in relation to cognition may be important. Sudden changes in circulating estradiol, such as in chemotherapy-induced amenorrhea in premenopausal women, may exacerbate the cognitive effects of subsequent ET. In postmenopausal women who begin their ET with low levels of circulating estradiol, the effects are likely different, but may be confounded by age-related cognitive decline that might also be exacerbated by chemotherapy, as suggested in the report by Wagner et al.²¹ Understanding cognitive symptoms in patients with breast cancer remains an important research priority, given the large number of long-term survivors. Future research on cognitive impairment in this setting should be aimed at understanding the complex interactions between breast cancer adjuvant treatments and women’s normal aging, including menopausal status.

AUTHOR CONTRIBUTIONS

Conception and design: All authors

Data analysis and interpretation: Patricia A. Ganz

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Cognitive Impairment in Patients With Breast Cancer: Understanding the Impact of Chemotherapy and Endocrine Therapy

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/jco/authors/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).

Patricia A. Ganz

Leadership: Intrinsic LifeSciences (I)

Stock and Other Ownership Interests: Xenon Pharma (I), Intrinsic LifeSciences (I), Silarus Therapeutics (I), Teva, Novartis, Merck, Johnson & Johnson, Pfizer, GlaxoSmithKline, Abbott Laboratories

Consulting or Advisory Role: Keryx (I), Silarus Therapeutics (I), InformedDNA, Vifor Pharma (I), Gilead Sciences (I), La Jolla Pharma (I), Ambys (I), Bioverativ (I), Global Blood Therapeutics (I), GSK (I), Ionis (I), Akebia (I), Protagonist Therapeutics (I), Regeneron (I), Sierra Oncology (I)

Research Funding: Keryx (I), Sierra Oncology (I)

Patents, Royalties, Other Intellectual Property: Related to iron metabolism and the anemia of chronic disease (I), UptoDate royalties for section editor on survivorship

Travel, Accommodations, Expenses: Intrinsic LifeSciences (I), Keryx (I)

No other potential conflicts of interest were reported.