A Prospective Study on Effect of Systemic Adjuvant Therapy on Cognitive and Brain Function of Breast Cancer Patients

Breast cancer is the most common cancer amongst women worldwide, with a lifetime risk of 7-8% in women and incidence rates ranging from 19.3-89.7 per 100,000 women in developing and developed countries. More than 1,383,500 cases of invasive breast cancer are estimated to be diagnosed worldwide,and approximately 1490 women were diagnosed to have breast cancer in Singapore every year. However, the mortality rates are lesser at 6-19 per 100,000 women because of the more favourable survival of breast cancer in high-incidence developed regions. In Southeast Asian Chinese 5-year overall survival is about 75.8%, although lower for Indians and Malays at 68% and 58.5% respectively. These improving outcomes are largely due to earlier detection and the availability of more successful treatment options. With rising prevalence of this disease and improving survivorship, the long term effects of current treatment options are becoming increasingly survivorship issues.An early Early Breast Trialists' Collaborative Group (EBCTCG) meta-analyses, reported that adjuvant chemotherapy produces an absolute 10-year survival improvement of 7-11% for those \< 50 years of age, and 2- 3% for those aged 50 -69 years. This benefit was most recently confirmed by an update which reported that adjuvant chemotherapy trials demonstrated recurrence and mortality reductions compared to no-chemotherapy trials.Hence, chemotherapy remains an important strategy in the adjuvant setting. However, several studies have reported reduced cognitive function in a proportion of breast cancer patients receiving chemotherapy, also known as the "chemobrain" effect. In these patients, cognitive impairments ranging from deficits in memory, attention, concentration, executive and psychomotor functioning to deficits in nearly all cognitive domains have been reported. Other studies however have disputed this association, and it thus remains unclear if cognitive impairment is indeed a potential risk of systemic treatment. There are a number of difference adjuvant chemotherapy regimens which comprise anthracyclines and/or taxanes in various regimens. An adjuvant regimen utilising weekly paclitaxel after standard doxorubicin/cyclophosphamide has been shown to improve disease-free and overall survival over 3-weekly taxanes given after the same standard anthracycline-based regimen In order to minimise heterogeneity, this study will only recruit patients utilising this adjuvant chemotherapy regimen.Anti-hormonal therapy has become a mainstay in the adjuvant treatment of hormone receptor-positive patients. Tamoxifen for 5 years reduces recurrence rates throughout the first 10 years (RR 0·53 during years 0-4 and RR 0·68 \[0·06\] during years 5-9 \[both 2p\<0·00001\]; and breast cancer mortality by about a third throughout the first 15 years. In addition, the use of aromatase inhibitors (AIs) in post-menopausal breast cancers has been reported to confer a significant 2.9% decrease in recurrence compared with tamoxifen (9.6% for AI v 12.6% for tamoxifen; 2P \< .00001) and a non-significant absolute 1.1% decrease in breast cancer mortality (4.8% for AI v 5.9% for tamoxifen; 2P = .1) in a recent meta-analyses by Dowsett et al.Compared to chemotherapy, there is a relative paucity of data pertaining to endocrine therapy-induced cognitive changes. Nonetheless, there have been several pertinent studies examining datasets within the larger context of pre-existing large multicenter adjuvant hormonal therapy trials which are outlined below.A growing body of literature suggests that adjuvant systemic therapy for breast cancer may be associated with cognitive impairment, which may persist for years and can impair decision-making abilities and disrupt the ability of women to fulfill family, career, and community responsibilities.The magnitude of these deficits may be moderated by depression, anxiety, fatigue, concomitant medications, genetics, education, intelligence, and menopause.Studies associating cognitive changes with chemotherapy had been reported since the mid 1970s;but systematic research did not come about until the mid 1990's to early 2000's.However, early studies were limited by lack of a pre-treatment assessment, the importance of which is illustrated by a study by Wefel et al.This was one of the first prospective longitudinal studies comparing pre- and post-treatment cognitive measures. Although there were no mean differences between patients who received 5- fluorouracil/doxorubicin/ cyclophosphamide (FAC) chemotherapy and normal controls, intra-individual analyses in patients demonstrated a 61% cognitive decline for learning, attention and processing speed. Without a baseline assessment, 46% of patients who had normal post-treatment scores would have their cognitive impairments missed, highlighting that cognitive dysfunctions can be subtle and small differences may be clinically significant. Furthermore, although acute cognitive changes during chemotherapy are common,long-term post-treatment cognitive changes seem to persist in only 17-34% of cancer survivors.Several additional longitudinal studies since then have assessed cancer or treatment-related cognitive difficulties using standard neuropsychological assessments.About 12-82% of patients had detected cognitive impairments in the domains of executive function, memory, psychomotor speed and attention, these being the most frequent measures affected. These studies used a variety of cognitive assessment tools and control groups, different treatment regimens and varying testing timepoints. A study comparing breast cancer patients, majority of which were exposed to adjuvant anthracycline and/or taxane-based chemotherapy, to those breast cancer patients who were chemonaive or healthy controls, reported the greatest deficits in processing speed and verbal ability domains for the chemotherapy-exposed group. Therefore, research is needed needed to look at the factors contributing to a decline and thereafter, a long-lasting impairment.Of interest, cognitive deficits have been detected in cancer patients pre-chemotherapy in several studies. For instance, Wefel et al described 33% of women in the current cohort exhibiting cognitive impairment even before systemic therapy.Another study reported a 23% impairment in cognition prior to adjuvant breast cancer chemotherapy.These could be related to psychological variables such as stress, anxiety or depression, or to other factors such as differing socioeconomic, educational or intrinsic cognitive levels. Research is needed to look at the factors contributing to a decline and thereafter, a long-lasting impairment. These data also suggest that systemic therapy-associated cognitive changes need to be examined within the broader context of risk factors and biological processes associated with cancer development.Inclusion of assessments of confounding factors which could play a role in influencing cognitive function is vital. Fan et al in their study reported that fatigue, menopausal symptoms and cognitive were important adverse effects of chemotherapy that improved in most patient over a period of 1-2 years.In comparison, hormonal therapy had minimal impact on them. Self-perceived cognitive deficits may not be reflective of objective test scores and hence is interesting to evaluate. In a population-based study in Denmark consisting of women \<60 years old who received adjuvant anthracycline-based chemotherapy or tamoxifen compared with healthy controls, it was shown that the patients were up to 3 times more likely than healthy controls to rate themselves cognitively impaired at 6 months, although neuropsychological tests did not reveal any evident or consistent pattern of cognitive change in any of the patient groups.Most clinical data on cognitive function and anti-hormonal therapy have been on selective estrogen receptor modulators (SERMS) such as tamoxifen, although focus on aromatase inhibitors has been increasing in recent years. A substudy (designated the CoSTAR trial) pooled subjects from the pre-existing National Surgical Adjuvant Breast and Bowel Project (NSABP) Study of Tamoxifen and Raloxifene (STAR) randomizing women =35 years old with increased risk of breast cancer by Gail model evaluation to either tamoxifen or raloxifene.The substudy required women to perform an 83-minute standardized test battery covering multiple domains. Outcomes on cognitive testing were similar in both groups. In the landmark NSABP Breast Cancer Prevention Trial (P-1) which included 13,388 women at increased breast cancer risk, treatment with 5 years of tamoxifen resulted in a reduced risk of breast cancer by 49% relative to placebo.A separate report documented results of the P-1 Symptom Checklist assessment encompassing a few self-report domains spanning cognition, musculoskeletal pain, vasomotor symptoms, nausea, sexual problems, bladder problems, body image and vaginal symptoms.The cognitive assessment included questions related to concentration, forgetfulness, avoidance of social affairs, and tendency towards accidents. There was little difference in cognitive dysfunction between groups. Notably, for these 2 studies, evaluation was conducted during the period of active therapy. To address this issue, a study by Paganini-Hill et al in a population-based case-control study examined 1163 women between 55-75 years old diagnosed with primary breast cancer.Previous tamoxifen users were classified as short-term users (\<4 years) or long-term users (=6 years) and were compared with never users. Women who had used tamoxifen for 4-5 years had more memory problems reported to their physician compared to never users (3.8% vs 1.5%, p=0.04), especially for women currently on tamoxifen. However, no differences were seen among never and past users suggesting limited long-term sequelae. Even among the premenopausal population on tamoxifen, significantly lower processing speeds which appeared to interfere with executive functioning have been reported in a cross-sectional study comparing women on adjuvant tamoxifen to healthy controls.Clinical data on the effects of aromatase inhibitors on cognitive function are emerging. The largest dataset comes from the International Breast Intervention Study II (IBIS II) studying postmenopausal women at increased risk of breast cancer receiving either anastrozole or placebo for 5 years, testing being done at 0, 6 and 24 months. No differences in cognitive function between the 2 groups were detected. However the neurocognitive testing was done during the period of treatment as in other previously described studies and in a postmenopausal hypoestrogenic population. As such, the effects of AIs could have been blunted.In a substudy of the Anastrazole, Tamoxifen and Combined (ATAC) trial which recruited patients with localized breast cancer to adjuvant tamoxifen, anastrozole or combination therapy, 94 patients from the ATAC trial were compared to 35 non-cancer controls. Their mean length of study time was 36 months. The patient group was significantly impaired compared to the control group on measures of verbal memory (p=0.026) and processing speed (p=0.032). Comparisons between anastrozole and tamoxifen have been done in a cross-sectional and longitudinal study, both of which demonstrated the anastrozole exposed group experiencing worse cognitive impairments.There were poorer verbal and visual learning and memory in the study by Bender et al,44 while the prospective longitudinal study showed a nine-fold and five-fold increase in risk of cognitive decline in the anastrozole and tamoxifen groups respectively relative to healthy controls; the most common domains affected being processing speed and verbal memory.A neuropsychological substudy was carried out from the pivotal Tamoxifen and Exemestane Multinational (TEAM) trial which randomized patients to adjuvant exemestane for 5 years or adjuvant tamoxifen for 2.5-3 years followed by exemestane for 2-2.5 years .The neuropsychological assessment was carried out pre-endocrine therapy and after a year of treatment, and included 80 tamoxifen and 99 exemestane users. Interestingly, exemestane users did not perform statistically worse than healthy controls on any cognitive domain, but tamoxifen users had significantly worse verbal memory and executive functioning compared to healthy controls, and significantly worse information processing speed compared to exemestane users. However, it has been postulated that exemestane's mildly androgenic properties may ultimately preserve cognition.A recent cognitive testing report of the Breast International Group (BIG)1-98 study randomizing women to adjuvant monotherapy of tamoxifen or letrozole for 5 years, or sequential therapy with letrozole followed by tamoxifen or vice versa revealed that women taking letrozole in the 5th year had better overall cognitive function than those taking tamoxifen (p=0.04).A significant improvement in composite cognitive function for both groups was observed in years 5-6 when treatment was ceased. In summary, with repect to AIs, a greater cognitive decline has been observed with anastrozole as compared to tamoxifen, while a lesser decline is seen with exemestane and letrozole.Larger studies are clearly need to validate these findings. Results from many of these studies need to be interpreted cautiously, as there have been various limitations in studies pertaining to this field. Many studies have been hampered by a variety of inherent methodological problems. These include lack of a baseline assessment, lack of a longitudinal design, and failure to control for physical and psychosocial confounding factors such as hormonal factors, depression, anxiety, fatigue and educational level.Furthermore, studies have also differed considerably with respect to regimens and doses of chemotherapy received, previous exposure to anti-hormonal therapy and length of time post-treatment. In addition, a limitation of several studies is the lack of pre-treatment evaluations of cognitive function.This is especially problematic because it makes it impossible to determine whether deficits were present before treatment or to detect changes from baseline. The timing of cognitive function assessment after the completion of therapy was also not uniform. Most studies lacked non-treated comparison groups and failed to measure potential moderators of cognitive function. Several studies failed to include a comprehensive assessment of the multiple domains of cognitive function, and occasionally relied on the use of a self-report questionnaire. Importantly, cognitive studies in the Asian population have been grossly limited. Certain biological pathways, genetic polymorphisms and epigenetic changes could lead to alterations in cognitive functioning in only a subgroup of people exposed to chemotherapy. For example, several genetic polymorphisms of multidrug resistance 1 (MDR1) have been identified that may influence P-glycoprotein (P-gp) function, one of the most studied polymorphisms being C3435T in exon26.P-gp which is present in the blood brain barrier affects the amount of drug uptake into the brain by actively transporting them out of the cells. Polymorphisms associated with lower expression or functionality of P-gp may cause increased vulnerability to chemotherapy-induced cognitive changes. Oxidative stress is the most frequent cause of DNA damage in neuronal cells, and has been associated with various neurodegenerative diseases including Alzheimer's and Parkinson's disease. DNA damage and repair systems have been linked to neurodegeneration such as in ataxia telengectasia and xeroderma pigmentosum. So far the relation of cognitive functions to the level of DNA damage post-chemotherapy has not been well studied. As mentioned, certain studies have detected higher than expected rates of cognitive impairment pre-chemotherapy. In addition higher levels of oxidative DNA damage have been found in women with breast cancer.These data are consistent with research linking certain DNA repair polymorphisms with a decreased DNA repair capacity and thus an increased cancer risk. Hence, changes in cognitive function following chemotherapy can be evaluated within the context of genetic factors that increase cancer risk but also increase risk of cognitive dysfunction pre-treatment. Certain polymorphisms in the base-excision pathway may also be worthwhile studying because of their importance in modulating oxidative stress and cancer risk.Cytokines play a role in central nervous system function, including modulation of neuronal and glial cell functioning, neural repair and metabolism of dopamine and serotonin, which are important neurotransmitters for cognition. Longitudinal studies of interferon-a and IL2 in cancer populations have shown decrements in cognitive performance, particularly in the domains of information processing speed, executive function, spatial ability and reaction time.Both chemotherapy and persistent fatigue have been linked to increased cytokine levels.Genetic polymorphisms have been identified that influence cytokine activity and are associated with Alzheimer's disease and depression. However, so far the relationships between these polymorphisms and chemotherapy-induced cognitive changes have not been well studied. The catechol-O-methyltransferase (COMT) valine genotype has been linked to chemotherapy-related cognitive impairment in breast cancer survivors and is worthwhile of further study. Chemotherapy in mouse models has also been demonstrated to decrease hippocampal cell proliferation, increase histone acetylation and decrease histone deacetylase activity. Hence, it would therefore be relevant to study the particular candidate mechanisms which are likely to influence cognitive changes in certain susceptible individuals. Inclusion of assessments of confounding factors which could play a role in influencing cognitive function is vital. Fan et al in their study reported that fatigue, menopausal symptoms and cognitive were important adverse effects of chemotherapy that improved in most patient over a period of 1-2 years. In comparison, hormonal therapy had minimal impact on them.But, even with the use of increasingly sophisticated performance-based assessments, subtle differences are often being missed along with the obscurity of the neural circuits and structures underlying the cognitive changes. For this reason, neuro-imaging tools have gained advantage because they provide the opportunity to examine the effects of chemotherapy on brain and cognition. However, such studies have been very limited in the field and therefore the need for further brain imaging studies. MRI has emerged as a noninvasive method with the potential to produce high resolution and contrast images of the brain. From the anatomical findings, adjuvant chemotherapy was associated with long-term injury to white matter and also damage to gray matter with associated functional deficits. An early study by Inagaki et al. comparing 51 adjuvant chemotherapy treated breast cancer patients with 55 patients who did not receive chemotherapy, showed that chemotherapy patients had smaller volumes in key areas involved in cognitive processing. The patients showed a strong positive correlation between volume loss and performance on the WMSR (Wechsler Memory Scale Revised) test of attention and memory. A serious weakness of this study was the failure to consider the effect of adjuvant endocrine therapy and the lack of prechemotherapy baseline assessment. Only three studies to date have investigated the integrity of white matter tracts and structural connectivity using diffusion tensor imaging (DTI). Results revealed structural neural changes over time as well as indicated region specific differences inherent to most of the cognitive deficits.The corpus callosum, an area important for communication between hemispheres, exhibited reduced white matter integrity which may explain the reduced processing speed reported in the chemotherapy-treated patients. The effect of endocrine treatment was not considered in the analysis which provides a drawback for proper interpretation. Another study showed decreased Fractional Anisotropy (FA) in frontal and temporal white matter tracts as well as increased Mean Diffusivity (MD) in frontal white matter. A significant correlation was also seen between the FA scores and the neuropsychological tests (attention and processing speed) in the chemotherapy-treated patients. White matter organization, particularly in the frontal, parietal and occipital white matter tracts, was negatively impacted by chemotherapy and this correlated strongly with cognitive functioning scores. This suggests that microstructural white matter changes or abnormalities may underlie reported cognitive dysfunctions found in chemotherapy treated cancer patients, especially that cerebral white matter is vulnerable to neurotoxins.In addition to the small number of structural imaging studies, a limited number of functional imaging studies have assessed neural changes in chemotherapy treated cancer patients. Combining the anatomical/structural assessment with functional imaging techniques provides a wider window of evaluation onto the functional changes associated with such cognitive deficits. To understand the underlying functional changes associated with poor executive function and processing in chemotherapy patients, a study by Saykin et al was one of the first to incorporate working memory assessments with functional MRI readouts. This was one of the first prospective longitudinal studies comparing pre- and post-treatment on a working memory N-back task. Although the authors reported no group differences in task performance with the expected main effect of working memory load on performance at baseline, patients assessed one month post-chemotherapy showed increased activation in posterior frontal and parietal regions compared to controls with less bilateral activity in more anterior frontal regions. This study illustrates that the cognitive changes associated with chemotherapy may be too subtle to reliably be detected with standard neuropsychologial assessments and underscore the importance of and potential for using functional MRI to elucidate underlying neurobiological changes.Interestingly, while investigating declarative memory in 14 breast cancer patients three years following chemotherapy, patients showed less activity in bilateral frontal gyri and postcentral gyrus with greater activation in parahippocampal, bilateral cerebellum, cingulate and precuneus while exhibiting normal cognitive assessment and cortisol levels (a measure of distress) compared to control. The increased activation in several brain areas compared to controls indicates that patients required greater and more global neural effort than controls when attempting to recall task information. More interestingly, this study showed that the type of chemotherapy regimen contributed to differential patient verbal memory impairments; CMF treated patients showed lower prefrontal cortex activity during encoding compared to patients on the adriamycine/cyclophosphamide/taxol regimen. This indeed highlights the importance of distinguishing between different types of chemotherapy treatment. Although functional MRI is the gold standard for evaluation of brain function, it is an indirect measure and functional readouts do not represent the underlying neuronal mechanisms involved. Other methods have been investigated to assess mere brain activity and more directly detect the underlying mechanism of cognitive dysfunction in chemotherapy patients. Only two PET (Positron Emission Tomography) imaging studies have been conducted to date. A recent study was conducted in 5 year post-chemotherapy patients. Patients exhibited lower resting metabolism in the frontal cortex and basal ganglia. A major drawback of this study was the lack of baseline assessment and the inhomogeneous study group where eight of the patients were on tamoxifen treatment.Another PET/CT study to assess brain metabolism after chemotherapy in 128 patients is still ongoing.Results from many of these studies need to be interpreted cautiously, as there have been various limitations in studies pertaining to this field. Many studies have been hampered by a variety of inherent methodological problems. These include lack of a baseline assessment, lack of a longitudinal design, and failure to control for physical and psychosocial confounding factors such as hormonal factors, depression, anxiety, fatigue and educational level. Furthermore, studies have also differed considerably with respect to regimens and doses of chemotherapy received, previous exposure to anti-hormonal therapy and length of time post-treatment. In addition, a limitation of several studies is the lack of pre-treatment evaluations of cognitive function. This is especially problematic because it makes it impossible to determine whether deficits were present before treatment or to detect changes from baseline. The timing of cognitive function assessment after the completion of therapy was also not uniform. Most studies lacked non-treated comparison groups and failed to measure potential moderators of cognitive function. Several studies failed to include a comprehensive assessment of the multiple domains of cognitive function, and occasionally relied on the use of a self-report questionnaire. Importantly, cognitive studies in the Asian population have been grossly limited. Taking these shortcomings into consideration, we designed a study intended to minimise biases mentioned as above. The chemotherapy regimen utilised will be limited to that of anthracyclines and taxanes; specifically we will use patients given doxorubicin/cyclophosphamide (AC) for 4 cycles followed by weekly paclitaxel for 12 cycles as per the adjuvant breast cancer study by Sparano et al as this is the chemotherapy regimen most commonly used in our division. Furthermore, we will have a baseline pre-treatment assessment for both CANTAB and imaging assessments, a longitudinal study design, and a more uniform evaluation taking 6 months and one year post-chemotherapy or after starting one year of anti-hormonal therapy as a cut-off point for CANTAB and imaging assessments. We will also attempt to minimize biases by analyzing confounders such as anxiety, depression and hormonal fluctuations using questionnaires and also recruit a control group of healthy subjects with similar age, gender and social background.