Springer Nature is making SARS-CoV-2 and COVID-19 research free. View research | View latest news | Sign up for updates

Longitudinal investigation of cognitive deficits in breast cancer patients and their gray matter correlates: impact of education level

  • 244 Accesses

  • 2 Citations

Abstract

Cognitive deficits are a major complaint in breast cancer patients, even before chemotherapy. Comprehension of the cerebral mechanisms related to cognitive impairment in breast cancer patients remains difficult due to the scarcity of studies investigating both cognitive and anatomical imaging changes. Furthermore, only some of the patients experienced cognitive decline following chemotherapy, yet few studies have identified risk factors for cognitive deficits in these patients. It has been shown that education level could impact cognitive abilities during the recovery phase following chemotherapy. Our main aim was to longitudinally evaluate cognitive and anatomical changes associated with cancer and chemotherapy in breast cancer patients. Our secondary aim was to assess the impact of education level on cognitive performances and gray matter (GM) atrophy in these patients. Twenty patients were included before chemotherapy (T1), 1 month (T2) and 1 year (T3) after chemotherapy. Twenty-seven controls without a history of cancer were assessed at T1 and T3 only. Cluster groups based on education level were defined for both groups and were further compared. Comparison between patients and controls revealed deficits in patients on verbal episodic memory retrieval at T1 and T3 and on executive functions at T3. After chemotherapy, breast cancer patients had GM atrophy that persisted or recovered 1 year after chemotherapy depending on the cortical areas. Increase in GM volumes from T1 to T3 were also found in both groups. At T2, patients with a higher level of education compared to lower level exhibited higher episodic memory retrieval and state anxiety scores, both correlating with cerebellar volume. This higher level of education group exhibited hippocampal atrophy. Our results suggest that, before chemotherapy, cancer-related processes impact cognitive functioning and that this impact seems exacerbated by the effect of chemotherapy on certain brain regions. Increase in GM volumes after chemotherapy were unexpected and warrant further investigations. Higher education level was associated, 1 month after the end of chemotherapy, with greater anxiety and hippocampal atrophy despite a lack of cognitive deficits. These results suggest, for the first time, the occurrence of compensation mechanisms that may be linked to cognitive reserve in relationship to state anxiety. This identification of factors, which may compensate cognitive impairment following chemotherapy, is critical for patient care and quality of life.

Introduction

Complaints about cognitive impairments in breast cancer patients are common and well described both during and after chemotherapy as the chemobrain effect (Wefel and Schagen 2012). This chemobrain effect is known to affect processing speed, memory and executive functions (O’Farrell et al. 2013; Vardy 2009) and seems specific to a subgroup of patients (Falleti et al. 2005; Stewart et al. 2006). One year or more after the end of chemotherapy, patients can display a partial recovery of their cognitive abilities (Ahles et al. 2010; Lepage et al. 2014; Lyon et al. 2016; Zheng et al. 2014). However, this recovery is not systematic (Collins et al. 2014; Jim et al. 2012; Pereira et al. 2015). In contrast to the extensive literature on the chemobrain effects, few studies investigated cognitive deficits prior to adjuvant therapy. Yet, prospective studies have found cognitive deficits even before adjuvant therapy has begun (Ahles et al. 2008; Cimprich et al. 2010; Lange et al. 2014; Quesnel et al. 2009; Wefel et al. 2010; Wefel and Schagen 2012). Specifically, studies highlighted that 20–30% of patients had cognitive impairments prior to adjuvant therapy (Ahles et al. 2012) and that such impairments were mainly related to verbal learning and executive functions (Lange et al. 2014; Wefel et al. 2004a, b). Explanation for this phenomenon is, for now, unclear and warrants further investigations (Ahles et al. 2012).

In order to better understand the neural basis of such cognitive impairments, previous studies have evaluated structural differences between breast cancer patients and controls using voxel-based morphometry (VBM) (Ashburner and Friston 2005). Studies investigating gray matter (GM) atrophy in breast cancer patients have shown diffuse GM atrophy in patients 1 month after the end of chemotherapy including frontal, temporal and occipital gyri, cerebellum and thalamus (McDonald et al. 2010, 2013; Lepage et al. 2014). One year or more after the end of chemotherapy, partial recovery has been found in superior frontal gyri, superior temporal gyrus and cerebellum, while persistent atrophy was found in middle and superior frontal gyri, cerebellum, hippocampus and thalamus (Bergouignan et al. 2011; Koppelmans et al. 2012; McDonald et al. 2013; Menning et al. 2015). Prior to chemotherapy, previous reports did not show atrophy in breast cancer patients compared to controls (See, McDonald and Saykin 2013; Joly et al. 2015; Saykin et al. 2013).

Cognitive and structural changes associated with chemotherapy in breast cancer patients are thus becoming well known. However, until now, a limited number of prospective longitudinal studies have been conducted in breast cancer patients that could help to distinguish the potential impact of the cancer disease process and of treatments including chemotherapy. Recent recommendations from the International Cancer and Cognition Task Force (ICCTF) highlighted the importance of conducting prospective longitudinal studies with pre- and post-chemotherapy neuroimaging (Deprez et al. 2018; Wefel et al. 2011). To our knowledge, only two studies have combined longitudinal investigations of cognitive abilities and their anatomical correlates (McDonald et al. 2013; Lepage et al. 2014). Additional prospective longitudinal investigations are needed in order to broaden understanding and to elucidate both cognitive and anatomical changes in breast cancer patients relative to controls.

A number of reviews (Ahles et al. 2012; Falleti et al. 2005) also highlighted that only some of the patients included in the studies experienced cognitive decline following chemotherapy. Therefore, there is a need for identification of risk factors for cognitive impairment in breast cancer patients (Ahles and Saykin 2007; Hurria et al. 2007). Among the impact of potential factors, state anxiety (Morel et al. 2015; Ramalho et al. 2017) and cognitive reserve (Ahles 2012; Ahles et al. 2010) seem to play an important role in cognitive abilities not only after cancer diagnosis, but also during the recovery phase following chemotherapy. The links between levels of anxiety and cognitive impairment (Wefel et al. 2004a, b; Hermelink et al. 2007; Ando-Tanabe et al. 2014) or GM atrophy (Inagaki et al. 2004, p. 200; Nakano et al. 2002; López Zunini et al. 2012) have been previously investigated, although leading to inconsistent results. The impact of cognitive reserve in breast cancer patients is less known. One of the proxies of cognitive reserve is education level and previous results suggest that some participants with high education level could cope better with the impact of cancer on cognition (Ahles et al. 2012). This assumption is concordant with positive associations between higher level of education and better cognitive performances, previously described in normal and pathological aging (Fratiglioni and Wang 2007; Hindle et al. 2014; Meng and D’Arcy 2012). In addition, previous reports have suggested interactions between cognitive reserve and depression or anxiety on cognition (McLaren et al. 2015), particularly with respect to hippocampal vulnerability to glucocorticoids (Freret et al. 2015). To our knowledge, such impact of an interaction between education level, as a marker of cognitive reserve, and anxiety on cognitive impairment and anatomical changes has not yet been studied in breast cancer patients.

In this context, our main aim was to describe the longitudinal cognitive and anatomical changes associated with the disease process and with chemotherapy in breast cancer patients. Our secondary aim was to assess the impact of education level, and its interaction with anxiety, on cognitive performances and GM atrophy in these patients. We expect cognitive impairments in breast cancer patients, not only following chemotherapy, but also after cancer diagnosis. Given previous findings, we expect to find GM atrophy after chemotherapy, possibly persisting 1 year after the end of chemotherapy. Finally, we predict that after chemotherapy, cognitive decline and GM atrophy will be related to education and anxiety levels.

Participants and methods

Participants

The patient group consisted of women who had been newly diagnosed with breast cancer, and the control group consisted of healthy volunteers matched on education, age and sex. Informed consent was obtained from all individual participants included in the study. All procedures performed in the current study that involved human participants were approved and in accordance with the local ethical standards research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Patient selection is displayed in Fig. 1. Among the 23 patients enrolled in the study, three did not participate in every MRI session and were thus excluded from further analyses. In total, the patient group consisted of 20 women with breast cancer.

Fig. 1
figure1

Distribution of women newly diagnosed with breast cancer across MRI scanning sessions

Patient inclusion criteria were (i) less than 70 years old, (ii) no metastatic breast cancer, (iii) already undergone surgical treatment for the cancer (tumorectomy or mastectomy) and scheduled to receive chemotherapy (fluorouracil, epirubicin, cyclophosphamide and docetaxel), followed by hormone therapy if necessary, (iv) no major psychiatric disorder according to DSM-IV criteria (Diagnostic and Statistical Manual of Mental Disorders) before or during breast cancer diagnosis, (v) no neurological disease, (vi) no drug use or alcohol abuse, and (vii) no overall cognitive impairment according to the criteria of the Mini-Mental Status Examination, i.e. participants with a score less than 26 out of 30 were not eligible (Folstein et al. 1975; Kalafat et al. 2003).

The control group consisted of 27 women without history of cancer. Inclusion criteria were the same as for patients, excluding cancer history (past or present). All participants were fluent in French and right-handed.

Examination of the participants took place at the Cyceron research center (Caen, France). The pretreatment assessment occurred after surgery, but before initiation of adjuvant therapy (T1). Follow-up assessments were conducted 1 month (T2) and 1 year (T3) after the end of chemotherapy for the patient group. Assessments for the control group were frequency matched to the interval for the chemotherapy patients (only T1 and T3).

Brief description of the analyses conducted

A brief explanation of the steps of analyses conducted is given below. For details regarding each analysis, see the Behavioral analyses and the Anatomical MRI analyses paragraphs:

Firstly, we conducted a main effects analysis for both behavioral and anatomical MRI data. Regarding behavioral data, we choose to use a linear mixed models (LME) approach that has several advantages over the more commonly used repeated-measures analysis of variance (ANOVA) (See Behavioral analyses paragraph). Regarding anatomical MRI analyses (groups comparisons and longitudinal evaluation for patients), the significance level was set at p < 0.05 FWE (family-wise error) corrected for multiple comparisons both at the voxel- and cluster-levels (See Anatomical MRI analyses paragraph).

Secondly, we performed a last exploratory analysis based on education level for both behavioral and anatomical MRI data. This analysis was considered exploratory because of the sample sizes of our groups. In addition, we used a covariate of interest in MRI statistical models (i.e. anxiety, STAI-A) that can limit the chances of finding significant results. Therefore, for MRI analyses, we choose to use a less stringent threshold (p < 0.001 uncorrected for multiple comparisons, instead of FWE). However, in order to avoid false negative results, the Monte Carlo approach has been used to consider only significant results corrected for multiple comparisons at the cluster level (See Behavioral analyses paragraph).

Behavior

The assessment battery included measures of psychopathological status, and standardized neuropsychological tests (Table 1).

Table 1 Description of assessments for psychopathological status and standardized neuropsychological tests divided in cognitive domains

Psychopathological assessments

Depression was assessed with the Beck Depression Inventory Short Form (BDI-SF), which is a validated 13-item self-report instrument developed to measure severity of depressive symptoms (Furlanetto et al. 2005). A higher score is representative of larger depressive symptoms. A total score (sum of all items) greater than 9–10 suggests the presence of clinically relevant depressive symptoms. However, Furlanetto and colleagues indicated that if a high specificity is desired, a 13–14 cut-off score is warranted. BDI-SF scores were below the clinical cut-off for pathological depression (high specificity, 13–14).

Anxiety was measured with the State-Trait Anxiety Inventory, which is a widely used to measure state and trait symptoms of anxiety (Spielberger and Sydeman 1994). The first 20 items assess state anxiety (STAI-A), or how the participant feels right now; the second 20 items assess trait anxiety (STAI-B), or how the participant generally feels. Scores for state and trait version were reported separately with higher values indicating higher anxiety symptoms. In the current study, six patients and one control were above the cut-off of 39–40 that has been suggested to detect clinically significant symptoms for the State Anxiety scale (Addolorato et al. 1999; Knight et al. 1983).

Neuropsychological assessments

To obtain more robust proxies of cognitive abilities and minimize the issue of multiple statistical testing, scores obtained with the following neuropsychological assessments were computed in five composite cognitive scores using the procedure detailed in La Joie et al. (2014). Such composite scores have been previously used in other studies in breast cancer patients (Lange et al. 2014, 2016). Firstly, performances were transformed into z scores, on the basis of the means and standard deviations of the scores of the control group (i.e. scores of controls at T1 were used for T1 scores of patients and scores of controls at T3 were used for T3 scores of patients). One can argue that results would be limited in highlighting normal changes over time by using each time point to calculate z-scores instead of using scores of controls at baseline (T1). However, we decided to follow this approach because control participants were not assessed at T2 and we wanted to limit the effects of repetition in our results. In addition, considering the age of some participants (close to 70 years old), we also wanted to limit the findings related to a potential cognitive decline.

Secondly, after z-score transformations, we combined z-scores into the five following cognitive domains (note that before averaging, z-scores derived from completion times and errors were reversed, such that increasing values always indicate better performances): encoding and retrieval in episodic memory (verbal and visual), working memory, executive functions and processing speed.

Verbal episodic memory was evaluated with the Encoding, Storage, Retrieval (ESR) paradigm (Eustache et al. 1998). Participants are instructed to recall as many words as possible from two distinct 16-word lists after either a superficial or a deep encoding phase. The sum of the two immediate free recalls is used as the measure of verbal episodic memory. Visual episodic memory was evaluated with an adaptation of the “Batterie d’Efficience Mnésique” (Signoret and Benoit 1991, BEM 144). This test includes 12 graphic signs from the BEM 144 memory battery and new items to mimic the conditions of the ESR paradigm. Two lists of eight items are presented to the participants, after either a superficial or a deep encoding phase. Immediately after this presentation, participants have to recall as many items as possible (immediate free recall). The sum of the two immediate free recalls corresponds to the visual episodic memory score. Importantly, ESR and BEM tests assess both encoding and retrieval, while most of previous episodic memory investigations assessed or reported only retrieval results. Indeed, depending of the nature of encoding (incidental or intentional), difficulty in retrieval will vary: intentional encoding will lead to better performance than incidental encoding regarding free recall, whereas recognition will be similar in both types of encoding.

Working memory was assessed with the digit span (WAIS-III, forward and backward subtests, Wechsler 1997) to evaluate attentional span and working memory. In the forward subtest, a list of numbers should be repeated in the correct order immediately after presentation. The backward subtest involves recalling numbers in the reverse order.

Executive functions were evaluated with the trail making test (TMT) (Reitan 1979) and verbal fluency (Cardebat et al. 1990) to assess alternating attention and inhibitory control. The TMT includes two parts (A and B). In part A, participants are asked to link numbered points randomly distributed on a sheet of paper in ascending order according to numbers. In part B, they have to link numbers and letters alternately. The outcome measures are the following: 1) difference between time needed to perform part B and part A (Time B-A); 2) the number of errors during part B and the total scores in 2 min. Verbal fluency requires participants to say as many words as possible from a specific category in 2 min. In the current study, the semantic category was animals and the phonemic category was words beginning with the “p” letter. Processing speed was calculated as corresponding to the time needed to complete part A of the TMT.

Behavioral analyses

Statistical analyses were done using R 3.1.2 software (www.r-project.org).

For comparisons of two (unpaired) groups on the socio-demographic and psychological variables, the t-test was used if the normality hypothesis could not be rejected by the Shapiro-Wilk test. For non-Gaussian samples, preference was given to the Wilcoxon-Mann-Whitney test.

The longitudinal neuropsychological assessments were analyzed using linear mixed model (LME) with selection of the time series typical AR(1) error structure. Heteroscedasticity was integrated into the residual variance function. Tuckey HSD test has been used for post hoc analysis in case of group or time effect. Mixed-effects modelling has several advantages over the more commonly used repeated-measures analysis of variance (ANOVA) (Bagiella et al. 2000; Detry and Ma 2016; Gueorguieva and Krystal 2004; Krueger and Tian 2004). Case-wise deletion of missing observations is not necessary, which allows for the analysis of all available data. It is also a superior approach to handling the correlation structures of repeated measures nested within participants, which circumvents the need to make adjustments for heteroscedascity and sphericity assumption violations. Several factors including age, education, state anxiety and behavioral performances were also entered as fixed-factors to test for relationship between behavioral performances and psycho-social variables. Indeed, LME models allow a systematic approach to incorporate both fixed-effect and random-effect terms to deal with the categorical grouping factor, between-subject baseline differences in the multiple measures, and the correlational structure among the predictor variables (Koerner and Zhang 2017).

To assess whether a variable had a significant effect, we followed the approach of Pinheiro and Bates (“Linear Mixed-Effects Models” 2000) and compared models with and without the respective variable by means of a likelihood ratio test (LRT), among other secondary results (e.g. AIC, BIC). The statistical result of this model comparison is represented as the likelihood ratio L.r. value and the associated p value. When the LRT indicated a significant effect of a variable, the coefficients of the model were further examined and represented as the t and p values associated with each variable tested.

We performed three separated analyses using LME approach. Firstly, we conducted analyses to compare neuropsychological performances in the five cognitive domains over time (T1 and T3) between controls and breast cancer patients. Secondly, we assessed cognitive performances and state anxiety (STAI-A) over time (T1, T2, T3) in breast cancer patients. Finally, in order to investigate interaction between groups and education level, the control and patient groups were both separated in two clusters based on education level. For this purpose, the standard k-means method was used to summarize both groups’ (patients and controls) data into two sets of homogeneous groups called clusters (Lloyd 1957; MacQueen et al. 1967). Comparisons between clusters were performed on neuropsychological performances and state anxiety (STAI-A) at T1 (four clusters; two in patients and two in controls), T2 (two clusters in patients only) and T3 (four clusters; two in patients and two in controls) separately. Such cluster analyses have been previously used to evaluate cognitive performances of sub-groups of participants (Schnakenberg Martin et al. 2016; Souza et al. 2016).

For all analyses, statistical significance was set at p ≤ 0.05.

Neuroimaging

MRI acquisition

For each participant, a high-resolution T1-weighted anatomical image was acquired on a Philips Achieva 3 T scanner using a three-dimensional fast-field echo sequence (sagittal; repetition time, 20 ms; echo time, 4.6 ms; flip angle, 20°; 170 slices; slice thickness, 1 mm; field of view, 256 × 256 mm2; matrix, 256 × 256).

Anatomic MRI preprocessing

For voxel-based morphometry (VBM) analysis, we used the CAT12 toolbox (C. Gaser, Structural Brain Mapping group, Jena University Hospital, Jena, Germany) implemented in SPM12 (Statistical Parametric Mapping, Institute of Neurology, London, UK). All T1- weighted images were corrected for bias field inhomogeneities, then spatially normalized using the DARTEL algorithm (Ashburner 2007) and segmented into GM, white matter (WM) and cerebrospinal fluid (CSF) (Ashburner and Friston 2005). The segmentation process was further extended by accounting for partial volume effects (Tohka et al. 2004), applying adaptive maximum a posteriori estimations (Rajapakse et al. 1997) and using a hidden Markov Random Field model (Cuadra et al. 2005). Following these steps, we performed the modulation step on the normalized, segmented images that allows comparisons of the absolute amount of tissue such as “volume” of gray matter (Good et al. 2001) in order to compensate for the effects of spatial normalization. After pre-processing (and in addition to visual checking for artefacts) all scans passed through an automated quality check protocol. Finally, the images were smoothed with a smoothing kernel of 8 mm (FWHM).

Anatomical MRI analyses

The modulated, smoothed, normalized GM maps were subjected to statistical parametric mapping on a voxel-by-voxel basis. As for the neuropsychological assessments, we performed three separated analyses.

Firstly, in order to evaluate group differences in GM volume between controls and breast cancer patients, two sample t-tests were used at T1 and T3. In addition, a flexible design was used with subject and time as factors, to determine if local GM volume modifications over time (T1, T3) differed between patients and controls. The contrast examining greater GM decline from T1 to T3 in patients than in controls was built by typing [1 -1 -1 1] in the SPM contrast manager.

Secondly, in order to assess longitudinal modifications of GM volume in patients, especially related to chemotherapy effects, a flexible design, with subject and time as factors, was used to determine if local GM volume differed over time (T1, T2, T3) in breast cancer patients. In the longitudinal investigation, persistence versus recovery of atrophy at T3 was evaluated using contrasts previously described in McDonald et al. (2010). Persistence contrast (i.e. atrophy at T2 persisting at T3) was built by typing [2 -1 -1] in the SPM contrast manager and the recovery contrast (i.e. atrophy at T2 recovered at T3) was built by typing [1 -2 1] in the SPM contrast manager.

For these first two analyses, nuisance covariates included years of education considering our working hypothesis (i.e. impact of education level on GM atrophy), and total intracranial volume (TIV) in order to correct for the whole brain size. Anxiety and age did not differ between groups and were thus not used as nuisance covariates. For statistical comparisons, the significance level was set at p < 0.05 FWE (family-wise error) corrected for multiple comparisons. In order to avoid spurious findings driven by noise, we applied a cluster extent threshold of 20.

Finally, in order to investigate interactions between groups and education level on GM volume, comparisons between clusters of participants were made at each time point of assessments separately with four clusters (two in patients and two in controls) at T1 and T3, and two clusters (patients only) at T2. State anxiety (STAI-A) and neuropsychological scores were added as covariates of interest for the interaction with group factor. Total intracranial volume (TIV) was added as a nuisance covariate to correct for whole brain sizes. Given the exploratory nature of these analyses, a statistical threshold of p < 0.001 uncorrected was used, with a corrected cluster extent threshold of 170 voxels. This combination of activation thresholds was determined using a Monte Carlo simulation program (AlphaSim, AFNI), with a cumulative proportion criterion of less than 0.05 (Cox 1996) to correspond to a false positive rate of p < 0.05, corrected for multiple comparisons.

Results

The demographic and clinical characteristics of the breast cancer patients and controls are displayed in Table 2. No significant difference has been found at baseline indicating good matching between groups.

Table 2 Demographic, clinical and psychological characteristics of patients with breast cancer and healthy controls at T1

Between-group and within-group analyses

Neuropsychological assessments

The main results are displayed in Table 3. A main group effect was found for verbal episodic memory retrieval (L.r = 21.47; p < 0.0001) and executive functions (L.r = 7.05; p = 0.0079). Moreover, group effect interacted significantly both with education level and state anxiety for verbal episodic memory retrieval (L.r = 12.71; p = 0.013) and with education level for executive functions (L.r = 6.76; p = 0.034), indicating that both state anxiety and education level influenced performances. This part is investigated later on using cluster analyses (see “Education level effect” section of the Results). Patients had more negative z-scores than controls at T1 and T3 for verbal memory retrieval (ps < 0.0001). Regarding executive functions, patients tended to have more positive z-scores than controls at T1 (p = 0.063) and had more negative z-scores than controls at T3 (p = 0.014) indicating lower performances in patients compared to controls for verbal memory retrieval at T1 and T3 and for executive functions at T3 only. Patients and controls did not differ on other neuropsychological performances.

Table 3 Adjusted domain scores by group and across assessment time points and models p values associated

Anatomical MRI

The main results are displayed in Table 4. Supplemental material, including additional results from MRI analyses, is provided as Table 7 (in the Supplemental material section). The results did not show any significant difference when direct comparisons were made between patients and controls at T1 and T3. However, group by time interaction investigation showed reduced GM volume in patients compared to controls at T3 relative to T1 in the left middle frontal gyrus, anterior cingulate cortex and right inferior occipital gyrus. In addition, within-group comparisons (See Table 7, Supplemental materials) showed increased GM volume between T1 and T3 in patients in the right occipital inferior gyrus, left precentral cortex, right cerebellum crus 1, left cerebellum VIII, right postcentral and precentral cortex, left and right cingulum middle cortex, right supplementary motor area and left parietal inferior gyrus. Increased GM volumes between T1 and T3 were found in controls in the left frontal superior medial and superior orbital gyri.

Table 4 Regional gray matter (GM) changes between and within groups depicting number of voxels (k), t values (T) and MNI coordinates, at p < 0.05 FWE corrected

Within the patient group, at T2 relative to T1 (GM decline from T1 to T2 in Table 4), patients showed decreased GM volumes in the left cerebellum crus1 and right lingual gyrus. In addition, significant atrophy at T2 recovered at T3 (GM decline from T1 to T2, recovery at T3 in Table 4 and Fig. 2) was found in the left middle temporal, and left inferior parietal gyri, cerebellum crus 1 bilaterally and right cerebellum VI. Finally, atrophy at T2 persistent at T3 (GM decline from T1 to T2, persistent at T3 in Table 4 and Fig. 3) was found in the right supplementary motor area (SMA), right lingual gyrus and vermis VI.

Fig. 2
figure2

Regions showing significant GM atrophy at T2 recovered at T3 within the patient group at p < 0.05 FWE corrected with a cluster extent threshold of 20. The top image shows the cerebellum crus 1 bilaterally and left inferior parietal gyrus. The lower image on the left shows the left inferior parietal gyrus and the lower image on the right shows the left middle temporal gyrus. See Table 4 for detailed information

Fig. 3
figure3

Regions showing significant GM atrophy at T2 persistent at T3 within the patients group at p < 0.05 FWE corrected with a cluster extent threshold of 20. The image on the left shows the vermis VI and the image on the right shows the supplementary motor area (SMA) and the right lingual gyrus. See Table 4 for detailed information

Education level effect

Neuropsychological assessments

Clusters based on education level were built and further analyses were conducted on the four clusters (high and low education level in both groups) obtained, see Table 5 for a description of the demographic and psychological characteristics of group clusters based on education level. Cluster analyses have been conducted only for scores showing interaction with education level using LME (see Table 3). Supplemental material is provided as Table 8 and includes scores and statistical results for cluster analyses. Cluster by group interactions are described below for verbal episodic memory retrieval and executive functions.

Table 5 Demographic and psychological characteristics (Mean ± SD) of group clusters based on education level, for breast cancer and healthy controls at T1

Cluster by group interaction has been found at T1 and T3 both for verbal episodic memory retrieval (T1: L.r = 8.55; p = 0.0035; T3: L.r = 8.57; p = 0.0034) and executive functions (T1: L.r = 6.21; p = 0.013; T3: L.r = 4.13; p = 0.042), meaning that differences between clusters of patients were not the same than differences between clusters of controls.

At T1, regarding verbal episodic memory retrieval, post hoc results showed that clusters of controls (C-high versus C-low) and clusters of patients (P-high versus P-low) did not differ in performances (p = 0.45 and p = 0.31, respectively). Regarding executive functions, clusters of controls (C-high versus C-low) did not differ in performances (p = 0.88), while P-high performed better than P-low (p = 0.01).

At T2, a significant cluster effect was found for verbal episodic memory retrieval (t16 = −2.33; p = 0.033) and state anxiety (t18 = 2.13; p = 0.048), indicating that P-high had better performance (less negative z-scores) on verbal memory retrieval and higher anxiety scores compared to P-low at T2. Moreover, state anxiety significantly interacted with education clustering for verbal episodic memory retrieval (t16 = 2.53; p = 0.023) meaning that P-high had better performance (less negative verbal episodic memory retrieval z-scores) (p = 0.054) and greater state anxiety scores than P-low. Both clusters of patients based on education level did not differ on executive functions at T2 (t18 = 1.61; p = 0.13).

At T3, regarding verbal episodic memory retrieval, post hoc results revealed that C-high tended to perform better than C-low (p = 0.076). Post hoc results revealed significant differences in performance between clusters of patients: P-high performed better than P-low (p = 0.041). Regarding executive functions, clusters of controls (C-high versus C-low) and clusters of patients (P-high versus P-low) did not differ in terms of performance (p = 0.51 and p = 0.31, respectively). There was no significant cluster by group interaction for state anxiety at T1 and T3.

Anatomical MRI

Results did not show any difference in atrophy between the four clusters of participants (patients and controls) based on education level at T1 and T3. At T2, both clusters of patients differed in terms of GM volume (Table 6). Considering the interaction between education level and anxiety using mixed-models investigations, we used state anxiety as a covariate of interest in interaction with group factor. Results revealed that P-high showed GM atrophy in the left posterior hippocampus compared to P-low (Fig. 4).

Table 6 Regional gray matter (GM) changes and correlations based on education level clustering at T2 using state anxiety as a covariate of interest, depicting number of voxels (k), t values (T) and MNI coordinates, at p < 0.001 uncorrected with a corrected cluster extent threshold of 170 voxels
Fig. 4
figure4

Regions showing significant GM atrophy at T2 in the left hippocampus in P-high compared to P-low at at p < 0.001 uncorrected with a corrected cluster extent threshold of 170 voxels. See Table 6 for detailed information

Correlations between neuropsychological scores and GM volumes at T2 were performed either with or without state anxiety as a covariate of interest. For P-high, executive function scores were positively correlated with GM volumes in cerebellum VIII bilaterally. This correlation was significant irrespective of whether state anxiety was entered as covariate of interest or not, suggesting no impact of state anxiety on executive function performances in P-high. In P-low, verbal memory recall scores were negatively correlated with GM volumes in right cerebellum VIII and left parahippocampal gyrus. This correlation was only significant when using state anxiety as covariate of interest; however, when state anxiety was removed from the design, this correlation was no longer present. This result indicates that state anxiety impacted verbal memory retrieval in P-low. No other significant correlations were found between neuropsychological scores and GM volumes.

Discussion

Our study aimed to describe the longitudinal cognitive and anatomical changes associated with disease process and with chemotherapy in breast cancer patients. We also wished to assess the impact of education level, in interaction with anxiety, on cognitive performances and GM atrophy in patients.

Briefly, patients had lower performances than controls on verbal episodic memory retrieval at T1 and T3, and on executive functions at T3. Longitudinal investigations revealed greater GM decline between T1 and T3 in patients than in controls in several regions (middle frontal and inferior occipital gyri, anterior cingulate cortex). Similarly, within-patient group comparisons revealed GM decline from T1 to T2 in the vermis and lingual gyrus, the latter persisting at T3. In contrast, we found a larger number of regions showing atrophy at T2 followed by recovery at T3 (i.e., temporal and parietal gyri, cerebellum). Finally, clustering based on education level analyses showed better performances on verbal episodic memory retrieval at T2, higher anxiety scores and hippocampal atrophy related to anxiety in P-high compared to P-low. Results also revealed significant correlations between GM densities in cerebellum and neuropsychological performances (executive functions for P-high and verbal memory retrieval for P-low).

At baseline (T1), performances were lower in patients compared to controls on verbal memory retrieval. However, neither group differed in GM volumes. These results confirm previous reports showing memory deficits (Wefel et al. 2004a, b; López Zunini et al. 2012; Ahles et al. 2012; Lange et al. 2014) despite a lack of GM atrophy before chemotherapy (López Zunini et al. 2012; McDonald et al. 2010, 2013; McDonald and Saykin 2013; Lepage et al. 2014; Joly et al. 2015). This lack of atrophy in breast cancer patients concurrent with verbal memory impairments at baseline confirms that cognitive deficits may be linked to yet unknown cancer disease effects unrelated to GM changes. The cancer disease may affect other brain mechanisms that can be revealed using fMRI instead of aMRI. As an example, previous findings have revealed prefrontal hyperactivation during a working memory task before chemotherapy, suggesting compensatory processes (McDonald et al. 2012). Future studies on brain connectivity during cognitive tasks are needed at baseline to well understand this process.

Neuropsychological deficits in patients were found at T3 for verbal memory retrieval and executive functions partially concordant with previous reports (Lyon et al. 2016). Indeed, some neuropsychological investigations have shown possible recovery 1 year or more after chemotherapy (Ahles et al. 2010; Lepage et al. 2014; Zheng et al. 2014), but not systematically (Jim et al. 2012; Collins et al. 2014; Pereira et al. 2015). More specifically, previous reports have shown an acute effect of chemotherapy on verbal ability still present 3 months (Quesnel et al. 2009) and 1 year (Ahles et al. 2010) after chemotherapy, but that resolved over time, from 18 to 36 months following cancer diagnosis in breast cancer receiving treatments (Zheng et al. 2014). We suppose that deficits already present at baseline may be exacerbated by chemotherapy and thus persist at T3, since chemotherapy is known to have strong effects on these cognitive domains (Deprez et al. 2012; Lange et al. 2014; Quesnel et al. 2009) and that these effects can be long lasting, even 20 years after chemotherapy (Koppelmans et al. 2012). The assumption according to deficits at T1 may be exacerbated by chemotherapy at T2 and thus persist at T3 is supported by anatomical results. Indeed, although direct comparison between groups at T3 did not reveal any difference, we found greater GM decline between T1 and T3 in patients than in controls in several regions (middle frontal and inferior occipital gyri, anterior cingulate cortex). Similarly, within-patient group comparisons revealed GM decline from T1 to T2 in the vermis and lingual gyrus, the latter persisting at T3. This atrophy partly confirms previous reports (McDonald et al. 2010, 2013) showing larger GM atrophy in breast cancer patients than in controls in frontal gyrus 1 month after chemotherapy compared to baseline. Therefore, although cancer-related cognitive deficits do not seem to be associated with anatomical changes, persistence of cognitive deficits at T3 was found in the current study and could be related to GM atrophy in occipital areas, as the right lingual gyrus appears to be involved in object naming (Kiyosawa et al. 1996) and verbal episodic memory retrieval (Dupont et al. 2001; Wiggs et al. 1998).

In contrast, we found a larger number of regions showing atrophy at T2 followed by recovery at T3 (i.e., temporal and parietal gyri, cerebellum). More specifically, GM in the left cerebellum crus 1 declined from T1 to T2 and recovered at T3. The cerebellum crus 1 is known to be part of the executive network (Habas et al. 2009; O’Reilly et al. 2010; Buckner et al. 2011). Thus, this result is concordant with the executive function differences found between groups at T3 (p = 0.069) which was less obvious than for episodic retrieval differences.

Additional contrasts reported in Table 7 (Supplemental materials) highlighted increase in GM volumes from T1 to T3 particularly in patients (mainly in cerebellum and occipital regions) and to some extent in controls (in frontal regions). These results were unexpected considering previous findings consistently showing GM decrease in BC patients following chemotherapy (McDonald et al. 2010, 2013). However, these studies either did not assess (McDonald et al. 2013) or did not report (McDonald et al. 2010) within group comparisons for GM volumes between baseline and 1 year after the end of chemotherapy. Therefore, our results need to be replicated and future prospective studies in BC patients should report within-groups comparisons for this purpose. Although a definitive explanation for this phenomenon is lacking, we propose several hypotheses regarding increase GM volumes over time. This GM increase (e.g. cerebellum in patients) could be a compensative process to account for atrophy elsewhere as cerebellum has been previously cited in relationship with compensation in functional neuroimaging studies (Cheng et al. 2017; O’Farrell et al. 2013). Regarding the mechanisms that may explain this phenomenon, it has been shown that changes in dendritic length are sufficient to increase GM volumes without necessarily reflecting addition of synapses (Anderson 2011), which could have led to cognitive changes. These additional results should be taken with caution and warrant further investigations.

Interestingly, we did not find any performance deficits in processing speed in contrast to previous studies (Ahles et al. 2010; Lepage et al. 2014). This discrepancy may be explained by the tests used to assess this ability that differed between the current study and others. Particularly, processing speed scores were based on one assessment (TMT, Time of part A), while other studies used multiple tests combined in one z-score (Ahles et al. 2010; Lepage et al. 2014). Therefore, assessments in the current study may be less sensitive than in previous investigations. It is important to note that longitudinal investigations of anatomical changes in breast cancer patients compared to controls are scarce and most results have been reported using low levels of correction (i.e., p < 0.001 uncorrected for multiple comparisons in Lepage et al. 2014; McDonald et al. 2010, 2013; Nudelman et al. 2016), while the current study reports main effects with an FWE correction enabling to correct for false discoveries. Altogether, our results, using correction for false discoveries and a longitudinal design, confirms previous findings suggesting atrophies following chemotherapy with partial recovery 1 year or more after treatment in breast cancer survivors (Bergouignan et al. 2011; de Ruiter et al. 2012; McDonald and Saykin 2013; Lepage et al. 2014; Joly et al. 2015; Nudelman et al. 2016). Our results also extend previous ones suggesting possible links between GM atrophy and cognitive ability changes over time in breast cancer patients.

Previous reports have shown that subgroups of breast cancer patients could emerge in relationship to various psychosocial factors that predict long term effects of chemotherapy (Vardy 2009; Ahles et al. 2010; Ahles 2012; Joly et al. 2015; Ramalho et al. 2017). Chemotherapy is known to affect only some of the patients included in studies (Falleti et al. 2005) and partial recovery is particularly dependent on age and education level (Ahles et al. 2010) and anxiety (Ramalho et al. 2017). Considering this literature and the strong interaction between education level and pathology in our neuropsychological results, we performed additional investigations using clustering for both groups based on education level.

Results revealed that P-high were more anxious than P-low at T2. P-high may be more aware of the cognitive deficits induced by chemotherapy due to their environment and, therefore, may have larger anxiety to develop cognitive problems in contrast to P-low (Arndt et al. 2014; Schagen et al. 2012). This larger anxiety in P-high at T2 was linked with a significant atrophy in the left posterior hippocampus in P-high compared to P-low. Previous reports have shown either no interaction between hippocampal atrophy in breast cancer patients and depressive/anxiety symptoms (Bergouignan et al. 2011; Inagaki et al. 2004) or significantly smaller hippocampal volume in participants with a history of distressing cancer-related recollections (Nakano et al. 2002). Thus, our results confirm and extend those of Nakano et al. (2002) by showing an interaction between education level and anxiety on GM volume in the left posterior hippocampus. Previous reports have shown that hypersecretion of cortisol in the hippocampus following hyperactivation of the hypothalamo-pituitary adrenal (HPA) axis could lead to volume alterations (Travis et al. 2016). We can suppose that, in the current study, higher education level could have led to larger anxiety that may have influenced hippocampal atrophy in P-high. Although anxiety is considered as one of the most important psychopathological comorbidities of cancer patients (Frick et al. 2007), cortisol secretion alterations in breast cancer patients are not clear (Schmidt et al. 2016; Zeitzer et al. 2014). It would be worth investigating in future studies the link between cortisol as a marker of anxiety and hippocampal atrophy in breast cancer patients to confirm our hypothesis.

Numerous studies have shown the crucial role of hippocampus in episodic memory storage and retrieval (Eichenbaum 2017). Especially, the left posterior hippocampus is known to be more involved in verbal episodic memory retrieval (Greicius et al. 2003; Viard et al. 2012). Therefore, considering posterior left hippocampal atrophy in P-high, we expected to find lower performances on the verbal memory task. However, results revealed that P-high performed better on the verbal memory task than P-low. In order to further understand this result, we performed correlations between neuropsychological scores and GM volumes.

Regarding verbal memory retrieval, at T2, we found significant negative correlations in P-low in the right cerebellum VIII and left parahippocampal gyrus. This result was present only using state anxiety as a covariate of interest and thus taking into account anxiety differences between both subgroups of patients. Regarding executive functions, performances were not significantly different between clusters at T2 and results showed a positive correlation in P-high in the cerebellum VIII bilaterally, irrespective of using state anxiety as a covariate of interest or not. Cognitive scores did not correlate with the same regions than those showing atrophy at T2 (i.e. hippocampus). This result suggests that performances were not directly related to hippocampal atrophy and that other mechanisms could be implicated. Especially, cognitive reserve in P-high may have led to strategies of compensation as shown in aging, for example (Franzmeier et al. 2018). This result is concordant with previous reports showing that post-treatment performance decline is related to cognitive reserve in breast cancer patients (Ahles et al. 2010). However, to our knowledge, our findings are the first to report an interaction between education level and state anxiety. In line with this hypothesis, a recent report, using resting state fMRI, revealed that greater functional connectivity between the hippocampus and cerebellum was negatively correlated to prospective memory scores in breast cancer patients (Cheng et al. 2017). Cheng and colleagues proposed that this result could represent a maladaptive compensation mechanism in breast cancer patients. This identification of factors, which may compensate cognitive impairment following chemotherapy, is critical for patient care and quality of life.

One of the limitations of the current study is the lack of assessment of controls and/or patients that did not receive chemotherapy at T2. Yet, the longitudinal investigation in the patient group at T1 and T3 still provides important results regarding chemotherapy effects on GM atrophy particularly, and extends previous reports. In addition, our secondary analysis, related to education level, provides preliminary findings showing an interaction between anxiety and education level on GM atrophy and performances in breast cancer patients following chemotherapy. However, given the small sample sizes of our groups, future studies are needed to confirm and/or extend our results.

In conclusion, using a longitudinal design, our results confirm that certain GM atrophies following chemotherapy in breast cancer patients persist 1 year after chemotherapy, while others recover, depending on the cortical areas. Increase in GM volumes after chemotherapy were unexpected and warrant further investigations. Interestingly, we found specific interactions between anxiety and education level both on neuropsychological scores and GM volumes. In patients with higher education level, greater anxiety was associated with hippocampal atrophy. However, hippocampal atrophy was not related to lower performances on verbal episodic memory suggesting the occurrence of compensation mechanisms that may be linked to cognitive reserve. Further functional MRI investigations in breast cancer patients, using for example the hippocampus as seed region, could provide insightful prospects on the connectivity between different large-scale networks.

Abbreviations

VBM:

Voxel Based Morphometry

GM:

Gray Matter

ICCTF:

International Cancer and Cognition Task Force

MRI:

Magnetic Resonance Imaging

LME:

Linear Mixed Model

FEW:

Family-wise error

STAIA:

State-Trait Anxiety

BDI-SF:

Beck Depression Inventory – Short Form

ESR:

Encoding, Storage, Retrieval

BEM:

Battérie D’Efficience Mnésique

TMT:

Trail Making Test

LRT:

Likehood Ratio Test

AIC:

Akaike Information Criterion

BIC:

Bayesian Information Criterion

WM:

White Matter

CSF:

Cerebro-spinal Fluid

FWHM:

Full Width at Half Maximum

TIV:

Total intracranial volume

MMSE:

Mini Mental State Examination

SD:

Standard Deviation

SMA:

Supplementary Motor Area

MNI:

Montreal Neurological Institute

References

  1. Addolorato, G., Ancona, C., Capristo, E., Graziosetto, R., Di Rienzo, L., Maurizi, M., & Gasbarrini, G. (1999). State and trait anxiety in women affected by allergic and vasomotor rhinitis. Journal of Psychosomatic Research, 46(3), 283–289.

  2. Ahles, T. A. (2012). Brain vulnerability to chemotherapy toxicities. Psycho-Oncology, 21(11), 1141–1148. https://doi.org/10.1002/pon.3196.

  3. Ahles, T. A., & Saykin, A. J. (2007). Candidate mechanisms for chemotherapy-induced cognitive changes. Nature Reviews Cancer, 7(3), 192–201. https://doi.org/10.1038/nrc2073.

  4. Ahles, T. A., Saykin, A. J., McDonald, B. C., Furstenberg, C. T., Cole, B. F., Hanscom, B. S., et al. (2008). Cognitive function in breast cancer patients prior to adjuvant treatment. Breast Cancer Research and Treatment, 110(1), 143–152. https://doi.org/10.1007/s10549-007-9686-5.

  5. Ahles, T. A., Saykin, A. J., McDonald, B. C., Li, Y., Furstenberg, C. T., Hanscom, B. S., et al. (2010). Longitudinal assessment of cognitive changes associated with adjuvant treatment for breast cancer: impact of age and cognitive reserve. Journal of Clinical Oncology: Official Journal of the American Society of Clinical Oncology, 28(29), 4434–4440. https://doi.org/10.1200/JCO.2009.27.0827.

  6. Ahles, T. A., Root, J. C., & Ryan, E. L. (2012). Cancer- and cancer treatment-associated cognitive change: an update on the state of the science. Journal of Clinical Oncology: Official Journal of the American Society of Clinical Oncology, 30(30), 3675–3686. https://doi.org/10.1200/JCO.2012.43.0116.

  7. Anderson, B. J. (2011). Plasticity of gray matter volume: the cellular and synaptic plasticity that underlies volumetric change. Developmental Psychobiology, 53(5), 456–465. https://doi.org/10.1002/dev.20563.

  8. Ando-Tanabe, N., Iwamitsu, Y., Kuranami, M., Okazaki, S., Yasuda, H., Nakatani, Y., et al. (2014). Cognitive function in women with breast cancer receiving adjuvant chemotherapy and healthy controls. Breast Cancer (Tokyo, Japan), 21(4), 453–462. https://doi.org/10.1007/s12282-012-0405-7.

  9. Arndt, J., Das, E., Schagen, S. B., Reid-Arndt, S. A., Cameron, L. D., & Ahles, T. A. (2014). Broadening the cancer and cognition landscape: the role of self-regulatory challenges. Psycho-Oncology, 23(1), 1–8. https://doi.org/10.1002/pon.3351.

  10. Ashburner, J. (2007). A fast diffeomorphic image registration algorithm. NeuroImage, 38(1), 95–113. https://doi.org/10.1016/j.neuroimage.2007.07.007.

  11. Ashburner, J., & Friston, K. J. (2005). Unified segmentation. NeuroImage, 26(3), 839–851. https://doi.org/10.1016/j.neuroimage.2005.02.018.

  12. Bagiella, E., Sloan, R. P., & Heitjan, D. F. (2000). Mixed-effects models in psychophysiology. Psychophysiology, 37(1), 13–20.

  13. Beck, A. T., Epstein, N., Brown, G., & Steer, R. A. (1988). An inventory for measuring clinical anxiety: Psychometric properties. Journal of Consulting and Clinical Psychology, 56(6), 893.

  14. Bergouignan, L., Lefranc, J. P., Chupin, M., Morel, N., Spano, J. P., & Fossati, P. (2011). Breast cancer affects both the hippocampus volume and the episodic autobiographical memory retrieval. PLoS One, 6(10), e25349. https://doi.org/10.1371/journal.pone.0025349.

  15. Buckner, R. L., Krienen, F. M., Castellanos, A., Diaz, J. C., & Yeo, B. T. T. (2011). The organization of the human cerebellum estimated by intrinsic functional connectivity. Journal of Neurophysiology, 106(5), 2322–2345. https://doi.org/10.1152/jn.00339.2011.

  16. Cardebat, D., Doyon, B., Puel, M., Goulet, P., & Joanette, Y. (1990). Evocation lexicale formelle et sémantique chez des sujets normaux. Performances et dynamiques de production en fonction du sexe, de l’âge et du niveau d’étude. Acta Neurologica Belgica, 90(4), 207–217.

  17. Cheng, H., Li, W., Gong, L., Xuan, H., Huang, Z., Zhao, H., et al. (2017). Altered resting-state hippocampal functional networks associated with chemotherapy-induced prospective memory impairment in breast cancer survivors. Scientific Reports, 7, 45135. https://doi.org/10.1038/srep45135.

  18. Cimprich, B., Reuter-Lorenz, P., Nelson, J., Clark, P. M., Therrien, B., Normolle, D., et al. (2010). Prechemotherapy alterations in brain function in women with breast cancer. Journal of Clinical and Experimental Neuropsychology, 32(3), 324–331. https://doi.org/10.1080/13803390903032537.

  19. Collins, B., Mackenzie, J., Tasca, G. A., Scherling, C., & Smith, A. (2014). Persistent cognitive changes in breast cancer patients 1 year following completion of chemotherapy. Journal of the International Neuropsychological Society: JINS, 20(4), 370–379. https://doi.org/10.1017/S1355617713001215.

  20. Cox, R. W. (1996). AFNI: software for analysis and visualization of functional magnetic resonance neuroimages. Computers and Biomedical Research, An International Journal, 29, 162–173. https://doi.org/10.1006/cbmr.1996.0014.

  21. Cuadra, M. B., Cammoun, L., Butz, T., Cuisenaire, O., & Thiran, J.-P. (2005). Comparison and validation of tissue modelization and statistical classification methods in T1-weighted MR brain images. IEEE Transactions on Medical Imaging, 24(12), 1548–1565. https://doi.org/10.1109/TMI.2005.857652.

  22. de Ruiter, M. B., Reneman, L., Boogerd, W., Veltman, D. J., Caan, M., Douaud, G., et al. (2012). Late effects of high-dose adjuvant chemotherapy on white and gray matter in breast cancer survivors: converging results from multimodal magnetic resonance imaging. Human Brain Mapping, 33(12), 2971–2983. https://doi.org/10.1002/hbm.21422.

  23. Deprez, S., Amant, F., Smeets, A., Peeters, R., Leemans, A., Van Hecke, W., et al. (2012). Longitudinal assessment of chemotherapy-induced structural changes in cerebral white matter and its correlation with impaired cognitive functioning. Journal of Clinical Oncology: Official Journal of the American Society of Clinical Oncology, 30(3), 274–281.

  24. Deprez, S., Kesler, S. R., Saykin, A. J., Silverman, D. H. S., de Ruiter, M. B., & McDonald, B. C. (2018). International cognition and cancer task force recommendations for neuroimaging methods in the study of cognitive impairment in non-CNS cancer patients. Journal of the National Cancer Institute, 110(3), 223–231. https://doi.org/10.1093/jnci/djx285.

  25. Detry, M. A., & Ma, Y. (2016). Analyzing repeated measurements using mixed models. JAMA, 315(4), 407–408. https://doi.org/10.1001/jama.2015.19394.

  26. Dupont, S., Samson, Y., Van de Moortele, P.-F., Samson, S., Poline, J.-B., Adam, C., et al. (2001). Delayed verbal memory retrieval: a functional MRI study in epileptic patients with structural lesions of the left medial temporal lobe. NeuroImage, 14(5), 995–1003. https://doi.org/10.1006/nimg.2001.0908.

  27. Eichenbaum, H. (2017). Prefrontal-hippocampal interactions in episodic memory. Nature Reviews. Neuroscience, 18(9), 547–558. https://doi.org/10.1038/nrn.2017.74.

  28. Eustache, F., Desgranges, B., & Lalevée, C. (1998). [Clinical evaluation of memory]. Revue Neurologique, 154(Suppl 2), S18–S32.

  29. Falleti, M. G., Sanfilippo, A., Maruff, P., Weih, L., & Phillips, K.-A. (2005). The nature and severity of cognitive impairment associated with adjuvant chemotherapy in women with breast cancer: a meta-analysis of the current literature. Brain and Cognition, 59(1), 60–70. https://doi.org/10.1016/j.bandc.2005.05.001.

  30. Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12(3), 189–198.

  31. Franzmeier, N., Hartmann, J., Taylor, A. N. W., Araque-Caballero, M. Á., Simon-Vermot, L., Kambeitz-Ilankovic, L., et al. (2018). The left frontal cortex supports reserve in aging by enhancing functional network efficiency. Alzheimer’s Research & Therapy, 10(1), 28. https://doi.org/10.1186/s13195-018-0358-y.

  32. Fratiglioni, L., & Wang, H.-X. (2007). Brain reserve hypothesis in dementia. Journal of Alzheimer’s disease: JAD, 12(1), 11–22.

  33. Freret, T., Gaudreau, P., Schumann-Bard, P., Billard, J.-M., & Popa-Wagner, A. (2015). Mechanisms underlying the neuroprotective effect of brain reserve against late life depression. Journal of Neural Transmission (Vienna, Austria: 1996), 122(Suppl 1), S55–S61. https://doi.org/10.1007/s00702-013-1154-2.

  34. Frick, E., Tyroller, M., & Panzer, M. (2007). Anxiety, depression and quality of life of cancer patients undergoing radiation therapy: A cross-sectional study in a community hospital outpatient centre. European Journal of Cancer Care, 16(2), 130–136. https://doi.org/10.1111/j.1365-2354.2006.00720.x.

  35. Furlanetto, L. M., Mendlowicz, M. V., & Romildo Bueno, J. (2005). The validity of the Beck depression inventory-short form as a screening and diagnostic instrument for moderate and severe depression in medical inpatients. Journal of Affective Disorders, 86(1), 87–91. https://doi.org/10.1016/j.jad.2004.12.011.

  36. Good, C. D., Johnsrude, I. S., Ashburner, J., Henson, R. N., Friston, K. J., & Frackowiak, R. S. (2001). A voxel-based morphometric study of ageing in 465 normal adult human brains. NeuroImage, 14(1 Pt 1), 21–36. https://doi.org/10.1006/nimg.2001.0786.

  37. Greicius, M. D., Krasnow, B., Boyett-Anderson, J. M., Eliez, S., Schatzberg, A. F., Reiss, A. L., & Menon, V. (2003). Regional analysis of hippocampal activation during memory encoding and retrieval: fMRI study. Hippocampus, 13(1), 164–174. https://doi.org/10.1002/hipo.10064.

  38. Gueorguieva, R., & Krystal, J. H. (2004). Move over ANOVA: progress in analyzing repeated-measures data and its reflection in papers published in the archives of general psychiatry. Archives of General Psychiatry, 61(3), 310–317. https://doi.org/10.1001/archpsyc.61.3.310.

  39. Habas, C., Kamdar, N., Nguyen, D., Prater, K., Beckmann, C. F., Menon, V., & Greicius, M. D. (2009). Distinct cerebellar contributions to intrinsic connectivity networks. The Journal of Neuroscience: The Official Journal of the Society for Neuroscience, 29(26), 8586–8594. https://doi.org/10.1523/JNEUROSCI.1868-09.2009.

  40. Hermelink, K., Untch, M., Lux, M. P., Kreienberg, R., Beck, T., Bauerfeind, I., & Münzel, K. (2007). Cognitive function during neoadjuvant chemotherapy for breast cancer: results of a prospective, multicenter, longitudinal study. Cancer, 109(9), 1905–1913. https://doi.org/10.1002/cncr.22610.

  41. Hindle, J. V., Martyr, A., & Clare, L. (2014). Cognitive reserve in Parkinson’s disease: a systematic review and meta-analysis. Parkinsonism & Related Disorders, 20(1), 1–7. https://doi.org/10.1016/j.parkreldis.2013.08.010.

  42. Hurria, A., Somlo, G., & Ahles, T. (2007). Renaming “chemobrain”. Cancer Investigation, 25(6), 373–377. https://doi.org/10.1080/07357900701506672.

  43. Inagaki, M., Matsuoka, Y., Sugahara, Y., Nakano, T., Akechi, T., Fujimori, M., et al. (2004). Hippocampal volume and first major depressive episode after cancer diagnosis in breast cancer survivors. The American Journal of Psychiatry, 161(12), 2263–2270. https://doi.org/10.1176/appi.ajp.161.12.2263.

  44. Jim, H. S. L., Phillips, K. M., Chait, S., Faul, L. A., Popa, M. A., Lee, Y.-H., et al. (2012). Meta-analysis of cognitive functioning in breast cancer survivors previously treated with standard-dose chemotherapy. Journal of Clinical Oncology: Official Journal of the American Society of Clinical Oncology, 30(29), 3578–3587. https://doi.org/10.1200/JCO.2011.39.5640.

  45. Joly, F., Giffard, B., Rigal, O., De Ruiter, M. B., Small, B. J., Dubois, M., et al. (2015). Impact of cancer and its treatments on cognitive function: advances in research from the Paris international cognition and cancer task force symposium and update since 2012. Journal of Pain and Symptom Management, 50(6), 830–841. https://doi.org/10.1016/j.jpainsymman.2015.06.019.

  46. Kalafat, M., Hugonot-Diener, L., & Poitrenaud, J. (2003). Standardisation et étalonnage français du “mini mental state”(MMS) version GRECO. Revue de Neuropsychologie, 13(2), 209–236 Accessed 2 June 2014.

  47. Kiyosawa, M., Inoue, C., Kawasaki, T., Tokoro, T., Ishii, K., Ohyama, M., et al. (1996). Functional neuroanatomy of visual object naming: a PET study. Graefe’s Archive for Clinical and Experimental Ophthalmology = Albrecht Von Graefes Archiv Fur Klinische Und Experimentelle Ophthalmologie, 234(2), 110–115.

  48. Knight, R. G., Waal-Manning, H. J., & Spears, G. F. (1983). Some norms and reliability data for the state--trait anxiety inventory and the Zung self-rating depression scale. The British Journal of Clinical Psychology, 22(Pt 4), 245–249.

  49. Koerner, T. K., & Zhang, Y. (2017). Application of linear mixed-effects models in human neuroscience research: a comparison with Pearson correlation in two auditory electrophysiology studies. Brain Sciences, 7(3), 26. https://doi.org/10.3390/brainsci7030026.

  50. Koppelmans, V., de Ruiter, M. B., van der Lijn, F., Boogerd, W., Seynaeve, C., van der Lugt, A., et al. (2012). Global and focal brain volume in long-term breast cancer survivors exposed to adjuvant chemotherapy. Breast Cancer Research and Treatment, 132(3), 1099–1106. https://doi.org/10.1007/s10549-011-1888-1.

  51. Krueger, C., & Tian, L. (2004). A comparison of the general linear mixed model and repeated measures ANOVA using a dataset with multiple missing data points. Biological Research for Nursing, 6(2), 151–157. https://doi.org/10.1177/1099800404267682.

  52. La Joie, R., Landeau, B., Perrotin, A., Bejanin, A., Egret, S., Pélerin, A., et al. (2014). Intrinsic connectivity identifies the hippocampus as a main crossroad between Alzheimer’s and semantic dementia-targeted networks. Neuron, 81(6), 1417–1428. https://doi.org/10.1016/j.neuron.2014.01.026.

  53. Lange, M., Giffard, B., Noal, S., Rigal, O., Kurtz, J.-E., Heutte, N., et al. (2014). Baseline cognitive functions among elderly patients with localised breast cancer. European Journal of Cancer (Oxford, England: 1990), 50(13), 2181–2189. https://doi.org/10.1016/j.ejca.2014.05.026.

  54. Lange, M., Heutte, N., Rigal, O., Noal, S., Kurtz, J.-E., Lévy, C., et al. (2016). Decline in cognitive function in older adults with early-stage breast cancer after adjuvant treatment. The Oncologist. https://doi.org/10.1634/theoncologist.2016-0014.

  55. Lepage, C., Smith, A. M., Moreau, J., Barlow-Krelina, E., Wallis, N., Collins, B., et al. (2014). A prospective study of grey matter and cognitive function alterations in chemotherapy-treated breast cancer patients. SpringerPlus, 3, 444. https://doi.org/10.1186/2193-1801-3-444.

  56. Lloyd, S. (1957). Least squares quantization in PCM’s bell telephone labs.

  57. López Zunini, R. A., Scherling, C., Wallis, N., Collins, B., Mackenzie, J., Bielajew, C., & Smith, A. M. (2012). Differences in verbal memory retrieval in breast cancer chemotherapy patients compared to healthy controls: a prospective fMRI study. Brain Imaging and Behavior. https://doi.org/10.1007/s11682-012-9213-0.

  58. Lyon, D. E., Cohen, R., Chen, H., Kelly, D. L., Starkweather, A., Ahn, H.-C., & Jackson-Cook, C. K. (2016). The relationship of cognitive performance to concurrent symptoms, cancer- and cancer-treatment-related variables in women with early-stage breast cancer: a 2-year longitudinal study. Journal of Cancer Research and Clinical Oncology, 142(7), 1461–1474. https://doi.org/10.1007/s00432-016-2163-y.

  59. MacQueen, J., & others. (1967). Some methods for classification and analysis of multivariate observations. In Proceedings of the fifth Berkeley symposium on mathematical statistics and probability (Vol. 1, pp. 281–297). Oakland, CA, USA.

  60. McDonald, B. C., & Saykin, A. J. (2013). Alterations in brain structure related to breast cancer and its treatment: chemotherapy and other considerations. Brain Imaging and Behavior, 7(4), 374–387. https://doi.org/10.1007/s11682-013-9256-x.

  61. McDonald, B. C., Conroy, S. K., Ahles, T. A., West, J. D., & Saykin, A. J. (2010). Gray matter reduction associated with systemic chemotherapy for breast cancer: a prospective MRI study. Breast Cancer Research and Treatment, 123(3), 819–828. https://doi.org/10.1007/s10549-010-1088-4.

  62. McDonald, B. C., Conroy, S. K., Ahles, T. A., West, J. D., & Saykin, A. J. (2012). Alterations in brain activation during working memory processing associated with breast cancer and treatment: a prospective functional magnetic resonance imaging study. Journal of Clinical Oncology, 30(20), 2500–2508. https://doi.org/10.1200/JCO.2011.38.5674.

  63. McDonald, B. C., Conroy, S. K., Smith, D. J., West, J. D., & Saykin, A. J. (2013). Frontal gray matter reduction after breast cancer chemotherapy and association with executive symptoms: a replication and extension study. Brain, Behavior, and Immunity, 30(Suppl), S117–S125. https://doi.org/10.1016/j.bbi.2012.05.007.

  64. McLaren, M. E., Szymkowicz, S. M., Kirton, J. W., & Dotson, V. M. (2015). Impact of education on memory deficits in subclinical depression. Archives of Clinical Neuropsychology, 30(5), 387–393. https://doi.org/10.1093/arclin/acv038.

  65. Meng, X., & D’Arcy, C. (2012). Education and dementia in the context of the cognitive reserve hypothesis: a systematic review with meta-analyses and qualitative analyses. PLoS One, 7(6), e38268. https://doi.org/10.1371/journal.pone.0038268.

  66. Menning, S., de Ruiter, M. B., Veltman, D. J., Koppelmans, V., Kirschbaum, C., Boogerd, W., et al. (2015). Multimodal MRI and cognitive function in patients with breast cancer prior to adjuvant treatment--the role of fatigue. NeuroImage. Clinical, 7, 547–554. https://doi.org/10.1016/j.nicl.2015.02.005.

  67. Morel, N., Dayan, J., Piolino, P., Viard, A., Allouache, D., Noal, S., et al. (2015). Emotional specificities of autobiographical memory after breast cancer diagnosis. Consciousness and Cognition, 35, 42–52.

  68. Nakano, T., Wenner, M., Inagaki, M., Kugaya, A., Akechi, T., Matsuoka, Y., et al. (2002). Relationship between distressing cancer-related recollections and hippocampal volume in cancer survivors. The American Journal of Psychiatry, 159(12), 2087–2093. https://doi.org/10.1176/appi.ajp.159.12.2087.

  69. Nudelman, K. N. H., McDonald, B. C., Wang, Y., Smith, D. J., West, J. D., O’Neill, D. P., et al. (2016). Cerebral perfusion and gray matter changes associated with chemotherapy-induced peripheral neuropathy. Journal of Clinical Oncology: Official Journal of the American Society of Clinical Oncology, 34(7), 677–683. https://doi.org/10.1200/JCO.2015.62.1276.

  70. O’Farrell, E., MacKenzie, J., & Collins, B. (2013). Clearing the air: a review of our current understanding of “chemo fog”. Current Oncology Reports, 15(3), 260–269. https://doi.org/10.1007/s11912-013-0307-7.

  71. O’Reilly, J. X., Beckmann, C. F., Tomassini, V., Ramnani, N., & Johansen-Berg, H. (2010). Distinct and overlapping functional zones in the cerebellum defined by resting state functional connectivity. Cerebral Cortex (New York, N.Y.: 1991), 20(4), 953–965. https://doi.org/10.1093/cercor/bhp157.

  72. Pereira, S., Fontes, F., Sonin, T., Dias, T., Fragoso, M., Castro-Lopes, J. M., & Lunet, N. (2015). Neurological complications of breast cancer: a prospective cohort study. Breast (Edinburgh, Scotland), 24(5), 582–587. https://doi.org/10.1016/j.breast.2015.05.006.

  73. Pinheiro, J. C., & Bates, D. M. (2000). Linear mixed-effects models: Basic concepts and examples. In Mixed-effects models in S and S-PLUS (pp. 3–56). New York: Springer. https://doi.org/10.1007/0-387-22747-4_1.

  74. Quesnel, C., Savard, J., & Ivers, H. (2009). Cognitive impairments associated with breast cancer treatments: results from a longitudinal study. Breast Cancer Research and Treatment, 116(1), 113–123. https://doi.org/10.1007/s10549-008-0114-2.

  75. Rajapakse, J. C., Giedd, J. N., & Rapoport, J. L. (1997). Statistical approach to segmentation of single-channel cerebral MR images. IEEE Transactions on Medical Imaging, 16(2), 176–186. https://doi.org/10.1109/42.563663.

  76. Ramalho, M., Fontes, F., Ruano, L., Pereira, S., & Lunet, N. (2017). Cognitive impairment in the first year after breast cancer diagnosis: a prospective cohort study. Breast (Edinburgh, Scotland), 32, 173–178. https://doi.org/10.1016/j.breast.2017.01.018.

  77. Reitan, R. M. (1979). Trail making test: TMT. Göttingen: Testzentrale.

  78. Saykin, A. J., de Ruiter, M. B., McDonald, B. C., Deprez, S., & Silverman, D. H. S. (2013). Neuroimaging biomarkers and cognitive function in non-CNS cancer and its treatment: current status and recommendations for future research. Brain Imaging and Behavior, 7(4), 363–373. https://doi.org/10.1007/s11682-013-9283-7.

  79. Schagen, S. B., Das, E., & Vermeulen, I. (2012). Information about chemotherapy-associated cognitive problems contributes to cognitive problems in cancer patients. Psycho-Oncology, 21(10), 1132–1135. https://doi.org/10.1002/pon.2011.

  80. Schmidt, M. E., Semik, J., Habermann, N., Wiskemann, J., Ulrich, C. M., & Steindorf, K. (2016). Cancer related fatigue shows a stable association with diurnal cortisol dysregulation in breast cancer patients. Brain, Behavior, and Immunity, 52, 98–105. https://doi.org/10.1016/j.bbi.2015.10.005.

  81. Schnakenberg Martin, A. M., Bonfils, K. A., Davis, B. J., Smith, E. A., Schuder, K., & Lysaker, P. H. (2016). Compared to high and low cannabis use, moderate use is associated with fewer cognitive deficits in psychosis. Schizophrenia Research. Cognition, 6, 15–21. https://doi.org/10.1016/j.scog.2016.09.001.

  82. Signoret, J. L., & Benoit, N. (1991). [Examination and memory]. La Revue du Praticien, 41(10), 866–868.

  83. Souza, C. P., Oliveira, G. N., Foss, M. P., & Tumas, V. (2016). Cluster analysis of cognitive performance in a sample of patients with Parkinson’s disease. Dementia & Neuropsychologia, 10(4), 315–319. https://doi.org/10.1590/s1980-5764-2016dn1004010.

  84. Spielberger, C. D., & Sydeman, S. J. (1994). State-trait anxiety inventory and state-trait anger expression inventory.

  85. Stewart, A., Bielajew, C., Collins, B., Parkinson, M., & Tomiak, E. (2006). A meta-analysis of the neuropsychological effects of adjuvant chemotherapy treatment in women treated for breast cancer. The Clinical Neuropsychologist, 20(1), 76–89. https://doi.org/10.1080/138540491005875.

  86. Tohka, J., Zijdenbos, A., & Evans, A. (2004). Fast and robust parameter estimation for statistical partial volume models in brain MRI. NeuroImage, 23(1), 84–97. https://doi.org/10.1016/j.neuroimage.2004.05.007.

  87. Travis, S. G., Coupland, N. J., Hegadoren, K., Silverstone, P. H., Huang, Y., Carter, R., et al. (2016). Effects of cortisol on hippocampal subfields volumes and memory performance in healthy control subjects and patients with major depressive disorder. Journal of Affective Disorders, 201, 34–41. https://doi.org/10.1016/j.jad.2016.04.049.

  88. Vardy, J. (2009). Cognitive function in breast cancer survivors. Cancer Treatment and Research, 151, 387–419. https://doi.org/10.1007/978-0-387-75115-3_24.

  89. Viard, A., Desgranges, B., Eustache, F., & Piolino, P. (2012). Factors affecting medial temporal lobe engagement for past and future episodic events: an ALE meta-analysis of neuroimaging studies. Brain and Cognition, 80(1), 111–125. https://doi.org/10.1016/j.bandc.2012.05.004.

  90. Wechsler, D. (1997). WAIS-III, Wechsler adult intelligence scale: Administration and scoring manual. Psychological Corporation.

  91. Wefel, J. S., & Schagen, S. B. (2012). Chemotherapy-related cognitive dysfunction. Current Neurology and Neuroscience Reports, 12(3), 267–275. https://doi.org/10.1007/s11910-012-0264-9.

  92. Wefel, J. S., Kayl, A. E., & Meyers, C. A. (2004a). Neuropsychological dysfunction associated with cancer and cancer therapies: a conceptual review of an emerging target. British Journal of Cancer, 90(9), 1691–1696. https://doi.org/10.1038/sj.bjc.6601772.

  93. Wefel, J. S., Lenzi, R., Theriault, R., Buzdar, A. U., Cruickshank, S., & Meyers, C. A. (2004b). “Chemobrain” in breast carcinoma?: a prologue. Cancer, 101(3), 466–475. https://doi.org/10.1002/cncr.20393.

  94. Wefel, J. S., Saleeba, A. K., Buzdar, A. U., & Meyers, C. A. (2010). Acute and late onset cognitive dysfunction associated with chemotherapy in women with breast cancer. Cancer, 116(14), 3348–3356. https://doi.org/10.1002/cncr.25098.

  95. Wefel, J. S., Vardy, J., Ahles, T., & Schagen, S. B. (2011). International cognition and cancer task force recommendations to harmonise studies of cognitive function in patients with cancer. The Lancet Oncology, 12(7), 703–708. https://doi.org/10.1016/S1470-2045(10)70294-1.

  96. Wiggs, C. L., Weisberg, J., & Martin, A. (1998). Neural correlates of semantic and episodic memory retrieval. Neuropsychologia, 37(1), 103–118. https://doi.org/10.1016/S0028-3932(98)00044-X.

  97. Zeitzer, J. M., Nouriani, B., Neri, E., & Spiegel, D. (2014). Correspondence of plasma and salivary cortisol patterns in women with breast cancer. Neuroendocrinology, 100(0), 153–161. https://doi.org/10.1159/000367925.

  98. Zheng, Y., Luo, J., Bao, P., Cai, H., Hong, Z., Ding, D., et al. (2014). Long-term cognitive function change among breast cancer survivors. Breast Cancer Research and Treatment, 146(3), 599–609. https://doi.org/10.1007/s10549-014-3044-1.

Download references

Acknowledgements

Authors would like to thank the clinical research department of the Centre François Baclesse (Caen; Dr. Clarisse, Mrs. Rieux), the medical oncology department of the Centre François Baclesse (Dr Delcambre, Dr. Ollivier, Dr. Berthet, Dr. Segura and Dr. Swisters) for their help in patient recruitment, and the participants for their active contribution to these results.

Funding

This work was supported by the ARC foundation – for cancer research (2017–2020), the association “Ligue contre la cancer” (both national and Calvados department), the association “Cancer du sein, parlons en” (2011), and the Cancéropôle Nord-Ouest.

Author information

Correspondence to Joy Perrier.

Ethics declarations

Conflict of interest

The authors report no financial interests or potential conflict of interest.

Informed consent

Informed consent was obtained from all individual participants included in the study. All procedures performed in the current study that involved human participants were approved and in accordance with the local ethical standards research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

ESM 1

(DOCX 28 kb)

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Perrier, J., Viard, A., Levy, C. et al. Longitudinal investigation of cognitive deficits in breast cancer patients and their gray matter correlates: impact of education level. Brain Imaging and Behavior 14, 226–241 (2020). https://doi.org/10.1007/s11682-018-9991-0

Download citation

Keywords

  • Breast cancer
  • Cognition
  • Education level
  • Anxiety
  • Magnetic resonance imaging