Gender-disease interaction on brain cerebral metabolism in cancer patients with depressive symptoms

In the present study, 49 cancer patients (35 males, 14 females, age 50.918 ± 9.338) and 48 normal controls (NCs) (30 males, 18 females, age 50.604 ± 7.721) were recruited. PET images were gathered between August 2014 and December 2015 and the privacy of participants was guaranteed. The inclusion criteria for cancer patients were as follows: (a) age at 18 or older at the time of diagnosis, (b) ability to tolerate 18F-fluorodeoxyglucose injection, (c) no cancer type which could be found only in males or females, (d) well-informed about his or her own condition, (e) knew about his or her disease within 1 year, (f) absence of clear focal brain lesions or mental diseases, (g) did not undergo chemotherapy, (h) no more than 2 years of cancer. The NCs should have no psychiatric disorder or mental disease, were all confirmed with no depression or depressive symptom by the doctors. We used Beck Depression Inventory (BDI)-II to assess the mental stress of depressive emotion in cancer patients [8]. All participants were given written informed consent at the time of enrollment for PET image scanning, according to the Declaration of Helsinki (1991).

The demographic and clinical measures between cancer patients and NCs were listed in Table 1. The types of cancer patients were as follows: lung cancer (n = 9, female = 3), bowel cancer (n = 7, female = 2), lymph cancer (n = 7, female = 4), gastric cancer (n = 6, female = 1), renal cancer (n = 2, female = 0), esophagus cancer (n = 2, female = 0), nasopharynx cancer (n = 2, female = 1), other types of cancer (n = 14, female = 3). The cancer patients included 28 minimal depressions (score 1–13, mean value: 6.50; SD: 3.45), 11 mild depressions (score 14–19, mean value: 15.55; SD: 1.63) and 10 moderate/severe depressions (score 20–63, mean value: 25.20; SD: 3.49). Chi-square test and two-sample T test was used, respectively, to examine the difference in gender and age between the two groups. No significant difference were found in age and gender (p > 0.05).

PET images of all participants were gathered using a Siemens Biograph TruePoint 64 PET/CT (Siemens Healthcare, Erlangen, Germany) in three-dimensional mode at Affiliated Lanzhou General Hospital of Lanzhou Military Area Command. The image collection of coregistered CT and PET images was performed together. Preparation rules of all participants were strictly followed during the acquisition process. In order to keep the blood glucose level within the scope of 3.9–6.1 mmol/L, all the participants were required to fast for at least 6 h. Seven minutes scan was acquired in 40–60 min after intravenous injection of 3.7 MBq/kg (maximum dose 370 MBq) of 18F-FDG. In this study, the CT data on the combined scanner were used for PET attenuation correction. The FDG-PET data were reconstructed with ordered subset expectation maximization iterative algorithm.

All the images were preprocessed using the Statistical Parametric Mapping software (SPM8, http://www.fil.ion.ucl.ac.uk/spm/software ) in MATLAB 7.14, and were normalized to the standard image data for further analysis. The spatial normalization with a 12-parameter affine transformation was used, followed by nonlinear iterative spatial transformation. PET voxel-wise analysis was used to analyze the cerebral metabolism, and the voxel size after normalization was 1 mm × 1 mm × 1 mm. Finally, the processed images were smoothed by using 8 mm FWHM Gaussian kernel.

In order to explore the correlation between cerebral metabolism and depression state, we calculated the Pearson correlation coefficient voxel-by-voxel between the voxels restricted to significant clusters and BDI scores (a voxel-wise analysis) in the cancer group, and then counted the number of significant voxels that were related to the BDI scores. A false discovery rate (FDR) correction was performed at a p-value of 0.05 for multiple comparison correction. Locations of the significant brain regions were conducted with xjview Toolbox ( http://www.alivelearn.net/xjview/ ) and BrainNet Viewer Toolbox [9].

As is shown in Tables 2, 8 clusters with significant cancer main effect were found by using two-way ANOVA, in which the largest cluster located in the left superior frontal gyrus (BA 6;MNI coordinates:x = − 21, y = − 12, z = 63). Other clusters were located in right superior parietal gyrus, left supplementary motor area, left middle part of orbitofrontal gyrus, left inferior parietal gyrus, left superior frontal gyrus: medial, and left median cingulate and paracingulate gyrus. The clusters with significant cancer main effect were shown in Fig. 1. We also found 7 clusters with significant gender main effect, in which the largest cluster located in bilateral cerebellum posterior lobe (right MNI coordinates: x = 22, y = − 68, z = − 30, left MNI coordinates: x = − 17, y = − 67, z = − 32), other clusters were located in left inferior temporal gyrus, right cuneus, left median cingulate and paracingulate gyrus, left postcentral gyrus, and right superior frontal gyrus. We also found 4 clusters with significant gender-by-cancer interaction effects, which mainly located in bilateral superior frontal gyrus, right postcentral gyrus, left precuneus, and left superior frontal gyrus(the largest cluster, BA 6; MNI coordinates: x = − 20, y = − 12, z = 61). The information of clusters were shown in Fig. 1 and Table 2.

Results from post-hoc test showed significant decrease of metabolic intensity in superior frontal gyrus and postcentral of the cancer patients relative to controls. No significant hypermetabolic cluster was found in cancer patients. We found 5 clusters with significant decrease of metabolic intensity in female compared to male, which were located in right parahippocampal gyrus, right cingulate gyrus, left median cingulate and paracingulate gyrus, left anterior cingulate and paracingulate gyrus and right thalamus. Whereas, 8 clusters were found with significant increase of metabolic intensity in female compared to male that were located in right cuneus, bilateral inferior parietal lobules, right median cingulate and paracingulate gyrus, left postcentral, right precuneus, right supplementary motor area, and left superior occipital gyrus (see Table 3 and Fig. 2).

The correlation analysis showed significant correlation between BDI score and voxels within the clusters of cancer main effect, which mainly located in right superior parietal gyrus, middle orbitofrontal gyrus and left middle frontal gyrus (q < 0.05,FDR corrected). Voxels within the clusters of gender main effect that showed significant correlations with BDI score mainly located in right cuneus and left inferior temporal gyrus(q < 0.05, FDR corrected). The detailed information were shown in the Table 4 and Fig. 3.

Details of the regions affected by gender and cancer factors are shown in Table 2 and Fig. 1. Significant cancer main effects have been found primarily in the frontal lobe, parietal lobe and the cingulate gyrus. Post hoc analysis showed the hypermetabolic regions primarily located in the frontal gyrus and postcentral in cancer patients compared with NCs. Our results were consistent with the previous studies showing the metabolic abnormalities in cingulate gyrus, orbitofrontal gyrus and dorsolateral prefrontal gyrus in cancer patients [7, 17]. The abnormal metabolism in cingulate gyrus and dorsolateral prefrontal gyrus in cancer patients was consistent with the previous findings that cingulate gyrus and dorsolateral prefrontal gyrus were more objective and validate biomarkers in emotional processing and depression diagnosis [10]. Besides, these regions might be associated with the depth of depression [7]. The abnormal metabolic regions in the frontal lobe may help to explain the dysfunction of the brain activities after the diagnose of cancer [11]. Our results suggested that the cancer factor may result in the abnormal metabolic intensity in these regions which might be associated with the mental stress of depressive emotion.

As is shown in Table 2 and Fig. 1, significant gender main effects were found primarily in the bilateral cerebellum posterior lobe, the temporal lobe, cuneus and frontal lobe, with the hypermetabolic regions primarily located in the parahippocampal gyrus and the cingulate gyrus, and the hypometabolic regions primarily located in the cuneus and the parietal lobe in the male group compared with the female group. Previous studies showed that males have significant higher glucose metabolism in bliateral inferior temporal gyrus and frontal lobe, and lower glucose metabolism in patietal lobe, frontal lobe and cingulate gyrus, our results were in line with these findings [12]. Researchers have suggested that cerebellum may play an important role in the pathophysiology of MDD [18, 19]. Wu and Baeken showed that depressive patients with longer depression duration represented significant less metabolic activity in bilateral cerebellum posterior lobe [20], which supported our findings that the metabolic difference existed in cerebellum and the difference was related to depression. Additionally, gender difference in gray matter volumes in hippocampus, orbitofrontal and cingulate cortex were also reported [21‐23], which may serve as the structure basis of the gender-related metabolic abnormalities.

To our knowledge, interaction of gender and cancer factors on cerebral metabolic pathology remains unclear. A previous MRI study have shown gender-disease interaction effects in the frontal lobe in MDD patients, which were implicated in the dysfunctional regulation of mood and emotion [24]. Hence, the metabolic abnormality we found in frontal lobe and cuneus may represent disorder of brain metabolic activity based on the pathological function of these regions.

We computed correlation coefficients between the BDI scores and the metabolic intensity in the clusters affected by the gender and cancer factors. The results in Table 4 and Fig. 3 showed a relationship between the brain regions and the mental stress of depressive emotion in parietal lobe, the frontal lobe, the cuneus and the left temporal lobe. The depression-related regions include three clusters with cancer main effect and two clusters with gender main effect. No cluster with gender-by-cancer interaction effect was found significantly related to the mental stress of depressive emotion. In previous studies, researchers focused on the regional metabolism [7], which omitted the voxel-wise information. In this paper, we explored the relationship between voxel metabolic intensity and the mental stress of depressive emotion in cancer patients, and voxels with significant correlation were marked. In addition, the comparison results were corrected by FDR correction (p < 0.05, voxel level) instead of FWE correction (p < 0.05), as FWE correction may be excessively strict criterium, by which no voxel could survive from the correction.

Similar to the previous research, our results showed that the abnormal metabolism in the parietal lobe and frontal lobe was related to the depression state [7, 25‐28]. The results from two-way ANOVA results showed that the metabolism in parietal lobe and frontal lobe was affected by cancer factor. These results suggested that the abnormal metabolism in cancer patients starts at very mild stages in depression [7], which should be paid more attention to as a depressive indicator in clinical nursing care of cancer patients. Regions with gender main effects and abnormal metabolism were found in the cuneus and inferior temporal gyrus, the metabolic intensity of which were related to mental stress of depressive emotion. The previous studies reported that the metabolism in cuneus was positively correlated with severity of mental stress of depressive emotion [27, 28], and the inferior temporal gyrus was related to mental stress of depressive emotion when exposed to negative emotion activation [29]. The abnormal metabolism affected by cancer factor might pose as a risk factor of energy disorder or psychomotor retardation. Which may be the pathological cause of depression; and the gender differences in cerebral metabolism might reflect psychological resilience difference facing the diagnosis of cancer or preventing depression [11, 30‐33]. The present work complemented the study of gender differences in cerebral metabolism using the neuroimaging method in cancer patients, and provided support evidence for abnormal cerebral glucose metabolism in cancer patients with mental stress of depressive emotion.

There are some limitations that needs to be addressed in our future work. Firstly, we only used the BDI scale to evaluate the mental status of the cancer patients, but not for the NCs. The mental state (with/without depression) of the NCs was evaluated by psychiatrist. More comprehensive scale tests will performed and provided to fix this problem. Secondly, the sample size in present study is not enough, and varied cancer type may influence the analysis. A replication study needs to be performed in the future on more patients with single type of cancer to examine our findings. Thirdly, some other variables were not included in the present study (such as the coping styles of subjects and hormonal treatments). These limitations may result in a lack of power to demonstrate the relationship between the gender-cancer factors and the mental stress of depressive emotion. In future studies, the gender effects in cancer patients with mental stress of depressive emotion needs to be investigated on more scales and larger sample.