Altered Network Efficiency Of Functional Brain Net Works In Patients With Breast Cancer After Chemotherapy

Background: As a common malignant tumor,breast cancer is the third most common malignant disease and ranks first among female tumors.Although the comprehensive treatment based on chemotherapy(surgical treatment,chemotherapy,radiation therapy,endocrine therapy)has significantly improved the treatment effect of breast cancer and prolonged the survival of patients,chemotherapy has resulted in various kinds of side effects.In recent years,more and more studies have found that chemotherapy may be one of the causes of cognitive dysfunction.Since most patients with breast cancer have a long survival,their quality of life seems even more important.There are reports that about 20% to 50% of breast cancer survivors have multiple cognitive damage.The concept of chemotherapy-induced cognitive impairment(CICI)was first proposed in 1970 s,whichrefers a series of cognitive impairment in patients during or after cancer chemotherapy,including memory,reasoning,learning,attention,and executive function,information processing speed and visual spatial function etc.A large number of studies have shown that breast cancer CICI has significant clinical heterogeneity,and its clinical manifestations are diverse and the onset time and duration are different.This poses great challenges for the study of its pathogenesis,diagnosis and treatment.Several influential journals,including the Journal of Clinical Oncology(JCO)have published relevant findings.In 2014,CICI became one of the most important hot topicsin the National Comprehensive Cancer Network(NCCN).The pathogenesis of breast cancer CICI is explored from different perspectives.In the recent ten years,the rapid development of functional MRI has become an important method for the study of the mechanism of CICI.The first emerging methods are voxel-based morphometry(VBM)and diffusion tensor imaging(DTI).It has been found that the volume and microstructure of grey matter and white matter in the frontal,parahippocampal,thalamus,cingulate,and anterior prefrontal areas of breast cancer CICI have changed.This provides a direct anatomical basis for the pathogenesis of CICI,but its mechanism of brain network change is not clear.In recent years,the development of Resting-functional magnetic resonance imaging(R-f MRI)has provided a new approach for the study of the brain network mechanism of breast cancer CICI.Graph-based network analyses of the human brain has been proposed by Sporns in 2005.Since then great progress has been made in the brain network study in many central nervous system diseases(such as ADHD,schizophrenia and early Alzheimer’s disease),but it is still rare in the study of breast cancer CICI.Objective: To explore the mechanism of breast cancer CICI brain network,analyze the relationship between breast cancer CICI network and its clinical manifestations,to find effective clues for later study,and eventuallyprovide useful information for early detection and early diagnosis.In this study,we enrolled 28 patients with breast invasive ductal carcinoma(Breast cancer group,BC group)and 40 age-and education-matched healthy controls(healthy controls group,HC group).The 28 breast cancer patients were the first to receive chemotherapy,and the standard TAC regimen was used.A month after the end of the six course of chemotherapy,we carried out several tests in all subjects,including Digital span test(DST),Verbal fluency test(VFT),Mini Mental State Examination(MMSE),Prospectivememory(PM)and Retrospective memory(RM),source memory test,item memory test etc.And R-f MRI scans were performed in all patients.90 Regions of interest(ROIs)were analyzed by Graph-based network analysesmethod,including the global efficiency,the local efficiency,the node efficiency and the distribution of the core nodes.Pearson correlation analysis was also used to analyze the relationship between brain network changes and clinical neuropsychological variables.Methods:Results: 1.BC group:MMSE score 22 ~ 30(25.89±2.41);VFTscore 3 ~ 6.5(5.43±1.07);DST score 6 ~ 8(6.64±0.91);Item memoryscore 0.51 ~ 0.66(0.60±0.06);Sourcememoryscore 0.45 ~ 0.68(0.53±0.04);PMscore 9 ~ 27(20.07±6.16);RM score 8 ~ 26(20.79±4.86)。HC group:MMSE score 27 ~ 30(28.35±1.46);VFTscore 5 ~ 8(6.60±1.03);DST score 8 ~ 16(11.73±2.09);Item memoryscore 0.58 ~ 0.69(0.64±0.04);Sourcememoryscore 0.67 ~ 0.81(0.75±0.04);PMscore 8 ~ 20(13.43±3.49);RM score 8 ~ 20(12.45±3.08)。 There were no significant differences in age or education between the patient and control groups(both P > 0.05).However,the patients had significantly lower scores on the digit span test,the verbal fluency test,the Mini-Mental State Examination(MMSE)scores,and the source memory task compared to the HC(all P < 0.001).2.Both groups exhibited small-world organization of their functional brain networks over the whole sparsity range studied(i.e.,normalized global efficiency ~ 1 and normalized local efficiency > 1).Further statistical comparisons of the AUCs revealed that the patients showed significantlyincreased global(P = 0.007)and local(P < 0.001)efficiency compared with the HC.However,after the normalization by random networks,these two measures exhibited significantly lower values in the patients compared to the HC(both P < 0.001).3.In the HC group,we identified 14 hub regions,including 10 association cortex regions,two paralimbic cortex regions and two primary cortex regions.These hubs were mainly located in frontal/prefrontal(the bilateral middle frontal gyri,the opercular part of the right inferior frontal gyrus,the orbital part of the right inferior frontal gyrus,the right rolandic operculum,and the left medial superior frontal gyrus),parietal(the bilateral postcentral gyri,the bilateral inferior parietal,but the supramarginal and angular gyri and the bilateral supramarginal gyri),and temporal(the left superior temporal gyrus and the right temporal pole: superior temporal gyrus)regions.In the BC group,we identified fifteen hub regions,including eight association cortex regions,five paralimbic cortex regions,and two primary cortex regions.These hubs were predominantly located in parietal(the bilateral postcentral gyri,the bilateral superior parietal gyri,the left inferior parietal,but the supramarginal and angular gyri,the right angular gyrus,and the left precuneus),prefrontal(the orbital part of the bilateral superior frontal gyrus,the orbital part of the right middle frontal gyrus,and the bilateral gyrus rectus),and occipital(the bilateral superior occipital gyri and the right middle occipital gyrus)regions.Notably,only three regions were commonly identified as hubs in both groups,including the bilateral postcentral gyrus and the left inferior parietal,but the supramarginal and angular gyri).Further between-group comparisons revealed that nine regions showed increased nodal efficiency in the patients with BC compared to the HC.The regions included the orbital part of the bilateral superior frontal gyrus,the bilateral gyrus rectus,and the bilateral superior occipital gyrus,the left superior parietal gyrus and the left precuneus,and the right middle occipital gyrus.Notably,all these regions were identified as hubs in the patient group.There were no regions showing decreased nodal efficiency in the patients compared to the HCs.4.We identified two connected components that exhibited increased functional connectivity in the patients with BC compared with the HC(P < 0.001,corrected).The first component was mainly situated in parietal and occipital regions.The other waspredominantly located in prefrontal regions.It is worth mentioning that all the regions showing increased nodal efficiency described above were included in these two components.No components were found to show significantly decreased functional connectivity in the patients.5.Significantly positive correlations were observed between normalized global efficiency and the MMSE scores(r = 0.398,P = 0.036)and between normalized local efficiency and the source memory scores(r = 0.497,P = 0.01)in the patients.Of note,these correlations did not survive after correcting for multiple comparisons(P > 0.05,corrected by the False Discovery Rate procedure).No significant correlations were found between other network/connectivity measures andcognitive variables(all P > 0.05).Conclusion: Breast cancer CICI has a significant cognitive decline in multiple cognitive aspects.The abnormal breast cancer CICI brain network is characterized by the increase of node efficiency and functional connectivity.The absolute value of local efficiency and global efficiency is increased,but the local efficiency and global efficiency is decreased after standardization.In addition,partial network global indicators are positively related to neuropsychological variables.This suggests that there may be an association between abnormal brain network alterations and cognitive decline in breast cancer CICI.