| Objective1.To study the changes of hippocampal volume(after 1 month of radiotherapy)and cognitive function(after 6 months of radiotherapy)of nasopharyngeal carcinoma patients after IMRT radiotherapy,and analyze the relationship between hippocampal volume and cognitive impairment in radiotherapy avoiding hippocampus,so as to provide clinical basis for cognitive impairment in radiotherapy avoiding hippocampus,and to evaluate the predictive value of hippocampal volume in cognitive impairment in radiotherapy avoiding hippocampus.2.The changes of cognitive function and hippocampal damage in nasopharyngeal carcinoma patients in radiotherapy avoiding hippocampus were analyzed by imaging texture evidence,so as to further clarify the imaging evidence of hippocampal damage during avoiding radiotherapy and its relationship with cognition,and to evaluate the predictive value of hippocampal texture features in cognitive impairment after avoiding radiotherapy.Methods1.A total of 164 patients with nasopharyngeal carcinoma who received IMRT and underwent hippocampal avoidance protection were collected.The general data collected included age,sex,education level and pathological type.All subjects were assessed by Montreal Cognitive Assessment Chinese Version(Mo CA-B)at baseline and after 6 months of radiotherapy,and were scanned by MR,and the hippocampal volume at baseline and after 1 month of radiotherapy was collected and obtained.The subjects were divided into cognitive impairment group and non-cognitive impairment group.The general data,Mo CA-B score and hippocampal volume of the two groups were compared.Pearson correlation analysis was used to evaluate hippocampal volume and Mo CA-B score.Logistic regression and ROC curve were used to calculate the predictive ability of hippocampal volume on cognitive impairment.2.164 patients with nasopharyngeal carcinoma were the same as above,and 30 matched healthy control groups(HC group)were collected.All subjects were assessed by Montreal Cognitive Assessment Chinese Version(Mo CA-B)at baseline and 6months after radiotherapy,and the patients with nasopharyngeal carcinoma were divided into cognitive impairment group and cognitive non-impairment group.All subjects were scanned by MR,and the hippocampal volumes at baseline and one month after radiotherapy were collected and obtained.T1 WI was used as the sequence for texture feature analysis.The region of interest(ROI)of individual hippocampus is extracted based on the brain template of Montreal Neurological Institute(MNI).Texture features are calculated based on image omics developed by C++ and Delphi,and feature dimensionality is reduced by using intra-class correlation coefficients(ICCs),average direction,machine learning(random forest)and autocorrelation matrix.Univariate analysis of variance and generalized linear model were used to compare the differences between groups.Pearson correlation analysis was used to evaluate the relationship between important texture features and clinical data.Logistic regression and ROC curve were used to calculate the predictive ability of texture features to cognitive impairment.Results1.The Mo CA-B scale of nasopharyngeal carcinoma patients was 24.8 ±3.8 after6 months of IMRT radiotherapy avoiding hippocampus,which was significantly lower than the baseline period(26.6±4.3),and the difference was statistically significant(P<0.01).According to the reliable change index of cognitive assessment,164 patients were divided into 56 patients with cognitive impairment and 108 patients with cognitive impairment,with cognitive impairment accounting for 33.33% of the total subjects.Compared with cognitive impairment group,there are differences in age,sex and Mo CA-B score after 6 months of radiotherapy between the two groups(P<0.05),but there are no significant differences in education level,pathological types and baseline Mo CA-B score(P>0.05).However,after 6 months of treatment,the score of Mo CA-B in cognitive impairment group was significantly lower than that in baseline period(P<0.05).2.There was no significant difference in the hippocampal volume(total volume,left and right hippocampus)between the two groups at baseline(P > 0.05).The total volume of hippocampus,the volume of left hippocampus and the volume of right hippocampus in the non cognitive impairment group were higher than those in the cognitive impairment group one month after radiotherapy(P < 0.05).The volume of hippocampus in the non cognitive impairment group was not significantly abnormal compared with that in the baseline period(P > 0.05),and the volume of hippocampus(total volume,left hippocampus and right hippocampus)in the cognitive impairment group was decreased compared with that in the baseline period one month after radiotherapy(P < 0.05).3.The results of correlation analysis showed that there was a positive correlation between the total volume of hippocampus and Mo CA-B score one month after radiotherapy(r = 0.517,P < 0.05);The left and right hippocampal volumes were also positively correlated with Mo CA-B score(r = 0.475 and 0.551,respectively,P <0.05).4.Logistic regression showed that the total volume of hippocampus at 1 month after radiotherapy was statistically significant(OR = 1.994,P = 0.008);The area under ROC curve was 0.517,P = 0.109.5.There were significant differences in age,gender,hippocampal volume and Mo CA-B score after 6 months among the three groups(P < 0.05),but there were no differences in education level,hippocampal radiation dose,hippocampal volume at baseline and Mo CA-B score among the three groups(P > 0.05).Within group pairwise comparison results showed that there were significant differences in age,gender,hippocampal volume and Mo CA-B score after 6 months of radiotherapy between the two groups(P < 0.05),while there were no significant differences in education level and baseline Mo CA-B score between the two groups(P > 0.05).The hippocampal volume(1 month)and Mo CA-B score(6 months)in cognitive impairment group were significantly lower than those in baseline period(P < 0.05).6.The results of texture feature screening showed that there were significant differences in the five features between IMRT radiotherapy group and HC group(P <0.05).Further comparison of the above five features between cognitive impairment group and non cognitive impairment group showed that only glevnonu,rlnonuni and contrast showed significant differences between cognitive impairment group and non cognitive impairment group(P < 0.05),There was no significant difference in variance and entropy between the two groups(P > 0.05).7.The results of correlation analysis showed that there was a positive correlation between variance and radiation dose in hippocampus(r = 0.448,P < 0.05);There was a negative correlation between entropy and radiation dose in hippocampus(r =-0.461,P < 0.05);However,there was no correlation between glivnonu,rlnonuni and contrast and radiation dose in hippocampus(P > 0.05).Glevnonu and rlnun were positively correlated with Mo CA-B score(r = 0.503,P < 0.05;r=0.527,P<0.05);There was no correlation between the two groups(P > 0.05).In HC group,there was no significant correlation between age and glevnonu,rlnonuni and contrast(P > 0.05);There was no significant difference between male and female(P > 0.05).8.Logistic regression analysis showed that glevnonu(OR = 0.978,95% CI:0.950-1.008)and contrast(OR = 0.651,95% CI: 0.496-0.853)were independent predictors of hippocampal cognitive impairment after radiotherapy.The area under ROC curve was 0.636 and 0.592 respectively,and the combined AUC of the two indexes was 0.704.Conclusion1.Hippocampal protective radiation therapy can not completely avoid the occurrence of cognitive impairment,and some patients may have hippocampal volume atrophy one month after radiotherapy,which may be associated with cognitive impairment of nasopharyngeal carcinoma patients after hippocampal protective radiation therapy,but it can not be used to predict cognitive impairment.2.The results of hippocampal texture analysis can be used as microscopic imaging evidence of hippocampal injury in radiotherapy,and it can also effectively predict the occurrence of cognitive impairment. |
PDF Full Text Request