The Correlation And Preliminary Mechanisms Between Symptom Clusters And Cancer-related Cognitive Impairment In Patients With Colorectal Cancer

Objectives:In this study,the network analysis was used to determine the symptom clusters and the core symptom clusters of colorectal cancer patients.The relationship between the symptom clusters and cancer-related cognitive impairment(CRCI)among patients with colorectal cancer was explored.The potential molecular mechanism and metabolomic characteristics of this relationship were preliminary conducted.This study aimed to provide new ideas for the identification and intervention of symptom clusters and CRCI,and better improve the theoretical basis for symptom cluster management.Methods:(1)This cross-sectional study collected 174 patients with colorectal cancer in the medical oncology department of a tertiary hospital in Wuxi.The participants were selected by convenient sampling.This survey included demographic data,disease-related data,the Chinese version of the Memory Symptom Assessment Scale(MSAS-Ch),and the Functional Assessment of Cancer Treatment-Cognitive Function(FACT-Cog),Wechsler Adult Intelligence Scale-Ⅲ(WAIS-Ⅲ),Trail Making Test(TMT),and Hopkins Verbal Learning Test-Revised(HVLT-R).The descriptive statistical analysis presented the symptom incidence,severity score,distress score,subjective cognitive impairment score,and the neuropsychological score of colorectal cancer patients.Network analysis was used to explore the core symptom clusters of different symptom dimensions.Differences in symptom clusters with different cognitive functions were compared;the relationship between CRCI and symptom clusters was explored by Spearman’s correlation analysis.(2)The fasting blood samples of the subjects were collected,and the levels of oxidative stress factors and heparan sulfate(HS)were determined.Correlation analysis was used to explore the relationship among symptom clusters,CRCI,oxidative stress factors,HS concentration,and common clinical common blood test indexes.The linear relationship between symptom clusters and oxidative stress factors,HS,and common clinical blood indexes was explored by multiple regression analysis.By moderating effect model,restricted cubic spline regression,and causal mediation analysis,the preliminary mechanism of CRCI,symptom clusters,oxidative stress factors,HS,and common clinical blood indicators in colorectal cancer patients were explored.(3)The analysis of serum metabolite components was carried out by ultra-high performance liquid chromatography-mass spectrometry system.Metabolomics data were processed by QI software,and SIMCA-P+13.0 software was used for multivariate statistical analysis.The variable importance in projection(VIP)threshold of the first principal component of the OPLS-DA model was>1,and the P-value of the t-test result was<0.05 to screen for significant difference metabolites.Differential metabolites were compared and screened through the human metabolite database and the KEGG database.Then,the correlation analysis of differential metabolites and symptom clusters,CRCI was carried out by correlation analysis,and the mechanism of serum metabolomics was preliminarily explored.Results:(1)Fatigue(70.1%)was the most common symptom in 174 patients with colorectal cancer.The symptom with the highest symptom frequency score was the loss of interest in sex(3.21±1.01),the highest score for symptom severity was feeling"I don’t look like myself"(2.89±1.09),the symptom with the highest symptom distress score was difficulty urinating(3.13±1.10).The nature and structure of symptom interactions in the three symptom dimensions of symptom occurrence,severity,and distress were analyzed using pair Markov random field models.The results showed that the associations between symptoms in colorectal cancer patients were different owing to the symptom dimension that created the network.Among them,according to the evaluation of the centrality index,pain is a more important node in the three network structures.The symptom clusters of each dimension were determined by the Walktrap algorithm.There were 6 symptom clusters in the dimension of symptom occurrence:chemotherapy-related symptom cluster,nutritional symptom cluster,disease symptom cluster,digestive tract symptom cluster,neuropsychological symptom cluster,pain and sleep symptom cluster.There were 5 symptom clusters of symptom severity:chemotherapy-related symptom cluster,autonomic nervous symptom cluster,sickness symptom cluster,neuropsychological symptom cluster,and digestive tract symptom cluster.There were 5 symptom clusters for symptom distress:digestive tract symptom cluster,pain and sleep symptom cluster,respiratory system symptom cluster,neuropsychological symptom cluster,and other symptom clusters.Among them,the core symptom cluster of the symptom occurrence network was the neuropsychological symptom cluster.The core symptom cluster of the symptom severity network was the sickness symptom cluster.The core symptom cluster of the symptom distress network was the pain and sleep symptom cluster.The relationship between CRCI and each symptom cluster was negatively correlated(P<0.05).With a higher score in symptom cluster,cancer-related cognitive function was worse,(2)The severity of symptoms was closely related to HS and was positively correlated(r=0.275,P<0.05).The chemotherapy-related symptom cluster was significantly positively correlated with SOD.CRP,CA125,and CA199 were positively correlated with multiple symptom clusters.Cancer-related cognitive impairment score played a mediating role in the relationship between chemotherapy-related symptom clusters and HS,with an average causal mediating effect of 0.825(95%CI:0.16 to 1.71,P<0.05).(3)A total of 14 differential metabolites were found,including 6 phosphatidylcholines,2 phosphorylcholines,3 phosphoethanolamines,and 1 sphingomyelin.Among them,phosphatidylcholine was significantly positively correlated with sickness symptom cluster in symptom severity,neuropsychological symptom cluster in symptom distress,chemotherapyrelated symptom cluster in symptom occurrence,and sickness symptom cluster symptom occurrence.Metabolites such as phosphatidylethanolamine were associated with multiple symptom clusters.Conclusions:(1)There were differences in the composition of symptom clusters of different dimensions in patients with colorectal cancer,suggesting that multi-dimensional assessment of symptoms and symptom clusters should be emphasized in symptom management.Especially,the intervention and management of pain and sleep symptom clusters,and neuropsychological symptom clusters should be paid more attention.And these symptoms should be incorporated into a patient-specific assessment tool for patients with colorectal cancer.(2)Symptom clusters of all dimensions were negatively correlated with cancer-related cognitive function.Symptom management should pay attention to evaluating the relationship between cognitive impairment and symptom clusters in patients with colorectal cancer,and collaboratively formulate symptom intervention measures to explore the relationship between symptom clusters and CRCI,and the underlying mechanism of this relationship.(3)Oxidative stress factors and HS were closely related to symptom clusters and cognitive function.In the study of symptom clusters and cognitive impairment,the roles of HS and oxidative stress factors can be further explored to provide a scientific basis for symptom manage ment.(4)14 different metabolites were screened out in the serum of colorectal cancer patients with different cognitive functions.These metabolites were mainly closely related to neurodevelopment and cognitive function.Among them,phosphatidylcholine may be inextricably linked with symptom clusters and CRCI.Further researches need to explore the potential mechanism of phosphatidylcholine in symptom clusters and CRCI,providing ideas for formulating interventions for symptom clusters and CRCI in the future.