| Part 1 Genetic analysis of variants associated with frontotemporal degeneration in the PUMCH dementia cohort[Objective]Frontotemporal lobar degeneration(FTLD)is a common cause of early-onset neurodegenerative disease.Genetic factors play an important role in the pathogenesis of FTLD.Many genes have been recognized to be the causative gene of FTLD including MAPT,GRN,C9orf72 and so on.There are considerable geographical and ethnic variability in the distribution of genetic FTLD.The purpose of this study was to analyze the gene variants associated with FTLD in a Chinese single-center dementia cohort.[Methods]This study was a retrospective study,patients were enrolled consecutively from the PUMCH dementia cohort,the Dementia Clinic of the Department of Neurology of PUMCH between 2007 to 2019.Genetic testing was conducted using use next-generation sequencing.Gene variants associated with FTLD were identified and pathogenicity rating was performed for each variant according to 2015 ACMG guideline.[Results]In the PUMCH dementia cohort,a total of 1277 dementia patients completed nextgeneration sequencing,and 82 patients(6.4%)carried gene variants associated with FTLD.There were 27 patients with MAPT mutations,10 with GRN mutations,9 with TBK1 mutations,7 with CHMP2B mutations,5 with VCP mutations,4 with FUS mutations,4 with SQSTM1 mutations,3 with C9orf72 repeat expansions,2 with CHCHD10 mutations,2 with TARDBP mutations,2 with UBQLN2 mutations,and 1 with TREM2 mutation.A total of 67 rare variants associated with FTLD were detected in 82 patients,including 18 pathogenic or probable pathogenic variants and 49 variants of uncertain significance.The clinical phenotypes of patients with gene variants associated with FTLD are heterogeneous.In addition to being diagnosed with FTLD spectrum diseases such as behavioral variant frontotemporal dementia,non-fluent variant primary progressive aphasia,semantic variant primary progressive aphasia,frontotemporal dementia with motor neuron disease,and progressive supranuclear palsy,there are also patients diagnosed with Alzheimer’s disease,vascular dementia,dementia with Lewy body and other neurodegenerative diseases.[Conclusion]This study expanded genotypic and clinical phenotypic spectra of FTLD in Chinese dementia patients.Part 2 Analysis of clinical characteristics of patients with gene variants associated with FTLD:the PUMCH dementia cohort[Objective]Frontotemporal lobar degeneration(FTLD)is a neurodegenerative syndrome with clinical phenotypic,genetic,and neuropathological heterogeneity.FTLD is divided into four different protein subtypes:FTLD-Tau,FTLD-TDP,FTLD-FET and FTLD-UPS.Studies have shown that gene variants of FTLD are correlated with protein subtypes.For example,FTLD-tau is often present in patients with MAPT mutation.The purpose of this study was to analyze the clinical characteristics of genetic FTLD cohort.We divide them into groups based on FTLD protein subtypes to compare their clinical characteristics.[Methods]This study was a retrospective study,patients were enrolled consecutively from the PUMCH dementia cohort,the Dementia Clinic of the Department of Neurology of PUMCH between 2007 to 2019.All the enrolled patients had gene variants associated with FTLD and met the DSM-5 criteria for dementia diagnosis.Detailed clinical data were collected and cognitive assessment,head imaging(CT,MRI,or PET),and related laboratory tests were performed.These patients were divided into FTLD-Tau group,FTLD-TDP group,FTLD-FET group and FTLD-other group according to the gene variants carried,and the clinical,neuropsychological and imaging manifestations of the four subgroups were compared.[Results]A total of 78 patients with gene variants associated with FTLD were enrolled in this study,including 35 males and 43 females,with an average age of 63.63±12.19 years and an average onset age of 60.75±12.14 years.69.3%had early-onset dementia,and 48.7%had a positive family history.FTLD patients most frequently reported behavioral abnormalities(60.3%),followed by emotional symptoms(51.3%),language impairments(48.7%),motor disorders(38.5%)and neuropsychiatric symptoms(21.8%).Almost half of the patients reported the first symptom was memory disorder.The mean MMSE and ADL scores of FTLD patients were 15.40±9.08 and 35.86±10.24,respectively.The impairment of executive function and memory function is predominant in mild dementia.Brain atrophy was more prominent in the frontal insula(53%),lateral temporal lobe(60.6%),temporal insula(69.7%),and medial temporal lobe(47%)in FTLD patients.CSF biomarker detection was performed in 17 patients,with an average of Aβ42 598.42±223.92pg/ml,T-tau 319.55±342.70pg/ml and P-tau 47.72±25.60pg/ml.CSF in 4 patients suggested evidence of the pathophysiological process of Alzheimer’s disease.52.6%of patients carried ApoE ε3/ε3 homozygote,17.9%were ApoE ε2/ε3,2.6%were ApoE ε2/ε4,24.4%were ApoE ε3/ε4.The onset age and cognitive level of FTLD patients were not significantly affected by ApoE ε4.FTLD-tau,FTLD-TDP,FTLD-FET and FTLD-other four subgroups showed no significant differences in gender,age,years of education,age of onset,family history,clinical characteristics,cognitive examination,imaging and ApoE genotype.[Conclusion]Carriers with gene variants associated with FTLD had strong heterogeneity in genotype,clinical phenotype,neuropsychological evaluation,imaging,CSF biomarkers,and so on.There were no significant differences in clinical characteristics between different subgroups in our single-center cohort.Part 3 Random Forest Model in the Diagnosis of Dementia Patients with Normal Mini-Mental State Examination Scores[Objective]Mini-Mental State Examination(MMSE)is the most widely used tool in cognitive screening for dementia.However,MMSE has a ceiling effect in distinguishing mild cognitive impairment(MCI)and dementia,and some individuals with normal scores were found to have extensive cognitive impairment.Therefore,it is necessary to establish an appropriate model to help correctly identify the cognitive function of individuals with normal MMSE scores.The purpose of this study was to use machine learning to screen out several of the most effective neuropsychological tests,simplify the systematic neuropsychological test battery currently in use,and develop a classification diagnostic model for distinguishing MCI and dementia among individuals with MMSE≥26.[Methods]Patients with MMSE≥26 in the Neurology Outpatient Department of Peking Union Medical College Hospital from May 2009 to April 2021 were included in the study.The patients can be classified into the cognitively unimpaired(CU)group,the MCI group and the dementia group.We compared five machine learning algorithms,including Logistic regression,decision tree,SVM,XGBoost,and random forest,to classify MCI and dementia based on neuropsychological characteristics.[Results]A total of 375 patients were recruited,161 males(42.9%)and 214 females(57.1%),aged 65.51±11.46 years.There were 62 cases(16.5%)in CU group,178 cases(47.5%)in MCI group,and 135 cases(36.0%)in dementia group.MMSE and MOCA-P scores:CU group<MCI group<dementia group(P<0.001).The ADL score of the CU group and MCI group was significantly lower than that of the dementia group(P<0.001),but there was no significant difference in ADL score between the CU group and the MCI group(P=0.986).The random forest model performed best in identifying MCI and dementia.Six neuropsychological subtests with high-importance features were selected to form a simplified neuropsychological test battery,and the test time was cut in half.The ROC-AUC of the random forest model based on the six subtests was 0.89 for distinguishing MCI from CU,and 0.84 for distinguishing dementia from non-dementia.[Conclusion]In this study,random forest algorithm was applied to neuropsychological testing to develop a simple cognitive assessment model for the classification of CU,MCI and dementia in individuals with normal MMSE.It not only optimizes the content of cognitive evaluation,but also improves diagnosis and reduces missed diagnosis. |
PDF Full Text Request