述职报告.ppt

《述职报告.ppt》由会员分享，可在线阅读，更多相关《述职报告.ppt（45页珍藏版）》请在淘文阁 - 分享文档赚钱的网站上搜索。

1、SymptomBasedClusteringofWomenintheLURNObservationalCohortStudy任洁琼 Journal of Urology,2018,200(6):1323-1331The Official Journal of the American Urological Association(AUA)covers the wide scope of urologyVictor P.AndreevArbor Research Collaborative for Health,Ann Arbor,Michiganmathematical modeler,com

2、putational biologist,bioinformatician,and physical chemist.https:/ IntroductionThe current paradigm for treating patients with LUTS is to assign a diagnosis based on a predefinedsymptomcomplexsuch as OAB or on asinglepredominantsymptomsuch as nocturia or SUI.Treatments are then administered based on

3、 these diagnoses.7 However,patients frequently present with other urinary symptoms in addition to those being treated.Diagnosis and treatment based solely on the patient predominant symptoms may be unsatisfactory since they disregard other presenting symptoms.Introduction8 IntroductionPrevious inves

4、tigations based on a relatively small number of self-reported symptom data in community dwelling cohorts.Another study of a large cohort of treatment seeking women focused only on patients with overactive bladder and sought to identify groups of highly correlated symptoms.9 IntroductionGoalsTo impro

5、ve the treatment of patients with LUTS by improving our understanding of the types of patients presenting with LUTS.By clustering groups of individuals with similar patterns of data,in this case LUTS,our main objective was to identify clusters that may better represent LUTS subtypes.02Methods11 TheS

6、ymptomsofLowerUrinaryTractDysfunctionResearchNetwork（LURN）funded by the NIDDK(National Institute of Diabetes and Digestive and Kidney Diseases)a multicenter study:at 1 of 6 LURN clinical centers in the United Statesrecruited patients over the age of 18 presented to a urologist or urogynecologist for

7、 treatment of their LUTS between June 2015 and January 2017Studydesignandparticipants12 Participants:Included criteria：the first time for treatment of LUTSaged 18 years oldAll participants provided written informed consent.Studydesignandparticipants13 Gross hematuria.Significant neurologic disease o

8、r injury.Primary complaint is pelvic pain.Diagnosis of interstitial cystitis.Pelvic or endoscopic GU surgery within the preceding 6 months.Current sexually transmitted infection.Ongoing symptomatic urethral stricture.History of lower urinary tract or pelvic malignancy.Current chemotherapy or other c

9、ancer therapy.Pelvic device or implant complication.Current functioning neurostimulator.Botox injection to the bladder or pelvic structures within the preceding 12 months.pregnancy.History of cystitis caused by tuberculosis,radiation therapy,or Cytoxan/cyclophosphamide therapy.Augmentation cystoplas

10、ty or cystectomy.Presence of urinary tract fistula.Current major psychiatric disorder or other psychiatric or medical issues that would interfere with study participationInability to relay valid information,actively participate in the study,or provide informed consentDifficulty reading or communicat

11、ing in English.Exclusion Criteria:14 15 DatacollectedcollectionmethodBy a prior to treatment by a LURN physiciandemographic databirth,gender,race,ethnicity,education,employment,marital status,medical and surgical history,diet,use of alcohol,tobacco and caffeine,obstetric history,menopausal status an

12、d use of hormone therapy PFDIPelvic Floor Distress InventoryTheLUTSToolhas44items,includingquestionsonseverityandbotherforeachsymptomTheAUASI7itemsandthequalityoflifequestion3-day fluid intake and urinary diary within 4 weeksfluid intake,volume of urine voided,episodes of urinary leakage and whether

13、 leakage occurred with activity or urgency.Blood,saliva,urine and biological flora PRO questionnairesthe PROMIS gastrointestinal constipation,diarrhea and bowel incontinence subsetspsychological healththe PROMIS Depression and Anxiety Short Forms,the PSS and the PROMIS Sleep Disturbance Short Formth

14、e GUPI urological pain 16 聚类分析数据集数据集模型模型1模型模型2模型模型3模型模型417 聚类分析聚类分析聚类分析的定义聚聚类类分分析析(Cluster Analysis)是一个将数据集中的所有数据，按照相相似似性性划分为多个类类别别（Cluster,簇簇）的过程；簇簇是相似数据的集合。聚聚类类分分析析是一种无监督(Unsupervised Learning)分类方法：数据集中的数据没有预定义的类别标号（无训练集和训练的过程）。要求：聚类分析之后，应尽可能保证类别相同的数据之间具有较较高高的的相相似似性性，而类别不同的数据之间具有较较低的低的相似性相似性。18 常用

15、的聚类分析方法：划分法划分法（Partitioning Methods）：以距离作为数据集中不同数据间的相似性度量，将数据集划分成多个簇。属于这样的聚类方法有：k-means、k-medoids等。层层次法次法（Hierarchical Methods）：对给定的数据集进行层次分解，形成一个树形的聚类结果。属于这样的聚类方法有：自顶向下法、自底向上法。聚类分析19 划分法划分法聚类分析年龄年龄80类别类别1类别类别2类别类别3收入收入6000030000 15000020 层层次法次法（Hierarchical Methods）聚类分析Step 0Step 1Step 2Step 3Step

16、4bdceaa bd ec d ea b c d eStep 4Step 3Step 2Step 1Step 0agglomerative(AGNES)divisive(DIANA)21 K 均值（K-means）聚类在数据划分上不考虑类的分层结构问题步骤：随机选择初始类中心将所有的表达谱分配到 K 个类中,根据欧氏距离，反复计算每个类的类中心，然后将每个表达谱分配到类中心与之最接近的类中,形成新的类再计算新的类中心。直至类中心保持不变,或达到最大叠代次数。该算法使待聚类的所有向量到聚类中心的距离的平方和最小。欧氏距离(Euclidean distance)是一个通常采用的距离定义，它是在空间

17、中两个点之间的真实距离。22 23 MethodsTo avoid clustering predominantly by the overall severity of LUTS we normalized the data by participant overall severity.Ani is the answer to the ith question by the nth study participant.Each response was then normalized by Ln,the Euclidean length of the participants respon

18、se vector,resulting in normalized responsesCni=Ani/Ln 24 MethodsWe also accounted for correlation among the items,partly due to redundancy in the questions.m-number of questions and rij correlation coefficients between questions i and jpresented25 MethodsClustering was performed using the resampling

19、 based consensusclusteringmethod.（k-means）IVEware,version 2.0.26 the best clustering is indicated by the highest values of thegeneralizedDunn（Dunn指数）andthepoint-biserialcriteria（相关系数）,and by the lowest value of the Davies-Bouldincriterion（DBI指数）.Methods27 28 Examiningpropertiesoftheidentifiedcluster

20、sDifferences across clusters in demographics,LUTS and other PROs as described were examined by 1-wayANOVA,the chi-squaretestand multiplelinearregression.03Results30 31 32 33 34 04DiscussionPPT模板下载： Discussioncluster F1having OAB dry because there was minimal incontinenceIn addition,urgency is the de

21、fining diagnostic criterion for OAB,which is not central tothis cluster.37 cluster F2 closely resembled the classical definition of patients with wet OAB.Unlike the women in cluster F1,those had urgency incontinence as well as urgency and frequency,suggesting that OAB wet and dry are different clini

22、cal entities.Discussion38 cluster F3 had several kinds of incontinence along with urgency and frequency.This suggests that these women might have poor urethral function,given that storage symptoms were only modest and voiding symptoms were nonexistent.Discussion39 cluster F4 had all LUTS,including v

23、oiding,storage and incontinence.These conditions were reported at a severe degree,suggesting that these women might have poor bladder function as well as poor outlet function.Discussion40 The strength we ignored the preconceived clinical notions and dogma surrounding LUTS in favor of using purely pa

24、tient reported symptoms to derive objective clusters.a large sample sizeThese data are available to further refine and test the LUTS subtypesDiscussion41 Limitations1.Clustering without clinical reasoning could result in associated but unrelated symptoms.We do not believe this to be the case,given t

25、hat these clusters revealed several groups of patients that were clearly distinguished from each other and are common in clinical practice.Discussion42 2.Our analysis only included female subjects and not all patients.However,all patients were selected.In addition,clustering of male patients is curr

26、ently under way.3.Finally,we consider this work preliminary.To our knowledge the clinical significance of these subtypes is currently unknown and requires further validation.Discussion05ConclusionsPPT模板下载： The 4 distinct subtypes of women seeking care for LUTS which we identified in this study.We will continue to refine these subtypes by adding multidimensional data from other LURN studies as well as longitudinal symptom data at 3 and 12 months.Future validation in the independent cohort will determine the generalizability of these clusters.ConclusionsThankyou