One-Stop Shop Quantitative Imaging Method For Differentiation Between Hypertrophic Cardiomyopathy And Hypertensive Heart Disease

Objective: Differentiating hypertrophic cardiomyopathy(HCM)from hypertensive heart disease(HHD)unavoidably encounters diagnostic challenges especially in patient of suspected HCM with history of hypertension.Diverse and overlapping forms of HCM can often lead to ambiguity when diagnosis is based on a single genetic or morphological index.Combining feature tracking(FT)and late gadolinium enhancement(LGE)using cardiac magnetic resonance(CMR)imaging,we aim to identify the difference between the two disorders.Methods: First of all,documented HCM,HHD and control group were compared using variance analysis for identifying the differences of baseline and imaging indices.Second,clinical case analysis were performed for single index of its role in differential diagnosis by while correlation analysis of imaging indices were performed.Finally,proposing and validation procedures were proposed in HCM and HHD.Principal component analysis(PCA)was conducted on CMR indices selected by univariate analysis to model an integrated algorithm(IntA),the cut-off point of which was derived from receiver operating curve(ROC)analysis.K-folding validation and second phase recruited subjects were to validate the algorithm.Results: Two hundred and thirty subjects were prospectively included: In the first part,52 control,75 HCM,33 HHD were included and left ventricular ejection fraction(LVEF),left ventricular volume,maximum left ventricular wall thickness(MLVWT)were found to be significantly different among groups(p<0.001).Global strain showed statistical difference(endocardial circumferential:-33.0%±6.0% vs-31.2%±7.0% vs-35.7%±5.3%,p=0.001,epicardial:-13.7%±3.0% vs-10.3%±2.8% vs-11.7%±2.4%,p<0.001;endocardial longitudinal:-22.7%±4.2% vs-20.0%±4.3% vs-19.8%±4.4%,p=0.001,epicardial :-17.3%±3.1% vs-13.9%±3.3% vs-13.1%±3.3%,p<0.001).Secondly,single index like LVEF,MLVWT and global strains failed in differentiation diagnosis.There is an strong positive correlation(r=0.76)between MLVWT and left ventricular mass index and an strong negative correlation(r=-0.78)between LVEF and endocardial global circumferential strain in HHD and HCM groups.In the last part,75 HCM and 33 HHD were to generate algorithm by PCA.Indices including LVEF,LV volume,MLVWT and global circumferential strain were weighed to deduce IntA.Notably,for late gadolinium enhancement(LGE)-positive subjects,the cut-off point of IntA?81 indicates HCM(83% sensitivity,91% specificity),and the area under the curve(AUC)reached 0.900.In LGE-negative subjects,a cut-off point of IntA ?84(100% sensitivity,82% specificity)indicates HHD,with the AUC of 0.947.IntA also exhibited better diagnostic value compared with single indices in HCM sub-group with hypertension or non-obstructive hypertrophy.Another 55 HCM,16 HHD were for second verification of IntA,which witnessed the AUC of 0.846 in LGE-negative subjects,and 0.86 in LGE-positive subjects.Conclusions: A per-patient based IntA formula was deduced from CMR exam containing FT,LGE and conventional parameters.The workflow provides a useful approach for differentiating between HCM and HHD and might contribute to more accurate identification of disease,especially in clinically ambiguous status.