1.Automatic evaluation of left ventricular systolic function in rats with myocardial infarction by myocardial contrast echocardiography based on mathematical morphology algorithm
Shuang CHEN ; Guoqing DU ; Jingyi XUE ; Pei DU ; Yan WU ; Hairu LI ; Jiawei TIAN
Chinese Journal of Ultrasonography 2014;23(10):897-901
Objective To evaluate the feasibility and value of quantitatively assessing left ventricular systolic function in rats with myocardial infarction (MI) by myocardial contrast echocardiography (MCE) based on mathematical morphology algorithm.Methods SD rats model of MI were mede,then MCE were performed before MI and at 1,3 weeks after MI.In vitro study:based on the principle of binary morphology,segmentation of endocardium was acquired using dilation,erosion,closed operations and connected domain object tag-based.The automatic segmentation of endocardial contour was compared with the manual segmentation boundary.In vivo study:the left ventricular area was calculated using the software,and the area variation fraction (AVF),a new index of left ventricular systolic function,were acquired.The correlation between AVF and left ventricular ejection fraction (LVEF) was analyzed.The value of AVF for diagnosis for left ventricular dysfunction was evaluated by ROC curve.Results ① The cross-correlation coefficient (CCC) and percent error (PE) between automatic contours and manual boundary were more than 0.90 and less than 9%,respectively.② AVF correlated positively with LVEF (r =0.934,P < 0.001).③ ROC analyses showed the area under curves for AVF diagnosing left ventricular dysfunction was 0.834.The best cutoff value was 27.5 % for diagnosing left ventricular dysfunction with the sensitivity of 85 % and specificity of 60 %.Conclusions Left ventricular endocardium can be identificated automatically and rapidly by MCE based on mathematical morphology algorithm.AVF can assessed quantitatively left ventricular systolic function in rats with myocardial infarction.
2.Evaluation of pulmonary arterial hypertension with simple congenital heart disease using tissue Doppler imaging
Pei DU ; Guoqing DU ; Hairu LI ; Shuang CHEN ; Yan WU ; Jiawei TIAN
Chinese Journal of Ultrasonography 2014;23(8):645-650
Objective To evaluate the value of tissue Doppler imaging (TDI) in diagnosing pulmonary arterial hypertension(PAH) with common simple congenital heart disease(CHD).Methods 104 patients of CHD approved by operation or intervention were retrospectively analyzed.Both TDI and right cardiac catheterization were performed in all patients,whose were divided into PAH and non-PAH groups according to mean pulmonary arterial pressure (mPAP) on diagnostic right cardiac catheterization:60 PAH patients[mPAP ≥25 mmHg(1 mmHg =0.133 kPa) and pulmonary arterial capillary wedge pressure ≤15 mmHg],44 non-PAH patients (mPAP <25 mmHg).The PAH group was divided into mild (25 mmHg≤ mPAP≤35 mmHg),moderate (36 mmHg≤mPAP≤50 mmHg) and severe (mPAP>50 mmHg) subgroups.The tricuspid annular systolic peak velocity (TS'),early diastolic peak velocity (TE'),late diastolic peak velocity (TA') were measured by TDI and TE'/TA' and Tei index were calculated.The correlation between TDI parameters and the cardiac catheterization findings (sPAP,mPAP and dPAP) were assessed.Results ①TS' and TE' in PAH group were significantly lower than those in non-PAH(P < 0.05),but Tei in PAH group were significantly higher than that in non-PAH (P <0.001),however,there were no significant difference in TA' and TE'/TA' between two groups (P > 0.05).②Tei correlated positively with pulmonary arterial pressure (P <0.001),TS' was found to be negatively correlated with pulmonary arterial pressure(P <0.05),but no correlation between TE',TA',TE'/TA' and pulmonary arterial pressure(P >0.05).③ROC analysis showed the area under curves for Tei index and TS' diagnosis for PAH was 0.893 and 0.699,respectively.At the cutoff of Tei>0.49 and TS'<16 cm/s for indicating PAH,the sensitivity were 86.67 % and 80.00%,respectively and the specificity were 79.55 % and 45.45 %,respectively.④Tei in the moderate and severe PAH were significantly higher than that in mild PAH (P < 0.05),TS' and TE' in severe PAH were significantly lower than those in mild and moderate PAH (P < 0.05),At the cutoff of Tei>0.55 for indicating moderate or severe PAH,the sensitivity and specificity were 95.24% and 72.22%.Conclusions Tei index and TS' have high application value in assessing PAH and classification of PAH.Tei>0.49 indicates PAH,while Tei>0.55 suggests moderate or severe PAH.
3.Advancing automated pupillometry: a practical deep learning model utilizing infrared pupil images
Guangzheng DAI ; Sile YU ; Ziming LIU ; Hairu YAN ; Xingru HE
International Eye Science 2024;24(10):1522-1528
AIM:To establish pupil diameter measurement algorithms based on infrared images that can be used in real-world clinical settings.METHODS:A total of 188 patients from outpatient clinic at He Eye Specialist Shenyang Hospital from Spetember to December 2022 were included, and 13 470 infrared pupil images were collected for the study. All infrared images for pupil segmentation were labeled using the Labelme software. The computation of pupil diameter is divided into four steps: image pre-processing, pupil identification and localization, pupil segmentation, and diameter calculation. Two major models are used in the computation process: the modified YoloV3 and Deeplabv3+ models, which must be trained beforehand.RESULTS:The test dataset included 1 348 infrared pupil images. On the test dataset, the modified YoloV3 model had a detection rate of 99.98% and an average precision(AP)of 0.80 for pupils. The DeeplabV3+ model achieved a background intersection over union(IOU)of 99.23%, a pupil IOU of 93.81%, and a mean IOU of 96.52%. The pupil diameters in the test dataset ranged from 20 to 56 pixels, with a mean of 36.06±6.85 pixels. The absolute error in pupil diameters between predicted and actual values ranged from 0 to 7 pixels, with a mean absolute error(MAE)of 1.06±0.96 pixels.CONCLUSION:This study successfully demonstrates a robust infrared image-based pupil diameter measurement algorithm, proven to be highly accurate and reliable for clinical application.