A deep learning approach for fully automated cardiac shape modeling in tetralogy of Fallot

Table 1 Summary characteristics of study participants in the training/validation and test sets

Characteristic	Training/validation set (n = 123)	Test set (n = 30)	p-value
Sex (m/f)	73/50	15/15	0.35
Age (y)	17 (12–26)	22 ± 13	0.47
Height (cm)	161 (150–168)	163 (155–176)	0.07
Weight (kg)	58 ± 25	63 ± 18	0.31
BSA (m²)	1.57 ± 0.42	1.72 (1.52–1.82)	0.27
LV EDV (mL)	128 ± 44	119 ± 36	0.31
LV ESV (mL)	66 (51–83)	60 (47–70)	0.19
LV SV (mL)	59 ± 21	57 ± 16	0.52
LV EF (%)	48 (41–52)	48 ± 7	0.23
LV mass (g)	118 ± 35	111 ± 33	0.28
RV EDV (mL)	205 ± 67	197 ± 51	0.54
RV ESV (mL)	127 ± 46	121 ± 37	0.48
RV SV (mL)	78 ± 28	76 ± 23	0.78
RV EF (%)	38 ± 7	39 ± 7	0.72
RV mass (g)	58 (43–77)	53 ± 24	0.16

Data are reported as mean ± standard deviation or as median (interquartile range), depending on the distribution, for continuous variables and as the count for categorical variables. Normality was tested using Shapiro-Wilks. Differences between the training/validation set and test set were assessed using two-sample t-tests or Wilcoxon rank-sum tests, depending on the distribution, for continuous variables and Pearson’s chi-squared tests for categorical variables. BSA body surface area, LV left ventricular; RV right ventricular, EDV end-diastolic volume, ESV end-systolic volume, SV stroke volume, EF ejection fraction

ISSN: 1532-429X