Skip to main content
. 2019 Apr 26;2:29. doi: 10.1038/s41746-019-0104-2

Table 1.

The clinical characteristics of the study datasets at patient and image level

Variables Patients (N = 1297) Ultrasound Images (N = 4505) Ultrasound Images in nontesting dataset (N = 4,010) Ultrasound Images in testing dataset (N = 495) p valuea
Demographics, median (IQR)
Age (years) 65 (53, 74) 65 (52, 74) 65 (53, 74) 63 (51, 74) 0.269
Male, n (%) 715 (55.1) 2464 (54.7) 2201 (54.9) 263 (53.1) 0.459
Comorbidity, n (%)
Cardiovascular disease 1001 (77.2) 3504 (77.8) 3117 (77.7) 387 (78.2) 0.820
Hypertension 968 (74.6) 3405 (75.6) 3024 (75.4) 381 (77.0) 0.446
Diabetes 533 (41.1) 1888 (41.9) 1697 (42.3) 191 (38.6) 0.112
Biochemical value, median (IQR)
Serum creatinine (mg/dL) 2.1 (1.4, 4.4) 2.1 (1.4, 4.3) 2.1 (1.4, 4.4) 2.0 (1.4, 3.6) 0.003
eGFR (mL/min/1.73 m2) 30.0 (12.3, 48.6) 30.1 (12.6, 48.7) 29.9 (12.3, 48.2) 31.6 (15.7, 55.5) 0.001
Sonographic parameter, median (IQR)
Kidney length (cm) 9.72 (8.76, 10.53) 9.62 (8.69, 10.51) 9.63 (8.66, 10.51) 9.54 (8.80, 10.48) 0.703

ap values denotes probability for difference between the nontesting and testing datasets and are calculated by Wilcoxon rank sum test for continuous variables and Chi-square test (or Fisher’s exact test as appropriate) for categorical variables