Table 1 Baseline characteristics of patients enrolled in training (Asian patients) and validation (European patients) cohort.

From: Training and external validation of machine learning supervised prognostic models of upper tract urothelial cancer (UTUC) after nephroureterectomy

Feature type

Feature – median (IQR) / n (%)

All (n = 3129)

Training cohort (n = 637)

Validation cohort (n = 2492)

patient-related

Gender – Male

1951 (62.3)

270 (42.4)

1681 (67.5)

patient-related

Gender – Female

1178 (37.7)

367 (57.5)

811 (32.5)

patient-related

Age

68 (61–76)

68 (61–76)

68 (61–76)

tumor-related

Grading – G1/Low Grade

127 (4)

112 (17.6)

15 (0.6)

tumor-related

Grading – G2

394 (12.6)

5 (0.8)

389 (15.6)

tumor-related

Grading – G3/High grade

2608 (83.3)

520 (81.6)

2088 (83.8)

tumor-related

pT – pT0

17 (0.5)

2 (0.3)

15 (0.6)

tumor-related

pT – pTa

631 (20.2)

104 (16.4)

527 (21.2)

tumor-related

pT – pTis

55 (1.8)

7 (1.1)

48 (1.9)

tumor-related

pT – pT1

713 (22.8)

158 (24.8)

555 (22.2)

tumor-related

pT – PT2

612 (19.6)

139 (21.8)

473 (19.0)

tumor-related

pT – pT3

945 (30.2)

191 (30.0)

754 (30.2)

tumor-related

pT – pT4

156 (4.9)

36 (5.7)

120 (4.9)

tumor-related

pN – pNx

2179 (69.6)

504 (79.1)

1675 (67.2)

tumor-related

pN – pN0

699 (22.3)

104 (16.3)

595 (23.8)

tumor-related

pN – pN1

236 (7.5)

14 (2.2)

222 (9)

tumor-related

pN – pN2

15 (0.5)

15 (2.4)

 

tumor-related

multi-focality – Yes

795 (25.4)

205 (32.2)

590 (23.7)

tumor-related

multi-focality – No

2334 (74.6)

432 (67.8)

1902 (76.3)

tumor-related

CIS – Yes

633 (20.2)

60 (9.4)

573 (23.3)

tumor-related

CIS – No

2496 (79.8)

577 (90.6)

1919 (76.7)

tumor-related

Lymphovascular invasion – Yes

1078 (34.5)

500 (78.5)

578 (23.2)

tumor-related

Lymphovascular invasion – No

2051 (65.5)

137 (21.5)

1914 (76.8)

  1. IQR: Inter Quartile Range, CIS: Carcinoma In Situ.