Skip to main content
. 2023 Feb 1;14:21. doi: 10.1186/s13244-023-01365-1

Table 4.

Subgroup analyses of total radiomics quality score assessed by the three independent readers

Reader 1 Reader 2 Reader 3
Group N RQS total p value RQS total p value RQS total p value
Journal type 0.542 0.302 0.562
 Imaging 18 6.5 (0–14) 8.5 (4.5–13) 8 (1–14)
 Clinical 15 12 (8–13) 8 (4–13) 11 (5–13)
 Computer science 5 9 (1–9) 2 (−1 to 9.5) 11 (−2 to 11.5)
Publication year 0.772 0.322 0.352
 2021 21 9 (2–13) 9 (4.5–12.5) 12 (5.5–14)
 2013–2020 17 8 (0.5–14.5) 7 (0.5–12.5) 7 (3–13.5)
Journal quartile 0.787 0.510 0.814
 Q1 17 8 (0.5–14.5) 7 (2.5–13) 9 (3.5–14)
 Q2 14 11.5 (0.5–13) 10 (4.5–14) 8 (0.7–13.2)
 Q3 7 9 (2–9) 6 (2–10) 11 (6–14)
Impact factor 0.048 0.035 0.224
  ≤ 4 15 4 (−2 to 10) 6 (−1 to 9) 7 (−1 to 14)
  > 4 23 11 (3–15) 11 (5–13) 11 (6–14)
Cholangiocarcinoma 0.074 0.124 0.152
 Intrahepatic 29 8 (−0.5 to 12.5) 6 (2–12) 9 (2–13.5)
 Perihilar/extrahepatic 9 12 (8–15) 11 (7–13.5) 13 (7.5–14.5)
Imaging modality 0.361 0.206 0.067
 CT 20 11 (0.2–13.7) 10.5 (3.5–13.7) 13 (1.5–14)
 MRI/MRI and CT 18 8 (1.7–10.5) 7 (1.2–11.2) 6.5 (4.7–11.2)
Number of patients  < 0.001  < 0.001  < 0.001
 ≤ 100 12 0.5 (−2 to 6.7) 2.5 (−1.7 to 5) 0.5 (−2.7 to 8.2)
 > 100 26 11.5 (8.7–15) 11 (7.5–13.2) 12.5 (7–14)

Continuous variables are expressed as medians and interquartile range (25th to 75th  percentile) in parenthesis. Continuous variables were compared using the Kruskal–Wallis or the Mann–Whitney U test. Statistically significant values (p < 0.05) are highlighted in bold

CT, Computed Tomography; MRI, Magnetic Resonance Imaging