Table 1.

Agreement statistics for human annotators and computational technique

ComparisonnObserved κLower 95% CIUpper 95% CI
RP1 versus GT, scheme: T, data: patients480.480.320.64
RP2 versus GT, scheme: T, data: patients480.680.500.86
C versus GT, scheme: T, data: patients540.550.500.60
C versus GT, scheme: T, data: sections1210.90.870.93
C versus GT, scheme: F, data: patients330.310.170.45
C versus GT, scheme: F, data: sections850.980.941
Baseline versus GT, scheme: T, data: patients54<0n/an/a
  • Reported values include linear weighted Cohen’s κ and upper and lower 95% CIs. T refers to classifications according to the Tervaert scheme; F refers to classifications according to the Fogo scheme. Data description identifies whether the experiment was performed using separate patients or separate sections as individual data. RP, renal pathologist; GT, ground truth; C, computer; n/a, not applicable.