Skip to main content

Table 2 Comparison between the PAIDM and pathologists

From: An artificial intelligence model for the pathological diagnosis of invasion depth and histologic grade in bladder cancer

 

AUC

(95% CI)

Accuracy

(95% CI)

Sensitivity

(95% CI)

Specificity

(95% CI)

PPV

(95% CI)

NPV

(95% CI)

PAIDM

0.847

(0.779–0.905)

0.793

(0.721–0.865)

0.797

(0.721–0.874)

0.899

(0.860–0.937)

0.816

(0.759–0.872)

0.902

(0.869–0.935)

Junior pathologist 1

0.752

(0.644–0.846)

0.676

(0.586–0.766)

0.680

(0.595–0.766)

0.838

(0.788–0.887)

0.709

(0.634–0.783)

0.842

(0.803–0.882)

Junior pathologist 2

0.792

(0.697–0.877)

0.743

(0.667–0.820)

0.734

(0.658–0.811)

0.865

(0.820–0.910)

0.747

(0.680–0.813)

0.869

(0.835–0.903)

Intermediate pathologist 1

0.822

(0.741–0.897)

0.779

(0.703–0.856)

0.779

(0.703–0.856)

0.890

(0.851–0.928)

0.803

(0.749–0.856)

0.891

(0.858–0.924)

Intermediate pathologist 2

0.877

(0.826–0.935)

0.856

(0.793–0.919)

0.860

(0.802–0.919)

0.928

(0.901–0.955)

0.862

(0.815–0.908)

0.931

(0.905–0.958)

Senior pathologist 1

0.918

(0.876–0.957)

0.901

(0.856–0.946)

0.901

(0.847–0.955)

0.951

(0.923–0.978)

0.905

(0.860–0.951)

0.953

(0.928–0.977)

Senior pathologist 2

0.930

(0.865–0.976)

0.910

(0.856–0.964)

0.910

(0.856–0.964)

0.955

(0.928–0.982)

0.920

(0.876–0.964)

0.957

(0.931–0.982)

  1. PAIDM  pathological artificial intelligence diagnostic model, AUC  area under the curve, PPV  positive predictive value, NPV  negative predictive value, CI  confidence interval