Table 2 Accuracy of differential diagnosis between eyelid sebaceous gland carcinoma and squamous cell carcinoma by five DL model-assisted pathologists

From: AI assistance enhances histopathologic distinction between sebaceous and squamous cell carcinoma of the eyelid

Pathologists

SGC and SC classification accuracy (+improvement)

SGC classification accuracy (+improvement)

SC classification accuracy (+improvement)

Resident pathologist 1

0.640 (+0.120)

0.733 (+0.667)

0.500 (+0.200)

Attending pathologist 2

0.720 (+0.240)

0.867 (+0.267)

0.500 (+0.200)

Attending pathologist 3

0.760 (+0.120)

0.933 (+0.133)

0.500 (+0.100)

Associate chief pathologist 4

0.840 (+0.200)

0.933 (+0.200)

0.700 (+0.200)

Associate chief pathologist 5

0.880 (+0.160)

0.933 (+0.267)

0.800 (0)

Total pathologists (mean ± standard deviation)

0.768 ± 0.096 (+0.168 ± 0.052)

0.880 ± 0.087 (+0.186 ± 0.087)

0.600 ± 0.141 (+0.140 ± 0.089)

  1. DL deep learning, SGC sebaceous gland carcinoma, SC squamous cell carcinoma.