Primary histologic diagnosis using automated whole slide imaging: a validation study

BMC Clinical Pathology

Table 2 Discrepancies between individual study pathologists' WSI-based reports

Case	Part	Specimen	At Issue	Discrepancy Type
37	1	Bladder Biopsy	Tumor Grade*	Type 3
38	1	Bladder Biopsy	Dysplasia v Benign	Type 4*
			Granulomatous Inflammation?	Type 2
	2	Bladder Biopsy	Granulomatous Inflammation?	Type 2
40	1	Bladder Biopsy	Tumor Grade	Type 3
			Superficial Invasion?	Type 3
45	1	Bladder Biopsy	Muscle Invasion?	Type 1
50	2	Skin Biopsy	Atypical Melanocytic Lesion: Invasion?	Type 4*
52	1	Skin Biopsy	Actinic Keratosis versus Seborrheic	Type 3
			Keratosis with HPV infection
54	1	Bladder Biopsy	AK versus Artifact	Type 3
55	1	Skin Biopsy	Junctional versus Compound Nevus	Type 2
57	2	Prostate Bx	Atypical Glands (Cancer v HGPIN)	Type 4*

Description: There were 12 discrepancies between study pathologists' WSI-based reports. Of these, sub-optimal image quality was implicated in 4 instances (38, 50 and 57). See text for the definition of the "discrepancy types".

ISSN: 1472-6890