Making cytological diagnoses on digital images using the iPath network

Area of Science:

• Digital pathology
• Cytopathology
• Telemedicine

Background:

• The iPath telemedicine platform in Basel facilitates histological and cytological consultations.
• It also serves as an effective educational resource.

Purpose of the Study:

• To assess the diagnostic accuracy of experienced cytopathologists using still images.
• To identify factors limiting diagnostic precision.
• To create a cytological image series for educational purposes.

Main Methods:

• Four cytopathologists evaluated 167 cytological specimens via the iPath platform.
• Only well-stained, wet-fixed specimens were included.
• Diagnoses were categorized as benign, suspicious, or malignant.

Main Results:

• Overall diagnostic accuracy ranged from 88-90% with substantial interobserver agreement (κ = 0.791).
• Specificity and sensitivity were 83-92% and 85-93%, respectively.
• Lowest concordance was observed in urine/bladder washings and identifying benign lesions.

Conclusions:

• Digital image diagnostics, even under optimal conditions, have accuracy limits of 80-90% due to the absence of ancillary tests.
• Digital images are still suitable for teleconsulting and as learning aids.
• The study's image series is publicly available for educational use.