Please login or create an account. If you do not have access to this content, you will be shown a 30 second preview and licensing options.
- Presentation
Telepathology and AI-Assisted Digital Pathology: Implementation, Workflow, and Clinical Impact
Description
The speaker, a pathologist from Nebraska, explains telepathology as the transfer of digital whole-slide images for primary diagnosis or second opinions, and frames digital pathology as important for patient care, workflow efficiency, archival access, and especially AI-assisted diagnostics. He reviews practical implementation issues such as scanner performance, image focus, storage options, LIS/IMS integration, DICOM standards, monitor requirements, encryption, licensing, credentialing, and CLIA-related regulations. He emphasizes that quality control and validation are essential, citing CAP guidance on washout periods and concordance thresholds, and describes a phased rollout strategy beginning with retrospective scanning for education and tumor boards, then limited prospective use, and finally routine sign-out. Examples include remote consultation in Mohs surgery, same-day clinic review of slides, and efficient multi-site research collaboration. He also highlights global health applications in China, Zambia, and Botswana, showing that telepathology can improve access and turnaround time in underserved settings. The second half focuses on AI-assisted pathology, where algorithms can classify cases, count mitoses, quantify tumor-infiltrating lymphocytes, measure tumor area, standardize immunohistochemistry metrics, and even support prognosis prediction in melanoma. He concludes that digital infrastructure will soon be necessary as AI-driven tools expand, while current billing and reimbursement remain in development through Category 3 CPT codes.
View moreConclusions
- Telepathology can improve patient care by enabling faster, more flexible access to pathology expertise, archives, and consultations across locations.
- Digital pathology is most valuable when it is integrated into real clinical workflows such as triage, second opinions, frozen sections, and routine sign-out rather than used only for retrospective review.
- The main barriers to adoption are upfront cost, storage demands, staffing, and IT support, but these challenges are outweighed by the need to prepare for AI-driven pathology.
- Successful telepathology requires careful attention to scanner performance, secure data handling, interoperable standards, and legal licensure and credentialing requirements.
- Quality assurance and rigorous validation are essential, with digital diagnoses needing high concordance with glass slides and ongoing internal and external validation.
- A phased implementation strategy beginning with retrospective scanning and limited consults is a practical way to introduce telepathology safely before full routine deployment.
- Telepathology can substantially speed research and education by allowing multi-site image review in days rather than months.
- Remote digital consultation is feasible in difficult settings like Mohs surgery, but comprehensive routine review of very large cases may not be practical without assistance.
- AI can make telepathology more feasible by focusing attention on regions of interest and by helping triage complex or subtle cases.
- In pathology, AI performs well at quantification tasks such as mitotic counts, tumor-infiltrating lymphocytes, tumor area, and PRAME scoring, often improving standardization and reducing bias.
- AI-based diagnostic models can achieve high accuracy in skin pathology and may help prioritize cases for ancillary testing or expert review.
- AI-enhanced prognostic models can add information beyond traditional staging variables and improve prediction of recurrence or metastasis.
- Telepathology has strong global health potential by extending specialist interpretation to underserved regions and improving management in a large share of cases.
- Overall, telepathology appears to be moving from a niche tool toward a foundational platform for future digital, collaborative, and AI-assisted pathology practice.
- American Telemedicine Association clinical guidelines for telepathology. J Pathol Inform. 2014.
- Pantanowitz L, et al. Telepathology guidelines. (Referenced on slide).
- Kumar, Gupta, and Gupta. Whole slide imaging in pathology. 2020.
- Clunie D, et al. Journal of Pathology Informatics. 2018.
- Stathonikos N, Veta M, Huisman A, van Diest PJ. Going fully digital: Perspective of a Dutch academic pathology lab. J Pathol Inform. 2013 Jun 29;4:15. doi:10.4103/2153-3539.114206.#10.4103/2153-3539.114206
- Zhang DY, Venkat A, Khasawneh H, Sali R, Zhang V, Pei Z. Implementation of Digital Pathology and Artificial Intelligence in Routine Pathology Practice. Lab Invest. 2024 Sep;104(9):102111.#10.1016/j.labinv.2024.102111
- Evans, Brown, Bui, et al. Telepathology validation recommendations. (Referenced on slide).
- Hsu, Ahmed, Phlamon, Carment-Baker, Chan, Prassas, Weiser, Zeidan, Clarke, Yousef. An adapted & improved validation protocol for digital pathology implementation. Semin Diagn Pathol. 2025.#10.1016/j.semdp.2025.150905
- McKenna and Florell. 2007.
- Emanuel, et al. 2021.
- Murphee, et al. Weakly supervised classification of Mohs surgical sections using artificial intelligence. Modern Pathology.#10.1016/j.modpat.2024.100653
- Varra, et al. Weakly supervised classification of Mohs surgical sections using artificial intelligence. Modern Pathology.#10.1016/j.modpat.2024.100653
- Ahsan M, Jahangir F, Rathore S, Mumtaz M, Zaman S. Evaluation of whole-slide imaging for diagnosing frozen sections. Ann Diagn Pathol. 2025 Apr;75:152431.#10.1016/j.anndiagpath.2024.152431
- Olsen, et al. Pathology informatics paper. 2018.
- Sturm B, Creytens D, Smits J, Ooms AHAG, Eijken E, Kupershoeck E, Küsters-Vandevelde HVN, Wauters C, Blox WAM, van der Laak JAWM. Computer-Aided Assessment of Melanocytic Lesions by Means of a Mitosis Algorithm. Diagnostics (Basel). 2022 Feb 8;12(2):436.#10.3390/diagnostics12020436
- Rosenbaum, et al. Modern Pathology. 2017.
- Enevolsen, et al. Digital quantification of PRAME to distinguish melanoma from nevi.
- Multimodal AI and tumour microenvironment integration predicts metastasis in cutaneous melanoma. Nature Communications.#10.1038/s41467-025-65051-0