Truace
Subscribe

AI-assisted pathology diagnosis and prognosis of cutaneous melanoma

Source article: Artificial Intelligence applied to surgical pathology in the management of cutaneous melanoma

Surgical pathology remains the cornerstone for the diagnosis, staging and prognostic stratification of cutaneous melanoma. The emergence of digital pathology and artificial intelligence (AI), notably deep learning, is transforming this field. Convolutional neural networks trained on large repositories of digitised histological slides demonstrate diagnostic capabilities comparable, and in specific tasks superior, to board-certified dermatopathologists. Current applications encompass the automated classification o…

By the Truvace desk2 min read1 sourcePeer-reviewedTRV-2026-1436· cite & verify
Artificial Intelligence applied to surgical pathology in the management of cutaneous melanoma
"Cytomegalovirus infection - Case 301" by Pulmonary Pathology is licensed under CC BY-SA 2.0. To view a copy of this license, visit https://creativecommons.org/licenses/by-sa/2.0/.
Trace impact readingContested
P 68The P score combines the specificity and measured human impact of the grounded problem claim with the strength of this Trace’s cited sources.

Both sides are scored from claims and sources, not community votes.

G 67The G score combines the specificity and measured human impact of the grounded gain claim with the strength of this Trace’s cited sources.

In brief

This review describes how digital pathology and deep learning are being applied to cutaneous melanoma. Neural networks trained on large collections of digitised histology slides can sort melanocytic lesions, including the hard distinction between benign naevi and early melanoma, and can estimate molecular alterations directly from standard stained slides.

The work matters because pathology underpins melanoma diagnosis, staging and prognosis, so better tools could refine adjuvant treatment and surveillance decisions. The review stresses that bias, workflow standardisation, interpretability and legal accountability remain unresolved, and that regulation is still evolving, so routine clinical use is not yet established.

Main points

  1. Deep learning models are reported as comparable to board-certified dermatopathologists, and superior in specific tasks.
  2. Algorithms can predict BRAF mutational status from haematoxylin-eosin-stained slides, acting as virtual biomarkers.
  3. AI may extract morphological and spatial features that the human eye cannot perceive, aiding prognostic stratification.
  4. The EU Artificial Intelligence Act classifies medical AI as high-risk, and regulation is rapidly evolving.

The gain

Convolutional neural networks trained on digitised slides can classify melanocytic lesions and infer molecular features such as BRAF status from routine stained slides, with performance comparable to dermatopathologists.

The problem

Use of pathology AI in melanoma is held back by dataset bias, inconsistent slide-imaging workflows between laboratories, limited model interpretability, and unresolved medico-legal accountability.

The rundown

Surgical pathology is the cornerstone of melanoma diagnosis, staging and prognostic stratification. The review says digital pathology and deep learning are changing this field, with networks trained on large repositories of digitised slides.

Current applications include automated classification of melanocytic lesions. Beyond morphology, algorithms act as virtual biomarkers, predicting molecular changes such as BRAF status and extracting features beyond human perception to stratify prognosis.

Remaining obstacles include dataset bias, inter-laboratory standardisation of whole slide imaging, model interpretability and medico-legal accountability. The EU AI Act treats medical AI as high-risk. The 2024 FDA de novo authorization of DermoSensor reflects wider dermatologic AI progress but does not concern surgical pathology.

The authors expect these tools to augment, not replace, the pathologist's central role in precision oncology, refining adjuvant treatment decisions and surveillance strategies in routine practice.

What this doesn’t fix

The source is a narrative overview, and it names unresolved issues including dataset bias, lab-to-lab slide workflow standardisation, interpretability and medico-legal accountability that limit routine adoption.

Sources

  1. Peer-reviewedLa Presse Médicale2026-10-01

The debate