The future of melanoma diagnosis is here. DiaDeep, a trailblazer in AI-powered pathology, has just published a groundbreaking study in the prestigious Journal of the European Academy of Dermatology and Venereology (JEADV). This isn't just another AI study—it's a game-changer in how we identify high-risk melanoma patients and guide life-saving treatments.

‍

A Paradigm Shift in Melanoma Prognostication

The study, "Risk Score Stratification of Cutaneous Melanoma Patients Based on Whole Slide Images Analysis by Deep Learning," introduces DiaSurv Melanoma (formally branded as SmartProg-MEL), an AI-driven risk assessment tool that has the potential to transform how oncologists and pathologists predict patient outcomes.

For decades, melanoma staging has relied on static histopathological factors that often fail to capture the full complexity of the disease. But what if AI could see beyond what the human eye perceives? DiaSurv does exactly that.

Using deep learning to analyze routine hematoxylin and eosin (HE)-stained whole slide images (WSIs), this AI model delivers unprecedented prognostic accuracy—achieving a concordance index of 0.78 and surpassing traditional pathology-based assessments.

Key Findings from the Study

The study was conducted across multiple independent datasets, confirming the robustness and clinical applicability of DiaSurv Melanoma. Key results include:

• DiaSurv independently outperformed all traditional prognostic factors, achieving a hazard ratio (HR) of 1.84 (p < 0.005) in multivariate analysis.
• The model effectively stratified patients into low-risk and high-risk groups, with a significant survival difference between the two categories (p < 0.001).
• Sensitivity and specificity remained high across external validation cohorts, demonstrating the model’s generalizability.
• The AI system identified previously undetectable morphological patterns in melanoma tissues that correlate with patient survival, redefining risk assessment in clinical oncology.

What This Means for Oncologists and Their Patients

Imagine being able to pinpoint exactly which early-stage melanoma patients are at the highest risk of recurrence—and who might be spared unnecessary treatment. That’s the power of DiaSurv Melanoma.

• Early detection of high-risk patients: Oncologists can now intervene sooner with targeted therapies.
• Fewer unnecessary treatments: Patients with low-risk scores can avoid the physical and financial burden of overtreatment.
• Validated across multiple cohorts: The AI model was rigorously tested across three independent datasets, proving its robustness and clinical applicability.

‍

“The ability to extract meaningful prognostic information directly from routine pathology slides is a major step forward,” says Pr. Céline Bossard, MD, PhD, lead author and pathologist at IHP Group. “This research helps us better identify which patients require closer follow-up and which might avoid unnecessary treatments.”

‍

Pr. Gaëlle Quéreux, dermatologist and melanoma researcher, comments: “This study opens new perspectives in melanoma research by unveiling patterns in tumor morphology that were previously undetectable. It’s a step toward a deeper understanding of the disease and how we can refine patient management.”

The Beginning of a New Standard in Cancer Care

This study represents more than just an AI breakthrough—it’s a call to action for oncologists, dermatologists, and pathologists worldwide. The future of precision oncology depends on integrating AI into routine clinical practice, and DiaDeep is leading the charge.

“This is the kind of study that makes you rethink everything you thought you knew about melanoma,” said an independent oncology expert reviewing the findings. “The implications are massive.”

About DiaDeep

DiaDeep is a French Startup co-founded by two pathologists and two former AI research professors. DiaDeep aims to revolutionize cancer care by using AI to enhance diagnostic accuracy, helping pathologists and oncologists improve patient survival.

DiaDeep is backed by 34 pathologist-investors from various specialties, who actively contribute their clinical expertise to refine the company’s AI technologies. This unique collaboration ensures that DiaDeep’s tools seamlessly integrate AI’s scientific rigor with real-world medical needs, offering practical and effective solutions for cancer diagnosis and patient monitoring.