← Evidence library

BIB_408

Laiouar-Pedari 2026 JAMA Dermatology (DKFZ Heidelberg) — prospective-only meta-analysis quantifying AI-assisted dermoscopy for melanoma: ANCHORS the brief's currently-unevidenced "consumer AI skin apps = variable accuracy, not a substitute" caution (TASK17 §300) and complements BIB075 (DERM deployment). KEY: in prospective settings AI alone ≈ dermatologist (NOT superior); the gain is as clinician DECISION-SUPPORT (AI-assisted sens 91.9% / spec 83.7% in the single such study); frequent risk of bias + selected populations → not yet ready as standalone or as a consumer-app substitute. Patient-p…

Evidence grade

A

Tier

1 (JAMA Dermatolo

Cited by tasks

5, 17, 22a

Identifiers

—

Full extracted findings & patient-implication analysis: bibliography/BIB_408/findings.md (research corpus). This page is a short context summary — not individualised medical advice.