Teledermatology agrees with in-person diagnosis 76% of the time, a review finds

Teledermatology is the practice of diagnosing skin conditions from photographs rather than in person, most often in a "store-and-forward" model: a patient or a primary-care clinician takes images, uploads them with a short history, and a dermatologist reviews them later and sends back a diagnosis and plan. A meta-analysis published in Frontiers in Medicine on 2 March, pooling 155 studies, found that this agrees with a face-to-face diagnosis about three-quarters of the time for skin conditions in general — high enough to support the model for everyday dermatology, but lower and more conditional for skin cancer [s1].

What the pooled numbers say

The review screened 30,412 records and included 155 studies, 139 of them in the quantitative pooling [s1]. Diagnostic concordance with in-person assessment was 76% across all skin conditions (95% CI 73–79%), 73% for skin cancers (95% CI 67–79%), and 76% for pigmented lesions (95% CI 67–83%) [s1].

The single most important modifier was dermoscopy — the use of a magnifying lens with standardised lighting to image a lesion. Adding it lifted concordance for skin cancers from 67% (95% CI 58–74%) to 80% (95% CI 73–85%) [s1]. By contrast, the review found no significant difference by communication type, platform, or photography device [s1]. In other words, what the image shows matters far more than whether it travels over one app or another, or is captured on a dedicated camera versus a phone.

Two secondary findings speak to why health systems adopt this at all. The mean time to reach a teledermatology diagnosis was 1.05 minutes per case (95% CI 0.98–1.12), and patient satisfaction was high at 82% (95% CI 76–87%) [s1].

An in-person trial points the same way

Pooled observational data can flatter a technology, because the studies feeding it vary in quality. A randomised cohort study at the American University of Beirut Medical Center, published in September 2025, tested the store-and-forward model prospectively against in-clinic dermatology as the reference standard [s2]. It enrolled 360 adults, each of whom contributed three photo sets — unassisted patient-taken, patient-taken after brief standardised training, and resident-taken under clinical conditions — with teledermatologists randomly assigned one set per patient [s2].

Diagnostic concordance rose with image quality: 79% for unassisted patient images, 84% after brief patient training, and 87% for resident-taken images [s2]. Treatment concordance was lower and followed the same gradient, 38% to 44% to 45%, and the need to request additional history fell from 64% to 47% as images improved [s2]. Acne showed the highest diagnostic match across all image types [s2]. The lesson is consistent with the meta-analysis: the diagnosis is only as good as the picture, and teaching patients how to take the picture is a cheap way to improve it.

Skin cancer is where the caveats concentrate

The higher-stakes question is melanoma, where a missed diagnosis costs the most. A systematic review in Cancers, searching the literature to December 2024, found that teledermatology showed high sensitivity and specificity for melanoma — but "particularly when dermoscopic images and expert interpretation were available" [s3]. Some studies reported reduced Breslow thickness (a measure of how deep a melanoma has grown at diagnosis) and shorter delays compared with face-to-face care, and both patients and clinicians were generally satisfied [s3].

That review also flagged the limits. Formal economic analyses were sparse, and the artificial intelligence tools layered onto some teledermatology pathways "yielded mixed results and [were] generally perceived with caution in the absence of clinical supervision" [s3]. This is the same signal running through the pooled data: teledermatology is a way to move a dermatologist's judgment to where the patient is, not a way to replace that judgment with an algorithm. It sits alongside the separate, and weaker, evidence for consumer skin-check AI, where accuracy has been shown to vary by skin tone.

What it means for a reader

The case for teledermatology is access — the same argument that underpins other remote front doors to specialist care, from phone-based eye screening to audio-only telehealth, whose evidence is thinner. For common rashes and lesions, the evidence now says a well-taken photo reviewed by a dermatologist lands close to a clinic visit, quickly and with high patient satisfaction [s1][s2].

For a suspicious mole, the picture is more demanding. Concordance for skin cancer is meaningfully lower than for skin disease overall, and the gap closes mainly when a dermatoscope and an expert reader are in the loop [s1][s3]. A teledermatology service that triages possible skin cancers on phone snapshots alone is operating at the low end of that range — which is why the strongest programmes still route uncertain or high-risk lesions to an in-person examination and, when needed, a biopsy. None of this changes the standard advice to have a changing or new mole assessed in person.

What to watch

Whether teledermatology services publish their own concordance and cancer-miss rates rather than citing the pooled literature, and whether dermoscopy-equipped imaging becomes standard for lesions referred as possible cancer rather than an optional extra.

Sources

1. [s1] Diagnostic accuracy of teledermatology for skin diseases: a systematic review and meta-analysis. Frontiers in Medicine, 2 March 2026. https://doi.org/10.3389/fmed.2026.1739592
2. [s2] Teledermatology Diagnostic Accuracy: A Randomized Cohort Study Comparing Three Image Acquisition Techniques. International Journal of Telemedicine and Applications, 24 September 2025. https://doi.org/10.1155/ijta/5789165
3. [s3] Teledermatology vs. Face-to-Face Dermatology for the Diagnosis of Melanoma: A Systematic Review. Cancers, 29 August 2025. https://doi.org/10.3390/cancers17172836