DermaSensor achieved 96% sensitivity for skin cancer in primary care physicians — here is a full review of the technology, the trial data, and what it means for skin cancer screening.
By Dr. Elena Marsh, Health Technology Correspondent
Published: November 18, 2026 · 9 min read
DermaSensor is the first FDA-cleared AI skin cancer screening device for use by non-dermatologist clinicians (cleared January 2024). It uses elastic scattering spectroscopy to analyze skin lesion tissue composition and outputs a biopsy/monitor recommendation in seconds. In its pivotal trial, it achieved 96.0% sensitivity and 97.9% NPV for all skin cancers; 95.5% sensitivity for melanoma specifically. It significantly outperformed primary care physicians evaluating lesions by clinical inspection alone.
Skin cancer is the most common cancer in the United States, with approximately 9,500 new diagnoses every day. Melanoma — the most dangerous form — kills approximately 8,290 Americans annually, with the majority of deaths occurring in patients where the cancer was detected at advanced stages. The bottleneck in skin cancer detection is not diagnostic technology — it is access to dermatologists, who currently perform most skin cancer triage.
The average wait time to see a dermatologist in the US is 32–40 days. Most patients with suspicious lesions first present to primary care physicians, who face a difficult decision: refer everyone (driving costs and dermatology backlogs) or rely on clinical inspection (risking missed cancers in a low-training-volume environment). Primary care physicians achieve approximately 60–80% sensitivity for melanoma detection by clinical inspection — meaningfully lower than the 90–95% achieved by trained dermatologists. DermaSensor was designed to close this accuracy gap without requiring specialist training.
Elastic scattering spectroscopy (ESS) exploits a physical phenomenon: when broad-spectrum light enters tissue, it scatters off cellular structures and chromophores (pigment molecules), and the specific wavelength pattern of the backscattered light encodes information about the tissue's structural and chemical composition at a depth of 0.5–2 mm — the depth where early skin cancers develop.
Malignant tissue has a characteristic spectral signature different from benign tissue: cancer cells have enlarged, irregular nuclei that scatter light differently from normal nuclei; melanoma lesions contain different proportions of oxidised and reduced melanin; angiogenic changes in cancer tissue alter haemoglobin absorption signatures. DermaSensor's machine learning algorithm learned to distinguish these signatures from a training set of over 20,000 biopsied lesions — combining spectroscopic features that no single expert clinician could reliably identify by eye.
The device takes a measurement by pressing the handheld sensor gently against the lesion surface for approximately 2 seconds. No patient preparation, gel, or additional light source is required. Results appear in the companion app within seconds as either 'Consider Biopsy' or 'Monitor' with a confidence indicator.
The DermaSensor pivotal trial enrolled 931 lesions across 752 patients at 22 primary care sites. All lesions were biopsied regardless of DermaSensor result, providing ground-truth histopathology for every evaluation. Key results:
| Metric | DermaSensor | PCP unaided inspection |
|---|---|---|
| Overall skin cancer sensitivity | 96.0% | 79.9% |
| Melanoma sensitivity | 95.5% | 73.8% |
| Basal cell carcinoma sensitivity | 97.1% | 81.2% |
| Negative predictive value | 97.9% | 88.1% |
| Specificity (benign lesions) | 34.4% | 62.1% |
DermaSensor's 34.4% specificity means that roughly 65% of benign lesions are flagged as 'consider biopsy' — a high false positive rate that will generate unnecessary biopsies and referrals in primary care settings. This is an inherent trade-off when tuning a device for near-maximum sensitivity: in a cancer screening context, missing a melanoma is far more costly than performing an unnecessary biopsy, so the device is intentionally biased toward sensitivity. Nevertheless, widespread deployment could significantly increase the number of biopsies performed in primary care, with associated costs, patient anxiety, and procedure complications.
DermaSensor provides a risk signal, not a diagnosis. A 'consider biopsy' result does not confirm cancer; a 'monitor' result does not confirm benignity, particularly in high-risk patients (personal or family history of melanoma, immunosuppression, numerous atypical nevi). Clinicians must integrate the DermaSensor recommendation with clinical history, visual appearance, and patient risk factors.
Deep Dive
Our comprehensive guide covers the science, protocols, and tools in detail — with FAQs answered by our research team.
Read the Full Guide →© 2026 QuanMed - All rights reserved