Over 20,000 spectral scans, and more than 4,500 skin lesions
DermaSensor combines advanced spectroscopy and machine learning algorithms to help clinicians evaluate skin lesions quickly, non-invasively, and objectively.
A few words from the organizations we’ve supported across strategy, operations, and digital transformation.
At NoSweat, we combine advanced technology with experienced, patient-focused care to create treatment plans that feel comfortable, personalized, and designed around realistic, natural-looking outcomes.
We’ve compiled answers to the most common questions our clients ask to help you get a better understanding of how we can assist you.
Elastic Scattering Spectroscopy (ESS) is a form of sub-diffuse reflectance spectroscopy that detects disease-associated changes at the cellular and subcellular level by analyzing how light interacts with tissue architecture. Unlike taking images of lesions, ESS translates tissue morphology directly into spectral features that correlate with histopathologic characteristics. ESS has been evaluated with 30+ peer-reviewed publications on clinical studies that show its ability to distinguish malignant from benign tissue across different tissue types.
DermaSensor delivers short pulses (~30 microseconds) of broadband white light spanning 300-900 nm wavelengths. When this light penetrates tissue, photons scatter from refractive-index gradients associated with micro- and nano-scale structures, including nuclear size, chromatin condensation, cellular architecture, and collagen organization. An adjacent collection fiber captures backscattered photons and conveys them to a microspectrometer.
The system records a spectral signature of backscattered intensity versus wavelength. Malignant transformation alters cellular and subcellular structures, creating distinct optical signatures in the ESS spectrum. Lesions with different morphology, cellular, and subcellular features have different densities. The recorded spectrum undergoes preprocessing (smoothing and downsampling) before analysis by the proprietary neural network machine learning algorithm, which compares the pattern against training data from over 2,000 histopathologically confirmed lesions.
Time-gated detection enables the device to function in ambient room light without requiring darkening, making it practical for clinical workflows. The entire measurement takes seconds and requires no tissue preparation.
Once captured, the spectral data is analyzed by a trained algorithm. The algorithm has been trained and validated using:
Over 20,000 spectral scans, and more than 4,500 skin lesions
Histologically confirmed melanoma, BCC, and SCC
Benign lesions, including unbiopsied lesions diagnosed by board-certified dermatologists