When a burn patient arrives at a trauma center, the hardest question is often the simplest to ask and the hardest to answer: which parts of the wound will heal on their own, and which need surgery now? At the American College of Surgeons Clinical Congress 2026, running September 26 through 29 in Washington, D.C., Spectral AI is presenting data on a device designed to answer that question in under half a minute, using multi-spectral imaging and an artificial intelligence model trained on thousands of burn images.
A Camera That Predicts Healing, Not Just Photographs It
The company’s DeepView System captures an image of a wound in roughly 0.2 seconds, then runs it through a proprietary AI classifier that takes about 20 to 25 seconds to render a verdict, according to Spectral AI’s presentation materials. The output is not a diagnosis in the traditional sense but a probability map: it flags areas within a burn that are unlikely to heal within 21 days without significant medical intervention, such as grafting. That 21-day window matters clinically, because burns that fail to close within three weeks carry sharply higher risk of hypertrophic scarring and prolonged disability, and burn surgeons have historically relied on visual judgment and clinical experience to make that call.
Wake Forest Surgeon to Present Validation Data
The presentation, titled ‘Artificial Intelligence Training and Validation of Multi-Spectral Imaging Device for Evaluation of Burn Wounds,’ will be delivered by James H. Holmes IV, MD, director of the Burn Center at Atrium Health Wake Forest Baptist Medical Center and a professor of surgery and regenerative medicine at Wake Forest University, according to Spectral AI’s announcement. Holmes’ involvement gives the data an academic surgical pedigree rather than a purely commercial one, and his center is among the sites that contributed imaging data used to train and validate the underlying algorithm.
Already Cleared, Now Chasing Adoption
DeepView is not experimental in the regulatory sense. The FDA granted it De Novo classification in May 2026, a pathway reserved for novel, generally low-to-moderate-risk devices with no direct predicate on the market, and the system earlier secured UKCA authorization in the United Kingdom in February 2024. That means the ACS Congress presentation is less about winning approval and more about winning over the burn surgeons who will decide whether to actually use the tool at the bedside. Spectral AI has also been showcasing DeepView at the American Burn Association’s own meetings earlier in 2026, part of a push to normalize the device across specialty conferences rather than treat it as a single launch event.
Why Burn Surgeons Are a Tough Room
Burn assessment has resisted automation for a straightforward reason: the visual cues surgeons use, including wound color, blanching and capillary refill, are notoriously inconsistent between observers, and even experienced clinicians disagree with each other a meaningful share of the time. Proponents argue that is exactly the gap an imaging-based AI tool is suited to close, by replacing subjective color judgment with a spectral signature and a trained model. Skeptics within the field counter that any tool making treatment-altering predictions needs extensive real-world validation across skin tones, wound depths and mechanisms of injury, cautioning that a device trained predominantly on one patient population could underperform on others until broader post-market data accumulates.
The Bigger Pattern in Surgical AI
DeepView’s rollout fits a broader trend of narrowly scoped, single-purpose AI devices working their way into surgical and trauma care, following imaging tools that have already gained footholds in radiology and dermatology. Rather than aiming for broad diagnostic autonomy, these tools focus on one high-stakes decision point, in this case whether to graft early, where speeding up a judgment by even a day or two can change outcomes and reduce costly, painful re-operations.
What Comes Next
The real test for DeepView will not be the September presentation itself but what burn centers do afterward: whether hospitals purchase the system, whether insurers reimburse its use, and whether independent, multi-center outcomes data eventually confirms the predictions match what actually happens to patients. Spectral AI’s repeated presence at both the ACS Congress and the American Burn Association’s meetings in the same year suggests the company is betting that surgeon-to-surgeon credibility, built one conference presentation at a time, will move faster than any single regulatory clearance toward getting the device into routine burn-unit workflows.
Photo: OsloMetX / PIXABAY via Pixabay
