Researchers from Tufts University’s Silklab, Helmholtz Munich, Ludwig Maximilian University, and the Technical University of Munich have developed a noninvasive, silk-based diagnostic patch designed to continuously monitor critical health metrics in premature neonates. Published in ACS Sensors, the coin-sized wearable addresses a longstanding challenge in neonatal intensive care: the physical toll of repeated blood draws and the burden of wired monitoring equipment on fragile infants. The patch operates by capturing trace volumes of interstitial fluid that naturally permeate the immature skin of preterm babies. Leveraging this physiological trait, the device employs a three-layer architecture. A silk fibroin base, harvested from moth cocoons, stabilizes temperature-sensitive enzymes for shelf stability. A wax-printed paper layer functions as a microfluidic network, routing fluid to discrete sensing dots. A flexible, waterproof medical adhesive secures the patch to the skin while maintaining a seal against incubator humidity. When bodily fluid contacts the patch, colorimetric indicators shift in response to temperature, pH, sodium, and glucose levels. To overcome the inherent difficulty of accurately reading color changes under variable hospital lighting and incubator conditions, the team integrated a deep-learning artificial intelligence model. This system automatically compensates for lighting, angle, and movement, translating visual shifts into precise numerical readings with over 91 percent accuracy across vital signs and exceeding 98 percent accuracy for detecting hypoglycemia. According to Fiorenzo Omenetto, director of the Silklab, the multi-parameter approach allows clinicians to interpret interconnected physiological trends rather than isolated data points. The technology eliminates the need for needles, external power, or refrigeration, reducing skin irritation and infection risk while enabling continuous surveillance between traditional laboratory tests. Researchers emphasize that the patch complements rather than replaces standard clinical diagnostics, intercepting subtle physiological drifts that might otherwise escalate into emergencies. With manufacturing costs measured in mere cents, the device is particularly suited for low-resource and remote settings where advanced neonatal monitoring infrastructure is scarce. The research team now describes the current iteration as a proof of concept. Upcoming phases involve expanded clinical trials in neonatal units to validate skin-fluid measurements against conventional blood draws and to refine the AI model across diverse environmental conditions. Long-term development aims to incorporate additional parameters such as oxygen saturation and carbon dioxide levels. If clinical validation confirms the technology’s reliability, the researchers project widespread integration into global neonatal care, potentially transforming how vulnerable infants are monitored in both high-tech hospitals and underserved communities.