In a paper published in the journal Science Advances, researchers present a paper-like PETAL sensor with five colorimetric sensors for holistic wound assessment. Using artificial intelligence (AI) and deep learning algorithms, the sensor accurately classifies healing versus nonhealing status. It also enables monitoring of wound progression in situ, providing early warning for timely clinical intervention in wound care management.
Background
Wound healing is a complex process involving four stages: hemostasis, inflammation, proliferation, and remodeling. The global health and economic burden caused by impaired wound repair, including chronic wounds and burn scars, is a substantial concern. The existing techniques used for wound monitoring suffer from a slow pace and inadequate holistic profiling and quantitative characterization, which hinders the accurate prediction of wound healing progression. Proposed as a solution to address these challenges, wearable wound sensor patches offer a promising approach.
Existing sensors focus on detecting single physical or biochemical markers using electrochemical sensors. Flexible electronics have enabled the integration of multiple sensors into a single platform, paving the way for wearable biosensors capable of detecting multiple wound parameters. However, these sensors still have limitations, such as bulky circuit boards and batteries, susceptibility to motion artifacts, and frequent calibrations. An alternative approach is an optical detection using paper-based colorimetric and fluorometric sensors.
Study: PETAL: AI-Enabled Sensor Patches for Wound Monitoring. Image Credit: Everything Possible / Shutterstock
This research paper introduces a paper-like, battery-free, AI-enabled multiplexed sensor patch (PETAL) for holistic wound healing monitoring. The patch incorporates five colorimetric sensors capable of measuring temperature, pH, trimethylamine (TMA), uric acid (UA), and moisture. This integration enables rapid and precise profiling of each biomarker within minutes. The PETAL sensor patch can be incorporated into wound dressings, allowing in situ analysis to be performed without needing removal. Using an AI algorithm based on deep learning and a smartphone, the sensor patch provides a comprehensive assessment of wound healing status and enables the classification of wound types and severity levels. This technology has the potential to provide early warning about adverse events and facilitate timely clinical intervention, thereby enhancing wound care management in various healthcare settings.
Results
Fluidic-based wound sensor patch: The PETAL sensor patch, a paper fluidic-based device, has been developed for monitoring wound conditions. It consists of a lightweight, flexible paper panel with five sensing regions arranged in a flower pattern, allowing an even flow of wound exudate through channels. The patch can be customized to match wound size and integrates with wound dressings. Wax patterns printed on the sensor panel create a continuous barrier, defining the fluidic pattern while optimizing the heating time prevents clogging and ensures efficient exudate flow within 2 to 3 minutes. This innovative patch offers a practical and adaptable solution for monitoring wound conditions.
Colorimetric sensor: Five colorimetric sensors are prepared for temperature, pH, moisture, trimethylamine (TMA), and uric acid (UA) detection. Each sensor uses specific sensing materials and formulations, and they are prepared sequentially on the sensor panel.
The temperature sensor uses cholesteric liquid crystals (CLCs) and exhibits distinct color changes within the temperature range of 31°C to 36°C. The TMA sensor employs Reichardt's dye and shows a color change from dark gray to light white in the presence of TMA. The pH sensor utilizes phenol red dye and displays a blue-to-green color ratio correlated with pH values from 6 to 10. The moisture sensor contains anhydrous cobalt chloride in a polyvinyl alcohol matrix, changing color from deep blue to pale violet to pink as moisture content increases. The UA sensor is based on a cascade enzymatic reaction using uricase and peroxidase, providing a linear calibration for UA concentrations ranging from 40 to 1000 μM.
Quantitative analysis: The PETAL sensor patch is capable of simultaneously analyzing the five wound biomarkers. Time-lapse images captured during the flow of simulated wound fluids demonstrate the patch's ability to differentiate between routine healing and infected nonhealing wounds. The color changes in the sensors indicate variations in temperature, TMA concentration, pH level, moisture content, and UA concentration.
In summary, the PETAL sensor patch offers a flexible and low-cost solution for monitoring wound conditions. Its design allows for efficient fluid flow, and the five colorimetric sensors provide quantitative information about temperature, pH, moisture, TMA, and UA levels in the wound. The patch has the potential for integration with wound dressings and could aid in wound management and healing.
AI analysis of PETAL sensor patch after being exposed to animal wound exudates
The PETAL sensor patch was tested with perturbed and burn wound exudates from rat models. The pH and UA sensors on the patch detected distinct color differences, with perturbed wounds showing higher pH and lower UA levels compared to the scaffold fluid. A neural network achieved 96.3% accuracy in identifying the wound state. Different degrees were simulated for burn wounds, with deep burns exhibiting compromised vascular permeability and increased cytokine release. Wound exudate collection became challenging as deep burns turned necrotic. The PETAL sensor patch holds promise for analyzing wound exudates and optimizing treatment.
Conclusions
To summarize, the researchers introduce a PETAL sensor patch that enables simultaneous sensing of five wound markers through adaptable wax printing and deep neural network analysis. It has been successfully tested for analyzing perturbed wound exudate and monitoring burn wound healing. The integration of this cost-effective patch into wound dressings offers the potential for real-time assessment and early intervention.
Tackling Text-to-Image AI Flaws