Integrated image and location analysis for wound classification: a deep learning approach.

The global burden of acute and chronic wounds presents a compelling case for enhancing wound classification methods, a vital step in diagnosing and determining optimal treatments. Recognizing this need, we introduce an innovative multi-modal network based on a deep convolutional neural network for categorizing wounds into four categories: diabetic, pressure, surgical, and venous ulcers. Our multi-modal network uses wound images and their corresponding body locations for more precise classification. A unique aspect of our methodology is incorporating a body map system that facilitates accurate wound location tagging, improving upon traditional wound image classification techniques. A distinctive feature of our approach is the integration of models such as VGG16, ResNet152, and EfficientNet within a novel architecture. This architecture includes elements like spatial and channel-wise Squeeze-and-Excitation modules, Axial Attention, and an Adaptive Gated Multi-Layer Perceptron, providing a robust foundation for classification. Our multi-modal network was trained and evaluated on two distinct datasets comprising relevant images and corresponding location information. Notably, our proposed network outperformed traditional methods, reaching an accuracy range of 74.79-100% for Region of Interest (ROI) without location classifications, 73.98-100% for ROI with location classifications, and 78.10-100% for whole image classifications. This marks a significant enhancement over previously reported performance metrics in the literature. Our results indicate the potential of our multi-modal network as an effective decision-support tool for wound image classification, paving the way for its application in various clinical contexts.

Dehydrated human amnion/chorion membrane to treat venous leg ulcers: a cost-effectiveness analysis.

OBJECTIVE: To evaluate the cost-effectiveness of dehydrated human amnion/chorion membrane (DHACM) in Medicare enrolees who developed a venous leg ulcer (VLU). METHOD: This economic evaluation used a four-state Markov model to simulate the disease progression of VLUs for patients receiving advanced treatment (AT) with DHACM or no advanced treatment (NAT) over a three-year time horizon from a US Medicare perspective. DHACM treatments were assessed when following parameters for use (FPFU), whereby applications were initiated 30-45 days after the initial VLU diagnosis claim, and reapplications occurred on a weekly to biweekly basis until completion of the treatment episode. The cohort was modelled on the claims of 530,220 Medicare enrolees who developed a VLU between 2015-2019. Direct medical costs, quality-adjusted life years (QALYs), and the net monetary benefit (NMB) at a willingness-to-pay threshold of $100,000/QALY were applied. Univariate and probabilistic sensitivity analyses (PSA) were performed to test the uncertainty of model results. RESULTS: DHACM applied FPFU dominated NAT, yielding a lower per-patient cost of $170 and an increase of 0.010 QALYs over three years. The resulting NMB was $1178 per patient in favour of DHACM FPFU over the same time horizon. The rate of VLU recurrence had a notable impact on model uncertainty. In the PSA, DHACM FPFU was cost-effective in 63.01% of simulations at the $100,000/QALY threshold. CONCLUSION: In this analysis, DHACM FPFU was the dominant strategy compared to NAT, as it was cost-saving and generated a greater number of QALYs over three years from the US Medicare perspective. A companion VLU Medicare outcomes analysis revealed that patients who received AT with a cellular, acellular and matrix-like product (CAMP) compared to patients who received NAT had the best outcomes. Given the added clinical benefits to patients at lower cost, providers should recommend DHACM FPFU to patients with VLU who qualify. Decision-makers for public insurers (e.g., Medicare and Medicaid) and commercial payers should establish preferential formulary placement for reimbursement of DHACM to reduce budget impact and improve the long-term health of their patient populations dealing with these chronic wounds. DECLARATION OF INTEREST: Support for this analysis was provided by MiMedx Group, Inc., US. JLD, and RAF are employees of MiMedx Group, Inc. WHT, BH, PS, BGC and WVP were consultants to MiMedx Group, Inc. VD, AO, MRK, JAN, NW and GAM served on the MiMedx Group, Inc. Advisory Board. MRK and JAN served on a speaker's bureau. WVP declares personal fees and equity holdings from Stage Analytics, US.

Treatment patterns and outcomes of Medicare enrolees who developed venous leg ulcers.

OBJECTIVE: To retrospectively evaluate the comorbidities, treatment patterns and outcomes of Medicare enrolees who developed venous leg ulcers (VLUs). METHOD: Medicare Limited Data Standard Analytic Hospital Inpatient and Outpatient Department Files were used to follow patients who received medical care for a VLU between 1 October 2015 and 2 October 2019. Patients diagnosed with chronic venous insufficiency (CVI) and a VLU were propensity matched into four groups based on their treatment regimen. Episode claims were used to document demographics, comorbidities and treatments of Medicare enrolees who developed VLUs, as well as important outcomes, such as time to ulcer closure, rates of complications and hospital utilisation rates. Outcomes were compared across key propensity-matched groups. RESULTS: In total, 42% of Medicare enrolees with CVI (n=1,225,278), developed at least one VLU during the study, and 79% had their episode claim completed within one year. However, 59% of patients developed another VLU during the study period. This analysis shows that only 38.4% of VLU episodes received documented VLU conservative care treatment. Propensity-matched episodes that received an advanced treatment or high-cost skin substitutes for a wound which had not progressed by 30 days demonstrated the best outcomes when their cellular, acellular, matrix-like product (CAMP) treatment was applied weekly or biweekly (following parameters for use). Complications such as rates of infection (33%) and emergency department visits (>50%) decreased among patients who received an advanced treatment (following parameters for use). CONCLUSION: Medicare enrolees with CVI have diverse comorbidities and many do not receive sufficient management, which contributes to high rates of VLUs and subsequent complications. Medicare patients at risk of a VLU who receive early identification and advanced CAMP treatment demonstrated improved quality of life and significantly reduced healthcare resource utilisation.

Cost-per-Response Analysis of Pure Hypochlorous Acid Among Patients with Chronic Venous Leg Ulcers: A Health Economic Analysis.

OBJECTIVE: To estimate the total cost-per-wound healing response (CPR) and the per-day CPR of patients with chronic leg ulcers treated with pure hypochlorous acid (pHA) as part of their overall would healing regimen. METHODS: The authors developed a deterministic decision-tree model to estimate the incremental CPR for pHA. The analysis was performed using clinical data from a published single-arm prospective study. The outcome of interest was re-epithelialization at 90 days. Economic data for pHA were based on public prices of pHA per dressing change from the wound care center perspective. The following time points were assessed: 90, 60, and 30 days. Dressing changes occurred every 2.5 days. Sensitivity analysis was performed to gauge the robustness of the results. RESULTS: A total of 31 patients (68% women) with 31 lesions (average age of wound, 29 months; range, 1-240 months) were included in the clinical study. Re-epithelialization occurred in 23 lesions (74%) at 90 days, 17 (55%) at 60 days, and 3 (10%) at 30 days. The total CPRs were $75.69, $68.27, and $193.44, and the per-day CPRs were $0.84, $1.13, and $6.45 at 90, 60, and 30 days, respectively. The sensitivity analysis revealed that CPRs ranged from $0.63 to $1.12 per day at 90 days. CONCLUSIONS: Incorporating pHA into standard wound healing protocols is a minimal added expense and may yield a substantial economic savings of $2,695 at 90 days.

Best practice for wound repair and regeneration use of cellular, acellular and matrix-like products (CAMPs).

There are currently over 80 biomaterials derived from autologous, allogeneic, synthetic and xenogeneic sources, or a combination of any or all these types of materials, available for soft-tissue coverage to effect wound closure. Often generically referred to as cellular and/or tissue-based products (CTPs), they are manufactured under various trade names and marketed for a variety of indications.

Multi-modal wound classification using wound image and location by deep neural network.

Wound classification is an essential step of wound diagnosis. An efficient classifier can assist wound specialists in classifying wound types with less financial and time costs and help them decide on an optimal treatment procedure. This study developed a deep neural network-based multi-modal classifier using wound images and their corresponding locations to categorize them into multiple classes, including diabetic, pressure, surgical, and venous ulcers. A body map was also developed to prepare the location data, which can help wound specialists tag wound locations more efficiently. Three datasets containing images and their corresponding location information were designed with the help of wound specialists. The multi-modal network was developed by concatenating the image-based and location-based classifier outputs with other modifications. The maximum accuracy on mixed-class classifications (containing background and normal skin) varies from 82.48 to 100% in different experiments. The maximum accuracy on wound-class classifications (containing only diabetic, pressure, surgical, and venous) varies from 72.95 to 97.12% in various experiments. The proposed multi-modal network also showed a significant improvement in results from the previous works of literature.

Evidence-based Potential Therapeutic Applications of Cannabinoids in Wound Management.

BACKGROUND: Although wound management is a major component of all domains of healthcare, conventional therapeutics have numerous limitations. The endocannabinoid system of the skin, one of the major endogenous systems, has recently been connected to wound healing. Cannabinoids and their interactions with the endogenous chemical signaling system may be a promising therapeutic option because they address some of the fundamental pathways for physiologic derangement that underpin chronic integumentary wounds. RECENT ADVANCES: The therapeutic applications of cannabinoids are increasing because of their legalization and resulting market expansion. Recently, their immunosuppressive and anti-inflammatory properties have been explored for the treatment of wounds that are not effectively managed by conventional medicines. CRITICAL ISSUES: Failure to manage wounds effectively is associated with reduced quality of life, disability, mortality, and increased healthcare expenditures. Therapeutic options that can manage wounds effectively and efficiently are needed. In this review, the authors summarize recent advances on the use of cannabinoids to treat skin disorders with an emphasis on wound management. FUTURE DIRECTIONS: Effective wound management requires medicines with good therapeutic outcomes and minimal adverse effects. Despite the promising results of cannabinoids in wound management, further controlled clinical studies are required to establish the definitive role of these compounds in the pathophysiology of wounds and their usefulness in the clinical setting.

Cost-effectiveness of dehydrated human amnion/chorion membrane allografts in lower extremity diabetic ulcer treatment.

OBJECTIVE: To evaluate the cost-effectiveness and budget impact of using standard care (no advanced treatment, NAT) compared with an advanced treatment (AT), dehydrated human amnion/chorion membrane (DHACM), when following parameters for use (FPFU) in treating lower extremity diabetic ulcers (LEDUs). METHOD: We analysed a retrospective cohort of Medicare patients (2015-2019) to generate four propensity-matched cohorts of LEDU episodes. Outcomes for DHACM and NAT, such as amputations, and healthcare utilisation were tracked from claims codes, analysed and used to build a hybrid economic model, combining a one-year decision tree and a four-year Markov model. The budget impact was evaluated in the difference in per member per month spending following completion of the decision tree. Likewise, the cost-effectiveness was analysed before and after the Markov model at a willingness to pay (WTP) threshold of $100,000 per quality adjusted life year (QALY). The analysis was conducted from the healthcare sector perspective. RESULTS: There were 10,900,127 patients with a diagnosis of diabetes, of whom 1,213,614 had an LEDU. Propensity-matched Group 1 was generated from the 19,910 episodes that received AT. Only 9.2% of episodes were FPFU and DHACM was identified as the most widely used AT product among Medicare episodes. Propensity-matched Group 4 was limited by the 590 episodes that used DHACM FPFU. Episodes treated with DHACM FPFU had statistically fewer amputations and healthcare utilisation. In year one, DHACM FPFU provided an additional 0.013 QALYs, while saving $3,670 per patient. At a WTP of $100,000 per QALY, the five-year net monetary benefit was $5003. CONCLUSION: The findings of this study showed that DHACM FPFU reduced costs and improved clinical benefits compared with NAT for LEDU Medicare patients. DHACM FPFU provided better clinical outcomes than NAT by reducing major amputations, ED visits, inpatient admissions and readmissions. These clinical gains were achieved at a lower cost, in years 1-5, and were likely to be cost-effective at any WTP threshold. Adoption of best practices identified in this retrospective analysis is expected to generate clinically significant decreases in amputations and hospital utilisation while saving money.

Image-Based Artificial Intelligence in Wound Assessment: A Systematic Review.

Significance: Accurately predicting wound healing trajectories is difficult for wound care clinicians due to the complex and dynamic processes involved in wound healing. Wound care teams capture images of wounds during clinical visits generating big datasets over time. Developing novel artificial intelligence (AI) systems can help clinicians diagnose, assess the effectiveness of therapy, and predict healing outcomes. Recent Advances: Rapid developments in computer processing have enabled the development of AI-based systems that can improve the diagnosis and effectiveness of therapy in various clinical specializations. In the past decade, we have witnessed AI revolutionizing all types of medical imaging like X-ray, ultrasound, computed tomography, magnetic resonance imaging, etc., but AI-based systems remain to be developed clinically and computationally for high-quality wound care that can result in better patient outcomes. Critical Issues: In the current standard of care, collecting wound images on every clinical visit, interpreting and archiving the data are cumbersome and time consuming. Commercial platforms are developed to capture images, perform wound measurements, and provide clinicians with a workflow for diagnosis, but AI-based systems are still in their infancy. This systematic review summarizes the breadth and depth of the most recent and relevant work in intelligent image-based data analysis and system developments for wound assessment. Future Directions: With increasing availabilities of massive data (wound images, wound-specific electronic health records, etc.) as well as powerful computing resources, AI-based digital platforms will play a significant role in delivering data-driven care to people suffering from debilitating chronic wounds.

Advances in Hemostatic Wound Dressings: Clinical Implications and Insight.

ABSTRACT: Hemorrhagic shock is one of the leading causes of prehospital death in the armed forces. In this state, the body begins to shut down because of blood volume depletion. In both civilian and military trauma, a significant number of hemorrhage deaths occur in the first several hours after injury. Researchers all over the globe are working to develop relatively inexpensive and easy-to-transport products to prevent soldiers from going into hemorrhagic shock. For example, many advances have been made during the last several years toward the development of ideal hemostatic dressings. No current hemostatic agents meet all of the requirements, but the ideal dressing would fulfill many important measures: minimizes or stops blood flow within minutes, contains hemostatic agents to enhance blood clotting, is easy to apply, does not need preapplication preparation, has a reasonably long shelf life, is safe to use, prevents bacterial or viral transmission, is stable at extreme temperatures, and is inexpensive. For this literature review, the authors conducted an extensive search of academic scientific databases for relevant keywords and assessed and summarized the results. This review aimed to identify recent advances in hemostatic wound dressings; summarize the currently available dressings and their supporting literature; and discuss the compositions, mechanisms of action, and clinical relevance of each category of dressing. In addition, case studies and suggestions for future research into hemorrhage control with new hemostatic agents are provided.

Synthesizing time-series wound prognosis factors from electronic medical records using generative adversarial networks.

Wound prognostic models not only provide an estimate of wound healing time to motivate patients to follow up their treatments but also can help clinicians to decide whether to use a standard care or adjuvant therapies and to assist them with designing clinical trials. However, collecting prognosis factors from Electronic Medical Records (EMR) of patients is challenging due to privacy, sensitivity, and confidentiality. In this study, we developed time series medical generative adversarial networks (GANs) to generate synthetic wound prognosis factors using very limited information collected during routine care in a specialized wound care facility. The generated prognosis variables are used in developing a predictive model for chronic wound healing trajectory. Our novel medical GAN can produce both continuous and categorical features from EMR. Moreover, we applied temporal information to our model by considering data collected from the weekly follow-ups of patients. Conditional training strategies were utilized to enhance training and generate classified data in terms of healing or non-healing. The ability of the proposed model to generate realistic EMR data was evaluated by TSTR (test on the synthetic, train on the real), discriminative accuracy, and visualization. We utilized samples generated by our proposed GAN in training a prognosis model to demonstrate its real-life application. Using the generated samples in training predictive models improved the classification accuracy by 6.66-10.01% compared to the previous EMR-GAN. Additionally, the suggested prognosis classifier has achieved the area under the curve (AUC) of 0.875, 0.810, and 0.647 when training the network using data from the first three visits, first two visits, and first visit, respectively. These results indicate a significant improvement in wound healing prediction compared to the previous prognosis models.

Observed impact of skin substitutes in lower extremity diabetic ulcers: lessons from the Medicare Database (2015-2018).

OBJECTIVE: To evaluate large propensity-matched cohorts to assess outcomes in patients receiving advanced treatment (AT) with skin substitutes for lower extremity diabetic ulcers (LEDUs) versus no AT (NAT) for the management of LEDUs. METHOD: The Medicare Limited Dataset (1 October 2015 through 2 October 2018) were used to retrospectively analyse people receiving care for a LEDU treated with AT or NAT (propensity-matched Group 1). Analysis included major and minor amputations, emergency department (ED) visits and hospital readmissions. In addition, AT following parameters for use (FPFU) was compared with AT not FPFU (propensity-matched Group 2). A paired t-test was used for comparisons of the two groups. For comparisons of three groups, the Kruskal-Wallis test was used. A Bonferroni correction was performed when multiple comparisons were calculated. RESULTS: There were 9,738,760 patients with a diagnosis of diabetes, of whom 909,813 had a LEDU. In propensity-matched Group 1 (12,676 episodes per cohort), AT patients had statistically fewer minor amputations (p=0.0367), major amputations (p<0.0001), ED visits (p<0.0001), and readmissions (p<0.0001) compared with NAT patients. In propensity-matched Group 2 (1131 episodes per cohort), AT FPFU patients had fewer minor amputations (p=0.002) than those in the AT not FPFU group. CONCLUSION: AT for the management of LEDUs was associated with significant reductions in major and minor amputation, ED use, and hospital readmission compared with LEDUs managed with NAT. Clinics should implement AT in accordance with the highlighted parameters for use to improve outcomes and reduce costs.

Multiclass wound image classification using an ensemble deep CNN-based classifier.

Acute and chronic wounds are a challenge to healthcare systems around the world and affect many people's lives annually. Wound classification is a key step in wound diagnosis that would help clinicians to identify an optimal treatment procedure. Hence, having a high-performance classifier assists wound specialists to classify wound types with less financial and time costs. Different wound classification methods based on machine learning and deep learning have been proposed in the literature. In this study, we have developed an ensemble Deep Convolutional Neural Network-based classifier to categorize wound images into multiple classes including surgical, diabetic, and venous ulcers. The output classification scores of two classifiers (namely, patch-wise and image-wise) are fed into a Multilayer Perceptron to provide a superior classification performance. A 5-fold cross-validation approach is used to evaluate the proposed method. We obtained maximum and average classification accuracy values of 96.4% and 94.28% for binary and 91.9% and 87.7% for 3-class classification problems. The proposed classifier was compared with some common deep classifiers and showed significantly higher accuracy metrics. We also tested the proposed method on the Medetec wound image dataset, and the accuracy values of 91.2% and 82.9% were obtained for binary and 3-class classifications. The results show that our proposed method can be used effectively as a decision support system in classification of wound images or other related clinical applications.

In Vitro Characterization of Pressure Redistribution Among Commercially Available Wound Dressings.

OBJECTIVE: Recent clinical evidence has suggested that certain wound dressings may play a significant role in protocols to prevent or reduce pressure injury (PI) in patients at risk by modifying the pressure, friction, and shear forces that can contribute to PI. The aim of this study was to investigate the pressure reduction properties of commercially available wound dressings in vitro. METHODS: Using a standardized protocol (1.7 kg, 7.5-cm sphere), testing was performed in a controlled environment by the same clinician using a pressure mapping device (XSENSOR LX205; XSENSOR Technology Corporation, Calgary, Alberta, Canada) to measure and compare the pressure mitigation properties in a variety of wound dressings. RESULTS: A total of 13 different commercially available dressings were tested in triplicate for changes in pressure redistribution as compared with the control. One dressing demonstrated the greatest reduction of pressure forces (OxyBand PR; 50.33 ± 1.45 mm Hg) compared with the control (302.7 ± 0.33 mm Hg) and the greatest surface area of all the study dressings tested. There was a negative correlation (R2 = 0.73) between the average pressure distribution of a wound dressing and its contact area. Further, the peak pressure for OxyBand PR (P ≤ .05) was significantly different from all other tested dressings. CONCLUSIONS: One dressing (OxyBand PR) provided superior pressure redistribution and significantly reduced peak pressure in this study when compared with currently available standard foam and silicone dressings that are marketed for the purpose of PI prevention.

Fully automatic wound segmentation with deep convolutional neural networks.

Acute and chronic wounds have varying etiologies and are an economic burden to healthcare systems around the world. The advanced wound care market is expected to exceed $22 billion by 2024. Wound care professionals rely heavily on images and image documentation for proper diagnosis and treatment. Unfortunately lack of expertise can lead to improper diagnosis of wound etiology and inaccurate wound management and documentation. Fully automatic segmentation of wound areas in natural images is an important part of the diagnosis and care protocol since it is crucial to measure the area of the wound and provide quantitative parameters in the treatment. Various deep learning models have gained success in image analysis including semantic segmentation. This manuscript proposes a novel convolutional framework based on MobileNetV2 and connected component labelling to segment wound regions from natural images. The advantage of this model is its lightweight and less compute-intensive architecture. The performance is not compromised and is comparable to deeper neural networks. We build an annotated wound image dataset consisting of 1109 foot ulcer images from 889 patients to train and test the deep learning models. We demonstrate the effectiveness and mobility of our method by conducting comprehensive experiments and analyses on various segmentation neural networks. The full implementation is available at https://github.com/uwm-bigdata/wound-segmentation .