Clinical validation of a smartphone application for automated wound measurement in patients with venous leg ulcers

Development of an explainable artificial intelligence model for Asian vascular wound images

Chronic wounds contribute to significant healthcare and economic burden worldwide. Wound assessment remains challenging given its complex and dynamic nature. The use of artificial intelligence (AI) and machine learning methods in wound analysis is promising. Explainable modelling can help its integration and acceptance in healthcare systems. We aim to develop an explainable AI model for analysing vascular wound images among an Asian population. Two thousand nine hundred and fifty-seven wound images from a vascular wound image registry from a tertiary institution in Singapore were utilized. The dataset was split into training, validation and test sets. Wound images were classified into four types (neuroischaemic ulcer [NIU], surgical site infections [SSI], venous leg ulcers [VLU], pressure ulcer [PU]), measured with automatic estimation of width, length and depth and segmented into 18 wound and peri-wound features. Data pre-processing was performed using oversampling and augmentation techniques. Convolutional and deep learning models were utilized for model development. The model was evaluated with accuracy, F1 score and receiver operating characteristic (ROC) curves. Explainability methods were used to interpret AI decision reasoning. A web browser application was developed to demonstrate results of the wound AI model with explainability. After development, the model was tested on additional 15 476 unlabelled images to evaluate effectiveness. After the development on the training and validation dataset, the model performance on unseen labelled images in the test set achieved an AUROC of 0.99 for wound classification with mean accuracy of 95.9%. For wound measurements, the model achieved AUROC of 0.97 with mean accuracy of 85.0% for depth classification, and AUROC of 0.92 with mean accuracy of 87.1% for width and length determination. For wound segmentation, an AUROC of 0.95 and mean accuracy of 87.8% was achieved. Testing on unlabelled images, the model confidence score for wound classification was 82.8% with an explainability score of 60.6%. Confidence score was 87.6% for depth classification with 68.0% explainability score, while width and length measurement obtained 93.0% accuracy score with 76.6% explainability. Confidence score for wound segmentation was 83.9%, while explainability was 72.1%. Using explainable AI models, we have developed an algorithm and application for analysis of vascular wound images from an Asian population with accuracy and explainability. With further development, it can be utilized as a clinical decision support system and integrated into existing healthcare electronic systems.

The clinical value of intelligent wound measurement devices in patients with chronic wounds: A scoping review

Chronic wounds are common in clinical practice, with long treatment cycle and high treatment cost. Changes in wound area can well predict the effectiveness of treatment and the possibility of healing. Therefore, continuous wound monitoring and evaluation are particularly important. Traditional manual wound measurement tends to overestimate wound area. Recently, various intelligent wound measurement devices have been introduced into clinical practice. This review aims to summarise the reliability, validity, types and measurement principles of different intelligent wound measurement devices, so as to analyse the clinical value and application prospect. Articles numbering 2610 were retrieved from the database, and 14 articles met the inclusion criteria. The results showed that the intelligent wound measurement devices included in the study reported good reliability and validity. Contact devices can lead to wound bed damage, wound deformation, patient pain, and is not convenient for electronic wound recording; partial contact devices can complete continuous monitoring and recording of wounds, but are not sensitive to wound depth measurement. Non-contact devices are more accurate in capturing wound images. In addition to wound measurement, they also have the function of wound assessment. In general, handheld and portable non-contact devices have great clinical value and promotion prospects.

Associations between self-care advice and healing time in patients with venous leg ulcer- a Swedish registry-based study

BACKGROUND

Venous leg ulcers take time to heal. It is advocated that physical activity plays a role in healing, and so does the patient's nutritional status. Additionally, malnutrition influences the inflammatory processes, which extends the healing time. Therefore, the staff's advising role is important for patient outcomes. Thus, this study aimed to investigate the associations between given self-care advice and healing time in patients with venous leg ulcers while controlling for demographic and ulcer-related factors.

METHODS

The sample consisted of patients registered in the Registry of Ulcer Treatment (RUT) which includes patient and ulcer-related and healing variables. The data was analyzed with descriptive statistics. Logistic regression models were performed to investigate the influence of self-care advice on healing time.

RESULTS

No associations between shorter healing time (less than 70 days) and the staff´s self-care advice on physical activity was identified, whilst pain (OR 1.90, CI 1.32-2.42, p < 0.001) and giving of nutrition advice (OR 1.55, CI 1.12-2.15, p = 0.009) showed an association with longer healing time.

CONCLUSIONS

Neither self-care advice on nutrition and/or physical activity indicated to have a positive association with shorter healing time. However, information and counseling might not be enough. We emphasize the importance of continuously and systematically following up given advice throughout ulcer management, not only when having complicated ulcers.

Intelligent Care Management for Diabetic Foot Ulcers: A Scoping Review of Computer Vision and Machine Learning Techniques and Applications

Ten percent of adults in the United States have a diagnosis of diabetes and up to a third of these individuals will develop a diabetic foot ulcer (DFU) in their lifetime. Of those who develop a DFU, a fifth will ultimately require amputation with a mortality rate of up to 70% within five years. The human suffering, economic burden, and disproportionate impact of diabetes on communities of color has led to increasing interest in the use of computer vision (CV) and machine learning (ML) techniques to aid the detection, characterization, monitoring, and even prediction of DFUs. Remote monitoring and automated classification are expected to revolutionize wound care by allowing patients to self-monitor their wound pathology, assist in the remote triaging of patients by clinicians, and allow for more immediate interventions when necessary. This scoping review provides an overview of applicable CV and ML techniques. This includes automated CV methods developed for remote assessment of wound photographs, as well as predictive ML algorithms that leverage heterogeneous data streams. We discuss the benefits of such applications and the role they may play in diabetic foot care moving forward. We highlight both the need for, and possibilities of, computational sensing systems to improve diabetic foot care and bring greater knowledge to patients in need.

Chronic Wound Management: From Gauze to Homologous Cellular Matrix

BACKGROUND

Chronic wounds are a significant health problem with devastating consequences for patients' physical, social, and mental health, increasing healthcare systems' costs. Their prolonged healing times, economic burden, diminished quality of life, increased infection risk, and impact on patients' mobility and functionality make them a major concern for healthcare professionals.

PURPOSE

This review offers a multi-perspective analysis of the medical literature focusing on chronic wound management.

METHODS USED

We evaluated 48 articles from the last 21 years registered in the MEDLINE and Global Health databases. The articles included in our study had a minimum of 20 citations, patients > 18 years old, and focused on chronic, complex, and hard-to-heal wounds. Extracted data were summarized into a narrative synthesis using the same health-related quality of life instrument.

RESULTS

We evaluated the efficacy of existing wound care therapies from classical methods to modern concepts, and wound care products to regenerative medicine that uses a patient's pluripotent stem cells and growth factors. Regenerative medicine and stem cell therapies, biologic dressings and scaffolds, negative pressure wound therapy (NPWT), electrical stimulation, topical growth factors and cytokines, hyperbaric oxygen therapy (HBOT), advanced wound dressings, artificial intelligence (AI), and digital wound management are all part of the new arsenal of wound healing.

CONCLUSION

Periodic medical evaluation and proper use of modern wound care therapies, including the use of plasma-derived products [such as platelet-rich plasma (PRP) and platelet-rich fibrin (PRF)] combined with proper systemic support (adequate protein levels, blood sugar, vitamins involved in tissue regeneration, etc.) are the key to a faster wound healing, and, with the help of AI, can reach the fastest healing rate possible.

Determining the measurement accuracy in assessing the progress of wound healing

Introduction

Wound management is a challenge in terms of the way, duration and cost of treatment both for the patient and health providers. The healing of skin wounds is a highly multi-step coordinated process. Objective monitoring of treatment at every stage is necessary to assess the applied therapy.

Aim

To show the possibility of using the AutoCad software (ACS) as a tool with a slight measurement error for accurate measurement of the venous leg ulcers on the lower limbs.

Material and methods

To determine the error of the measurement method Circle Templates For Drafting for four different sizes were used as ulcer models. Seventy-six wounds of various sizes from patients with venous leg ulcers (VLUs) were photographed and outlined with a marker on a transparent foil. The wounds were measured both using ACS and digital planimetry with C-Geo software (CGS). Data were analysed using Wilcoxon test, intraclass correlation coefficient (ICC) and Bland-Altman analysis.

Results

The mean relative error of the surface wound model area measured by the ACS was 0.30 ±0.31% (range: 0.004-1.25) and a median of 0.18%. Areas and perimeters measured with ACS were higher than areas and perimeters measured with CGS, and the difference was statistically significant.

Conclusions

The analysis of the wound images obtained in the ACS showed a very high potential of the software in terms of the accuracy of the analysed areas, which significantly increases the possibility of the analysis and reduces the measurement error in relation to planimetry using a digital digitizer.

Implementation of a patient-centered remote wound monitoring system for management of diabetic foot ulcers

Background

Regular clinical assessment is critical to optimize lower extremity wound healing. However, family and work obligations, socioeconomic, transportation, and time barriers often limit patient follow-up. We assessed the feasibility of a novel, patient-centered, remote wound management system (Healthy.io Minuteful for Wound Digital Management System) for the surveillance of lower extremity wounds.

Methods

We enrolled 25 patients from our outpatient multidisciplinary limb preservation clinic with a diabetic foot ulcer, who had undergone revascularization and podiatric interventions prior to enrollment. Patients and their caregivers were instructed on how to use the digital management system and asked to perform one at-home wound scan per week for a total of 8 weeks using a smartphone application. We collected prospective data on patient engagement, smartphone app useability, and patient satisfaction.

Results

Twenty-five patients (mean age 65.5 ± 13.7 years, 60.0% male, 52.0% Black) were enrolled over 3 months. Mean baseline wound area was 18.0 ± 15.2 cm2, 24.0% of patients were recovering from osteomyelitis, and post-surgical WiFi stage was 1 in 24.0%, 2 in 40.0%, 3 in 28.0%, and 4 in 8.00% of patients. We provided a smartphone to 28.0% of patients who did not have access to one that was compatible with the technology. Wound scans were obtained by patients (40.0%) and caregivers (60.0%). Overall, 179 wound scans were submitted through the app. The mean number of wound scans acquired per patient was 0.72 ± 0.63 per week, for a total mean of 5.80 ± 5.30 scans over the course of 8 weeks. Use of the digital wound management system triggered an early change in wound management for 36.0% of patients. Patient satisfaction was high; 94.0% of patients reported the system was useful.

Conclusion

The Healthy.io Minuteful for Wound Digital Management System is a feasible means of remote wound monitoring for use by patients and/or their caregivers.