AutoMated BUrn Diagnostic System for Healthcare (AMBUSH)

Overview

The primary objective of this study is to develop a high accuracy and automated system that can provide early assessment of burn injuries with at least 90% accuracy in absence of burn experts, using AI and FDA cleared harmonic ultrasound TDI data based on the analysis of mechanical and hemodynamic properties of the subcutaneous burned tissue. Data collected in this study will lead to the development of better diagnostic tools that could inform clinical burn practices by enabling doctors to determine burn depth and the need for surgery with greater speed and accuracy, resulting in better clinical outcomes.

Burn injuries are considered as one of the most complex type of traumatic injuries. In the United States (US), about 1.25 million people are treated each year for burns, and 40,000 are hospitalized for the treatment of these injuries resulting in high medical costs, approximately $7.9 billion per year. Early and accurate treatment of burns is critical to prevent infections and improve the possible outcomes of the patient, decreasing the mortality rate by about 36% In this regard, prioritizing burns that require a surgical procedure to heal, i.e. burn excision and skin grafting, is critical. This constitutes a challenging task because it involves the determination of the burn depth, for which experienced burn surgeons achieve an accuracy of 67-76%, value that decreases to 50% for inexperienced surgeons. Assessment of the burn depth is one of the most important aspects of burn care. It is a predictor of pathological scarring that occurs in 30%-91% of burn injuries. However, it continues to be an open clinical challenge for which an accurate solution has yet to be found. Burns are classified into three different categories: superficial burns, which involve only epidermis; partial-thickness burns, which affect epidermis and dermis; and full-thickness burns, which include deep structures such as subcutaneous, muscles, and bone. The first two are expected to heal without scarring by 14-21 days after the burn injury, while the last one will heal with scarring resulting in significant morbidity to the patient, including pain, loss of joint mobility, loss of function, and social isolation. For this reason, full-thickness burns are excised, and exposed body parts are covered with skin grafts to prevent further complications with scarring Currently, burn depth is determined by clinical assessment, based on appearance, blanching to pressure, sensation to pin prick, and bleeding on needle prick. This visual and tactile inspection approach introduces inter-subject variability, especially when partial-thickness burns are involved. Overestimation of burn depth leads to unnecessary surgery of excision of viable skin and replacement with skin grafts that look different from surrounding skin. These grafts are less pliable with lack of the ability to sweat for thermoregulation, while underestimation leads to surgical delay, long length of hospital stays, high treatment costs, scarring, and poor functional and aesthetic outcomes. This issue is exacerbated by a phenomenon known as burn conversion, which is not fully understood yet, and it is usually not accounted for in the assessment process or in the clinical decision support technologies. Burn injury is a dynamic process, the longer the delay in intervening, the more abundant the scar tissue formation and the more likely the patient is to have deformity with loss of mobility and function. Thus, early prediction of burn conversion is critical for surgical decision making and burn follow up. Several non-invasive light-based imaging technologies have been developed to support clinical assessment, which is currently the most common technique for burn depth assessment. They operate based on the existing correlation among blood perfusion, functional blood vessels, and burn depth. However, Laser Doppler Imaging (LDI) is the only one approved by FDA. It is the most widely adopted non-invasive technology in burn treatment facilities in the US, but it is used as the preferred modality in only 6% of them. Considering the many limitations of this technology, this is not unexpected. It uses light/optic principles to detect the active blood flow in the damaged tissue of the patient, and the results can be greatly altered by the curvature of tissues, ambient light and temperature, motion artifacts, topical wound dressings, non-debrided tissue, skin color, pigment from tattoos, and blisters. In addition, LDI performance is poor when burn injuries are less than 24 hours old and can be difficult to interpret for inexperienced users. To overcome the limitations of light-based technologies, the investigators will integrate novel techniques in Artificial Intelligence (AI), Computer Vision, with FDA approved Harmonic B-mode ultrasound (HUSD B-mode) and Harmonic Ultrasound Tissue Doppler Elastography imaging (TDI), medical expertise in the wound care, tissue injury management, and burn surgery domains. This will enhance the burn assessment accuracy, improving the patient's prognosis.