Hi-Res IR Thermography for the Diagnosis of Toddler's Fractures

High Resolution Infrared Thermal Imaging for the Diagnosis of Toddler's Fractures: A Pilot Diagnostic Study

This is a pilot proof of feasibility study to explore the efficacy of high resolution thermal imaging (HRTI) to screen for toddler's fractures. In HRTI, a sensitive thermal camera is used to capture the temperature profile of the suspected fracture region and then analyse and interpret the information using image processing. Toddler's fractures are a common type of tibial fracture in children under 5 years old that can be particularly difficult to detect as the bone typically appears normal on x-ray for the first 10 days and becomes detectable thereafter only as new bone forms. The child's leg is put in plaster in the interim and the diagnosis cannot be confirmed until 10-14 days later. Around 60% of children with suspected toddler's fracture will not have a fracture, and would therefore have had plaster, repeat attendances and repeat x-ray unnecessarily. There is also the inconvenience and general risks of plaster immobilisation for the child. In an earlier study, the investigators demonstrated that HRTI has potential in diagnosing limp in children. This study included two children with toddler's fracture where HRTI detected the fracture's location on the patient's first Emergency Department(ED) visit. In this study, the investigators aim to build on earlier work to explore HRTI for screening for toddler's fracture. On its successful completion, the investigators aim to develop the technology in a follow on larger diagnostic study as a tool that could be used to quickly exclude cases where the injury has not resulted in a fracture. This would allow more objective decision making during the initial assessment of the child. In addition, a reduction in the number of unnecessary x-rays, revisits, treatment as well as cost is anticipated.

High resolution thermal imaging (HRTI) imaging is based on recording and processing of a part of electromagnetic spectrum below visible light i.e. infrared (IR) band. Objects with a temperature above -273 °C (-459.7 °F) emit IR (i.e. thermal) radiation. In HRTI a highly sensitive thermal camera that operates to the mid (3-5 μm) and long (7-14 μm) IR bands of the electromagnetic spectrum is used for imaging. The method results in a series of images (capture rate typically 30 frames per second). The images are processed using specialised software to extract and interpret its information. This includes neural networks and similar artificial intelligence models. Detection of an injury using thermal imaging relies on underlying physiology of temperature differentials. Dermal temperature differentials usually do not exceed 0.25°C, while differentials in excess of 0.65 °C are consistently related to pathology. Therefore observation of a significant temperature differential can be an indication of an injury. The use of infrared HRTI in paediatric for musculoskeletal diagnosis and monitoring as well as physiological measurements have showed potential. Examples of the related studies are: Vertebral fractures were detected in osteogenesis imperfecta patients using thermal imaging . Thermal imaging assisted in diagnosis of limp, including bone fracture cases Thermal imaging showed potential for differentiating between wrist fracture and sprain. Thermal imaging could accurately quantify the temperature difference between inflamed and uninflamed knees thus assisting with the diagnosis of juvenile idiopathic arthritis. Infrared thermal imaging has proved valuable in detecting inflammatory intra-abdominal pathology in infants. Thermal imaging provided effective for measuring respiration rate in a non- contact manner, i.e. no sensing unit attached to the patient's body. Blood convection warms the skin by transfer of heat from the core and this process plays the major role in determining skin temperature. Skin's has a thermoregulatory role, i.e. it generates, absorbs, conducts and radiates heat. Changes in the skin surface temperature are valuable in detecting physiological and pathological states such as inflammation. With recent developments in thermal imaging devices, the use of infrared imaging for injury examination is expanding, with more evidence supporting its use. However, the data in children are still limited, with the investigators' research group undertaking significant development work in this field. In this study the IR emission (characterised by heat radiation) from the skin at the site of injury is imaged and analysed to screen for a toddler's fracture. The hypothesis is that the inflammation and blood perfusion in fracture and less severe injuries at the site of injury are distinct, leading to distinct temperature gradients. Toddler's fracture is characterised as a non-displaced spiral fracture of the tibial shaft in young children, usually between the ages of 9 months to 3 years. However other lower extremity injuries in young children can also have similar clinical appearance to the non-displaced spiral tibial fracture. Toddler's fracture usually results from an indirect innocuous twisting or rotational force applied to the foot and lower leg. The cause could be a stumble or fall or attempts to extricate the foot from between the bars of a crib for example. Innovations in the use of HRTI in screening toddler's fracture can be beneficial as this could reduce the number of unnecessary x-ray radiographs by filtering out cases where the bone is not fractured. Given that toddler's fracture may not be visible on the radiograph close to the time of the injury's occurrence and a repeat radiograph around 10 days is usually required, earlier identification of a fracture would be beneficial. HRTI may be able to detect a fracture at the index visit, thus allowing directed management.

All eligible children will have infrared imaging undertaken of both affected limb and unaffected limb simultaneously.

Diagnostic probability thresholds for Infrared data (delta-temperature thresholds) for identifying Toddler fractures.

Diagnostic performance (Sensitivity, Specificity) at different thresholds of infrared data measures (measure = delta-temperature; unit = degrees Celsius - as assessed by trained neural networks using temperature data from the high resolution IR camera). Gold standard is clinical diagnosis compared with neural network diagnosis in children with a toddlers fracture.

Feasibility of recruitment strategy and conversion rates (% eligible), reasons for non-inclusion/opting out of study (% of eligible). Attrition rates (%). To inform sample size calculation/feasibility of larger trial.

Patient/parent qualitative interviews to determine parent experience of care pathway and relative importance to design of future trial. Unvalidated questionnaire used in interviews for exploratory purposes, as no validated tool exists for this area of care.

Inclusion Criteria: Children aged 9 months-5 years (inclusive) Clinical suspicion of Toddler's fracture i.e x-rayed as part of their standard clinical management (confirmation of diagnosis). Injury within the preceding 72 hours. Ability to understand written and /or verbal consent and participant information. Exclusion Criteria: Significant pain or discomfort (requiring second line analgesia as defined by ED guidelines) Multiple injuries (not localised to a single lower limb)