Point-of-care device for Graves’ disease detection using Infrared Thermography

Abstract

IR Thermography (IRT) provides real-time, non-invasive sub-surface images of the thyroid region without any hazardous radiation exposures. Thermography represents thermal radiation in the form of an image and provides information about important physiological parameters using skin temperature distribution. Graves' disease is the most frequently diagnosed disease of women in India. Patients with graves' disease have different temperature distribution relative to the healthy ones. This paper aims to use thermography with deep learning and machine learning techniques to develop a GUI/non-invasive method for detecting graves' disease. A CNN-based app is developed and deployed in a clinical environment to add to the analysis so as to diminish sparsity in the available data set. For our study IR Images of the sixty persons with different age groups and gender has been taken and showed that control and hyperthyroid patient have different behaviour and lie in different quadrants, which can help clinicians to conclude whether the patient is healthy or not. It is expected that this imaging biomarker test will reduce the ethical burden on clinicians' shoulders before recommending scintigraphy and standard thyroid blood tests, thus saving costs at patients' end and pathology labs in populated countries such as India. An affordable Thermal plug-and-play attachment can be connected to any smartphone for faster diagnosis at the patient end, converting the smartphone into a point of care device.