Abstract

An Advanced Long-wave-infrared Imaging System (ALIAS) is being developed that utilizes infrared imaging to detect dynamic thermal signatures (DTS) of suspicious lesions with respect to healthy surrounding skin after the application of a temperature stimulus. This DTS is then used to classify lesions as malignant or benign. The ALIAS consists of an infrared and a visible camera, registration marker, cold air source, computer and software for controlling the image acquisition and analysis. Data is collected by first applying the registration marker over the lesion in question and acquiring a visible image of the lesion. The lesion and surrounding skin is then cooled and imaged as the lesion and surrounding skin warm up by natural convection with the air. The differences between the DTS of the lesions and healthy skin are then mathematically compared to distinguish malignant and benign conditions. This device was developed using an observational pilot study of 74 human subjects, 53% of which were male, and the average age was 55 years with a standard deviation of 17. Patients with suspicious lesions were given the opportunity to volunteer to have their lesions imaged prior to their biopsy. The biopsy results were then correlated with the ALIAS results in order to benchmark our approach. Malignant skin lesions were distinguished from benign lesions with a sensitivity of 96.97% and a specificity of 78.05%. The receiver operating curves (ROC) had an area under the curve (AUC) of 95.31%. It is noted that the malignant skin lesions include such cases as melanoma, squamous cell carcinoma (SCC), and basal cell carcinoma (BCC). 55% of lesions were benign, 34% were BCC, 7% were SCC, and 4% were melanoma. These results suggest the technique is promising as a non-invasive screening tool and is worth continued research and development to further improve the sensitivity, specificity, and statistical confidence.