Infrared Thermography for use as early Detection Method for Digital Dermatitis in Dairy Cattle

Abstract

Digital dermatitis (DD) is widely prevalent in cattle in the world and causes economic losses as it can cause lameness and a decreased milk yield. For adequate treatment, proper diagnostics are needed, which are preferably cost-effective and least labour intensive to assure producer implementation. Therefore a new, non-invasive way of detecting DD might be rewarding, such as infrared thermography (IRT), which detects heat as related to inflammation. The aim of this study was to investigate the use of infrared thermography (IRT) in the detection of DD in dairy cattle. The hypothesis was that a DD lesion goes with inflammation of the skin of the bulb of the heel and therefore radiates more heath than a not affected foot. A total of 850 cows (four farms visited five times, one farm visited once) were scored for DD, lameness and IRT images were taken before (pre-) and after (post-) hosing of the hooves. This resulted in 3,314 images, divided over two different IRT cameras: a cost-friendly and an expensive one. The means of the maximum temperatures from the images per DD stage were compared to each other and a statistical significance was found for the difference in the mean of the maximum temperatures between M0 and M2 stages for both pre and post images. Also, the differences in mean temperatures between absence of DD (M0) versus active DD lesions (M1 and M2), absence of DD versus chronic DD lesions (M3 and M4) and chronic versus active DD lesions were significantly different for pre images. For the post images only the mean temperatures for the absence of DD versus active DD lesions were significantly different. A sensitivity of 0.500 and a specificity of 0.697 was found in establishing a cutoff temperature to detect lameness using IRT on hind feet, which has therefore a limited value in lameness detection. Also there appeared to be a very strong correlation between the two cameras, meaning that the more expensive camera didn’t exceed the results of the cost-friendly camera and a strong correlation between pre- and post-images was found, which tackles the need of washing of the hooves. In conclusion, IRT was able to find significant differences between the mean maximum temperatures of the M stages of DD which is promising for a developing new method of early DD detection. Also the results support a future on-farm implementation as the cost-friendly camera performed equally well and hosing of the hooves didn’t improve any results and is therefore not adding extra labour.