Mathematicians at Cambridge University have used images from an ioLight field microscope to spot intestinal parasites, reducing drug overuse and increasing farm profitability.
Intestinal parasites are a growing problem in many animals including horses, cattle, pigs, sheep and chickens. These parasites cause illness and even death in infected animals and reduce the farmer’s productivity. Up until now the parasite worms were controlled by indiscriminate use of anti-parasitic drugs, treating all animals whether they were infected or not. Unfortunately, this has led to drug-resistant strains of parasite worms developing, which farmers and horse owners have no way of controlling. An infestation of one of these drug-resistant intestinal worms would have devastating consequences for farmers and horse owners.
The solution to this problem is to restrict the use of the anti-parasitic drugs so that only animals infected with harmful parasite worms are treated. In this way, only the harmful parasites are killed and other strains are allowed to live.
To treat only infected animals, farmers and horse owners need to perform a worm egg count. This test involves examining each animal’s faeces to see if the eggs of harmful parasitic worms are present in dangerous quantities. Today, this is done by sending samples off to a lab, where the sample is manually analysed using a lab microscope and the number of parasite eggs counted. ioLight has developed a pocket microscope which enables farmers and vets to do this in the field. This reduces the time taken to do the analysis so that the animals can be treated more rapidly and smaller doses of anti-parasitic drugs used.
Despite the ioLight field microscope making worm egg counts easier and more efficient, the vet or farmer still has to count the eggs manually by looking at the microscope images. This takes time, and is difficult unless you are an expert.
Margaret Duff of Cambridge University has produced proof-of-concept image analysis software using MATLAB to analyse the microscope images automatically and count parasite eggs. Margaret showed that using a combination of algorithms and machine learning, she was able to detect eggs correctly with approximately 85% success rate. This shows great potential to help vets and farmers increase productivity and win the fight against drug-resistant parasites.
Margaret’s work is a collaboration between The Cantab Capital Institute for the Mathematics of Information and The Centre for Mathematical Imaging in Healthcare (both at Cambridge University), The Mathworks LTD and Cancer Research UK. ioLight would like to thank Carola-Bibiane Schonlieb and Joana Grah of Cambridge University, Jasmina Lazic (Bayes Centre, University of Edinburgh), Sylvain Sauvage (Mathworks) and Stefanie Reichelt (Cancer Research UK) for their support and supervision, without which this work would not have been possible.
The details of Margaret’s work can be read on her GitHub repository.