Welcome to the website of HELiCoiD: HypErspectraL Imaging Cancer Detection.
HELICoiD was a European collaborative project funded by the Research Executive Agency, through the Future and Emerging Technologies (FET-Open) programme, under the 7th Framework Programme of the European Union. The project was a collaboration between four universities, three industrial partners and two hospitals, and ran from 1st January 2014 until 31st December 2016.
The aim of HELICoiD was to demonstrate the use of hyperspectral imaging for real-time identification of tumour margins during surgery, helping the surgeon to extract the entire tumour and to spare as much of the healthy tissue as possible. You can read more about the project outcomes
, and publications
Hyperspectral imaging is a non-contact, non-ionizing and minimally-invasive sensing technique. Whereas a conventional camera captures images in three colour channels (red, blue and green), a hyperspectral camera captures data over a large number of contiguous and narrow spectral bands, and over a wide spectral range. For this reason it is sometimes known as imaging spectroscopy. Rather than a single image, we instead obtain a data-cube: a stack of images collected at different wavelengths, with each pixel containing an entire spectrum of data. This cube contains much richer information than a simple colour image, but interpreting this huge dataset and extracting its clinical meaning becomes a huge challenge.
Real-time Cancer Detection
Focusing on neurosurgery, HELICoiD developed a methodology to discriminate between healthy and malignant tissues in real-time during surgical procedures.
Using a dataset of hyperspectral images of healthy and cancerous tissue, an algorithm was trained to detect differences in their spectral signature using machine learning techniques. The algorithm was then being applied to identify cancerous and healthy tissue in hyperspectral images acquired during surgery. In the future, this information could be provided to the surgeon in real time by overlaying conventional images with a colour-map that indicates the likelihood of that particular area being cancerous.
One of the major applications of hyperspectral imaging is likely to be in neurosurgery. While malignant primary brain tumours rank only 13th in the list of cancer incidence rates, their particularly poor prognosis results in them being the fifth most common cause of cancer deaths in under 65s. Among children, they are the second most common form of cancer and the most common cause of cancer death.
Brain tumours, more than other cancers, can resemble normal neurological tissue, making them difficult to differentiate. Unlike many tumours, they infiltrate the surrounding tissue and their borders are indistinct and difficult to identify. Yet in the brain it is essential to correctly identify the boundary so as to protect the surrounding brain tissue.
The HELICoiD demonstrator system, shown here, was assembled in 2015, and was used to capture neurosurgical images in our two partner hospitals. In 2016, dedicated processing hardware was added to allow the HELICoiD image analysis algorithms to run in real time.