Hyperspectral imaging was originally developed for military and surveillance purposes but is now being used for life science vision applications such as identifying tumors. As the healthcare industry shifts to more noninvasive diagnostics and therapies, hyperspectral imaging offers an alternative to traditional invasive testing methods.
How Hyperspectral Imaging is Used
Hyperspectral imaging uses a camera to capture electromagnetic spectral signals from the body. The information is combined to form a three-dimensional data cube for processing and analysis. Once the data is collected, support vector machines (SVMs) classify the images. Individual pixels have different electromagnetic spectrums and the SVMs mark the pixels accordingly. The SVM classification is combined with probability data to determine what cells are cancerous and what cells are healthy tissue.
Image classification is a challenge because of the high spatial resolution, high dimensionality of spectral bands, and high redundancy caused by overlapping adjacent bands. Therefore, the implementation of hyperspectral imaging depends on the development of accurate, complex image processing algorithms to analyze the data. Those algorithms are used to differentiate between malignant and non-malignant cells.
Advantages of Using Hyperspectral Imaging
By using hyperspectral imaging, doctors can now receive a high-resolution image of a patient’s body that allows them to find cancer earlier than current technologies. Current diagnosis methods require expensive imaging agents to be given to patients. These ionizing radiation and contrasting agents pose risks to the patient’s health. They also add inconvenience to the diagnostic procedure. Hyperspectral imaging requires no such agent.
Hyperspectral imaging has been able to detect head and neck cancers with a high degree of accuracy and sensitivity. Current technologies also require large pieces of equipment, but hyperspectral imaging machines are smaller, more mobile, and less expensive.
Cancer is often treated by surgical resection of the tumor. During the surgery, a surgeon needs to verify complete removal of the cancerous tissue. Current imaging techniques require equipment that’s not convenient for use in the operating room. Due to its portability, hyperspectral imaging is a better option for detecting marginal cancerous tissue left behind after removal of a tumor. Hyperspectral imaging technologies can be used in other areas as well, such as stained slide microscopy and pathology identification.
Hyperspectral Imaging Research Continues
Dr. Baowei Fei, Ph.D., is leading the effort to turn hyperspectral imaging into a clinical tool to speed up and simplify the process of identifying cancer in patients. Fei has tested tumor identification and resections in both mice and human surgical specimen. The study is now moving onto testing human subjects. Fei and his team will use the data they collect and the resected cancerous material to improve their imaging algorithms. Fei’s goal is to create a portable device that can be sold and distributed for use in hospitals and operating rooms anywhere. He also sees promise for hyperspectral imaging to be used in diagnostics of retinal diseases.
To read about more ways life science vision systems are being leveraged for medical applications, read the blog post Life Science Vision System Applications for Medical Imaging.