Spectroscopy in a nutshell is a technique of measuring any physical quantity in terms of wavelength or frequency. This is typically accomplished by using appropriate optics in either reflection or transmission mode.
Hyperspectral imaging is the extension of spectroscopy, where multiple images of the same scene are taken to identify specific signature spectra of materials from the multi-dimensional dataset. For the medical hyperspectral imaging under consideration, the device uses an external illumination source and a camera to image the oxyhemoglobin and de-oxyhemoglobin and thereby, provides valuable microcirculatory information on oxygen delivery and oxygen extraction. It must be emphasized that the ability to comprehend reflected spectra can be tuned dependant on the desired clinical outcome.
The choice of oxyhemoglobin and de-oxyhemoglobin stems from the underlying microvascular disease in patients with diabetes that predisposes tissue to ulceration. Therefore, hyperspectral imaging is a supplementary perfusion assessment tool for management of the diabetic foot. The device provides a whole field image of the anatomical site; is non-contact, objective, portable, safe and provides sub-clinical vascular indicators of metabolic status invisible to the naked eye.
When compared to alternative diagnostic modalities that assess impaired microcirculation such as isotope clearance, laser Doppler and TcPO2, hyperspectral imaging is non invasive, insensitive to motion and offers higher spatial resolution respectively. A typical assessment involves calibration of the device, identifying test site (dorsum or plantar foot), alignment of the camera head using laser beam and acquiring data. Data acquisition times are 15 seconds, during which the camera images signature spectra at specific wavelengths and combines all the acquired data into two separate images representing the oxyhemoglobin and de-oxyhemoglobin at each pixel. The spatial distribution of oxyhemoglobin and de-oxyhemoglobin provides a visual map, isolating areas of active ulceration. By drawing a region of interest around the ulcer, average values of both parameters can be determined to predict the healing outcome. Although, there is no consensus on the parameter values and healing index, preliminary studies have suggested that higher oxygenation suggests better healing. In a six-month follow up study of type 1 diabetic patients, sensitivity, specificity, positive and negative predictive values of 93%, 86%, 93% and 86% respectively have been reported. Unpublished data suggests that clinicians have effectively used the technique to carry out limb sparing amputations.
The other important application for the technique is assessment of postoperative physiological changes. Following compelling evidence from preliminary studies, further clinical studies are underway to develop a robust assessment protocol and knowledge base for furthering the role of hyperspectral imaging in the assessment of the diabetic foot. As the technique measures metabolic status, it may be capable of measuring pre-ulcerous inflammation and therefore, identify patients at high-risk of developing foot ulcers. This can be easily demonstrated by comparing hyperspectral dataset to thermometry assessments. In summary, the technique of hyperspectral imaging provides a supplementary assessment of functional changes in nutritive blood flow, which may be useful in predicting ulcer healing, identifying at risk patients and surgical planning.
