Speaker
Description
Several diseases, including rheumatoid arthritis and cancer, are associated with localized increases in tissue temperature. This characteristic suggests that thermal imaging could be exploited for diagnostic and monitoring purposes, which led to the development of the field of medical thermography. However, when a localized hot-spot is situated more than ~15 mm beneath the skin and its temperature exceeds that of the surrounding tissue by 1–3 K, as is the case in medical conditions, the resulting change in surface (skin) temperature is typically of the order of tens of mK. Such small temperature variations of skin temperature approach the sensitivity- stability limits of commercial uncooled thermal cameras. When combined with environmental, instrumental and systematic uncertainties, these factors can render conventional two-dimensional thermography and thermal tomographic imaging unreliable.
To overcome these limitations it is necessary to minimize instrumental errors and other experimental uncertainties. A controlled and well monitored environment, combined with a structured data acquisition and analysis protocol, is essential for obtaining measurements of the required precision. We have designed and constructed an experimental setup that maintains a stable environmental temperature, has adequate monitoring and calibrating sensors, and minimizes infrared radiation reflections from external sources.
The dedicated experimental space and instrumentation, in conjunction with the adopted calibration and data acquisition protocol, allowed for thermal images to be obtained with the required precision and accuracy, over extended intervals (up to a few hours) during which small ambient temperature variations between successive acquisitions were allowed. Furthermore, reference surfaces within the thermal camera’s field of view during image acquisition enabled post-processing corrections that further reduced spatial and temporal fluctuations arising from instrumental and environmental factors. Using this setup and approach, we achieved an overall uncertainty of 25 mK.
We tested the suitability of this setup and methodology for medical thermal tomography using hardware phantoms. In these experiments, we successfully detected hot-spots with small temperature differences (ΔT < 2 °K, above surrounding medium) at depths of up to 30 mm. These results demonstrate, that a controlled experimental setup, together with the suitable calibration and data acquisition protocol, provides the conditions required to make thermal tomography feasible.
