Imaging Radiotherapy-Induced Cherenkov Emission in Color

Abstract

<h3>Purpose/Objective(s)</h3> In the last decade, Cherenkov imaging has been demonstrated as a useful clinical technique both for real-time surface dose visualization during patient treatment and for various high-resolution QA measurements. However there remains much work to be done in the development of Cherenkov-based in vivo dosimetry. In this work, the concept of in vivo color Cherenkov imaging is introduced. It is hypothesized that Cherenkov images with sensitivity to the spectral quality of Cherenkov emission induced inside tissue can provide increased information about tissue composition as opposed to monochromatic, red-weighted images used in prior work. <h3>Materials/Methods</h3> A custom Color Cherenkov camera was created from three time-gated intensified CMOS cameras, each with varying spectral sensitivity of the photocathode, and sharing one imaging lens. A dichroic mirror assembly allowed for incoming Cherenkov light signals to be redirected according to wavelength to the appropriate camera resulting in three raw image stacks for each acquisition. These images were then reconstructed into three-channel color images via color calibration, summed, and background subtracted. In vivo color Cherenkov images of three right-sided breast patients were acquired for one treatment fraction each as part of an IRB-approved clinical trial. These images were interpreted alongside images of liquid tissue-simulating phantoms that matched tissue optical properties. Variations in oxygenation and blood volume were mimicked in this way, and the data was used to interpret the clinical images. Impacts on Cherenkov emission spectra from tissue under these various conditions were also estimated. <h3>Results</h3> The presence of oxygenated vs deoxygenated blood in tissue has a visible impact on the color of Cherenkov emission, leading to a redder hue due to decreased attenuation of red wavelengths by oxyhemoglobin, similar to erythema or blush of tissue. This was verified in the tissue phantoms. Variations in blood concentration produce a strong blue-red trend in color space (R² = 0.96), as verified by detailed analysis. Color Cherenkov images of right breast radiotherapy treatments displayed pink regions across the breast coincident with the treated area for all three patients, and the variations in color were interpreted relative to the phantom data. <h3>Conclusion</h3> This first investigation of color Cherenkov imaging uncovered the natural relationship between tissue composition and the spectral quality of emitted Cherenkov light. The correlation between blood content and color will be translated in future work to in vivo sensing of variations in fat and fibroglandular content in irradiated tissues. These findings will allow for the correction of Cherenkov images to produce accurate in vivo surface dose maps during patient treatments.