Widefield micro-camera integrated into the objective lens of a reflectance confocal microscope for concurrent image registration

Abstract

With millions of new skin disease cases reported annually, non-invasive imaging methods have been developed to diagnose skin disease accurately. Reflectance confocal microscopes (RCM) have led these new technologies with high sensitivity and specificity. However, current methods use multiple devices: a digital camera, a dermoscope, and an RCM, which are not co-registered. Therefore, locating small, microscopic RCM fields-of-view (0.5x0.5 mm) at specific suspicion sites within the larger dermoscopic field-of-view (10x10 mm) is extremely difficult. This 'blind' RCM imaging results in lower and more variable diagnostic accuracy, particularly sensitivity, where positive and negative predictive values can drop by up to 30%. Our team has designed a new objective lens with an integrated micro-camera to deliver a concurrent widefield image of the skin surface surrounding the location of microscopic RCM imaging. The widefield image can be used directly to provide context for RCM or can be registered to a previously stored high-resolution clinical image to show where RCM imaging is occurring. In this thesis, the micro-camera is characterized and tested in laboratory and clinical settings. In addition, this thesis investigates a co- and cross-polarized micro-camera and LED system. It compares them to the non-polarized system to explore whether the cross-polarized version enhances feature contrast and enables better dermoscopic imaging. Non-polarized, co-polarized, and cross-polarized mock-up probes of the objective lens with a micro-camera were designed and built for testing. Images of resolution targets, color charts, and skin were taken to obtain modulation transfer function (MTF) measurements, color analysis data, and representative skin images. The results showed improvement in the MTF for the cross- polarized probe when compared to the co- and non-polarized probes. It was also found that the polarization of the imaging system did not significantly affect the color quality of the images. When tested by scientists at Memorial Sloan Kettering Cancer Center, sub-surface features not seen with the co- and non-polarized probes were observed with the cross-polarized probe. The cross-polarized probe suppressed the surface reflections, allowing for sub-surface information to be captured.