Display Optimization of Indocyanine Green Cholangiography (ICG) Imaging

Jeff T Flinn, MSc1, Ramon Berguer, MD2, Caroline G Cao, PhD1. 1Wright State University, 2Walnut Creek Surgical Associates

INTRODUCTION: Intraoperative cholangiogram is frequently used to visualize the biliary system during laparoscopic cholecystectomies. However, an alternative imaging technique, Indocyanine Green (ICG) cholangiography, is potentially safer as it does not require radiation or catheterization of the cystic duct.

An ICG imaging system prototype has been developed that can be coupled directly to an existing laparoscope. Preliminary evaluation reveals that image quality varies over time. Further refinements of the prototype are possible only with increasing development costs. The goal of this research was to optimize the ICG imaging display, post-hoc, by systematically manipulating visual parameters of the display and measuring the effect on perceived image quality.

METHODS AND PROCEDURES: Three images, each with a similar view of the juncture where the cystic duct and the common bile duct (CBD) join, were extracted from the video produced by the ICG imaging prototype. Each image was selected based on its image quality (one low-, one medium-, and one high-quality image) and served as a standard for comparison. Comparison images were created from each standard image by digitally manipulating brightness and/or contrast in 8% increments in the range of 0-24%, resulting in 16 comparison images per standard.

Six observers viewed pairs of images, one after the other, on a computer monitor. One image was the standard image, and the other was a comparison image, presented in a random order. Each observer completed 16 blocks of trials, where a block consisted of a randomized ordering of the 48 standard-comparison combinations, for a total of 768 trials. The observers’ task for each trial was to decide which image was perceived as having better image quality. “Better” was defined as being the image in which the juncture where the cystic duct and CBD join could be most clearly differentiated from the surrounding tissue. The number of times that observers chose each comparison image over the standard image was measured.

RESULTS: When a comparison image was chosen more frequently than the standard, then the manipulation of contrast and/or brightness was perceived to enhance image quality. Likewise, when the standard was chosen more frequently, then the manipulation degraded perceived quality. In general, the results showed that increases in brightness (8-16%) and/or contrast (16-24%) enhanced both low- and medium-quality images. Corresponding manipulations to the high-quality images resulted in only modest enhancements or even slight degradation.

CONCLUSION: The low- and medium-quality images were the images most in need of improvement and were, in fact, the images most improved by the brightness and contrast manipulations. Although commensurate improvements were not found for the high-quality image, it was already of superior quality initially and thus its enhancement (or slight degradation) would have a lower impact on the overall image quality of the display. Therefore, a moderate increase in brightness and/or contrast can increase the overall clinical utility of ICG imaging and may be a more economical alternative to extensive development and refinement of the prototype.

« Return to SAGES 2015 abstract archive