Although color now pervades medical imaging, currently accepted calibration methods for imaging displays still rely on assessing gray levels. Display calibration may soon enter the color era, however, if a proposed extension to the DICOM Grayscale Standard Display Function (GSDF) is adopted.
In testing, the proposed extension -- called Color Standard Display Function (CSDF) -- yielded significant improvement in perceptual color linearity over high-end color displays calibrated just with DICOM GSDF, as well as consumer color displays utilizing the red-green-blue standard (sRBG), said Tom Kimpe, PhD, of display firm Barco.
"The study suggests that the current way color medical images are visualized is not optimal and not consistent," Kimpe told AuntMinnie.com. "It is very important that the medical community and radiologists in particular are aware of this, since they are using color information more and more for decision-making. Also, this work is a strong encouragement for the medical community to think about, agree upon, and standardize how color medical images need to be visualized."
He discussed the proposed CSDF method during a scientific session at the RSNA 2014 meeting in Chicago.
Adding color
While radiology in the past has mostly utilized grayscale images, that's no longer the case. The use of color images in medical imaging has become more the standard than the exception, according to Kimpe.
"The majority of images being produced are in fact color images, and this will only increase in the future," he said.
In many of these applications, radiologists are tasked with detecting subtle color differences on the images. Other types of color images, such as in contrast imaging applications, involve determining a quantification value that corresponds to color in the image.
Currently, there is no consensus standard for how color information on medical images needs to be visualized on color displays, and also for how medical color displays need to be calibrated to ensure standardized and consistent color behavior, Kimpe said.
"This results in a situation where there's a large variability of color appearance," he said. "It makes it very hard to have a consistent image, and it makes [quality assurance] a challenge."
At this time, state-of-the-art medical color displays are calibrated using DICOM GSDF. However, DICOM GSDF doesn't calibrate, stabilize, or alter at all the color behavior of the display.
"The color behavior of the display remains native," Kimpe said. "Whatever it is, it will stay the same after calibration to DICOM GSDF."
The researchers tested the color behavior of four new color displays from different vendors and found large variability in the color gamut and in the unstable color point of neutral gray.
Calibrating color behavior
Fortunately, there is much that can be learned from DICOM GSDF, a longtime standard that calibrates grayscale behavior not by imposing specific minimum or maximum luminance levels, but by ensuring that the display has perceptual linearization of gray levels, Kimpe said.
"It will try to make all steps between the gray scales equally large," he said. "In other words, it will also maximize the number of different shades of gray we can see on such a display."
This concept can be expanded to color calibration: Perceptual linear color behavior can be assessed using the International Commission on Illumination (CIE) DeltaE2000 metric for measuring color difference.
Based on the DeltaE2000 metric, CSDF could be added as a simple extension to DICOM GSDF to provide color calibration. In testing, the researchers found that CSDF yielded much more perceptually linear color behavior than GSDF and sRGB.
As a result, a CSDF-calibrated display will maximize the probability that a subtle color difference can be detected, according to Kimpe.
As an example of the clinical relevance of CSDF, the calibration method was found to boost the visibility of clinically relevant features on digital pathology images.
In addition, informal feedback from radiologists has demonstrated CSDF's potential for enhancing visualization of several color modalities such as PET/CT, ultrasound, and MRI. While it wasn't a formal clinical study, "all of them agreed that there is an improved visualization [from CSDF]," Kimpe said.
They also noted its possible usefulness for quantitative imaging applications, where color is typically quite important, he added.
"Preliminary results suggest that CSDF may have the potential to improve visualization and quantification of clinically relevant features," Kimpe said.
