Author: Diana Cardona, Casey C. Heirman, William Jeck, Kyle J. Lafata, Lauren M. Neldner, Jeffrey S. Nelson, Megan K. Russ, Ehsan Samei π¨βπ¬
Affiliation: Duke University, Department of Radiation Oncology, Duke University, Department of Pathology, Duke University, Clinical Imaging Physics Group, Department of Radiology, Duke University Health System π
Purpose: Display image accuracy is critical for digital diagnostic fields, such as radiology and digital pathology. While the AAPM TG-18 test patterns are established for grayscale radiology monitor QA, test patterns specific to digital pathology are lacking. Most well-established digital color test patterns donβt include colors commonly found in pathological images. This study evaluated whether color accuracy assessed using an established test pattern can predict the displayed accuracy of these pathology-relevant colors.
Methods: Eight consumer-grade monitors were evaluated for color accuracy using chromaticity (uβvβ) and luminance (L) color measurements made from the standard Macbeth ColorChecker pattern and five pure color patches (black, white, red, green, blue). Color measurements were also made using a custom test pattern of 24 colors found in hematoxylin and eosin-stained (H&E) slide images. Reference Luβvβ values were calculated for every test patch following CIE guidance. The relationship between measured and reference values for Macbeth and pure color patterns were empirically fitted and used to predict the measured values of the custom H&E pattern. The accuracy of this prediction was assessed by comparing Ξuβvβ from the measured, predicted, and reference values of the H&E pattern.
Results: The Macbeth Luβvβ color data were fitted to 2nd-order polynomial functions with strong goodness of fit. The differences between measured and predicted H&E chromaticity values were less than the Ξuβvβ threshold of perceivability for all pathology-relevant colors on all monitors. The absolute difference between reference and predicted color accuracy was also less than the threshold for perceivability.
Conclusion: These findings suggest that measurements made on the standard Macbeth ColorChecker provides a reliable means of evaluating color accuracy of digital pathology displays, even for colors not included in the pattern. This approach has the potential to streamline display QA processes and enhance the reliability of relevant color representation in digital pathology workflows.