Research On Dark-field Imaging As A Novel Immunohistochemical Method

Fluorescence immune hybrid method and immunoenzyme technique have been by far the most common tools in cancer detection.Immunoenzyme technique is a common screening method with low cost and simple operation,but it relies too heavily on the experience and subjective judgment to ensure accuracy.Fluorescence immune hybrid method is accurate,and can be used for quality detection,however,promotion and application were hindered by the high cost.3,3’-diaminobenzidim(DAB),as one of the most important chromogenic enzyme,whose oxidative product presents irregular crystal structure and can be observed as dark brown outline of tumor cells in light field,meanwhile the scattered light strengthen highly in dark field.According to Mie scattering law,the oxidative product scatters red light when the overexpression rate is high,and scatters yellow and green when the overexpression rate is low.According to this principle,a novel immunohistochemical interpretation method based on dark field imaging is proposed and preliminarily established.Immunohistochemical dark field imaging of tumor tissue was performed by using DAB.The color,brightness and staining area of the scattered light were used to establish interpretation criteria to comprehensively reflect the positive degree of samples.The research bottleneck of computer pathology diagnosis mainly lies in three aspects:sample collection,data processing and result interpretation.So the setting of detection hardware,image acquisition quality and the definition of diagnostic threshold play a vital role in the accuracy of detection.In order to reach the standard of practical clinical application,the stability and standardization of this method were systematically studied,and the optimal image acquisition scheme which can reflect the positive degree of samples to the greatest extent was successfully obtained.On this basis,computer automation interpretation with excellent accuracy and reliability were realized.We selected 160 HER2 samples of breast cancer,and the results interpreted independently by three diagnostic physicians were taken as reference to optimize microscopic view and red light collection amount respectively.Dark field imaging photos with clear structure and balanced color were obtained by adjusting the brightness levels of amplification factor,gain and RGB primary colors.We found the best detection performance was gained at 4.250 X,contrast at 1.2,red 1.5,green 1.0,blue 1.0.After tuning the parameters,we got a better result with the sensitivity and specificity reached95.16% and 100.00%,respectively,and the Youden index of 0.95.In order to expand the application of dark field immunohistochemical interpretation,we involved a new biomarker of lung cancer,PD-L1.Different from breast cancer HER2 interpretation method based on color intensity and area,PD-L1 TPS only observed the color area regardless of color intensity.Compared with TPS,H-Score was a more suitable method of interpretation for the quantitative software.The interpretation threshold was redefined according to the H-Score scoring method,and the computer automated interpretation of immunohistochemical detection for NSCLC with over-expression of PD-L1 was successfully realized.The results show that the novel immunohistochemical interpretation method based on dark field imaging has good applicability and potential to be applied to the CPS interpretation...