Differentiation of low grade non-Hodgkin's lymphoma by digital image processing

Area of Science:

• Hematology
• Oncology
• Digital Pathology

Background:

• Accurate classification of low-grade B-cell non-Hodgkin's lymphoma (NHL) is crucial for effective treatment.
• Existing classification systems, like the Revised European American Lymphoma (REAL) and Kiel classifications, require precise diagnostic methods.

Purpose of the Study:

• To evaluate the efficacy of digital image processing and texture analysis in classifying different subtypes of low-grade B-cell NHL.
• To compare the diagnostic performance of Feulgen-Azure A staining with Romanowsky-Giemsa staining in conjunction with image analysis.

Main Methods:

• Seventy-four lymphoma touch imprints were analyzed using digital image processing.
• Specimens were stained with either Romanowsky-Giemsa or Feulgen-Azure A.
• Thirty nuclear features, including size and chromatin texture, were extracted and evaluated.

Main Results:

• Romanowsky-Giemsa staining, coupled with texture analysis, enabled accurate differentiation of all investigated lymphoma types.
• Specific subtype diagnostic accuracies included: mantle cell lymphoma (87.5%), follicle center cell lymphoma (78%), chronic lymphocytic leukemia (78%), immunocytoma (75%), and marginal zone B-cell lymphoma (80%).
• Feulgen-Azure A staining showed a lower average correct identification rate of 59% and struggled to differentiate mantle cell lymphoma from chronic lymphocytic leukemia.

Conclusions:

• Texture analysis using digital image processing is a feasible method for classifying NHL subtypes.
• Romanowsky-Giemsa staining provides superior results compared to Feulgen-Azure A for this application.
• This approach holds promise for improving the accuracy and efficiency of lymphoma diagnosis.