Quantitative evaluation of morphologic changes in histologic samples by means of mathematical mor-phology
| Author | Affiliation |
|---|---|
Pečiulis, R. | Kauno technologijos universitetas |
Kiškis, Andrius | Kauno technologijos universitetas |
| Date |
|---|
2015-12-03 |
ISBN 978-9986-680-58-1.
The efficacy of histological analysis used for evaluation pathologies can be improved by means of digital imaging, giving quantitative estimates of known diagnostic features used by the experts, enriching them by mathe-matical morphology based estimates. This study illustrates proposed methodology by formation of diagnostic features set reflecting inflamma-tion severity in colon tissue and elaboration of automatic evaluation method for computer-aided diagnostics. Acute and chronic colon inflammation in the experimental animals was induced by oral administration of 3.5 % dextran sulphate sodium (DSS, TdB Consultancy, Uppsala, Sweden). Histological images were taken by means of OLYMPUS IX71 light microscope (×20 magnification) equipped with Q IMAGING EXI aqua camera at 1392 × 1040 pixels res-olution (0.6 μm/pixel). Image processing algorithms were realized as pro-grams in MATLAB computation environment. Image features formation was performed on 512 × 512 pixel colon image cutouts selected by the experts, representing as much as possible homo-geneous and typical tissue pattern without any gaps. Main diagnostic fea-tures in histologic images characterizing ulcerative colitis include crypt distortion, branching, and appearance of lymphoplasmacytic infiltrate deep in the crypts. Every feature has its representation in digital image: for example, crypts appear as elliptic white spots varying about 180–350 pixels long and 50–130 pixels wide, eosinophils, are rounded spots of 7–25 pixels etc. We used Gabor filters for detection and evaluation of such morphological changes. Principal component analysis was used to con-struct optimal representation of such morphological changes. First princi-pal component (PC1) was representing the major part of features variation (97 %) and was selected as a measure of inflammation severity. Valida-tion of new measure was performed by means of custom-made software realizing doub[...].