Semi-Automatic And Automatic Ki-67 Index Examination In Whole Slide Images Of Meningiomas
Histological examination of tissue subjects by immunohistochemical staining is the basic method of recognizing types of cancer and it provides valuable indicators concerning choice of optimal therapy or defining the prognosis. One of a most important markers is the mitotic receptor Ki-67, among others, in meningiomas . According to examination guidelines, ROI’s (Region of interest) whose fields correspond with the high positive receptors’ reaction should be selected.
The aim of this paper is a compare of Ki-67 index examination in meningioma specimens performed on the whole slide images(WSI) in two ways: with selection of hot-spot regions by the experts, and with automatic se- lection of hot-spots. Using both ways we have analyzed variability of results between two experts and between the experts and the automatic procedure, also in respect of Ki-67 level.
The fifty cases of meningiomas were stained with the ready-to-use FLEX Ki-67 antigen (Dako, code IR626) in Dako Autostainer Link. Acquisition of WSIs was carried out by the 3DHistech Pannoramic 250 Flash II scanner under the 20x magnification of lens. The selection of hot-spots was done manually by two experts and automatically with the proposed method of automatic hot-spot detection. The suggested WSI processing scheme was based on the following steps:
- defining the map of specimen using the thresholding procedure and morphological filtering,
- eliminating the areas containing blood cells (hemorrhages) by the texture analysis (Unser features) and classification,
- eliminating the specinem folds by the texture analysis (Unser and Local Binary Patterns) and classification,
- selecting sequential fields of the hot-spots based on cells segmentation and the punishment function to avoid excessive proximity, and it is the extention of idea presented in paper . The final analysis of Ki-67 index was performed on the full resolution images with the same procedure of image analysis.
The results indicated that the mean difference between the Ki-67 index of Expert A and Expert B was -0.6065% (SD ±1.27%). Comparison between the results of Automatic system and Expert A gives mean difference 0.5207% (SD 1.18%) whereas in relation to the Exert B, it was -0.0858% (SD 1.21%). No significant skewness was observed in any of Bland-Altman plots.
The determination analysis gives R2 equals 0.947 (Expert A to Expert B), 0.947 (System to Expert A), and 0.944 (System to Expert B), all p<0.000001. The automatic procedure for the hot-spot detection in meningioma WSI gives the high concordance of results with the expert’s examinations. The differences between the automatic and both experts’ results are included in the range of variability of experts’ results. The presented results confirm that the proposed automatic procedure can be introduced to the multicenter verification process for practical applicability in histopathological diagnosis in the near future.
This work has been supported by the National Centre for Research and Development (PBS2/ A9/21/2013 grant), Poland.
p16 in meningiomas tissue arrays, Pathol. Res. Pract 2008, 204:305-314.
 Swiderska, Z., Korzynska, A., Markiewicz, T., Lorent, M., Zak, J., Wesolowska, A., Roszkowiak, L., Slodkowska, J., Grala, B.,
Comparison of the manual, semiautomatic, and automatic selection and leveling of hot spots in whole slide images for Ki-67 quantification
in meningiomas., Anal Cell Pathol 2015, 2015:498746.
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Authors who publish with this journal agree to the following terms:
1. Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
2. Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
3. Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).
4. In case of virtual slide publication the authors agree to copy the article in a structural modified version to the journal's VS archive.