Tissue counter analysis of benign common nevi and malignant melanoma
Introduction
Automatic medical image analysis is successful when the structures, e.g. blood cells, are well separated and can be clearly defined [1], [2], [3]. In histological tissue, the structures are mostly arranged in a variety of patterns and the segmentation of different structures, such as cells, nuclei, cytoplasm, vessels etc., is difficult, case dependent and cannot be done in a general approach. This is one reason why compact tissue structures generally resist a fully automatic analysis [4].
We concentrated our attention on the automatic analysis of benign common nevi and malignant melanoma lesions. In this study, we tested the applicability of tissue counter analysis to the diagnostic discrimination of a set of melanoma and nevi [5]. The images of the test set show a broad variation in patterns in several cases. To enable the analysis of tissue structures, without being limited by the selection and detection of structures of interest, the images are divided into square elements of equal size. In this way, a priori definition and segmentation of the structures, which is the crucial point in automatic classification, was avoided. Features are calculated by extracting the digital information in each square element. The features from all the elements in the set of images were submitted as data set to a classification and regression trees (CART) analysis [6]. By the CART analysis the elements were automatically separated into more or less homogeneous nodes corresponding to different classes of structures.
Section snippets
Materials
All biopsies were routinely embedded in paraffin and cut at a section thickness of 4 μm. The slides were routinely stained with hematoxylin and eosin using a fully automated staining device (DRF, Fakura, Japan). The specimens were randomly sampled from the Dermatopathology files of the Department of Dermatology, University of Graz, Austria. All cases were diagnosed by two pathologists. In total, we sampled 40 biopsies of benign common nevi and 40 biopsies of malignant melanoma. For the study
Method
The tissue counter analysis consists of three steps: the feature extraction, the classification and the verification.
Results
The classification results of the CART analysis are given in Table 3. When all nodes were taken into account for the 40 cases of the learning set, the percentage of tissue elements classified as ‘malignant elements was 7.3±4.9% (range: 1.6–22.7%) for the cases of benign common nevi and 92.1±6.6% (range: 74.6–99.82%) for the cases of malignant melanoma. For the 40 cases of the test set (by considering again all the nodes from the CART tree), the percentage of tissue elements classified as
Discussion
There is an increasing number of melanocytic tumors. In the fight against skin cancer, researchers have high hopes in improved provisional screening methods, such as optimized computer aided diagnostic methods.
The texture features obtained from histogram and co-occurrence matrix for each square element contain sufficient information to facilitate the differentiation between malignant melanoma and benign common nevi elements. In contrast to other methods in medical image analysis no previous
Acknowledgements
This study was supported by the ‘Jubiläumsfond der Österreichischen Nationalbank’, project number 9297.
References (7)
- et al.
Segmentation of stained blood cell images measured at high scanning density with high magnification and high numerical aperture optics
Cytometry
(1992) - et al.
Differentiation between maligant melanomas and benign melanocytic nevi by computerized DNA cytometry of imprint specimens
J. Cutaneous Pathol.
(1994) - et al.
Automated histometry in quantitative prostate pathology
Anal. Quant. Cytol. Histol.
(1998)
Cited by (54)
Detection of u-serrated patterns in direct immunofluorescence images of autoimmune bullous diseases by inhibition-augmented COSFIRE filters
2019, International Journal of Medical InformaticsCitation Excerpt :Thus, dermatologists could focus on the final diagnosis of EBA and the subsequent treatment. Computer-aided diagnosis has been implemented in radiology images, immunofluorescence assays and stained tissue sections for pathology to improve diagnostic accuracy, such as discrimination of benign and malignant melanomas [7,8], detection of anti-nuclear antibodies in serum [9] or lymph node metastasis in breast cancer [10,11] or bacteria in microscope images [12]. So far, there are only a few automatic techniques for serration pattern analysis in DIF images.
Automated analysis and classification of melanocytic tumor on skin whole slide images
2018, Computerized Medical Imaging and GraphicsCitation Excerpt :This technique first divides the image into many non-overlapping subregions, and then extracts colors and textures in each subregion for statistical analysis and classification. Wiltgen et al. (2003) applied the TCA technique to classify benign nevus and malignant melanoma. Although a good classification accuracy (about 92%) is reported, its performance is sensitive to staining variations as the TCA technique is completely replied on color and texture analysis (Smolle et al., 2002).
Techniques and algorithms for computer aided diagnosis of pigmented skin lesions—A review
2018, Biomedical Signal Processing and ControlAutomatic measurement of melanoma depth of invasion in skin histopathological images
2017, MicronCitation Excerpt :Smolle (2000) proposed a tissue counter analysis technique that recognizes skin structures like epidermis and dermis by using color features and Haralick texture features of each tissue element. Wiltgen et al. (2003) further applied the tissue counter analysis technique on skin tissue classification, which classifies skin histological images as benign nevi or malignant melanoma using features from histogram and co-occurrence matrix. Miedema et al. (2012) reported an image and statistical analysis system for melanocytic histology classification, which distinguishes melanoma from nevus by using cytological and textural features in DE junction areas.
Computer vision-based limestone rock-type classification using probabilistic neural network
2016, Geoscience FrontiersCitation Excerpt :In this paper, color image histogram-based features were extracted. The reason for a color image histogram is that the content based image retrieval (Jadhav et al., 2012; Malik and Baharudin, 2013) and medical image analysis (Wiltgen et al., 2003), color image histogram feature plays an important role for classification. Medical image classification is considered to be a difficult task as it has a rich pattern in color and structure (Wiltgen et al., 2003).
Hepatoma cells recognition based on matrix absolute gray relational degree of B-mode
2014, OptikCitation Excerpt :Different features that describe the deviations in the cell structures (cell shape, cell size, color hue, cell regularity, density of staining, etc.) have been proposed for cellular-level cancer diagnosis. Such as, Morphological features [5,6], Textural features [7,8], Fractal-based features [9,10] and Intensity-based features [11]. In the classification step, the artificial neural networks [12,13], support vector machines [14], fuzzy systems [15] and the k-NN algorithm [16] are usually used for recognizing normal and malignant lesions.