A gamma-gaussian mixture model for detection of mitotic cells in breast cancer histopathology images

Abstract

In this paper, we propose a statistical approach for mitosis detection in breast cancer histological images. The proposed algorithm models the pixel intensities in mitotic and non-mitotic regions by a Gamma-Gaussian mixture model (GGMM) and employs a context aware post-processing (CAPP) in order to reduce false positives. Experimental results demonstrate the ability of this simple, yet effective method to detect mitotic cells (MCs) in standard H & E breast cancer histology images.Context:Counting of MCs in breast cancer histopathology images is one of three components (the other two being tubule formation, nuclear pleomorphism) required for developing computer assisted grading of breast cancer tissue slides. This is very challenging since the biological variability of the MCs makes their detection extremely difficult. In addition, if standard H & E is used (which stains chromatin rich structures, such as nucleus, apoptotic, and MCs dark blue) and it becomes extremely difficult to detect the latter given the fact that former two are densely localized in the tissue sections.Aims:In this paper, a robust MCs detection technique is developed and tested on 35 breast histopathology images, belonging to five different tissue slides.Settings and Design:Our approach mimics a pathologists’ approach to MCs detections. The idea is (1) to isolate tumor areas from non-tumor areas (lymphoid/inflammatory/apoptotic cells), (2) search for MCs in the reduced space by statistically modeling the pixel intensities from mitotic and non-mitotic regions, and finally (3) evaluate the context of each potential MC in terms of its texture.Materials and Methods:Our experimental dataset consisted of 35 digitized images of breast cancer biopsy slides with paraffin embedded sections stained with H and E and scanned at × 40 using an Aperio scanscope slide scanner.Statistical Analysis Used:We propose GGMM for detecting MCs in breast histology images. Image intensities are modeled as random variables sampled from one of the two distributions; Gamma and Gaussian. Intensities from MCs are modeled by a gamma distribution and those from non-mitotic regions are modeled by a gaussian distribution. The choice of Gamma-Gaussian distribution is mainly due to the observation that the characteristics of the distribution match well with the data it models. The experimental results show that the proposed system achieves a high sensitivity of 0.82 with positive predictive value (PPV) of 0.29. Employing CAPP on these results produce 241% increase in PPV at the cost of less than 15% decrease in sensitivity.Conclusions:In this paper, we presented a GGMM for detection of MCs in breast cancer histopathological images. In addition, we introduced CAPP as a tool to increase the PPV with a minimal loss in sensitivity. We evaluated the performance of the proposed detection algorithm in terms of sensitivity and PPV over a set of 35 breast histology images selected from five different tissue slides and showed that a reasonably high value of sensitivity can be retained while ...

Image Montaging for Creating a Virtual Pathology Slide: An Innovative and Economical Tool to Obtain a Whole Slide Image

“…If any researcher wants to identify and count specific type of cell(s), this can be done using the entire section rather than the conventional method of counting the cells in “per high power fields.” Several recent studies of The Cancer Genome Atlas data have illustrated important relationships between morphology observed in whole slide images, outcome, and genetic events [13]. Numerous studies using WSI relating to Identification of mitosis [14–16] are already published. Ultimately, we believe that it comes down to individual's intelligence and creativity to show how one can make use of such high resolution histopathological images.…”

Automated mitosis detection in histopathology based on non-gaussian modeling of complex wavelet coefficients

“…However, we also noted that our method yielded higher numbers of false positive compared to the ID-SIA, NEC, and CCIPD/MINDLAB methods. In addition, our model-based approach greatly outperformed the WARWICK method [44] (F-score=0.51), which also used statistical modeling and SVM. The same experimental results using different statistical models are plotted in the Precision-Recall plane in Figure 6.…”

“…The UTRECHT team extracted size, shape, color and texture features of candidate objects for automatically detecting mitotic figures [66]. The WARWICK approach modeled the pixel intensities of mitosis by a Gamma-Gaussian mixture model in conjunction with the SVM classifier [44]. …”

“…The DT-CWT and UDT-CWT have showed promise in various applications in medical image processing, which motivated our work in detecting mitosis and non-mitosis using wavelet-based method in histopathological images. Recently, Khan et al [44] presented a statistical approach using a GammaGaussian mixture model for mitosis detection, in which the pixel intensities were modeled as random variables via Gamma and Gaussian distributions. This method yielded 51% F-score during the ICPR 2012 Context [6].…”

Workflow for High-content, Individual Cell Quantification of Fluorescent Markers from Universal Microscope Data, Supported by Open Source Software

“…Also shown here is the use of nuclear DNA intensity measurements in their raw form as a continuous range of values in combination with the gated data. Other approaches to managing image analysis data should be considered, depending on the nature of the study; statistical alternatives to using gates for assigning cells to subpopulations have been reported 2 and systematic comparisons of strategies to summarize high-content data across large numbers of parameters have been reported 3 .…”