Detection of Mitosis in Multispectral Breast Cancer Imagery

Mitotic count is an important parameter for the prognosis of breast cancer. However, mitosis detection is a challenging problem and has not been addressed well in the literature. Indeed, mitosis detection is very challenging since mitosis appear in image as small objects with a large variety of shapes, and they can thus be confused with some other objects or artifacts present in the image. The “Mitosis Detection in Breast Cancer Images” was a contest hosted with the 22nd International Conference on Pattern Recognition (ICPR2012). The objective of the contest is to detect mitosis on H&E stained histological images of different breast cancers images. The contest data provided two set of images – RGB images and Multispectral images. More details available on contest site [HERE].

In our opinion, the multi-spectral images had more information than the normal RGB images, hence we had decided to use the multi-spectral images. In the contest our team (LNM-IIT) was ranked 3rd (out of 4 participants) amongst the participants who used the multi-spectral images. Detailed results of the contest can be accessed from [HERE].

After the contest, we extended the work and published this as a conference paper at ICIP-2013 (International Conference on Pattern Recognition), titled “2-SiMDoM: A 2-SIEVE MODEL FOR DETECTION OF MITOSIS IN MULTISPECTRAL BREAST CANCER IMAGERY”. Our 2-SiMDoM scheme is now ranked 1st amongst the participants using multispectral images. Also, as of Apr 2014, it is the best performing scheme amongst all the participants when taking into account all the metrics viz. F-measure, Recall and Precision.

Publications

Ardhendu Shekhar Tripathi, Atin Mathur, Mohit Daga, Manohar Kuse, Oscar C. Au, “2-SiMDoM: A 2-Sieve Model for Detection of Mitosis in Multispectral Breast Cancer Imagery”, Proceedings of IEEE ICIP, Melbourne, Australia, September 2013.

Abstract

In this paper, we propose a 2-Sieve model for the detection of mitosis in breast cancer multispectral images. Multiresolution wavelet features & Gray Level Entropy Matrix (GLEM) features have been computed for each candidate on all the spectral bands. A novel dimensionality election algorithm has been introduced and its performance compared with other existing algorithms. Data imbalance and data cleaning have been taken care of using classical data mining techniques. Furthermore, a Second Sieve classiﬁcation is performed to increase the Positive Predictive Value (PPV) with minimal loss in Sensitivity. A ﬁnal Sensitivity and PPV of 82.35% & 73.04% respectively was achieved over the testing set using the proposed scheme.

Biological Background

Nottingham Grading System is an international grading system for breast cancer recommended by the World Health Organization. It is derived from the assessment of three morphological features: tubule formation, nuclear pleomorphism and mitotic count. Several studies on automatic tools to process digitized slides have been reported focusing mainly on nuclei or tubule detection. Mitosis detection is a challenging problem and has not been addressed well in the literature.

Mitotic count is an important parameter in breast cancer grading as it gives an evaluation of the agressiveness of the tumour. Detection of mitosis is a very challenging task since mitosis are small objects with a large variety of shape configurations. The four main phases of a mitosis are prophase, metaphase, anaphase and telophase. The shape of the nucleus is very different depending on the phase of the mitosis. On its last stage, the telophase, a mitosis has two distinct nuclei, but they are not yet full individual cells. A mitosis in telophase must be counted as a single mitosis, it should not be miscounted as two mitosis. Artefacts are also common and should not be confused with mitosis. Figure below shows some examples of possible mitotic phases.