Abstract:Studies estimate that there will be 266,120 new cases of invasive breast cancer and 40,920 breast cancer induced deaths in the year of 2018 alone. Despite the pervasiveness of this affliction, the current process to obtain an accurate breast cancer prognosis is tedious and time consuming, requiring a trained pathologist to manually examine histopathological images in order to identify the features that characterize various cancer severity levels. We propose MITOS-RCNN: a novel region based convolutional neural network (RCNN) geared for small object detection to accurately grade one of the three factors that characterize tumor belligerence described by the Nottingham Grading System: mitotic count. Other computational approaches to mitotic figure counting and detection do not demonstrate ample recall or precision to be clinically viable. Our models outperformed all previous participants in the ICPR 2012 challenge, the AMIDA 2013 challenge and the MITOS-ATYPIA-14 challenge along with recently published works. Our model achieved an F-measure score of 0.955, a 6.11% improvement in accuracy from the most accurate of the previously proposed models.

What problem does this paper attempt to address?

This paper attempts to solve the problem of detecting mitotic figures in breast cancer histopathology images. Specifically, the authors propose a new Region - based Convolutional Neural Network (RCNN) model - MITOS - RCNN, aiming to accurately detect and count mitotic figures in breast cancer tissue sections to assist in breast cancer grading and prognosis assessment. ### Problem Background Breast cancer is one of the most common cancers worldwide, with a large number of new cases and deaths every year. Currently, the prognosis assessment of breast cancer relies on pathologists manually examining histopathology images to identify and count features such as mitotic figures. However, this method is both time - consuming and error - prone, especially when the number of mitotic figures in high - power fields (HPF) is small and their morphology is variable. ### Paper Goals To solve the above problems, the authors propose the MITOS - RCNN model, which is specifically optimized for the detection of small objects such as mitotic figures. Specific goals include: 1. **Improve Detection Accuracy**: Through an improved Region Proposal Network (RPN) and detection network, achieve high - precision detection of mitotic figures. 2. **Handle the Difficult Problem of Small - Object Detection**: Since mitotic figures are usually small in images, the traditional Faster - RCNN model has difficulty in effectively detecting them. The authors adjust the network structure to make it better able to handle small objects. 3. **Accelerate the Detection Process**: By optimizing the network structure and training methods, the model can complete the detection task in a shorter time, thereby increasing the feasibility of clinical applications. ### Main Contributions 1. **Innovative Two - Stage Multi - Scale Region Proposal Generation Method**: By combining feature maps of different convolutional layers, generate more accurate small - object proposals. 2. **Improved Faster - RCNN Architecture**: By removing certain convolutional layers and adjusting the feature extraction method, the model can better detect small objects. 3. **Excellent Performance**: Achieve results significantly better than existing methods on multiple public datasets, with an F - measure score of 0.955, which is 6.22% higher than the previous best method. Through these improvements, the MITOS - RCNN model not only achieves high - precision mitotic figure detection in breast cancer histopathology images but also shows its potential application value in other small - object detection tasks.

MITOS-RCNN: A Novel Approach to Mitotic Figure Detection in Breast Cancer Histopathology Images using Region Based Convolutional Neural Networks

Accurate and Fast Mitotic Detection Using an Anchor-Free Method Based on Full-Scale Connection with Recurrent Deep Layer Aggregation in 4D Microscopy Images

A Novel Two-Stage Deep Method for Mitosis Detection in Breast Cancer Histology Images

MaskMitosis: a Deep Learning Framework for Fully Supervised, Weakly Supervised, and Unsupervised Mitosis Detection in Histopathology Images

Multi CNN based automatic detection of mitotic nuclei in breast histopathological images

PartMitosis: A Partially Supervised Deep Learning Framework for Mitosis Detection in Breast Cancer Histopathology Images

A multi-phase deep CNN based mitosis detection framework for breast cancer histopathological images

DeepMitosis: Mitosis Detection Via Deep Detection, Verification and Segmentation Networks.

Multi-Scale Deep Neural Network for Mitosis Detection in Histological Images

Mitosis Detection, Fast and Slow: Robust and Efficient Detection of Mitotic Figures

Multi-Scale Deep Neural Network for Mitosis Detection in Breast Cancer Histological Images

CNN-based Mitosis Detection for Assisting Doctors to Diagnosis

Cascade RCNN for MIDOG Challenge

Mask RCNN Algorithm for Nuclei Detection on Breast Cancer Histopathological Images.

A Fuzzy Segmentation Method to Learn Classification of Mitosis

Deep Feature Fusion for Mitosis Counting

MITNET: a novel dataset and a two-stage deep learning approach for mitosis recognition in whole slide images of breast cancer tissue

SmallMitosis: Small Size Mitotic Cells Detection in Breast Histopathology Images

Weakly Supervised Mitosis Detection in Breast Histopathology Images Using Concentric Loss.

OMG-Net: A Deep Learning Framework Deploying Segment Anything to Detect Pan-Cancer Mitotic Figures from Haematoxylin and Eosin-Stained Slides

FF-CNN: an Efficient Deep Neural Network for Mitosis Detection in Breast Cancer Histological Images.