Field Of Interest Proposal for Augmented Mitotic Cell Count: Comparison of two Convolutional Networks
This work addresses the sampling problem in tumor prognostication for veterinary and human histopathology by aiding pathologists in selecting high-activity regions, though it is incremental as it compares existing CNN-based methods on a specific dataset.
The study tackled the problem of selecting fields of interest for mitotic cell counting in tumor grading by comparing two convolutional neural network methods for segmentation on whole slide images, finding that the fine-resolution approach had higher correlation to ground truth (0.94 vs. 0.83) but only marginally better field proposals, with both methods proposing fields in the upper quartile of mitotic counts.
Most tumor grading systems for human as for veterinary histopathology are based upon the absolute count of mitotic figures in a certain reference area of a histology slide. Since time for prognostication is limited in a diagnostic setting, the pathologist will often almost arbitrarily choose a certain field of interest assumed to have the highest mitotic activity. However, as mitotic figures are commonly very sparse on the slide and often have a patchy distribution, this poses a sampling problem which is known to be able to influence the tumor prognostication. On the other hand, automatic detection of mitotic figures can't yet be considered reliable enough for clinical application. In order to aid the work of the human expert and at the same time reduce variance in tumor grading, it is beneficial to assess the whole slide image (WSI) for the highest mitotic activity and use this as a reference region for human counting. For this task, we compare two methods for region of interest proposal, both based on convolutional neural networks (CNN). For both approaches, the CNN performs a segmentation of the WSI to assess mitotic activity. The first method performs a segmentation at the original image resolution, while the second approach performs a segmentation operation at a significantly reduced resolution, cutting down on processing complexity. We evaluate the approach using a dataset of 32 completely annotated whole slide images of canine mast cell tumors, where 22 were used for training of the network and 10 for test. Our results indicate that, while the overall correlation to the ground truth mitotic activity is considerably higher (0.94 vs. 0.83) for the approach based upon the fine resolution network, the field of interest choices are only marginally better. Both approaches propose fields of interest that contain a mitotic count in the upper quartile of respective slides.