Bayesian model for detection and classification of meningioma nuclei in microscopic images.

Image segmentation aims to determine structures of interest inside a digital picture in biomedical sciences. State-of-the art automatic methods however still fail to provide the segmentation quality achievable by humans who employ expert knowledge and use software to mark target structures on an image. Manual segmentation is time-consuming, tedious and suffers from interoperator variability, thus not serving the requirements of daily use well. Therefore, the approach presented here abandons the goal of full-fledged segmentation and settles for the localization of circular objects in photographs (10 training images and 20 testing images with several hundreds of nuclei each). A fully trainable softcore interaction point process model was hence fit to the most likely locations of nuclei of meningioma cells. The Broad Bioimage Benchmark Collection/SIMCEP data set of virtual cells served as controls. A 'colour deconvolution' algorithm was integrated to determine (based on anti-Ki67 immunohistochemistry) which real cells might have the potential to proliferate. In addition, a density parameter of the underlying Bayesian model was estimated. Immunohistochemistry results were 'simulated'for the virtual cells. The system yielded true positive (TP) rates in the detection and classification of real nuclei and their virtual counterparts. These hits outnumbered those obtained from the public domain image processing software ImageJ by 10%. The method introduced here can be trained to function not only in medicine and morphology-based systems biology but in other application domains as well. The algorithm lends itself to an automated approach that constitutes a valuable tool which is easy to use and generates acceptable results quickly.

Three-dimensional image analytical detection of intussusceptive pillars in murine lung.

A variety of diseases can lead to loss of lung tissue. Currently, this can be treated only symptomatically. In mice, a complete compensatory lung growth within 21 days after resection of the left lung can be observed. Understanding and transferring this concept of compensatory lung growth to humans would greatly improve therapeutic options. Lung growth is always accompanied by a process called angiogenesis forming new capillary blood vessels from preexisting ones. Among the processes during lung growth, the formation of transluminal tissue pillars within the capillary vessels (intussusceptive pillars) is observed. Therefore, pillars can be understood as an indicator for active angiogenesis and microvascular remodelling. Thus, their detection is very valuable when aiming at characterization of compensatory lung growth. In a vascular corrosion cast, these pillars appear as small holes that pierce the vessels. So far, pillars were detected visually only based on 2D images. Our approach relies on high-resolution synchrotron microcomputed tomographic images. With a voxel size of 370 nm we exploit the spatial information provided by this imaging technique and present the first algorithm to semiautomatically detect intussusceptive pillars. An at least semiautomatic detection is essential in lung research, as manual pillar detection is not feasible due to the complexity and size of the 3D structure. Using our algorithm, several thousands of pillars can be detected and subsequently analysed, e.g. regarding their spatial arrangement, size and shape with an acceptable amount of human interaction. In this paper, we apply our novel pillar detection algorithm to compute pillar densities of different specimens. These are prepared such that they show different growing states. Comparing the corresponding pillar densities allows to investigate lung growth over time.