A new method for automatic counting of ovarian follicles on whole slide histological images based on convolutional neural network

Overview of the Multispecies Ovary Tissue Histology Electronic Repository†

The Multispecies Ovary Tissue Histology Electronic Repository (MOTHER) is a publicly accessible repository of ovary histology images. MOTHER includes hundreds of images from nonhuman primates, as well as ovary histology images from an expanding range of other species. Along with an image, MOTHER provides metadata about the image, and for selected species, follicle identification annotations. Ongoing work includes assisting scientists with contributing their histology images, creation of manual and automated (via machine learning) processing pipelines to identify and count ovarian follicles in different stages of development, and the incorporation of that data into the MOTHER database (MOTHER-DB). MOTHER will be a critical data repository storing and disseminating high-value histology images that are essential for research into ovarian function, fertility, and intra-species variability.

Quantification of follicles in human ovarian tissue using image processing software and trained artificial intelligence†

Cancer survival rates in prepubertal girls and young women have risen in recent decades due to increasingly efficient treatments. However, many such treatments are gonadotoxic, causing premature ovarian insufficiency, loss of fertility, and ovarian endocrine function. Implantation of donor ovarian tissue encapsulated in immune-isolating capsules is a promising method to restore physiological endocrine function without immunosuppression or risk of reintroducing cancer cells harbored by the tissue. The success of this approach is largely determined by follicle density in the implanted ovarian tissue, which is analyzed manually from histologic sections and necessitates specialized, time-consuming labor. To address this limitation, we developed a fully automated method to quantify follicle density that does not require additional coding. We first analyzed ovarian tissue from 12 human donors between 16 and 37 years old using semi-automated image processing with manual follicle annotation and then trained artificial intelligence program based on follicle identification and object classification. One operator manually analyzed 102 whole slide images from serial histologic sections. Of those, 77 images were assessed by a second manual operator, followed with an automated method utilizing artificial intelligence. Of the 1181 follicles the control operator counted, the comparison operator counted 1178, and the artificial intelligence counted 927 follicles with 80% of those being correctly identified as follicles. The three-stage artificial intelligence pipeline finished 33% faster than manual annotation. Collectively, this report supports the use of artificial intelligence and automation to select tissue donors and grafts with the greatest follicle density to ensure graft longevity for premature ovarian insufficiency treatment.

An end-to-end pipeline based on open source deep learning tools for reliable analysis of complex 3D images of ovaries

Computational analysis of bio-images by deep learning (DL) algorithms has made exceptional progress in recent years and has become much more accessible to non-specialists with the development of ready-to-use tools. The study of oogenesis mechanisms and female reproductive success has also recently benefited from the development of efficient protocols for three-dimensional (3D) imaging of ovaries. Such datasets have a great potential for generating new quantitative data but are, however, complex to analyze due to the lack of efficient workflows for 3D image analysis. Here, we have integrated two existing open-source DL tools, Noise2Void and Cellpose, into an analysis pipeline dedicated to 3D follicular content analysis, which is available on Fiji. Our pipeline was developed on larvae and adult medaka ovaries but was also successfully applied to different types of ovaries (trout, zebrafish and mouse). Image enhancement, Cellpose segmentation and post-processing of labels enabled automatic and accurate quantification of these 3D images, which exhibited irregular fluorescent staining, low autofluorescence signal or heterogeneous follicles sizes. In the future, this pipeline will be useful for extensive cellular phenotyping in fish or mammals for developmental or toxicology studies.

Region-Based Segmentation and Classification for Ovarian Cancer Detection Using Convolution Neural Network

Ovarian cancer is a serious sickness for elderly women. According to data, it is the seventh leading cause of death in women as well as the fifth most frequent disease worldwide. Many researchers classified ovarian cancer using Artificial Neural Networks (ANNs). Doctors consider classification accuracy to be an important aspect of making decisions. Doctors consider improved classification accuracy for providing proper treatment. Early and precise diagnosis lowers mortality rates and saves lives. On basis of ROI (region of interest) segmentation, this research presents a novel annotated ovarian image classification utilizing FaRe-ConvNN (rapid region-based Convolutional neural network). The input photos were divided into three categories: epithelial, germ, and stroma cells. This image is segmented as well as preprocessed. After that, FaRe-ConvNN is used to perform the annotation procedure. For region-based classification, the method compares manually annotated features as well as trained feature in FaRe-ConvNN. This will aid in the analysis of higher accuracy in disease identification, as human annotation has lesser accuracy in previous studies; therefore, this effort will empirically prove that ML classification will provide higher accuracy. Classification is done using a combination of SVC and Gaussian NB classifiers after the region-based training in FaRe-ConvNN. The ensemble technique was employed in feature classification due to better data indexing. To diagnose ovarian cancer, the simulation provides an accurate portion of the input image. FaRe-ConvNN has a precision value of more than 95%, SVC has a precision value of 95.96%, and Gaussian NB has a precision value of 97.7%, with FR-CNN enhancing precision in Gaussian NB. For recall/sensitivity, SVC is 94.31 percent and Gaussian NB is 97.7 percent, while for specificity, SVC is 97.39 percent and Gaussian NB is 98.69 percent using FaRe-ConvNN.

An efficient real-time colonic polyp detection with YOLO algorithms trained by using negative samples and large datasets

Colorectal cancer (CRC) is one of the common types of cancer with a high mortality rate. Colonoscopy is the gold standard for CRC screening and significantly reduces CRC mortality. However, due to many factors, the rate of missed polyps, which are the precursors of colorectal cancer, is high in practice. Therefore, many artificial intelligence-based computer-aided diagnostic systems have been presented to increase the detection rate of missed polyps. In this article, we present deep learning-based methods for reliable computer-assisted polyp detection. The proposed methods differ from state-of-the-art methods as follows. First, we improved the performances of YOLOv3 and YOLOv4 object detection algorithms by integrating Cross Stage Partial Network (CSPNet) for real-time and high-performance automatic polyp detection. Then, we utilized advanced data augmentation techniques and transfer learning to improve the performance of polyp detection. Next, for further improving the performance of polyp detection using negative samples, we substituted the Sigmoid-weighted Linear Unit (SiLU) activation functions instead of the Leaky ReLU and Mish activation functions, and Complete Intersection over Union (CIoU) as the loss function. In addition, we present a comparative analysis of these activation functions for polyp detection. We applied the proposed methods on the recently published novel datasets, which are the SUN polyp database and the PICCOLO database. Additionally, we investigated the proposed models for MICCAI Sub-Challenge on Automatic Polyp Detection in Colonoscopy dataset. The proposed methods outperformed the other studies in both real-time performance and polyp detection accuracy.

Ovary Development: Insights From a Three-Dimensional Imaging Revolution

The ovary is an indispensable unit of female reproduction and health. However, the study of ovarian function in mammals is hindered by unique challenges, which include the desynchronized development of oocytes, irregular distribution and vast size discrepancy of follicles, and dynamic tissue remodeling during each hormonal cycle. Overcoming the limitations of traditional histology, recent advances in optical tissue clearing and three-dimensional (3D) visualization offer an advanced platform to explore the architecture of intact organs at a single cell level and reveal new relationships and levels of organization. Here we summarize the development and function of ovarian compartments that have been delineated by conventional two-dimensional (2D) methods and the limits of what can be learned by these approaches. We compare types of optical tissue clearing, 3D analysis technologies, and their application to the mammalian ovary. We discuss how 3D modeling of the ovary has extended our knowledge and propose future directions to unravel ovarian structure toward therapeutic applications for ovarian disease and extending female reproductive lifespan.

Deep Learning and Transfer Learning for Automatic Cell Counting in Microscope Images of Human Cancer Cell Lines

In biology and medicine, cell counting is one of the most important elements of cytometry, with applications to research and clinical practice. For instance, the complete cell count could help to determine conditions for which cancer cells could grow or not. However, cell counting is a laborious and time-consuming process, and its automatization is highly demanded. Here, we propose use of a Convolutional Neural Network-based regressor, a regression model trained end-to-end, to provide the cell count. First, unlike most of the related work, we formulate the problem of cell counting as the regression task rather than the classification task. This allows not only to reduce the required annotation information (i.e., the number of cells instead of pixel-level annotations) but also to reduce the burden of segmenting potential cells and then classifying them. Second, we propose use of xResNet, a successful convolutional architecture with residual connection, together with transfer learning (using a pretrained model) to achieve human-level performance. We demonstrate the performance of our approach to real-life data of two cell lines, human osteosarcoma and human leukemia, collected at the University of Amsterdam (133 training images, and 32 test images). We show that the proposed method (deep learning and transfer learning) outperforms currently used machine learning methods. It achieves the test mean absolute error equal 12 (±15) against 32 (±33) obtained by the deep learning without transfer learning, and 41 (±37) of the best-performing machine learning pipeline (Random Forest Regression with the Histogram of Gradients features).

A Validation Employing Convolutional Neural Network for the Radiographic Detection of Absence or Presence of Teeth

Dental radiography plays an important role in clinical diagnosis, treatment and making decisions. In recent years, efforts have been made on developing techniques to detect objects in images. The aim of this study was to detect the absence or presence of teeth using an effective convolutional neural network, which reduces calculation times and has success rates greater than 95%. A total of 8000 dental panoramic images were collected. Each image and each tooth was categorized, independently and manually, by two experts with more than three years of experience in general dentistry. The neural network used consists of two main layers: object detection and classification, which is the support of the previous one. A Matterport Mask RCNN was employed in the object detection. A ResNet (Atrous Convolution) was employed in the classification layer. The neural model achieved a total loss of 0.76% (accuracy of 99.24%). The architecture used in the present study returned an almost perfect accuracy in detecting teeth on images from different devices and different pathologies and ages.

C-ECi: a CUBIC-ECi combined clearing method for three-dimensional follicular content analysis in the fish ovary†

Deciphering mechanisms of oocyte development in the fish ovary still remain challenging, and a comprehensive overview of this process at the level of the organ is still needed. The recent development of optical tissue clearing methods has tremendously boosted the three-dimensional (3D) imaging of large size biological samples that are naturally opaque. However, no attempt of clearing on fish ovary that accumulates extremely high concentration of lipids within oocytes has been reported to date. To face with this ovarian-specific challenge, we combined two existing clearing methods, the nontoxic solvent-based ethyl cinnamate (ECi) method for efficient clearing and the Clear Unobstructed Brain Imaging Cocktails and Computational (CUBIC) method to enhance lipid removal and reduce nonspecific staining. The methyl green fluorescent dye was used to stain nuclei and delineate the follicular structures that include oocytes. Using this procedure (named CUBIC-ECi [C-ECi]), ovaries of both medaka and trout could be imaged in 3D and follicles analyzed. To our knowledge, this is the first procedure elaborated for clearing and imaging fish ovary in 3D. The C-ECi method thus provides an interesting tool for getting precise quantitative data on follicular content in fish ovary and promises to be useful for further developmental and morphological studies.

Combined iDISCO and CUBIC tissue clearing and lightsheet microscopy for in toto analysis of the adult mouse ovary†

At any given time, the ovary contains a number of follicles in distinct growth stages, each with a set of identifying characteristics. Although follicle counting and staging using histological stains on paraffin-embedded ovary sections has been the gold standard in assessing ovarian health in fertility studies, the final counts rely on extrapolation factors that diverge greatly among studies. These methods also limit our ability to investigate spatial aspects of ovary organization. Recent advances in optical tissue clearing and lightsheet microscopy have permitted comprehensive analysis of intact tissues. In this study, we set out to determine the best clearing and imaging methods to generate 3D images of the complete adult mouse ovary that could be used for accurate assessments of ovarian follicles. We found that a combination of iDISCO and CUBIC was the best method to clear the immunostained ovary. Using lightsheet microscopy, we generated 3D images of the intact ovary and performed qualitative assessments of follicles at all stages of development. This study is an important step toward developing quantitative computational models that allow rapid and accurate assessments of growing and quiescent primordial follicles, and to investigate the integrity of extrinsic ovarian components including vascular and neuronal networks.