Automated image analysis in histopathology: A valuable tool in medical diagnostics

UCD School of Biomolecular and Biomedical Science, Conway Institute, University College Dublin, Belfield, Dublin 4, Ireland.
Expert Review of Molecular Diagnostics (Impact Factor: 3.52). 12/2008; 8(6):707-25. DOI: 10.1586/14737159.8.6.707
Source: PubMed


Virtual pathology, the process of assessing digital images of histological slides, is gaining momentum in today's laboratory environment. Indeed, digital image acquisition systems are becoming commonplace, and associated image analysis solutions are viewed by most as the next critical step in automated histological analysis. Here, we document the advances in the technology, with reference to past and current techniques in histological assessment. In addition, the demand for these technologies is analyzed with major players profiled. As there are several image analysis software programs focusing on the quantification of immunohistochemical staining, particular attention is paid to this application in this review. Oncology has been a primary target area for these approaches, with example studies in this therapeutic area being covered here. Toxicology-based image analysis solutions are also profiled as these are steadily increasing in popularity, especially within the pharmaceutical industry. Reinforced by the phenomenal growth of the virtual pathology field, it is envisioned that the market for automated image analysis tools will greatly expand over the next 10 years.

140 Reads
  • Source
    • "Even though current systems are technically complex and not easily operated by the average practising pathologist, automated quantification will increasingly be a part of standard practice in the future (Leong and Leong, 2003; Park et al., 2012). Also, with the continuing rapid development of computer technologies and their reduced cost, it is suggested that successful computer-based automated image analysis will become commonplace and lie within the reach of individual pathologists , clinical researchers and other members of the wide-ranging histological community (Krenacs et al., 2010; Mulrane et al., 2008). At a time when automated morphological classification of histological images is increasingly prevalent in routine practice for other disciplines , including human anatomic and clinical pathology, the use of this methodology is still largely absent from fish health and veterinary applications. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Virtual histology, the process of assessing digital images of histological slides, is gaining momentum in modern histopathology and digital image acquisition systems are becoming commonplace. Associated image processing and analysis methods can potentially complement traditional histological assessment methodologies.Image analysis of digitised histological sections, can provide a practical means for quantifying and helping interpretation of functional alterations in an objective and reproducible fashion. This study focused on the development of a practical and time-efficient image capture, processing and analysis pipeline, employing advanced image analysis that was able to identify features of salmon intestine histological sections in a quantitative manner. Through standardisation of the sampling and preparation methodologies, staining protocols and digital capture thresholds and techniques, this assessment system has proven to be an efficient, accurate and objective method, having consistent data outputs when the analysis is performed by different observers with varying levels of expertise in histopathological assessment.
    Aquaculture 03/2015; 442. DOI:10.1016/j.aquaculture.2015.02.034 · 1.88 Impact Factor
  • Source
    • "However, both solutions can incur significant man-hours to derive an interpretation. Automated feature detection in images is an active area of research in image processing, including many applications such as road extraction (Shao et al., 2011; Treash and Amaratunga, 2000) and medical applications (Den Hertog et al., 2010; Mulrane et al., 2008; Onkaew et al., 2011). Image analysis techniques provide an effective and fast method of lineament detection and these techniques can extract lineaments which are difficult to recognise using the human eye alone (Wang and Howarth, 1990). "
    [Show abstract] [Hide abstract]
    ABSTRACT: Recent advances in data acquisition technologies, such as Unmanned Aerial Vehicles (UAVs), have led to a growing interest in capturing high-resolution rock surface images. However, due to the large volumes of data that can be captured in a short flight, efficient analysis of this data brings new challenges, especially the time it takes to digitise maps and extract orientation data. We outline a semi-automated method that allows efficient mapping of geological faults using photogrammetric data of rock surfaces, which was generated from aerial photographs collected by a UAV. Our method harnesses advanced automated image analysis techniques and human data interaction to rapidly map structures and then calculate their dip and dip directions. Geological structures (faults, joints and fractures) are first detected from the primary photographic dataset and the equivalent three dimensional (3D) structures are then identified within a 3D surface model generated by structure from motion (SfM). From this information the location, dip and dip direction of the geological structures are calculated. A structure map generated by our semi-automated method obtained a recall rate of 79.8% when compared against a fault map produced using expert manual digitising and interpretation methods. The semi-automated structure map was produced in 10 minutes whereas the manual method took approximately 7 hours. In addition, the dip and dip direction calculation, using our automated method, shows a mean±standard error of 1.9°±2.2° and 4.4°±2.6° respectively with field measurements. This shows the potential of using our semi-automated method for accurate and efficient mapping of geological structures, particularly from remote, inaccessible or hazardous sites.
    Computers & Geosciences 08/2014; 69. DOI:10.1016/j.cageo.2014.04.012 · 2.05 Impact Factor
  • Source
    • "The presented approach differs from other commercially available packages and previously proposed supervised methods, since a molecular marker is used to objectively select data for training purposes. Additionally, previously described methods for use with digital pathology pattern recognition applications, have focused on laborious image object identification by skilled pathologists, followed by the measurement of texture and morphological features of single or group of pixels with similar texture or color properties [9], [10], [43]. Methods characterizing epithelial nuclear characteristics, such as size, color and spatial distribution in order to distinguish tumor and lymphocyte infiltration has shown to be effective in solid tumors such as breast cancer [11], [44]. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Immunohistochemistry is a routine practice in clinical cancer diagnostics and also an established technology for tissue-based research regarding biomarker discovery efforts. Tedious manual assessment of immunohistochemically stained tissue needs to be fully automated to take full advantage of the potential for high throughput analyses enabled by tissue microarrays and digital pathology. Such automated tools also need to be reproducible for different experimental conditions and biomarker targets. In this study we present a novel supervised melanoma specific pattern recognition approach that is fully automated and quantitative. Melanoma samples were immunostained for the melanocyte specific target, Melan-A. Images representing immunostained melanoma tissue were then digitally processed to segment regions of interest, highlighting Melan-A positive and negative areas. Color deconvolution was applied to each region of interest to separate the channel containing the immunohistochemistry signal from the hematoxylin counterstaining channel. A support vector machine melanoma classification model was learned from a discovery melanoma patient cohort (n = 264) and subsequently validated on an independent cohort of melanoma patient tissue sample images (n = 157). Here we propose a novel method that takes advantage of utilizing an immuhistochemical marker highlighting melanocytes to fully automate the learning of a general melanoma cell classification model. The presented method can be applied on any protein of interest and thus provides a tool for quantification of immunohistochemistry-based protein expression in melanoma.
    PLoS ONE 05/2013; 8(5):e62070. DOI:10.1371/journal.pone.0062070 · 3.23 Impact Factor
Show more