High-throughput detection of prostate cancer in histological sections using probabilistic pairwise Markov models

Abstract

In this paper we present a high-throughput system for detecting regions of carcinoma of the prostate (CaP) in HSs from radical prostatectomies (RPs) using probabilistic pairwise Markov models (PPMMs), a novel type of Markov random field (MRF). At diagnostic resolution a digitized HS can contain 80K×70K pixels -far too many for current automated Gleason grading algorithms to process. However, grading can be separated into two distinct steps: 1) detecting cancerous regions and 2) then grading these regions. The detection step does not require diagnostic resolution and can be performed much more quickly. Thus, we introduce a CaP detection system capable of analyzing an entire digitized whole-mount HS (2×1.75 cm 2 ) in under three minutes (on a desktop computer) while achieving a CaP detection sensitivity and specificity of 0.87 and 0.90, respectively. We obtain this high-throughput by tailoring the system to analyze the HSs at low resolution (8 µm per pixel). This motivates the following algorithm:Step 1) glands are segmented, Step 2) the segmented glands are classified as malignant or benign, and Step 3) the malignant glands are consolidated into continuous regions. The classification of individual glands leverages two features: gland size and the tendency for proximate glands to share the same class. The latter feature describes a spatial dependency which we model using a Markov prior. Typically, Markov priors are expressed as the product of potential functions. Unfortunately, potential functions are mathematical abstractions, and constructing priors through their selection becomes an ad hoc procedure, resulting in simplistic models such as the Potts. Addressing this problem, we introduce PPMMs which formulate priors in terms of probability density functions, allowing the creation of more sophisticated models. To demonstrate the efficacy of our CaP detection system and assess the advantages of using a PPMM prior instead of the Potts, we alternately incorporate both priors into our algorithm and rigorously evaluate system performance, extracting statistics from over 6000 simulations run across 40 RP specimens. Perhaps the most indicative result is as © 2010 Elsevier B.V. All rights reserved. * Corresponding authors jpmonaco@rci.rutgers.edu (James P. Monaco), anantm@rci.rutgers.edu (Anant Madabhushi).Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. NIH Public Access Author ManuscriptMed Image Anal. Author manuscript; available in PMC 2011 August 1.

Automated classification of brain tumor type in whole-slide digital pathology images using local representative tiles

“…Since these methods are automatic, they are not subjective and do not require a time commitment from an individual with pathology expertise. This method has most commonly been applied to assigning a Gleason grade to prostate cancer samples (Doyle et al, 2012b;Monaco et al, 2008;Monaco et al, 2010;Naik et al, 2008Naik et al, , 2007Sparks and Madabhushi, 2013) though it has found applications in other cancers as well (Petushi et al, 2006). The Gleason scoring system is particularly well suited to this type of analysis because the system characterizes how ordered the glands are in a prostate tumor.…”

Structure and Context in Prostatic Gland Segmentation and Classification

“…Given the pathologist's annotation on each image, we manually label 525 artifacts, 931 normal glands and 1,375 cancer glands to form the (ground truth) gland dataset. We also implemented the methods in [2] and [6] to compare them with the proposed method. Since all three methods perform segmentation by starting at the same lumen objects (identified in section 2), and use the same ground truth (which is not affected by lumen objects), the comparison is unbiased.…”

“…The average JI value per gland segment obtained by the algorithms in [2], [6] and the proposed NLA algorithm are 0.31, 0.43 and 0.66, respectively. Since the NLA algorithm aims at detecting nuclei surrounding the lumen (while the algorithms in [2] and [6] mostly detect lumen and cytoplasm), it segments more complete gland regions than [2] and [6] (Fig. 2).…”

“…Monaco et al [2] (Table 1) segmented glands and classified individual glands into normal or cancer by (i) using gland size feature to assign initial gland labels and (ii) applying a probabilistic pairwise Markov model (PPMM) to update gland labels. On the other hand, Nguyen et al [3] and Doyle et al [4] detected cancer regions by classifying individual image patches and image pixels, respectively, by using a combination of cytological and textural features [3], or textural features alone [4].…”

“…On the other hand, Nguyen et al [3] and Doyle et al [4] detected cancer regions by classifying individual image patches and image pixels, respectively, by using a combination of cytological and textural features [3], or textural features alone [4]. Unlike [2,3,4], other studies focused on the tissue image classification problem, i.e. they classified every prostatic tissue image (which is a region in the whole slide tissue) into normal or cancer or into different cancer grades.…”