Dataset Information

Glandular Morphometrics for Objective Grading of Colorectal Adenocarcinoma Histology Images.

ABSTRACT: Determining the grade of colon cancer from tissue slides is a routine part of the pathological analysis. In the case of colorectal adenocarcinoma (CRA), grading is partly determined by morphology and degree of formation of glandular structures. Achieving consistency between pathologists is difficult due to the subjective nature of grading assessment. An objective grading using computer algorithms will be more consistent, and will be able to analyse images in more detail. In this paper, we measure the shape of glands with a novel metric that we call the Best Alignment Metric (BAM). We show a strong correlation between a novel measure of glandular shape and grade of the tumour. We used shape specific parameters to perform a two-class classification of images into normal or cancerous tissue and a three-class classification into normal, low grade cancer, and high grade cancer. The task of detecting gland boundaries, which is a prerequisite of shape-based analysis, was carried out using a deep convolutional neural network designed for segmentation of glandular structures. A support vector machine (SVM) classifier was trained using shape features derived from BAM. Through cross-validation, we achieved an accuracy of 97% for the two-class and 91% for three-class classification.

SUBMITTER: Awan R

PROVIDER: S-EPMC5715083 | biostudies-other | 2017 Dec

REPOSITORIES: biostudies-other

ACCESS DATA

Publications

Glandular Morphometrics for Objective Grading of Colorectal Adenocarcinoma Histology Images.

Awan Ruqayya R Sirinukunwattana Korsuk K Epstein David D Jefferyes Samuel S Qidwai Uvais U Aftab Zia Z Mujeeb Imaad I Snead David D Rajpoot Nasir N

Scientific reports 20171204 1

Determining the grade of colon cancer from tissue slides is a routine part of the pathological analysis. In the case of colorectal adenocarcinoma (CRA), grading is partly determined by morphology and degree of formation of glandular structures. Achieving consistency between pathologists is difficult due to the subjective nature of grading assessment. An objective grading using computer algorithms will be more consistent, and will be able to analyse images in more detail. In this paper, we measur ...[more]

PMID: 29203775

Similar Datasets

Project description:ObjectivesThe aim of this study was to assess the objective and subjective image characteristics of monoenergetic images (MEI[+]), using a noise-optimized algorithm at different kiloelectron volts (keV) compared to polyenergetic images (PEI), in patients with pancreatic ductal adenocarcinoma (PDAC).MethodsThis retrospective, institutional review board-approved study included 45 patients (18 male, 27 female; mean age 66 years; range, 42-96 years) with PDAC who had undergone a dual-energy CT (DECT) of the abdomen for staging. One standard polyenergetic image (PEI) and five MEI(+) images in 10-keV intervals, ranging from 40 to 80 keV, were reconstructed. Line-density profile analysis, as well as the contrast-to-noise ratio (CNR) of the tumor, the signal-to-noise ratio (SNR) of the regular pancreas parenchyma and the tumor, and the CNR of the three main peripancreatic vessels, was calculated. For subjective quality assessment, two readers independently assessed the images using a 5-point Likert scale. Reader reliability was evaluated using an intraclass correlation coefficient.ResultsLine-density profile analysis revealed the largest gradient in attenuation between PDAC and regular tissue in MEI(+) at 40 keV. Low-keV MEI(+)reconstructions at 40 and 50 keV increased CNR and SNR compared to PEI (40 keV: CNR 46.8 vs. 7.5; SNRPankreas 32.5 vs. 15.7; SNRLesion 13.5 vs. 8.6; p?<?0.001). MEI(+) at 40 keV and 50 keV were consistently preferred by the observers (p?<?0.05), showing a high intra-observer 0.937 (0.92-0.95) and inter-observer 0.911 (0.89-0.93) agreement.ConclusionMEI(+) reconstructions at 40 keV and 50 keV provide better objective and subjective image quality compared to conventional PEI of DECT in patients with PDAC.Key points• Low-keV MEI(+) reconstructions at 40 and 50 keV increase tumor-to-pancreas contrast compared to PEI. • Low-keV MEI(+) reconstructions improve objective and subjective image quality parameters compared to PEI. • Dual-energy post-processing might be a valuable tool in the diagnostic workup of patients with PDAC.

Project description:Neoadjuvant therapy is increasingly used to control local tumor spread and micrometastasis of pancreatic ductal adenocarcinoma (PDAC). Pathology assessments of treatment effects might predict patient outcomes after surgery. However, there are conflicting reports regarding the reproducibility and prognostic performance of commonly used tumor regression grading systems, namely College of American Pathologists (CAP) and Evans' grading system. Further, the M.D. Anderson Cancer Center group (MDA) and the Japan Pancreas Society (JPS) have introduced other grading systems, while we recently proposed a new, simple grading system based on the area of residual tumor (ART). Herein, we aimed to assess and compare the reproducibility and prognostic performance of the modified ART grading system with those of the four grading systems using a multicenter cohort. The study cohort consisted of 97 patients with PDAC who had undergone post-neoadjuvant pancreatectomy at four hospitals. All patients were treated with gemcitabine and S-1 (GS)-based chemotherapies with/without radiation. Two pathologists individually evaluated tumor regression in accordance with the CAP, Evans', JPS, MDA and ART grading systems, and interobserver concordance was compared between the five systems. The ART grading system was a 5-tiered system based on a number of 40× microscopic fields equivalent to the surface area of the largest ART. Furthermore, the final grades, which were either the concordant grades of the two observers or the majority grades, including those given by the third observer, were correlated with patient outcomes in each system. The interobserver concordance (kappa value) for Evans', CAP, MDA, JPS and ART grading systems were 0.34, 0.50, 0.65, 0.33, and 0.60, respectively. Univariate analysis showed that higher ART grades were significantly associated with shorter overall survival (p = 0.001) and recurrence-free survival (p = 0.005), while the other grading systems did not show significant association with patient outcomes. The present study revealed that the ART grading system that was designed to be simple and more objective has achieved high concordance and showed a prognostic value; thus it may be most practical for assessing tumor regression in post-neoadjuvant resections for PDAC.

Dataset Information

Glandular Morphometrics for Objective Grading of Colorectal Adenocarcinoma Histology Images.

Publications

Glandular Morphometrics for Objective Grading of Colorectal Adenocarcinoma Histology Images.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets