scispace - formally typeset
Search or ask a question
Proceedings ArticleDOI

Nuclear Segmentation and its Quantification in H&E Stained Images of Oral Precancer to Detect its Malignant Potentiality

TL;DR: This algorithm uses differential contrast enhancement and distance map transformations to segment out the cell nuclei in ImageJ Software and performed successfully on high magnification images with high speed and relative simplicity thus proving its credibility.
Abstract: Diagnosis of oral cancer using pathology is becoming more dependent on digital imaging. Since precancerous conditions like Oral submucous fibrosis originate in the basal layer of the tissue, it is very important to investigate the cell nuclei of the basal layer in Haematoxylin and Eosin stained tissue as it contains diagnostically important information. For that, accurate identification and segmentation of the nuclei is imperative. Our algorithm uses differential contrast enhancement and distance map transformations to segment out the cell nuclei in ImageJ Software. The algorithm performed successfully on high magnification images with high speed and relative simplicity thus proving its credibility. The nuclear attributes like entropy, polarity, and compactness are calculated and the values obtained are then statistically analyzed using Mann-Whitney U Test using SPSS Software to differentiate between normal and OSF(with severe dysplasia and without dysplasia). The results showed that in case of entropy, statistical significant difference $(\mathbf{p} is present between all the above mentioned three classes but in cases of compactness and polarity, statistical significant differences are present between normal and diseased classes, but not between OSF (without dysplasia) and OSF (with severe dysplasia) cases for both attributes $(\mathbf{p}=\pmb{0.1527}$ for compactness and $\mathbf{p}\pmb{=0.6965}$ for polarity).
Citations
More filters
Journal ArticleDOI
TL;DR: In this article , a deep learning approach was proposed for the development of prognostic models for malignant transformation and their association with clinical outcomes in histology whole slide images (WSIs) of OED tissue sections.
Abstract: Oral squamous cell carcinoma (OSCC) is amongst the most common cancers, with more than 377,000 new cases worldwide each year. OSCC prognosis remains poor, related to cancer presentation at a late stage, indicating the need for early detection to improve patient prognosis. OSCC is often preceded by a premalignant state known as oral epithelial dysplasia (OED), which is diagnosed and graded using subjective histological criteria leading to variability and prognostic unreliability. In this work, we propose a deep learning approach for the development of prognostic models for malignant transformation and their association with clinical outcomes in histology whole slide images (WSIs) of OED tissue sections. We train a weakly supervised method on OED cases (n = 137) with malignant transformation (n = 50) and mean malignant transformation time of 6.51 years (±5.35 SD). Stratified five‐fold cross‐validation achieved an average area under the receiver‐operator characteristic curve (AUROC) of 0.78 for predicting malignant transformation in OED. Hotspot analysis revealed various features of nuclei in the epithelium and peri‐epithelial tissue to be significant prognostic factors for malignant transformation, including the count of peri‐epithelial lymphocytes (PELs) (p < 0.05), epithelial layer nuclei count (NC) (p < 0.05), and basal layer NC (p < 0.05). Progression‐free survival (PFS) using the epithelial layer NC (p < 0.05, C‐index = 0.73), basal layer NC (p < 0.05, C‐index = 0.70), and PELs count (p < 0.05, C‐index = 0.73) all showed association of these features with a high risk of malignant transformation in our univariate analysis. Our work shows the application of deep learning for the prognostication and prediction of PFS of OED for the first time and offers potential to aid patient management. Further evaluation and testing on multi‐centre data is required for validation and translation to clinical practice. © 2023 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.
Posted ContentDOI
22 Feb 2023
TL;DR: In this article , a deep learning approach was proposed for the development of prognostic models for malignant transformation and their association with clinical outcomes in histology whole slide images (WSIs) of OED tissue sections.
Abstract: Abstract Oral squamous cell carcinoma (OSCC) is amongst the most common cancers worldwide, with more than 377,000 new cases worldwide each year. OSCC prognosis remains poor, related to cancer presentation at a late stage indicating the need for early detection to improve patient prognosis. OSCC is often preceded by a premalignant state known as oral epithelial dysplasia (OED), which is diagnosed and graded using subjective histological criteria leading to variability and prognostic unreliability. In this work, we propose a deep learning approach for the development of prognostic models for malignant transformation and their association with clinical outcomes in histology whole slide images (WSIs) of OED tissue sections. We train a weakly supervised method on OED (n= 137) cases with transformation (n= 50) status and mean malignant transformation time of 6.51 years (±5.35 SD). Performing stratified 5-fold cross-validation achieves an average AUROC of ∼0.78 for predicting malignant transformations in OED. Hotspot analysis reveals various features from nuclei in the epithelium and peri-epithelial tissue to be significant prognostic factors for malignant transformation, including the count of peri-epithelial lymphocytes (PELs) ( p < 0.05), epithelial layer nuclei count (NC) ( p < 0.05) and basal layer NC ( p < 0.05). Progression free survival using the Epithelial layer NC ( p < 0.05, C-index = 0.73), Basal layer NC ( p < 0.05, C-index = 0.70) and PEL count ( p < 0.05, C-index = 0.73) shown association of these features with a high risk of malignant transformation. Our work shows the application of deep learning for prognostication and progression free survival (PFS) prediction of OED for the first time and has a significant potential to aid patient management. Further evaluation and testing on multi-centric data is required for validation and translation to clinical practice.
References
More filters
Journal ArticleDOI
TL;DR: The recent state of the art CAD technology for digitized histopathology is reviewed and the development and application of novel image analysis technology for a few specific histopathological related problems being pursued in the United States and Europe are described.
Abstract: Over the past decade, dramatic increases in computational power and improvement in image analysis algorithms have allowed the development of powerful computer-assisted analytical approaches to radiological data. With the recent advent of whole slide digital scanners, tissue histopathology slides can now be digitized and stored in digital image form. Consequently, digitized tissue histopathology has now become amenable to the application of computerized image analysis and machine learning techniques. Analogous to the role of computer-assisted diagnosis (CAD) algorithms in medical imaging to complement the opinion of a radiologist, CAD algorithms have begun to be developed for disease detection, diagnosis, and prognosis prediction to complement the opinion of the pathologist. In this paper, we review the recent state of the art CAD technology for digitized histopathology. This paper also briefly describes the development and application of novel image analysis technology for a few specific histopathology related problems being pursued in the United States and Europe.

1,644 citations


"Nuclear Segmentation and its Quanti..." refers background in this paper

  • ...The available image analysis algorithms hence analyse low-resolution descriptions of the image, with the main emphasis on rough scale structural knowledge, since they are not able to meet the data processing necessities of the large images [19]....

    [...]

Book
01 Jan 1981
TL;DR: This text presents the chemical and physical principles of fixation, staining and histochemistry, and offers a practical guide to the preparation of specimens for light microscopy and includes detailed practical instructions of the techniques used.
Abstract: This text presents the chemical and physical principles of fixation, staining and histochemistry. It assumes a basic level of chemical and biological knowledge, and requires little mathematical skill. The relations of chemical structures and reactions to fixation, tissue processing, staining, enzyme location, immunohistochemistry and other procedures are explained. For this edition the author has updated the text with the latest techniques and developments within the field, whilst retaining the details of the classic techniques still in use. The book offers a practical guide to the preparation of specimens for light microscopy and includes detailed practical instructions of the techniques used. This text presents the chemical and physical principles of fixation, staining and histochemistry. It assumes a basic level of chemical and biological knowledge, and requires little mathematical skill. The relations of chemical structures and reactions to fixation, tissue processing, staining, enzyme location, immunohistochemistry and other procedures are explained. For this edition the author has updated the text with the latest techniques and developments within the field, whilst retaining the details of the classic techniques still in use. The book offers a practical guide to the preparation of specimens for light microscopy and includes detailed practical instructions of the techniques used.

1,454 citations

Journal ArticleDOI
TL;DR: This paper presents a robust and accurate novel method for segmenting cell nuclei using a combination of ideas, and presents an efficient semiautomated approach to editing automated segmentation results that requires two mouse clicks per operation.
Abstract: Automatic segmentation of cell nuclei is an essential step in image cytometry and histometry. Despite substantial progress, there is a need to improve accuracy, speed, level of automation, and adaptability to new applications. This paper presents a robust and accurate novel method for segmenting cell nuclei using a combination of ideas. The image foreground is extracted automatically using a graph-cuts-based binarization. Next, nuclear seed points are detected by a novel method combining multiscale Laplacian-of-Gaussian filtering constrained by distance-map-based adaptive scale selection. These points are used to perform an initial segmentation that is refined using a second graph-cuts-based algorithm incorporating the method of alpha expansions and graph coloring to reduce computational complexity. Nuclear segmentation results were manually validated over 25 representative images (15 in vitro images and 10 in vivo images, containing more than 7400 nuclei) drawn from diverse cancer histopathology studies, and four types of segmentation errors were investigated. The overall accuracy of the proposed segmentation algorithm exceeded 86%. The accuracy was found to exceed 94% when only over- and undersegmentation errors were considered. The confounding image characteristics that led to most detection/segmentation errors were high cell density, high degree of clustering, poor image contrast and noisy background, damaged/irregular nuclei, and poor edge information. We present an efficient semiautomated approach to editing automated segmentation results that requires two mouse clicks per operation.

683 citations


"Nuclear Segmentation and its Quanti..." refers methods in this paper

  • ...To step ahead of local knowledge, initial colour segmentation was refined using active contours [22], and then graph cuts method was used which was established on colour and Laplacian of Gaussian features [23]....

    [...]

Journal ArticleDOI
TL;DR: Histological and Histochemical Methods by Professor John A. Kiernan is a classic in the histochemical literature since its first edition, in 1981, when it was first published.
Abstract: Histological and Histochemical Methods by Professor John A. Kiernan is a classic in the histochemical literature since its first edition, in 1981.....

363 citations


"Nuclear Segmentation and its Quanti..." refers background in this paper

  • ...Tissue segments obtained were de-parrafinized and dampened for staining with H&E and lastly dried out through boosts of alcohol and set [20]....

    [...]

Journal ArticleDOI
15 Jul 1985-Cancer
TL;DR: In 271 breast cancer patients with adequate follow‐up for at least 5.5 and maximally 12 years, the value of morphometry to classic prognosticators of breast cancer was assessed and it was found that mitotic activity index is the best single predictor of the prognosis.
Abstract: In 271 breast cancer patients with adequate follow-up for at least 5.5 and maximally 12 years, the value of morphometry to classic prognosticators of breast cancer (tumor size and axillary lymph node status) was assessed. Previous studies had indicated the value of this quantitative microscopic technique. Apart from quantitative microscopic features, subjective qualitative features such as nuclear and histologic grade were assessed as well. Univariate life-table analysis showed the significance (p less than 0.001) of several features such as lymph node status, tumor size, nuclear and histologic grade, and several morphometric variables (mitotic activity index, mean and standard deviation of nuclear area). Cellularity index was also significant (p = 0.02). Survival analysis with Cox's regression model, using a stepwise selection as well as backwards elimination, pointed to three features: mitotic activity index, tumor size, and lymph node status. Mitotic activity was the most important prognostic feature, but the combination of these three features in a multivariate prognostic index had even more prognostic significance. Kaplan-Meier curves showed that the 5-year survival of lymph node-negative patients (n = 146) is 85%, versus 93% in patients with a "good prognosis index" (n = 150). For lymph node-positive patients (n = 125), 5-year survival was 55%, compared with 47% in the "high index" (poor prognosis) patients (n = 121). Logistic discriminant analysis with 5.5-year follow-up as a fixed endpoint (191 survivors and 80 nonsurvivors) essentially gave the same results. Application of two instead of one decision threshold (e.g., numerical classification probability 0.60 and 0.40) decrease the number of false-negative and false-positive outcomes, however, with a number of patients falling in the class "uncertain." Thus, in agreement with other studies, morphometry significantly adds to the prognosis prediction of lymph node status and tumor size. Mitotic activity index is the best single predictor of the prognosis. An additional index advantage is that the multivariate model results in a continuous index variable that can be subdivided in many classes with an increasing risk of recurrence, so that more refined clinical therapeutic decision making is possible in individual patient care. The morphometric techniques are inexpensive and fairly simple and therefore can be applied in most pathology laboratories.

337 citations