Spelling suggestions: "subject:"breast cancer histopathological"" "subject:"greast cancer histopathological""
1 |
Tumour localisation in histopathology imagesAkbar, Shazia January 2015 (has links)
Immunohistochemical (IHC) assessment in cancer research is important for understanding the distribution and localisation of biomarkers at the cellular level. However currently IHC analyses are predominantly performed manually, increasing workloads and introducing inter- and intra-observer variability. Automation shows great potential in clinical research to reduce pathologists' workloads and speed up cancer research in large clinical studies. Whilst recent advancements in digital pathology have enabled IHC measurements to be performed automatically, the acquisition of manual annotations of tumours in scanned digital slides is still a limiting factor. In this thesis, an automated solution to tumour localisation is explored with the aim of replacing manual annotations. As an exemplar, human breast tissue microarrays stained with estrogen receptor are considered. Methods for automated tumour localisation are described with a focus on capturing structural information in tissue by adopting superpixel properties in a rotation invariant manner, suitable for histopathology images. To incorporate essential contextual information, methods which utilise posterior tumour probabilities in an iterative manner are proposed. Results showed pixel-level agreements between automated and manual tumour segmentation masks (κ=0.811) approach inter-rater agreement between expert pathologists (κ=0.908). A large proportion of disagreements between automated and manual segmentations were shown to correlate to minor discrepancies, inconsequential for IHC assessment. IHC scores extracted from automated and manual tumour segmentation masks showed strong agreements (Allred: κˆ=0.911; Quickscore: κˆ=0.922), demonstrating the potential of automation in clinical practice across large clinical trials.
|
2 |
Deep YOLO-Based Detection of Breast Cancer Mitotic-Cells in Histopathological ImagesMaisun Mohamed, Al Zorgani,, Irfan, Mehmood,, Hassan,Ugail,, Al Zorgani, Maisun M., Mehmood, Irfan, Ugail, Hassan 25 March 2022 (has links)
yes / Coinciding with advances in whole-slide imaging scanners, it is become essential to automate the conventional image-processing techniques to assist pathologists with some tasks such as mitotic-cells detection. In histopathological images analysing, the mitotic-cells counting is a significant biomarker in the prognosis of the breast cancer grade and its aggressiveness. However, counting task of mitotic-cells is tiresome, tedious and time-consuming due to difficulty distinguishing between mitotic cells and normal cells. To tackle this challenge, several deep learning-based approaches of Computer-Aided Diagnosis (CAD) have been lately advanced to perform counting task of mitotic-cells in the histopathological images. Such CAD systems achieve outstanding performance, hence histopathologists can utilise them as a second-opinion system. However, improvement of CAD systems is an important with the progress of deep learning networks architectures. In this work, we investigate deep YOLO (You Only Look Once) v2 network for mitotic-cells detection on ICPR (International Conference on Pattern Recognition) 2012 dataset of breast cancer histopathology. The obtained results showed that proposed architecture achieves good result of 0.839 F1-measure.
|
3 |
IR imaging in breast cancer: from histopathological recognition to characterization of tumour microenvironment / Imagerie IR dans l'étude du cancer du sein: reconnaissance histopathologique et caractérisation du microenvironnement tumoralBenard, Audrey 15 June 2012 (has links)
Breast cancer is a global public health problem since it is the most frequently diagnosed cancer in women in Western countries. Clinical guidelines for breast cancer prognosis/diagnosis are currently based on tumour size, histological type and grade, lymph node status as well as the expression of various cellular receptors. Yet, current predictions remain unsatisfactory to identify the best treatment for the individual patient. The search for identifying new predictive and prognostic factors is ongoing. Furthermore, compelling evidences have solidified the notion that the evolving epithelial cells, founders of the breast disease, are helped in their malignant course by the tumour microenvironment. Better characterizing the dual effect of the immune regulation but also the epithelial-stromal cross-talk on both tumour-promotion and -suppression is essential for understanding patient uniqueness and their implication in disease outcome. Because of its potential to probe tissues and cells at the molecular level without requirement for extrinsic contrast agents, infrared spectroscopy was seen as an attractive tool for clinical and diagnostic analysis in order to complement the existing methods. <p>In a first step, recording and processing methodology had to be defined in order to optimally compare IR spectra. The methodology developed and the analysis tools tested on carcinoma cell lines, demonstrated that spectra could be distinguished based on the cell line phenotypic nature. <p>The potential of IR imaging for breast tissular structure differentiation was highlighted in this thesis, demonstrating that spectral signature can be correlated with the major histological cell types observed in breast disease tissues. In order to develop a robust algorithm translating spectral data into helpful histopathological information, a spectral database of histologically well-defined breast tissues was built and used for the development of a cell type classifier. This latter one was extensively validated on independent clinical cases. Firstly, the IR-based histopathological classifier correctly assigned spectra acquired on eleven breast disease samples based on their histological nature. Secondly, lymphocyte and Collagen & Fibroblasts spectral signatures were demonstrated to be independent from tissue type and organ since, although trained on reference spectra recorded into breast disease samples, the cell type classifier correctly assigned spectra acquired on lymph nodes/tonsils and scar tissues respectively. Thirdly, we concluded that spectroscopically, breast carcinoma cell lines in culture are well-suited tumour models since spectra acquired on these carcinoma cell lines were correctly recognized as epithelium by the IR-based histological classifier. <p>By spectral characterizing lymphocytes from lymph nodes and tonsils, we demonstrated that the spectra acquired contained enough information to statistically discriminate them according to their lymphocyte activation states. Although considered as activated, the breast disease lymphoid infiltrates were found to present distinct spectral signature from lymphocytes acquired on activated lymph nodes and tonsils. Furthermore, tumour microenvironment, characterized by IR-imaging was demonstrated to exhibit a distinct spectral signature from wound healing tissues. These studies proved the uniqueness of the signature of both lymphoid infiltrate and tumour microenvironment in breast disease context. Correlating these specific spectral signatures to patient outcome and therapeutics response could help better consider the uniqueness of the patient. In a last step, considering the epithelial signature of carcinomas of both low and high grades, we demonstrated that the biochemical information reflected in the IR micro-spectra was clinically relevant for grading purpose.<p><p> <p><p>Le cancer du sein est le cancer le plus fréquemment diagnostiqué chez les femmes dans les pays occidentaux. Jusqu’à peu, les cellules épithéliales tumorales étaient vues comme les seuls acteurs de la carcinogenèse ;processus se déroulant dans un milieu extracellulaire considéré au pire comme passif ou permissif à l’évolution tumorale des cellules épithéliales adjacentes. Cependant, de nombreuses études ont montré que ce microenvironnement tumoral pouvait soit promouvoir le processus de carcinogenèse soit le combattre empêchant par la même, l’occurrence de la maladie. <p>Ce projet de thèse s’inscrit dans une problématique actuelle, à savoir une meilleure compréhension de la maladie mais également une prise en charge plus individualisée des patientes. Nous abordons ici une voie de recherche novatrice basée sur la signature globale des molécules cellulaires via leur spectre infrarouge. La technologie utilisée, à savoir la spectroscopie infrarouge, nous fournit une observation quantitative et qualitative de milliers de vibrations moléculaires. L’adaptation de réseaux de plusieurs milliers de détecteurs indépendants aux microscopes infrarouges permet, grâce aux méthodes statistiques multivariées, d’investiguer l’architecture macromoléculaire des cellules au sein d’une coupe tissulaire et de corréler les informations spectrales ainsi obtenues à l’histopathologie des tissus. Par cette technologie, nous visons à mettre au point un outil diagnostique et pronostique pour le cancer du sein basé sur l’imagerie IR. <p>Durant ce projet, nous avons montré que les différents types cellulaires observés dans les carcinomes mammaires pouvaient être distingués par le biais de leur spectre IR, qu’un modèle de reconnaissance histologique pouvait être construit, validé et surtout automatisé et que ce modèle pouvait être transposé à l’étude d’autres tissus (ganglions, amygdales et cicatrices) et d’autres types d’échantillons (cellules épithéliales en culture). Nous avons également montré que les spectres de cellules épithéliales pouvaient être corrélés au grade histopathologique de la tumeur. Les spectres acquis de ganglions/amygdales ont montré que les profils spectraux pouvaient être corrélés à l’état d’activation lymphocytaire. De plus, l’étude de l’état d’activation lymphocytaire et fibroblastique a permis de mettre en avant un profil spectral propre et bien distinct des infiltrats lymphocytaires d’une part et de la matrice extracellulaire aux abords des tumeurs invasives d’autre part. <p> / Doctorat en Sciences agronomiques et ingénierie biologique / info:eu-repo/semantics/nonPublished
|
Page generated in 0.1101 seconds