Rotationally resliced 3D prostate segmentation of MR images using Bhattacharyya similarity and active band theory

PurposeIn this article, we propose a novel, semi-automatic segmentation method to process 3D MR images of the prostate using the Bhattacharyya coefficient and active band theory with the goal of providing technical support for computer-aided diagnosis and surgery of the prostate.MethodsOur method consecutively segments a stack of rotationally resectioned 2D slices of a prostate MR image by assessing the similarity of the shape and intensity distribution in neighboring slices. 2D segmentation is first performed on an initial slice by manually selecting several points on the prostate boundary, after which the segmentation results are propagated consecutively to neighboring slices. A framework of iterative graph cuts is used to optimize the energy function, which contains a global term for the Bhattacharyya coefficient with the help of an auxiliary function. Our method does not require previously segmented data for training or for building statistical models, and manual intervention can be applied flexibly and intuitively, indicating the potential utility of this method in the clinic.ResultsWe tested our method on 3D T2-weighted MR images from the ISBI dataset and PROMISE12 dataset of 129 patients, and the Dice similarity coefficients were 90.34 ± 2.21% and 89.32 ± 3.08%, respectively. The comparison was performed with several state-of-the-art methods, and the results demonstrate that the proposed method is robust and accurate, achieving similar or higher accuracy than other methods without requiring training.ConclusionThe proposed algorithm for segmenting 3D MR images of the prostate is accurate, robust, and readily applicable to a clinical environment for computer-aided surgery or diagnosis.     

Development of automatic surveillance of animal behaviour and welfare using image analysis and machine learned segmentation technique

In this paper the feasibility to extract the proportion of pigs located in different areas of a pig pen by advanced image analysis technique is explored and discussed for possible applications. For example, pigs generally locate themselves in the wet dunging area at high ambient temperatures in order to avoid heat stress, as wetting the body surface is the major path to dissipate the heat by evaporation. Thus, the portion of pigs in the dunging area and resting area, respectively, could be used as an indicator of failure of controlling the climate in the pig environment as pigs are not supposed to rest in the dunging area. The computer vision methodology utilizes a learning based segmentation approach using several features extracted from the image. The learning based approach applied is based on extended state-of-the-art features in combination with a structured prediction framework based on a logistic regression solver using elastic net regularization. In addition, the method is able to produce a probability per pixel rather than form a hard decision. This overcomes some of the limitations found in a setup using grey-scale information only. The pig pen is a difficult imaging environment because of challenging lighting conditions like shadows, poor lighting and poor contrast between pig and background. In order to test practical conditions, a pen containing nine young pigs was filmed from a top view perspective by an Axis M3006 camera with a resolution of 640×480 in three, 10-min sessions under different lighting conditions. The results indicate that a learning based method improves, in comparison with greyscale methods, the possibility to reliable identify proportions of pigs in different areas of the pen. Pigs with a changed behaviour (location) in the pen may indicate changed climate conditions. Changed individual behaviour may also indicate inferior health or acute illness.     

Fully automatic and robust segmentation of the clinical target volume for radiotherapy of breast cancer using big data and deep learning

PurposeTo train and evaluate a very deep dilated residual network (DD-ResNet) for fast and consistent auto-segmentation of the clinical target volume (CTV) for breast cancer (BC) radiotherapy with big data.MethodsDD-ResNet was an end-to-end model enabling fast training and testing. We used big data comprising 800 patients who underwent breast-conserving therapy for evaluation. The CTV were validated by experienced radiation oncologists. We performed a fivefold cross-validation to test the performance of the model. The segmentation accuracy was quantified by the Dice similarity coefficient (DSC) and the Hausdorff distance (HD). The performance of the proposed model was evaluated against two different deep learning models: deep dilated convolutional neural network (DDCNN) and deep deconvolutional neural network (DDNN).ResultsMean DSC values of DD-ResNet (0.91 and 0.91) were higher than the other two networks (DDCNN: 0.85 and 0.85; DDNN: 0.88 and 0.87) for both right-sided and left-sided BC. It also has smaller mean HD values of 10.5 mm and 10.7 mm compared with DDCNN (15.1 mm and 15.6 mm) and DDNN (13.5 mm and 14.1 mm). Mean segmentation time was 4 s, 21 s and 15 s per patient with DDCNN, DDNN and DD-ResNet, respectively. The DD-ResNet was also superior with regard to results in the literature.ConclusionsThe proposed method could segment the CTV accurately with acceptable time consumption. It was invariant to the body size and shape of patients and could improve the consistency of target delineation and streamline radiotherapy workflows.     

Quantitative proteome analysis of breast cancer cell lines using 18O-labeling and an accurate mass and time tag strategy

Proteome comparison of cell lines derived from cancer and normal breast epithelium provide opportunities to identify differentially expressed proteins and pathways associated with specific phenotypes. We employed 16O/18O peptide labeling, FT-ICR MS, and an accurate mass and time (AMT) tag strategy to simultaneously compare the relative abundance of hundreds of proteins in non-cancer and cancer cell lines derived from breast tissue. A cell line reference panel allowed relative protein abundance comparisons among multiple cell lines and across multiple experiments. A peptide database generated from multidimensional LC separations and MS/MS analysis was used for subsequent AMT tag-based peptide identifications. This peptide database represented a total of 2299 proteins, including 514 that were quantified in five cell lines using the AMT tag and 16O/18O strategies. Eighty-six proteins showed at least a threefold protein abundance change between cancer and non-cancer cell lines. Hierarchical clustering of protein abundance ratios revealed that several groups of proteins were differentially expressed between the cancer cell lines.     

Identification of breast cancer metastasis-associated proteins in an isogenic tumor metastasis model using two-dimensional gel electrophoresis and liquid chromatography-ion trap-mass spectrometry

To better understand the molecular mechanisms underlying breast cancer metastasis and search for potential markers for metastatic progression, we have developed a highly metastatic variant of human MDA-MB-435 breast cancer cell line through in vivo stepwise selection of pulmonary metastatic cells caused by parental MDA-MB-435 cells in the athymic mice. Comparative proteomic analysis using 2-DE and LC-IT-MS revealed that 102 protein spots were reproducibly altered more than three-fold between the selected variant and its parental counterpart. Eleven differentially expressed protein spots were identified with high confidence using SEQUEST with uninterpreted tandem mass raw data. Cathepsin D precursor, peroxiredoxin 6 (PDX6), heat shock protein 27 (HSP27), HSP60, tropomyosin 1 (TPM1), TPM2, TPM3, TPM4, 14-3-3 protein epsilon, and tumor protein D54 were up-regulated in the highly metastatic variant, whereas alpha B-crystalline (CRAB) was only detected in its parental counterpart. Differential expression was confirmed for four proteins including PDX6, CRAB, TPM4, and HSP60 by real-time quantitative PCR and Western blotting analysis in our model. Immunohistochemical analysis in 80 breast cancer donors demonstrated a significant association of TPM4 (p = 0.002), HSP60 (p = 0.001), PDX6 (p = 0.002) but not CRAB (p = 0.113) staining with the presence of lymph node metastasis. In addition, TPM4 staining was also associated with clinical stage (p = 0.000), but no significant association was found between TPM4, PDX6, CRAB, and HSP60 expression and tumor size, hormone receptor, and HER-2 status (p > 0.05). The functional implication of these identified proteins was also discussed. These proteomic data are valuable and informative for understanding breast cancer metastasis and searching for potential markers for metastatic progression.     

Proliferating cell nuclear antigen in normal urothelium and urothelial lesions of the urinary bladder: a quantitative assessment using a true color image analysis system

To evaluate proliferating cell nuclear antigen (PCNA) staining for assessing proliferative activity in routine pathology specimens of urinary bladder, the bladder carcinoma cell line J82 and a total of 122 specimens of normal bladder and urothelial lesions were stained with the antibody clone PC10 against proliferating cell nuclear antigen. In in vitro plateau cultures the proportion of PCNA-positive cells exceeded that of Ki-67-positive cells, and only very few cells were negative. In formalin-fixed tissues, the PCNA staining pattern, which should be confined to replicon units in the nucleus, was optimized by 1 h postfixation in an organic solvent (methacarn). Sections showed positive nuclear staining confined to basal and some suprabasal cells in normal urothelium and grade 1 dysplasias, but more generalized nuclear staining in all other neoplastic lesions. In addition, stromal cells adjacent to invasive tumors showed nuclear positivity in some instances. Using quantitative true color image analysis of sections counterstained with hemalum, the degree of brown staining of the PCNA reaction product is contrasted with the blue staining of the nuclear area. With this method low contrast specific staining not appreciated optically can be reliably detected. Image analysis data confirmed observations made on noncounterstained sections and showed significant differences between grade 1 and 2 dysplasias as well as between grade 1 dysplasia and all grades of papillary tumor. Furthermore, a significant difference in PCNA staining indices was found between grade 1 and 3 bladder carcinomas. The results indicate that PCNA staining using the PC10 antibody is not confined to the proliferative fraction of neoplastic urothelium. In contrast with data from normal tissue and malignant hematological neoplasms, the amount of PCNA is regulated differently in urothelial neoplasms, emphasizing the biological differences between the following two sets: mild dysplasia and moderate dysplasia; mild dysplasia and papillary carcinomas. The use of image analysis to standardize the detection process after controlled staining conditions is advisable in order to provide reliable data. Supported by the DFG project: Knuechel/Urothelcarcinom 263