Fully Automatic Segmentation of AP Pelvis X-rays via Random Forest Regression and Hierarchical Sparse Shape Composition

Knowledge of landmarks and contours in anteroposterior (AP) pelvis X-rays is invaluable for computer aided diagnosis, hip surgery planning and image-guided interventions. This paper presents a fully automatic and robust approach for landmarking and segmentation of both pelvis and femur in a conventional AP X-ray. Our approach is based on random forest regression and hierarchical sparse shape composition. Experiments conducted on 436 clinical AP pelvis x-rays show that our approach achieves an average point-to-curve error around 1.3 mm for femur and 2.2 mm for pelvis, both with success rates around 98%. Compared to existing methods, our approach exhibits better performance in both the robustness and the accuracy.

Do mature pulmonary lobes grow after transplantation into an immature recipient?

BACKGROUND The use of reduced-size adult lung transplants could help solve the profound pediatric donor lung shortage. However, adequate long-term function of the mature grafts requires growth in proportion to the recipient's development. METHODS Mature left lower lobes from adult mini-pigs (age: 7 months; mean body weight: 30 kg) were transplanted into 14-week-old piglets (mean body weight: 15 kg). By the end of the 14-week holding period, lungs of the recipients (n = 4) were harvested. After volumetric measurements, the lung morphology was studied using light microscopy, scanning, and transmission electron microscopy. Changes of alveolar airspace volume were determined using a computer aided image analysis system. Comparisons were made to age- and weight-matched controls. RESULTS Volumetric studies showed no significant differences (p = 0.49) between the specific volume (mL/kg body weight) of lobar grafts and left lower lobes of adult controls. Morphologic studies showed marked structural differences between the grafts and the right native lungs of the recipients, with increased average alveolar diameter of the grafts. On light microscopy and scanning electron microscopy, alveoli appeared dilated and rounded compared to the normal polygonal shape in the controls. The computer generated semi-quantitative data of relative alveolar airspace volume tended to be higher in transplanted lobes. CONCLUSIONS The mature pulmonary lobar grafts have filled the growing left hemithorax of the developing recipient. Emphysema-like alterations of the grafts were observed without evidence of alveolar growth in the mature lobar transplants. Thus, it can be questioned whether mature pulmonary grafts can guarantee sufficient long-term gas exchange in growing recipients.

Postmortem MR quantification of the heart for characterization and differentiation of ischaemic myocardial lesions.

OBJECTIVES Recently, an MRI quantification sequence has been developed which can be used to acquire T1- and T2-relaxation times as well as proton density (PD) values. Those three quantitative values can be used to describe soft tissue in an objective manner. The purpose of this study was to investigate the applicability of quantitative cardiac MRI for characterization and differentiation of ischaemic myocardial lesions of different age. MATERIALS AND METHODS Fifty post-mortem short axis cardiac 3 T MR examinations have been quantified using a quantification sequence. Myocardial lesions were identified according to histology and appearance in MRI images. Ischaemic lesions were assessed for mean T1-, T2- and proton density values. Quantitative values were plotted in a 3D-coordinate system to investigate the clustering of ischaemic myocardial lesions. RESULTS A total of 16 myocardial lesions detected in MRI images were histologically characterized as acute lesions (n = 8) with perifocal oedema (n = 8), subacute lesions (n = 6) and chronic lesions (n = 2). In a 3D plot comprising the combined quantitative values of T1, T2 and PD, the clusters of all investigated lesions could be well differentiated from each other. CONCLUSION Post-mortem quantitative cardiac MRI is feasible for characterization and discrimination of different age stages of myocardial infarction. KEY POINTS • MR quantification is feasible for characterization of different stages of myocardial infarction. • The results provide the base for computer-aided MRI cardiac infarction diagnosis. • Diagnostic criteria may also be applied for living patients.

Temperature dependence of postmortem MR quantification for soft tissue discrimination.

OBJECTIVES To investigate and correct the temperature dependence of postmortem MR quantification used for soft tissue characterization and differentiation in thoraco-abdominal organs. MATERIAL AND METHODS Thirty-five postmortem short axis cardiac 3-T MR examinations were quantified using a quantification sequence. Liver, spleen, left ventricular myocardium, pectoralis muscle and subcutaneous fat were analysed in cardiac short axis images to obtain mean T1, T2 and PD tissue values. The core body temperature was measured using a rectally inserted thermometer. The tissue-specific quantitative values were related to the body core temperature. Equations to correct for temperature differences were generated. RESULTS In a 3D plot comprising the combined data of T1, T2 and PD, different organs/tissues could be well differentiated from each other. The quantitative values were influenced by the temperature. T1 in particular exhibited strong temperature dependence. The correction of quantitative values to a temperature of 37 °C resulted in better tissue discrimination. CONCLUSION Postmortem MR quantification is feasible for soft tissue discrimination and characterization of thoraco-abdominal organs. This provides a base for computer-aided diagnosis and detection of tissue lesions. The temperature dependence of the T1 values challenges postmortem MR quantification. Equations to correct for the temperature dependence are provided. KEY POINTS • Postmortem MR quantification is feasible for soft tissue discrimination and characterization • Temperature dependence of the T1 values challenges the MR quantification approach • The results provide the basis for computer-aided postmortem MRI diagnosis • Diagnostic criteria may also be applied for living patients.

The molecular signature of the stroma response in prostate cancer-induced osteoblastic bone metastasis highlights expansion of hematopoietic and prostate epithelial stem cell niches

The reciprocal interaction between cancer cells and the tissue-specific stroma is critical for primary and metastatic tumor growth progression. Prostate cancer cells colonize preferentially bone (osteotropism), where they alter the physiological balance between osteoblast-mediated bone formation and osteoclast-mediated bone resorption, and elicit prevalently an osteoblastic response (osteoinduction). The molecular cues provided by osteoblasts for the survival and growth of bone metastatic prostate cancer cells are largely unknown. We exploited the sufficient divergence between human and mouse RNA sequences together with redefinition of highly species-specific gene arrays by computer-aided and experimental exclusion of cross-hybridizing oligonucleotide probes. This strategy allowed the dissection of the stroma (mouse) from the cancer cell (human) transcriptome in bone metastasis xenograft models of human osteoinductive prostate cancer cells (VCaP and C4-2B). As a result, we generated the osteoblastic bone metastasis-associated stroma transcriptome (OB-BMST). Subtraction of genes shared by inflammation, wound healing and desmoplastic responses, and by the tissue type-independent stroma responses to a variety of non-osteotropic and osteotropic primary cancers generated a curated gene signature ("Core" OB-BMST) putatively representing the bone marrow/bone-specific stroma response to prostate cancer-induced, osteoblastic bone metastasis. The expression pattern of three representative Core OB-BMST genes (PTN, EPHA3 and FSCN1) seems to confirm the bone specificity of this response. A robust induction of genes involved in osteogenesis and angiogenesis dominates both the OB-BMST and Core OB-BMST. This translates in an amplification of hematopoietic and, remarkably, prostate epithelial stem cell niche components that may function as a self-reinforcing bone metastatic niche providing a growth support specific for osteoinductive prostate cancer cells. The induction of this combinatorial stem cell niche is a novel mechanism that may also explain cancer cell osteotropism and local interference with hematopoiesis (myelophthisis). Accordingly, these stem cell niche components may represent innovative therapeutic targets and/or serum biomarkers in osteoblastic bone metastasis.

Classification of interstitial lung disease patterns using local DCT features and random forest.

Over the last decade, a plethora of computer-aided diagnosis (CAD) systems have been proposed aiming to improve the accuracy of the physicians in the diagnosis of interstitial lung diseases (ILD). In this study, we propose a scheme for the classification of HRCT image patches with ILD abnormalities as a basic component towards the quantification of the various ILD patterns in the lung. The feature extraction method relies on local spectral analysis using a DCT-based filter bank. After convolving the image with the filter bank, q-quantiles are computed for describing the distribution of local frequencies that characterize image texture. Then, the gray-level histogram values of the original image are added forming the final feature vector. The classification of the already described patches is done by a random forest (RF) classifier. The experimental results prove the superior performance and efficiency of the proposed approach compared against the state-of-the-art.

Fully automatic segmentation of AP pelvis X-rays via random forest regression with efficient feature selection and hierarchical sparse shape composition

In clinical practice, traditional X-ray radiography is widely used, and knowledge of landmarks and contours in anteroposterior (AP) pelvis X-rays is invaluable for computer aided diagnosis, hip surgery planning and image-guided interventions. This paper presents a fully automatic approach for landmark detection and shape segmentation of both pelvis and femur in conventional AP X-ray images. Our approach is based on the framework of landmark detection via Random Forest (RF) regression and shape regularization via hierarchical sparse shape composition. We propose a visual feature FL-HoG (Flexible- Level Histogram of Oriented Gradients) and a feature selection algorithm based on trace radio optimization to improve the robustness and the efficacy of RF-based landmark detection. The landmark detection result is then used in a hierarchical sparse shape composition framework for shape regularization. Finally, the extracted shape contour is fine-tuned by a post-processing step based on low level image features. The experimental results demonstrate that our feature selection algorithm reduces the feature dimension in a factor of 40 and improves both training and test efficiency. Further experiments conducted on 436 clinical AP pelvis X-rays show that our approach achieves an average point-to-curve error around 1.2 mm for femur and 1.9 mm for pelvis.

Postmortem quantitative 1.5-T MRI for the differentiation and characterization of serous fluids, blood, CSF, and putrefied CSF.

The purpose of the present study was to investigate whether serous fluids, blood, cerebrospinal fluid (CSF), and putrefied CSF can be characterized and differentiated in synthetically calculated magnetic resonance (MR) images based on their quantitative T 1, T 2, and proton density (PD) values. Images from 55 postmortem short axis cardiac and 31 axial brain 1.5-T MR examinations were quantified using a quantification sequence. Serous fluids, fluid blood, sedimented blood, blood clots, CSF, and putrefied CSF were analyzed for their mean T 1, T 2, and PD values. Body core temperature was measured during the MRI scans. The fluid-specific quantitative values were related to the body core temperature. Equations to correct for temperature differences were generated. In a 3D plot as well as in statistical analysis, the quantitative T 1, T 2 and PD values of serous fluids, fluid blood, sedimented blood, blood clots, CSF, and putrefied CSF could be well differentiated from each other. The quantitative T 1 and T 2 values were temperature-dependent. Correction of quantitative values to a temperature of 37 °C resulted in significantly better discrimination between all investigated fluid mediums. We conclude that postmortem 1.5-T MR quantification is feasible to discriminate between blood, serous fluids, CSF, and putrefied CSF. This finding provides a basis for the computer-aided diagnosis and detection of fluids and hemorrhages.

Lung Pattern Classification for Interstitial Lung Diseases Using a Deep Convolutional Neural Network

Automated tissue characterization is one of the most crucial components of a computer aided diagnosis (CAD) system for interstitial lung diseases (ILDs). Although much research has been conducted in this field, the problem remains challenging. Deep learning techniques have recently achieved impressive results in a variety of computer vision problems, raising expectations that they might be applied in other domains, such as medical image analysis. In this paper, we propose and evaluate a convolutional neural network (CNN), designed for the classification of ILD patterns. The proposed network consists of 5 convolutional layers with 2×2 kernels and LeakyReLU activations, followed by average pooling with size equal to the size of the final feature maps and three dense layers. The last dense layer has 7 outputs, equivalent to the classes considered: healthy, ground glass opacity (GGO), micronodules, consolidation, reticulation, honeycombing and a combination of GGO/reticulation. To train and evaluate the CNN, we used a dataset of 14696 image patches, derived by 120 CT scans from different scanners and hospitals. To the best of our knowledge, this is the first deep CNN designed for the specific problem. A comparative analysis proved the effectiveness of the proposed CNN against previous methods in a challenging dataset. The classification performance (~85.5%) demonstrated the potential of CNNs in analyzing lung patterns. Future work includes, extending the CNN to three-dimensional data provided by CT volume scans and integrating the proposed method into a CAD system that aims to provide differential diagnosis for ILDs as a supportive tool for radiologists.