Robust cranial cavity segmentation in CT and CT perfusion images of trauma and suspected stroke patients

Segmentation of Spontaneous Intracerebral Hemorrhage on CT With a Region Growing Method Based on Watershed Preprocessing

Intracerebral hemorrhage (ICH) poses a great threat to human life due to its high incidence and poor prognosis. Identification of the bleeding location and quantification of the volume based on CT images are of great significance for assisting the diagnosis and treatment of ICH. In this study, a region-growing algorithm based on watershed preprocessing (RG-WP) was proposed to segment and quantify the hemorrhage. The lowest points yielded by the watershed algorithm were used as seed points for region growing and then hemorrhage was segmented based on the region growing method. At the same time, to integrate the rich experience of clinicians with the algorithm, manual selection of seed points on the basis of watershed segmentation was performed. With the application of segmentation on CT images of 55 patients with ICH, the performance of the RG-WP algorithm was evaluated by comparing it with manual segmentations delineated by professional clinicians as well as the traditional ABC/2 method and the deep learning algorithm U-net. The mean deviation of hemorrhage volume of the RG-WP algorithm from manual segmentation was −0.12 ml (range: −1.05–1.16), while that of the ABC/2 from the manual was 1.05 ml (range: −0.77–9.57). Strong agreement of the algorithm and the manual was confirmed with a high intraclass correlation coefficient (ICC) (0.998, 95% CI: 0.997–0.999), which was superior to that of the ABC/2 and the manual (0.972, 95% CI: 0.953–0.984). The sensitivity (Sen), positive predictive value (PPV), dice similarity index (DSI), and Jaccard index (JI) of the RG-WP algorithm compared to the manual were 0.92 ± 0.04, 0.95 ± 0.04, 0.93 ± 0.02, and 0.88 ± 0.04, respectively, showing high consistency. Besides, the accuracy of the algorithm was also comparable to that of the deep learning method U-net, with Sen, PPV, DSI, and JI being 0.91 ± 0.09, 0.91 ± 0.06, 0.91 ± 0.05, and 0.91 ± 0.06, respectively.

Automated Quantification of Brain Lesion Volume From Post-trauma MR Diffusion-Weighted Images

Objectives

Determining the volume of brain lesions after trauma is challenging. Manual delineation is observer-dependent and time-consuming and cannot therefore be used in routine practice. The study aimed to evaluate the feasibility of an automated atlas-based quantification procedure (AQP) based on the detection of abnormal mean diffusivity (MD) values computed from diffusion-weighted MR images.

Methods

The performance of AQP was measured against manual delineation consensus by independent raters in two series of experiments based on: (i) realistic trauma phantoms (n = 5) where low and high MD values were assigned to healthy brain images according to the intensity, form and location of lesion observed in real TBI cases; (ii) severe TBI patients (n = 12 patients) who underwent MR imaging within 10 days after injury.

Results

In realistic TBI phantoms, no statistical differences in Dice similarity coefficient, precision and brain lesion volumes were found between AQP, the rater consensus and the ground truth lesion delineations. Similar findings were obtained when comparing AQP and manual annotations for TBI patients. The intra-class correlation coefficient between AQP and manual delineation was 0.70 in realistic phantoms and 0.92 in TBI patients. The volume of brain lesions detected in TBI patients was 59 ml (19–84 ml) (median; 25–75th centiles).

Conclusions

Our results support the feasibility of using an automated quantification procedure to determine, with similar accuracy to manual delineation, the volume of low and high MD brain lesions after trauma, and thus allow the determination of the type and volume of edematous brain lesions. This approach had comparable performance with manual delineation by a panel of experts. It will be tested in a large cohort of patients enrolled in the multicenter OxyTC trial (NCT02754063).

Towards subject-level cerebral infarction classification of CT scans using convolutional networks

Automatic evaluation of 3D volumes is a topic of importance in order to speed up clinical decision making. We describe a method to classify computed tomography scans on volume level for the presence of non-acute cerebral infarction. This is not a trivial task, as the lesions are often similar to other areas in the brain regarding shape and intensity. A three stage architecture is used for classification: 1) A cranial cavity segmentation network is developed, trained and applied. 2) Region proposals are generated 3) Connected regions are classified using a multi-resolution, densely connected 3D convolutional network. Mean area under curve values for subject level classification are 0.95 for the unstratified test set, 0.88 for stratification by patient age and 0.93 for stratification by CT scanner model. We use a partly segmented dataset of 555 scans of which 186 scans are used in the unstratified test set. Furthermore we examine possible dataset bias for scanner model and patient age parameters. We show a successful application of the proposed three-stage model for full volume classification. In contrast to black-box approaches, the convolutional network’s decision can be further assessed by examination of intermediate segmentation results.

Stacked Bidirectional Convolutional LSTMs for Deriving 3D Non-Contrast CT From Spatiotemporal 4D CT

The imaging workup in acute stroke can be simplified by deriving non-contrast CT (NCCT) from CT perfusion (CTP) images. This results in reduced workup time and radiation dose. To achieve this, we present a stacked bidirectional convolutional LSTM (C-LSTM) network to predict 3D volumes from 4D spatiotemporal data. Several parameterizations of the C-LSTM network were trained on a set of 17 CTP-NCCT pairs to learn to derive a NCCT from CTP and were subsequently quantitatively evaluated on a separate cohort of 16 cases. The results show that the C-LSTM network clearly outperforms the baseline and competitive convolutional neural network methods. We show good scalability and performance of the method by continued training and testing on an independent dataset which includes pathology of 80 and 83 CTP-NCCT pairs, respectively. C-LSTM is, therefore, a promising general deep learning approach to learn from high-dimensional spatiotemporal medical images.

Automated Segmentation of Tissues Using CT and MRI: A Systematic Review

RATIONALE AND OBJECTIVES

The automated segmentation of organs and tissues throughout the body using computed tomography and magnetic resonance imaging has been rapidly increasing. Research into many medical conditions has benefited greatly from these approaches by allowing the development of more rapid and reproducible quantitative imaging markers. These markers have been used to help diagnose disease, determine prognosis, select patients for therapy, and follow responses to therapy. Because some of these tools are now transitioning from research environments to clinical practice, it is important for radiologists to become familiar with various methods used for automated segmentation.

MATERIALS AND METHODS

The Radiology Research Alliance of the Association of University Radiologists convened an Automated Segmentation Task Force to conduct a systematic review of the peer-reviewed literature on this topic.

RESULTS

The systematic review presented here includes 408 studies and discusses various approaches to automated segmentation using computed tomography and magnetic resonance imaging for neurologic, thoracic, abdominal, musculoskeletal, and breast imaging applications.

CONCLUSION

These insights should help prepare radiologists to better evaluate automated segmentation tools and apply them not only to research, but eventually to clinical practice.

Intracerebral Haemorrhage Segmentation in Non-Contrast CT

A 3-dimensional (3D) convolutional neural network is presented for the segmentation and quantification of spontaneous intracerebral haemorrhage (ICH) in non-contrast computed tomography (NCCT). The method utilises a combination of contextual information on multiple scales for fast and fully automatic dense predictions. To handle a large class imbalance present in the data, a weight map is introduced during training. The method was evaluated on two datasets of 25 and 50 patients respectively. The reference standard consisted of manual annotations for each ICH in the dataset. Quantitative analysis showed a median Dice similarity coefficient of 0.91 [0.87-0.94] and 0.90 [0.85-0.92] for the two test datasets in comparison to the reference standards. Evaluation of a separate dataset of 5 patients for the assessment of the observer variability produced a mean Dice similarity coefficient of 0.95 ± 0.02 for the inter-observer variability and 0.97 ± 0.01 for the intra-observer variability. The average prediction time for an entire volume was 104 ± 15 seconds. The results demonstrate that the method is accurate and approaches the performance of expert manual annotation.

Three-dimensional orbital wall modeling using paranasal sinus segmentation

PURPOSE

Three-dimensional orbital wall modeling is a time-consuming process because of the presence of pseudoforamina. We developed an automated three-dimensional modeling software to characterize the orbital wall, and evaluated it using data from fracture patients.

METHODS

We first characterized the air and face regions using multiphase segmentation; the sinuses were segmented by applying morphological operations to air regions. Pseudoforamina of the orbital wall were offset with the segmented sinuses. Finally, the three-dimensional facial bone model, with orbital wall, was reconstructed from the segmented images.

RESULTS

Ten computed tomography data sets were used to evaluate the proposed method. Results were compared with those obtained using the active contour model and manual segmentation. The process took 31.7 ± 8.0 s, which was 30-60 times faster than other methods. The average distances between surfaces obtained with the proposed method and those obtained with manual segmentation (normal side: 0.20 ± 0.06 mm; fractured side: 0.28 ± 0.10 mm) were approximately half those obtained using the active contour model.

CONCLUSIONS

Three-dimensional orbital wall models, which were very similar to the manually segmented models, were archived within 1 min using the developed software, regardless of fracture presence. The proposed method might improve the safety and accuracy of surgical procedures.

PCSeg: Color model driven probabilistic multiphase level set based tool for plasma cell segmentation in multiple myeloma

Plasma cell segmentation is the first stage of a computer assisted automated diagnostic tool for multiple myeloma (MM). Owing to large variability in biological cell types, a method for one cell type cannot be applied directly on the other cell types. In this paper, we present PCSeg Tool for plasma cell segmentation from microscopic medical images. These images were captured from bone marrow aspirate slides of patients with MM. PCSeg has a robust pipeline consisting of a pre-processing step, the proposed modified multiphase level set method followed by post-processing steps including the watershed and circular Hough transform to segment clusters of cells of interest and to remove unwanted cells. Our modified level set method utilizes prior information about the probability densities of regions of interest (ROIs) in the color spaces and provides a solution to the minimal-partition problem to segment ROIs in one of the level sets of a two-phase level set formulation. PCSeg tool is tested on a number of microscopic images and provides good segmentation results on single cells as well as efficient segmentation of plasma cell clusters.

Robust Segmentation of the Full Cerebral Vasculature in 4D CT of Suspected Stroke Patients

A robust method is presented for the segmentation of the full cerebral vasculature in 4-dimensional (4D) computed tomography (CT). The method consists of candidate vessel selection, feature extraction, random forest classification and postprocessing. Image features include among others the weighted temporal variance image and parameters, including entropy, of an intensity histogram in a local region at different scales. These histogram parameters revealed to be a strong feature in the detection of vessels regardless of shape and size. The method was trained and tested on a large database of 264 patients with suspicion of acute ischemia who underwent 4D CT in our hospital in the period January 2014 to December 2015. Five subvolumes representing different regions of the cerebral vasculature were annotated in each image in the training set by medical assistants. The evaluation was done on 242 patients. A total of 16 (<8%) patients showed severe under or over segmentation and were reported as failures. One out of five subvolumes was randomly annotated in 159 patients and was used for quantitative evaluation. Quantitative evaluation showed a Dice coefficient of 0.91 ± 0.07 and a modified Hausdorff distance of 0.23 ± 0.22 mm. Therefore, robust vessel segmentation in 4D CT is feasible with good accuracy.