Digital Imaging

Evaluating Completeness of a Radiology Glossary Using Iterative Refinement Abstract A lay-language glossary of radiology, built to help patients better understand the content of their radiology reports, has been analyzed for its coverage and readability, but not for its completeness. We present an iterative method to sample radiology reports, identify "missing" terms, and measure the glossary's completeness. We hypothesized that the refinement process would reduce the number of missing terms to fewer than 1 per report. A random sample of 1000 radiology reports from a large US academic health system was divided into 10 cohorts of 100 reports each. Each cohort was reviewed in sequence by two investigators to identify terms (single words and multi-word phrases) absent from the glossary. Terms marked as new were added to the glossary and hence was shown as matched in subsequent cohorts. This HIPAA-compliant study was IRB-approved; informed consent was waived. The refinement process added a mean of 288.0 new terms per 100 reports in the first 5 cohorts vs. a mean of 66.0 new terms per 100 reports in the last 5 cohorts; the difference was statistically significant (p < .01). After reviewing 500 reports, the review process found fewer than 1 new term per report in each of 500 subsequent reports. The findings suggest that 500 to 1000 reports is adequate to test the completeness of a glossary, and that the glossary after iterative refinement achieved a high level of completeness to cover the vocabulary of radiology reports.

Improving Sensitivity on Identification and Delineation of Intracranial Hemorrhage Lesion Using Cascaded Deep Learning Models Abstract Highly accurate detection of the intracranial hemorrhage without delay is a critical clinical issue for the diagnostic decision and treatment in an emergency room. In the context of a study on diagnostic accuracy, there is a tradeoff between sensitivity and specificity. In order to improve sensitivity while preserving specificity, we propose a cascade deep learning model constructed using two convolutional neural networks (CNNs) and dual fully convolutional networks (FCNs). The cascade CNN model is built for identifying bleeding; hereafter the dual FCN is to detect five different subtypes of intracranial hemorrhage and to delineate their lesions. Using a total of 135,974 CT images including 33,391 images labeled as bleeding, each of CNN/FCN models was trained separately on image data preprocessed by two different settings of window level/width. One is a default window (50/100[level/width]) and the other is a stroke window setting (40/40). By combining them, we obtained a better outcome on both binary classification and segmentation of hemorrhagic lesions compared to a single CNN and FCN model. In determining whether it is bleeding or not, there was around 1% improvement in sensitivity (97.91% [± 0.47]) while retaining specificity (98.76% [± 0.10]). For delineation of bleeding lesions, we obtained overall segmentation performance at 80.19% in precision and 82.15% in recall which is 3.44% improvement compared to using a single FCN model.

Automatic Lumbar MRI Detection and Identification Based on Deep Learning Abstract The aim of this research is to automatically detect lumbar vertebras in MRI images with bounding boxes and their classes, which can assist clinicians with diagnoses based on large amounts of MRI slices. Vertebras are highly semblable in appearance, leading to a challenging automatic recognition. A novel detection algorithm is proposed in this paper based on deep learning. We apply a similarity function to train the convolutional network for lumbar spine detection. Instead of distinguishing vertebras using annotated lumbar images, our method compares similarities between vertebras using a beforehand lumbar image. In the convolutional neural network, a contrast object will not update during frames, which allows a fast speed and saves memory. Due to its distinctive shape, S1 is firstly detected and a rough region around it is extracted for searching for L1–L5. The results are evaluated with accuracy, precision, mean, and standard deviation (STD). Finally, our detection algorithm achieves the accuracy of 98.6% and the precision of 98.9%. Most failed results are involved with wrong S1 locations or missed L5. The study demonstrates that a lumbar detection network supported by deep learning can be trained successfully without annotated MRI images. It can be believed that our detection method will assist clinicians to raise working efficiency.

Towards Accurate Segmentation of Retinal Vessels and the Optic Disc in Fundoscopic Images with Generative Adversarial Networks Abstract Automatic segmentation of the retinal vasculature and the optic disc is a crucial task for accurate geometric analysis and reliable automated diagnosis. In recent years, Convolutional Neural Networks (CNN) have shown outstanding performance compared to the conventional approaches in the segmentation tasks. In this paper, we experimentally measure the performance gain for Generative Adversarial Networks (GAN) framework when applied to the segmentation tasks. We show that GAN achieves statistically significant improvement in area under the receiver operating characteristic (AU-ROC) and area under the precision and recall curve (AU-PR) on two public datasets (DRIVE, STARE) by segmenting fine vessels. Also, we found a model that surpassed the current state-of-the-art method by 0.2 − 1.0% in AU-ROC and 0.8 − 1.2% in AU-PR and 0.5 − 0.7% in dice coefficient. In contrast, significant improvements were not observed in the optic disc segmentation task on DRIONS-DB, RIM-ONE (r3) and Drishti-GS datasets in AU-ROC and AU-PR.

Integrating Wikipedia Articles and Images into an Information Resource for Radiology Patients Abstract Wikipedia—an open-access online encyclopedia—contains a large number of medically relevant articles and images that may help supplement glossaries of radiology terms. We sought to determine the extent to which concepts from a large online radiology glossary developed as part of the Patient-Oriented Radiology Reporter (PORTER) initiative could be mapped to relevant Wikipedia web pages and images using automated or semi-automated approaches. The glossary included 4090 concepts with their definitions; the concept's preferred name and lexical variants, such as plurals, adjectival forms, synonyms, and abbreviations, yielded a total of 13,030 terms. Of the 4090 concepts, 3063 (74.9%) had a corresponding English-language Wikipedia page identified by automated search with subsequent manual review. We applied the MediaWiki application programming interface (API) to generate web-service calls to identify the images from each concept's corresponding Wikipedia page; three reviewers selected relevant images to associate with the glossary's concepts. Licensing terms for the images were reviewed. For 800 randomly sampled concepts that had associated Wikipedia pages, 362 distinct images were identified from the MediaWiki library and matched to 404 concepts (51%). Three images (1%) had unspecified licensing terms; the rest were in the public domain or available via a Creative Commons license. Wikipedia and the MediaWiki library offer a large collection of medical articles and images that can be incorporated into an online lay-language glossary of radiology terms though a semi-automated approach.

A Platform Integrating Acquisition, Reconstruction, Visualization, and Manipulator Control Modules for MRI-Guided Interventions Abstract This work presents a platform that integrates a customized MRI data acquisition scheme with reconstruction and three-dimensional (3D) visualization modules along with a module for controlling an MRI-compatible robotic device to facilitate the performance of robot-assisted, MRI-guided interventional procedures. Using dynamically-acquired MRI data, the computational framework of the platform generates and updates a 3D model representing the area of the procedure (AoP). To image structures of interest in the AoP that do not reside inside the same or parallel slices, the MRI acquisition scheme was modified to collect a multi-slice set of intraoblique to each other slices; which are termed composing slices. Moreover, this approach interleaves the collection of the composing slices so the same k-space segments of all slices are collected during similar time instances. This time matching of the k-space segments results in spatial matching of the imaged objects in the individual composing slices. The composing slices were used to generate and update the 3D model of the AoP. The MRI acquisition scheme was evaluated with computer simulations and experimental studies. Computer simulations demonstrated that k-space segmentation and time-matched interleaved acquisition of these segments provide spatial matching of the structures imaged with composing slices. Experimental studies used the platform to image the maneuvering of an MRI-compatible manipulator that carried tubing filled with MRI contrast agent. In vivo experimental studies to image the abdomen and contrast enhanced heart on free-breathing subjects without cardiac triggering demonstrated spatial matching of imaged anatomies in the composing planes. The described interventional MRI framework could assist in performing real-time MRI-guided interventions.

Multi-objective Parameter Auto-tuning for Tissue Image Segmentation Workflows Abstract We propose a software platform that integrates methods and tools for multi-objective parameter auto-tuning in tissue image segmentation workflows. The goal of our work is to provide an approach for improving the accuracy of nucleus/cell segmentation pipelines by tuning their input parameters. The shape, size, and texture features of nuclei in tissue are important biomarkers for disease prognosis, and accurate computation of these features depends on accurate delineation of boundaries of nuclei. Input parameters in many nucleus segmentation workflows affect segmentation accuracy and have to be tuned for optimal performance. This is a time-consuming and computationally expensive process; automating this step facilitates more robust image segmentation workflows and enables more efficient application of image analysis in large image datasets. Our software platform adjusts the parameters of a nuclear segmentation algorithm to maximize the quality of image segmentation results while minimizing the execution time. It implements several optimization methods to search the parameter space efficiently. In addition, the methodology is developed to execute on high-performance computing systems to reduce the execution time of the parameter tuning phase. These capabilities are packaged in a Docker container for easy deployment and can be used through a friendly interface extension in 3D Slicer. Our results using three real-world image segmentation workflows demonstrate that the proposed solution is able to (1) search a small fraction (about 100 points) of the parameter space, which contains billions to trillions of points, and improve the quality of segmentation output by × 1.20, × 1.29, and × 1.29, on average; (2) decrease the execution time of a segmentation workflow by up to 11.79× while improving output quality; and (3) effectively use parallel systems to accelerate parameter tuning and segmentation phases.

The Role of the Integrated Digital Radiology System in Assessing the Impact of Patient Load on Emergency Computed Tomography (CT) Efficiency Abstract Time-critical management is of particular significance in the trauma and emergency setting, where intervals from patient arrival to diagnostic imaging and from imaging to radiology report are key determinants of outcome. This study, based in the Trauma and Emergency Unit of a large, tertiary-level African hospital with a fully digital radiology department, assessed the impact of increased workload on computerised tomography (CT) efficiency. Sequential, customised searches of the institutional radiology information system (RIS) were conducted to define two weekends in 2016 with the lowest and highest emergency CT workloads, respectively. The electronic RIS timestamps defining the intervals between key steps in the CT workflow were extracted and analysed for each weekend. With the exception of radiologist reporting time, workflow steps were significantly prolonged by increased workload. This study highlights the potential role of the integrated digital radiology system in enabling a detailed analysis of imaging workflow, thereby facilitating the identification and appropriate management of bottlenecks.

Probabilistic Modeling of Exam Durations in Radiology Procedures Abstract In this paper, we model the statistical properties of imaging exam durations using parametric probability distributions such as the Gaussian, Gamma, Weibull, lognormal, and log-logistic. We establish that in a majority of radiology procedures, the underlying distribution of exam durations is best modeled by a log-logistic distribution, while the Gaussian has the poorest fit among the candidates. Further, through illustrative examples, we show how business insights and workflow analytics can be significantly impacted by making the correct (log-logistic) versus incorrect (Gaussian) model choices.

Content-Based Image Retrieval System for Pulmonary Nodules Using Optimal Feature Sets and Class Membership-Based Retrieval Abstract Lung cancer manifests itself in the form of lung nodules, the diagnosis of which is essential to plan the treatment. Automated retrieval of nodule cases will assist the budding radiologists in self-learning and differential diagnosis. This paper presents a content-based image retrieval (CBIR) system for lung nodules using optimal feature sets and learning to enhance the performance of retrieval. The classifiers with more features suffer from the curse of dimensionality. Like classification schemes, we found that the optimal feature set selected using the minimal-redundancy-maximal-relevance (mRMR) feature selection technique improves the precision performance of simple distance-based retrieval (SDR). The performance of the classifier is always superior to SDR, which leans researchers towards conventional classifier-based retrieval (CCBR). While CCBR improves the average precision and provides 100% precision for correct classification, it fails for misclassification leading to zero retrieval precision. The class membership-based retrieval (CMR) is found to bridge this gap for texture-based retrieval. Here, CMR is proposed for nodule retrieval using shape-, margin-, and texture-based features. It is found again that optimal feature set is important for the classifier used in CMR as well as for the feature set used for retrieval, which may lead to different feature sets. The proposed system is evaluated using two independent databases from two continents: a public database LIDC/IDRI and a private database PGIMER-IITKGP, using three distance metrics, i.e., Canberra, City block, and Euclidean. The proposed CMR-based retrieval system with optimal feature sets performs better than CCBR and SDR with optimal features in terms of average precision. Apart from average precision and standard deviation of precision, the fraction of queries with zero precision retrieval is also measured.