Quality-of-service improvements from coupling a digital chest unit with integrated speech recognition, information, and Picture Archiving and Communications Systems

Speech recognition reporting for chest examinations was introduced and tightly integrated with a Radiology Information System (RIS) and a Picture Archiving and Communications System (PACS). A feature of this integration was the unique one-to-one coupling of the workstation displayed case and the reporting via speech recognition for that and only that particular examination and patient. The utility of the resulting, wholly integrated electronic environment was then compared with that of the previous analog chest unit and dedicated wet processor, with reporting of hard copy examinations by direct dictation to a typist. Improvements in quality of service in comparison to the previous work environment include (1) immediate release of the patient, (2) decreased rate of repeat radiographs, (3) improved image quality, (4) decreased time for the examination to be available for interpretation, (5) automatic hanging of current and previous images, (6) ad-hoc availability of images, (7) capability of the radiologist to immediately review and correct the transcribed report, (8) decreased time for clinicians to view results, and (9) increased capacity of examinations per room.     

Building and Querying RDF/OWL Database of Semantically Annotated Nuclear Medicine Images

As the use of positron emission tomography-computed tomography (PET-CT) has increased rapidly, there is a need to retrieve relevant medical images that can assist image interpretation. However, the images themselves lack the explicit information needed for query. We constructed a semantically structured database of nuclear medicine images using the Annotation and Image Markup (AIM) format and evaluated the ability the AIM annotations to improve image search. We created AIM annotation templates specific to the nuclear medicine domain and used them to annotate 100 nuclear medicine PET-CT studies in AIM format using controlled vocabulary. We evaluated image retrieval from 20 specific clinical queries. As the gold standard, two nuclear medicine physicians manually retrieved the relevant images from the image database using free text search of radiology reports for the same queries. We compared query results with the manually retrieved results obtained by the physicians. The query performance indicated a 98 % recall for simple queries and a 89 % recall for complex queries. In total, the queries provided 95 % (75 of 79 images) recall, 100 % precision, and an F1 score of 0.97 for the 20 clinical queries. Three of the four images missed by the queries required reasoning for successful retrieval. Nuclear medicine images augmented using semantic annotations in AIM enabled high recall and precision for simple queries, helping physicians to retrieve the relevant images. Further study using a larger data set and the implementation of an inference engine may improve query results for more complex queries.     

A Mammography Database and View System for African-American Patients

We have digitized mammography films of African-American patients treated in the Howard University Hospital Radiology Department and have developed a database using these images. Two hundred and sixty cases totaling more than 5,000 images have been scanned with a high resolution Kodak LS85 laser scanner. The database system and web-based search engine were developed using MySQL and PHP. The database has been evaluated by medical professionals, and the experimental results obtained so far are promising with high image quality and fast access time. We have also developed an image viewing system, D-Viewer, to display these digitized mammograms. This viewer is coded in Microsoft Visual C# and is intended to help medical professionals view and retrieve large data sets in near real time. Finally, we are currently developing an image content-based retrieval function for the database system to provide improved search capability for the medical professionals.     

"Robo-Rad": an inexpensive user-friendly multimedia report system for radiology.

BACKGROUND AND PURPOSE: The complex information obtained by CT, MR, and ultrasound examinations is often difficult to convey with a written report. Today's multimedia computer technology provides a medium within which the audio and the visual components of a radiologic consultation can be made available simultaneously, with the projected capability of remote access from any personal computer. A system was developed to run on low-end computer systems with image quality adequate for reporting purposes and prudent memory management (each report occupies < 4 MB). With this system-"Robo-Rad"-the image and radiologist are recorded simultaneously while he or she describes and points out (with a mouse) areas of interest. This dynamic report, along with patient data, can be retrieved and viewed by the consulting physician at his/her convenience using a low-end PC or Macintosh computer. MATERIALS AND METHODS: To assess the clinical utility of Robo-Rad, survey responses were solicited from clinical physicians at the Penn State University Hospital (41.5% faculty/fellows, 31.7% residents, 11.8% medical students, 2% clinical nursing; n = 101) during a hands-on demonstration using studies of 35 consecutive inpatients whose CT scans had been dictated into the system. RESULTS: In an average week, the surveyed professionals ordered 3.2 +/- 3.0 CT studies, reviewed 3.8 +/- 3.0 CTs, spent 1.5 +/- 2.0 hours locating Ct studies, and discussed 2.3 +/- 1.9 CT cases with a radiologist. The average time spent discussing a single CT case with a radiologist was reported as 9.4 +/- 5.9 minutes. On a five-point rating scale (1 = not at all to 5 = very much so), respondents indicated that the Robo-Rad report was helpful (4.3 +/- 0.7) and provided clinically important information that would be difficult to convey with current dictation systems (4.2 +/- 0.8). Desire to discuss the case with a radiologist in addition to viewing the Robo-Rad report scored 3.2 +/- 1.0. If such a system were readily available, 91.8% of the respondents indicated that they would use it in addition to the currently available written report and audio dictation system, and 96.6% would use it instead of the current system. Local area network and modems were the modalities of highest interest for remote access (69.3% and 44.6%, respectively). CONCLUSIONS: Judging by these data, the Robo-Rad system would be of benefit to clinicians. It provides a user-friendly, low-cost multimedia radiology report utilizing readily available technology to improve radiologist-clinician communication.     

Radiologist’s clinical information review workstation interfaced with digital dictation system

Efficient access to information systems integrated into the radiologist’s interpretation workflow will result in a more informed radiologist, with an enhanced capability to render an accurate interpretation. We describe our implementation of radStation, a radiologist’s clinical information review workstation that combines a digital dictation station with a clinical information display. radStation uses client software distributed to the radiologist’s workstation and central server software, both running Windows NT (Microsoft, Redmond, WA). The client sytem has integrated digital dictation software. The bar-code microphone (Boomerang, Dictaphone Corp, Stratford, CT) also serves as a computer input device forwarding the procedure’s accession number to the server software. This initiates multiple quaries to available lagacy databases, including the radiology information system (RIS), laboratory information system, clinic notes, hospital discharge, and operative report system. The three-tier architecture then returns the clinical results to the radStation client for display. At the conclusion of the dictation, the digital voice file is transferred to the dictation server and the client notifies the RIS to update the examination status. The system is efficient in its information retrieval, with queries displayed in about 1 second. The radStation client requires less than 5 minutes of radiologist training in its operation, given that its control interface integrates with the well-learned dictation process. The telephone-based dictation system, which this new system replaced, remains available as a back-up system in the event of an unexpected digital dictation system failure. This system is well accepted and valued by the radiologists. The system interface is quickly mastered. The system does not interrupt dictation workflow with the display of all information initiated with examination bar-coding. This system’s features could become an accepted model as a standard tool for radiologists.     

Early Experience Using an Online Reporting System for Interventional Radiology Procedure-Related Complications Integrated with a Digital Dictation System

The absence of user-friendly systems for reporting complications is a major barrier to improving quality assurance (QA) programs in interventional radiology (IR) services. We describe the implementation of a QA application that is completely integrated with the radiology dictation system. We implemented an IR QA process as a module within the electronic medical record and radiologist dictation system applications used at our institution. After a radiologist completes a dictation, he or she must select from a drop-down list of complications before proceeding to the next case. Delayed QA events can be entered using the same applications. All complication entries are sent to a database, which is queried to run reports. During the study period, all the 20,034 interventional procedures were entered in the QA database, 1,144 complications were reported, 110 (9.6%) of which were classified as major. Although majority of the complications (996) were entered at the time of dictation, 148 complications (12.9%) were entered afterwards. All major complications were referred to the IR peer review committee, and 30 of these were discussed in the morbidity and mortality meetings. We studied post-lung-biopsy pneumothorax and chest tube rates and initiated a quality improvement process based on the results.The integration of the IR QA reporting system into the workflow process and the mandatory requirements for completion has the potential to minimize the work effort required to enter complication data, and improve participation in the QA process.     

Automatic radiologic reporting system using speech recognition

A radiograph report is usually made from an oral dictation by a radiologist, which is then typed. Typing Japanese is rather inconvenient and consumes many hours. In this paper we introduce a computer-assisted reporting system for radiologic images using speech recognition. The hardware of the reporting system consists of a speech recognizer DP-200(NEC) and a personal computer PC-8801 or PC-9801. The DP-200 has the capability of storing 500 different words spoken by a radiologist. At present, three application programs have been designed. These are for the interpretation of a liver scintigram, a bone scintigram and a chest radiograph. Data entry is done by the radiologist at a CRT display terminal in a conversational manner with predefined and predetermined branching. The time required to make a normal report using the liver or bone scintigram system was within one minute. The reporting time was several minutes in the case of an abnormality report. It is suggested that the system is useful for making an imaging report, for constructing the data base for the interpretation of medical images and for the picture archiving and communication system.     

An automated, video tape-based image archiving system.

We have developed an image storage and retrieval system that makes use of a Super-VHS video tape recorder, and a personal computer fitted with an interface board and a video frame grabber. Under PC control, video images are acquired into the frame grabber, a numeric bar code is graphically superimposed for identification purposes, and the composite images are recorded on video tape. During retrieval, the bar code is decoded in real-time and the desired images are automatically retrieved. This video tape-based system, enables the images to be previewed and retrieved much faster than if stored in digital format.     

MyCases—A Portable Application for Radiologic Case Collections

Radiologists come across interesting patient cases almost every day. This work proposes a novel case database server for quick and easy storage of such cases including whole image series, patient data, and annotations. Cases can be added to the database by saving DICOM images into a predefined directory on the local network. The application automatically extracts patient and study data from the DICOM header and saves it in the database while images are stored as anonymized JPEG files. Users can mark their cases as private or public (visible to all users). Different data fields for annotations and categorization of a case are available. The user frontend also provides several retrieval mechanisms allowing for browsing the cases and performing different kinds of search queries. The stored series can be scrolled interactively in the form of scrollable image stacks. The project is realized as a web-based application using a portable web and database server software package (XAMPP). This makes the system very lightweight and easy to run on almost any desktop computer, even from a USB flash drive, without the need for deeper IT knowledge and administrative rights.     

Mapping LIDC,RadLex?, and Lung Nodule Image Features

Ideally, an image should be reported and interpreted in the same way (e.g., the same perceived likelihood of malignancy) or similarly by any two radiologists; however, as much research has demonstrated, this is not often the case. Various efforts have made an attempt at tackling the problem of reducing the variability in radiologists’ interpretations of images. The Lung Image Database Consortium (LIDC) has provided a database of lung nodule images and associated radiologist ratings in an effort to provide images to aid in the analysis of computer-aided tools. Likewise, the Radiological Society of North America has developed a radiological lexicon called RadLex. As such, the goal of this paper is to investigate the feasibility of associating LIDC characteristics and terminology with RadLex terminology. If matches between LIDC characteristics and RadLex terms are found, probabilistic models based on image features may be used as decision-based rules to predict if an image or lung nodule could be characterized or classified as an associated RadLex term. The results of this study were matches for 25 (74%) out of 34 LIDC terms in RadLex. This suggests that LIDC characteristics and associated rating terminology may be better conceptualized or reduced to produce even more matches with RadLex. Ultimately, the goal is to identify and establish a more standardized rating system and terminology to reduce the subjective variability between radiologist annotations. A standardized rating system can then be utilized by future researchers to develop automatic annotation models and tools for computer-aided decision systems.Key words: Chest CT, digital imaging, image data, image interpretation, imaging informatics, lung, radiographic image interpretation, computer-assisted, reporting, RadLex, semantic, LIDC     

Recommendations for image prefetch or film digitization strategy based on an analysis of an historic radiology image database

Picture archiving and communications systems (PACS) utilize short- and long-term storage to provide both rapid retrieval and large storage capacity. Owing to the practical limitations imposed on the size of the much faster short-term storage, it is important to use an effective algorithm in the retrieval of comparison images from long to short-term storage. A strategy must be used to maximize the likelihood that the relevant historic images have been previously retrieved into short-term memory. Data were collected with a database consisting of 754 consecutive examinations and 7,723 associated historic studies. The most frequent number of previous examinations was zero (11% of patients). In 45% of cases, no previous matching examinations had been performed. Two basic strategies of image retrieval were evaluated. The first algorithm retrieved the lastn studies in chronological order. The second strategy tested was retrieval based on a defined interval of time. This strategy was found to be less efficient. By using the former strategy, a 91% success rate (defined as successful retrieval of the previous matching exam) was achieved with retrieval of only 30% of the prior exams. The second approach required retrieval of 70% of the prior exams to achieve a 90% success rate for the previous matching exam. However, the data from this latter strategy suggest that examinations are often ordered in clusters. Thus, there was found to be a 72% likelihood that a previous matching exam, if present, would available on a PACS after only 1 week of operation, and an 80% chance after only 1 month of operation. The data therefore suggest that digitization of film in a new PACS environment might not be necessary owing to the relatively short period of time required to populate the database with historical studies.     

INDIAM—An e-Learning System for the Interpretation of Mammograms

We propose the design of a teaching system named Interpretation and Diagnosis of Mammograms (INDIAM) for training students in the interpretation of mammograms and diagnosis of breast cancer. The proposed system integrates an illustrated tutorial on radiology of the breast, that is, mammography, which uses education techniques to guide the user (doctors, students, or researchers) through various concepts related to the diagnosis of breast cancer. The user can obtain informative text about specific subjects, access a library of bibliographic references, and retrieve cases from a mammographic database that are similar to a query case on hand. The information of each case stored in the mammographic database includes the radiological findings, the clinical history, the lifestyle of the patient, and complementary exams. The breast cancer tutorial is linked to a module that simulates the analysis and diagnosis of a mammogram. The tutorial incorporates tools for helping the user to evaluate his or her knowledge about a specific subject by using the education system or by simulating a diagnosis with appropriate feedback in case of error. The system also makes available digital image processing tools that allow the user to draw the contour of a lesion, the contour of the breast, or identify a cluster of calcifications in a given mammogram. The contours provided by the user are submitted to the system for evaluation. The teaching system is integrated with AMDI—An Indexed Atlas of Digital Mammograms—that includes case studies, e-learning, and research systems. All the resources are accessible via the Web.     

Automatic on-line electronic portal image analysis with a wavelet-based edge detector

A fully automatic method for on-line electronic portal image analysis is proposed. The method uses multiscale edge detection with wavelets for both the field outline and the anatomical structures. An algorithm to extract and combine the information from different scales has been developed. The edges from the portal image are aligned with the edges from the reference image using chamfer matching. The reference is the first portal image of each treatment. The matching is applied first to the field and subsequently to the anatomy. The setup deviations are quantified as the displacement of the anatomical structures relative to the radiation beam boundaries. The performance of the algorithm was investigated for portal images with different contrast and noise level. The automatic analysis was used first to detect simulated displacements. Then the automatic procedure was tested on anterior-posterior and lateral portal images of a pelvic phantom. In both sets of tests the differences between the measured and the actual shifts were used to quantify the performance. Finally we applied the automatic procedure to clinical images of pelvic and lung regions. The output of the procedure was compared with the results of a manual match performed by a trained operator. The errors for the phantom tests were small: average standard deviation of 0.39 mm and 0.26 degrees and absolute mean error of 0.31 mm and 0.2 degrees were obtained. In the clinical cases average standard deviations of 1.32 mm and 0.6 degrees were found. The average absolute mean errors were 1.09 mm and 0.39 degrees. Failures were registered in 2% of the phantom tests and in 3% of the clinical cases. The algorithm execution is approximately 5 s on a 168 MHz Sun Ultra 2 workstation. The automatic analysis tool is considered to be a very useful tool for on-line setup corrections.     

Dynamic “Inline” Images: Context-Sensitive Retrieval and Integration of Images into Web Documents

Integrating relevant images into web-based information resources adds value for research and education. This work sought to evaluate the feasibility of using “Web 2.0” technologies to dynamically retrieve and integrate pertinent images into a radiology web site. An online radiology reference of 1,178 textual web documents was selected as the set of target documents. The ARRS GoldMiner™ image search engine, which incorporated 176,386 images from 228 peer-reviewed journals, retrieved images on demand and integrated them into the documents. At least one image was retrieved in real-time for display as an “inline” image gallery for 87% of the web documents. Each thumbnail image was linked to the full-size image at its original web site. Review of 20 randomly selected Collaborative Hypertext of Radiology documents found that 69 of 72 displayed images (96%) were relevant to the target document. Users could click on the “More” link to search the image collection more comprehensively and, from there, link to the full text of the article. A gallery of relevant radiology images can be inserted easily into web pages on any web server. Indexing by concepts and keywords allows context-aware image retrieval, and searching by document title and subject metadata yields excellent results. These techniques allow web developers to incorporate easily a context-sensitive image gallery into their documents.     

一个基于内容检索的医学图像数据库

介绍了一个基于图像内容检索的医学图像数据库系统。分别讨论了用于图像内容检索的颜色、纹理和形状特征以及在此特征集下的图像相似性度量 ;介绍了数据库的组织及该系统的界面     

Content-based histopathology image retrieval using a kernel-based semantic annotation framework

Large amounts of histology images are captured and archived in pathology departments due to the ever expanding use of digital microscopy. The ability to manage and access these collections of digital images is regarded as a key component of next generation medical imaging systems. This paper addresses the problem of retrieving histopathology images from a large collection using an example image as query. The proposed approach automatically annotates the images in the collection, as well as the query images, with high-level semantic concepts. This semantic representation delivers an improved retrieval performance providing more meaningful results. We model the problem of automatic image annotation using kernel methods, resulting in a unified framework that includes: (1) multiple features for image representation, (2) a feature integration and selection mechanism (3) and an automatic semantic image annotation strategy. An extensive experimental evaluation demonstrated the effectiveness of the proposed framework to build meaningful image representations for learning and useful semantic annotations for image retrieval.     

基于数字图像视觉特征自动检测超声图像中胎儿颈项透明层

目的利用计算机技术设计一种基于数字图像视觉特征的超声图像中胎儿颈项透明层（NT）的自动检测方法。方法建立基于数字图像视觉特征自动检测超声图像中胎儿NT的方法,包括胎儿超声图像提取、连通分量提取、目标检测和NT测量的实现共4个主要步骤。评价标准实验图像（n=35）、临床挑选图像（n=1208）和人工已标注图像（n=120）3组图像采用自动检测方法进行NT测量的实验结果,包括其定位准确性和测量误差,并记录最大误差值。结果应用自动检测方法检测标准实验图像、临床挑选图像和人工已标注图像中胎儿NT的定位准确率分别为100％（35／35）、90．7％（1096／1208）、90．8％（109／120）,NT的测量误差均小于0．03mm。结论成功建立了基于数字图像视觉特征自动检测超声图像中胎儿NT的方法,可减少人工测量的主观性和随机性,提高超声筛查的准确性。