Smart searching enables radiologists' e-learning

May 19, 2008

The Internet has become an invaluable tool to radiology. New media formats make it easy to stay up to date. In addition to browsing key websites, radiologists can sign up for e-mail newsletters, RSS feeds, and podcasts, and they can access these textual, visual, and auditory data via a variety of gadgets.

The Internet has become an invaluable tool to radiology. New media formats make it easy to stay up to date. In addition to browsing key websites, radiologists can sign up for e-mail newsletters, RSS feeds, and podcasts, and they can access these textual, visual, and auditory data via a variety of gadgets. User interaction with electronic media enables truly custom-tailored learning. Such practices are impossible when learning is focused on textbooks.

Increasingly sophisticated search engines help to target the retrieval of high-quality radiological reviews, images, and case reports. These "intelligent" engines are very different from standard web-based search engines. They will learn from users' searching behavior and produce search results that no longer contain volumes of irrelevant information.

Below is a selection of websites that qualified radiologists and trainees alike may find useful for daily practice. After all, in this era of evidence-based and time-pressured medicine, the ability to gather specific information quickly, to solve a given problem, is an essential survival tool.


  • Searching Radiology (Michael P. D'Alessandro, M.D.) allows users to search specifically for peer-reviewed radiological information. The website is powered by a Google custom search engine, which concentrates its search on various radiology journals, online textbooks, and atlases. Information retrieved must be peer-reviewed and publicly accessible.

Websites scanned by the engine include eMedicine (www.emedicine. com) and Medcyclopaedia (, as well as the online versions of Radiographics, Radiology, American Journal of Neuroradiology, American Journal of Roentgenology, and British Journal of Radiology. This targeting prevents users from being flooded with results. Users no longer need waste time sifting through links to irrelevant, untrustworthy, or inaccessible information.

  • GoldMiner (American Roentgen Ray Society)

The American Roentgen Ray Society's GoldMiner is a tool for the instant retrieval of radiological images. The search technology is very different from generic web-based image-searching algorithms. Sophisticated techniques from the U.S. National Library of Medicine are applied to retrieve medical concepts contained in free-text captions of images published in peer-reviewed journals. This includes the use of Medical Subject Heading (MeSH) terms. The efficient search algorithms closely resemble the way information is indexed on PubMed. GoldMiner is, in effect, the picture-search equivalent to Medline.

To date, GoldMiner has indexed over 170,000 images from nearly 90 peer-reviewed journals.

  • Yottalook (iVirtuoso)

Yottalook may be viewed as the new star among radiology-specific search sites, owing to its elegant combination of several key search strategies into one engine. The basis for free-text Internet and image searching is provided by Google's web indexing technology. Websites known to have radiological content are emphasized. This allows sites that are normally ranked low on regular commercial search engines to become visible.

"Natural query analysis" technology first checks for any misspellings in the radiologist's query. If such errors go unrecognized, then key sites may be missed from the results list. Any ambiguities are also challenged. For example, if the engine is asked to find results for "PE," then the enquirer may be asked if this refers to pulmonary embolism or preeclampsia. Information on placental imaging will be of little use to a radiologist seeking information on lung clots.

Yottalook has also developed a thesaurus of medical terminologies to identify synonyms and define relationships between terms. Inputting "phacomatosis," for example, will also trigger "tuberous sclerosis" and return results for all other subordinate diseases. This semantic ontology builds on existing medical ontologies, such as the Radiological Society of North America's indexing lexicon, RadLex.

Final search results can be refined further as appropriate. The option exists to restrict results to a certain modality. An automatically generated link list (Yottalinks) provides access to screened web content known to contain particularly relevant information on the search topic. For example, searching for "pulmonary nodule" brings up a link to the Fleischner Society's guidelines on follow-up and management of such findings.


  • MIRC (Radiological Society of North America)

The various case collections available on the web use different ways of storing and presenting data. This makes it impossible to query cases across all sites simultaneously. The Medical Imaging Resource Center (MIRC) offers a standardized way of structuring case reports, enabling cross-site querying. MyPacs (www. and MedPix (rad. are two well-known sites that use the MIRC format. A query from allows users to search across multiple servers, providing access to tens of thousands of cases. This makes the MIRC server community one of the largest collections of cases on the web.


  • CHORUS (Charles E. Kahn Jr., M.D., M.S.)

CHORUS (Collaborative Hypertext of Radiology) is a quick reference hypertext collection for radiologists. The documents describe diseases, findings, and their differential diagnoses as well as relevant anatomy, pathology, and physiology. Documents are interconnected by thousands of hyperlinks, allowing efficient navigation.

Many authors have contributed to the text via an anonymous peer-review process. The site consequently resembles a notebook belonging to multiple radiologists. The plain HTML-formatted pages are suitable for offline use with any basic personal digital assistant browser. Downloading links are provided on the site.

  • Wikipedia for Radiologists

"Wiki" is a Hawaiian term meaning fast. Wikipedias are online text collections that can be read, edited, and extended by anyone in the community. Some of the radiology wikis listed here started out as personal scrapbooks. They were subsequently put on the web and have been extended by committed users. The largest wikis contain more than 1000 entries.

Peer review for all wikis is provided by fellow readers, yet the quality of content is surprisingly high. Most authors provide references and links to the information they contribute.


  • Visible Human Browser (Ecole Polytechnique Fédérale de Lausanne)

visiblehuman.epfl.chThe famous U.S. National Library of Medicine Visible Human data set was captured as axial slices and stored on a server. Users wishing to view the data could browse through the 2D slices. The Visible Human Browser has now made it possible for any online user to view the data set in 3D. The result is a new way of exploring human anatomy.

  • Merck Medicus (Merck)

Merck offers free 3D anatomy models at this site. The models are precalculated from appropriate data sets and can be crudely rotated and stripped down stepwise to the bones. They are most useful for teaching purposes and for illustrating talks.

  • e-Anatomy (Campus Medica)

e-Anatomy is an interactive reference and tutorial site for human slice anatomy. Users can explore the human body via 1500 detailed, continuously labeled CT and MR images acquired in different planes. This means that every structure on every single slice is labeled, which is unique; other websites offer only labeled or unlabeled exemplary slices.

Users can cine seamlessly through the image series, much like sitting in front of a workstation. Schematic 3D anatomical drawings beside the viewing portal display slice cut lines. These drawings also assist with orientation in complex anatomical regions, such as the temporal bone. Exploring human anatomy cannot become much easier and more enjoyable than this.


  • e-MRI (Campus Medica)

The dry physics of MRI is presented in an interactive manner with many animations on e-MRI. This makes complex physics clear at a single glance and eliminates the need to read long sections of explanatory text. Experiments and quizzes provide additional entertainment. The aim is to let users have fun while learning MRI physics.

  • Pediatric Radiology (Cleveland Clinic Children's Hospital)

This site contains 80 interactive learning modules from the Cleveland Clinic Children's Hospital, covering many important aspects of pediatric radiology. Instead of simply consuming information, participants must generate their knowledge interactively. Info-r-mation is not just displayed but has to be actively gathered by clicking on images, taking quizzes, and clicking and dragging anatomical labels to the appropriate spots on images. Findings on images are highlighted by simply moving the mouse across them. This removes the need to scroll back and forth between images and textual descriptions.

  • Radiology Integrated Training Initiative (Royal College of Radiologists and Department of Health, U.K.)

The Radiology Integrated Training Initiative (R-ITI) was initiated in the U.K. as a way of increasing the number of radiology trainees without overloading the existing training infrastructure. The radiology teaching files that were subsequently produced can be accessed from the U.K. Department of Health's e-Learning for Health website.

The files provide hundreds of hours of high-quality teaching material. All content is endorsed by the Royal College of Radiologists. Access to the information is on a subscribers-only basis. Demonstrations are available for radiology departments that are interested in gaining access to the educational materials.

DR. ERTL is a radiology resident at the Zentralklinikum in Augsburg, Germany. DR. ADRIAENSEN is a radiology resident based in Amersfoort, the Netherlands.