Indulge in the fascinating world of

Diagnostic Imaging 2.0: Specific search engines August 2, 2008

If you read John Battelle’s book titled The Search: how Google and its rivals rewrote the rules of business and transformed our culture, it is clear that without search engines the World Wide Web and the experience of its users would not be the same as it is today.

Companies such as Goto.com and Altavista.com were among the first search engines, but today Google, Yahoo and several others are the ones we turn to when looking for information on the Web. The problem with these engines is that they are not too good at accommodating the work of diagnostic imaging professionals. They cover too much virtual space, therefore their results are enormous, but most of the time they are not too relevant.

Here is a list of diagnostic imaging-specific search engines. Some are a result of small projects, while others enjoy the comfort of large support groups. They are all different, and they all work a little bit different. You have to find the one that suits your needs best.

1. As the name implies this one is a search engine fine tuned for Radiology searches, and indeed it is. RadiologySearch.net was developed by Dr. Roland Talanow, a radiology resident at the Cleveland Clinic in Ohio. His goal was to develop an engine that turns up more information from radiology-specific peer-reviewed sources while filters non peer-reviewed content. As you can see on the image you can select from 13 options before hitting the “search” button. You can search in over a 1000 peer-reviewed websites, in over 150 scientific journals, and among thousands of images, radiology cases, teaching files, books, lectures, videos. The links to all the relevant journals and societies may be found under the search area, plus all the latest news and publications appear towards the bottom of the page. Therefore this website goes beyond the traditional search page role, and resembles more to a radiology portal. Registered users may also save their individual searches for quick access, and repeated search.

Further reading:

• RSNA article (pdf)
• AuntMinnie article (pdf)

2. Thursday, the 21st of December, 2006 was the release date of Yottalook, a radiology-specific search engine developed and actively maintained by four radiologists – Woojin Kim MD, Khan M. Siddiqui MD, William Boonn MD and Nabile Safdar MD – and powered by iVirtuoso’s algorithm. Yottalook offers four search engines: Yottalook Web, Yottalook Images, Yottalook Anatomy and Yottalook Book.

Yottalook Web has been designed to search online radiology sources only. The “Refine results” window offers a list of words every time you initiate a new search, and tries to help you narrow down your destination quicker.

Yottalook Images relies on peer-reviewed materials only, and works from a database of 700,000 images. Search results are presented in either “detailed” or “thumbnails only” mode. It supports PicLens, so that you can view all the images full screen on a 3D wall!

Yottalook Anatomy helps you find anatomical information and images with higher efficiency.

Yottalook Books is a customized and Yottalook optimized Google Book Search. It allows you to read full-text radiology books online.

According to Wikipedia Yotta is “an SI prefix in the SI (system of units) denoting 10²⁴ or 1 000 000 000 000 000 000 000 000.” That is a lot of zeros, but a very catching name for a search engine. Could a search engine give you that many results? Find out for yourself at Yottalook.com!

3. It will not find you gold, but it will dig up just about anything besides that: ARRS GoldMiner® was created for finding images in a select group of peer-reviewed journals and now it has access to 196,130 images published in 249 radiology journals.

After initiating a search, you have the option to further narrow your results by modality, patient age and sex. What is interesting about GoldMiner® is that it “understands” medical vocabulary! For example it “knows” that gallstones and “cholelithiasis” mean the same thing by incorporating the UMLS Metathesaurus and the Medical Subject Heading (MeSH) into its technology. So, if you enter “kidney stones” it will look for images that are labeled as “renal calculi” as well, giving you an extended list of options.

GoldMiner® Global beta allows users to search Goldminer’s English-language database in their own language. It manages to do so by translating the words the user types into the search box through the National Library of Medicine. It is available in German, Italian, French, Spanish, Russian, Portuguese and some other languages.

Read the journal article on ARRS GoldMiner here.

4. “A collaborative social search engine.” That is the Wikipedia definition of a swicki. The Radiology Search Engine swicki was created by Gerard Deib and is further improved by its user community.

Swickis are easy to edit search engines, that rely heavily on group feedback. When you set up a swicki, you need to give it a name, choose a topic, and start “training” your search engine by specifying a couple websites, blogs, and wikis as starting points.

Then you need to add a couple words to the tag cloud, which is like a collection of the most popular key words. The more often a key word is used, the bigger it gets in the tag cloud, and the less frequently it is used, the smaller is gets. The editor of the swicki may add or delete words from the tag cloud at any time, whenever she feels it is necessary. Here is the current tag cloud of Radiology Search Engine:

You can see that at the time of this snapshot most people were looking for information regarding: mri breast cancer, myocardial spect images, oligodenrogliomas, and some others. The interesting aspect of the swicki search results is that each individual result may be voted on, so that they either move up or down the list:

On the above image you can see that 4 people though that the article from radiologyinfo.org was the best for “MRI neck anatomy visualized.”

When your swicki is published, anybody may be able to use it just like a standard search engine, and anybody can copy the source code to their blog or website so that your swicki will become available from their site as well, and they all can contribute to the continued growth of the swicki.

According to the company – Eurekster – behind the swicki technology, there are over 100,000 swickis already, and this number is continually growing. If you are ready, you may begin building your new swicki here.

5. And last, but not least, here is SEEKRadiology. It is still in beta at the mean time, and you may find it here:

There are all sorts of search engines that are pertinent to the field of diagnostic imaging such as Goldminer, RadiologySearch and YottaLook. But these are all on a different server with different URL’s so if you want to use all of them at once, you have to have 3 or more windows or tabs open. TomographyBlogSearch is bound to solve just that problem. It is not a search engine in itself, but it is a clever use of some old HTML and new Java Script to ease your search needs.

It is very simple to use: just type your key words in the search box, and then click on the name of the search engine you wish to use. There is no need to type in your key words again if you wish to switch between the search engines. Your results will open in new windows, so that you can go back to your original search whenever you need.

SEEKRadiology is being developed by Péter, Herczeg and András Székely. Here is a sneak peak at SeekRadiology V1.0 coming very soon:

I hope you have enjoyed the first edition of Diagnostic Imaging 2.0. We are going to take a closer look at social communities for diagnostic imaging professionals next.

– Andras