Virtual slides in peer reviewed, open access medical publication

Open access publication has been developed after successful implementation of scientific electronic publication. In pathology, the Electronic Journal of Pathology and Histology (Elec J Pathol Histol) was the first scientific peer reviewed journal in pathology that has been solely electronically distributed to our knowledge [20]. It has been implemented in 1995 and was distributed via floppy-discs worldwide. Internet access and other electronic distribution media had not been available at that time. The main characteristics of Elec J Pathol Histol are displayed in <Figure 1>. Each floppy-disc was distributed with specific "reader" software, which also allowed the compilation and execution of programs. Tests of so - called interactive publication were also performed. These test allowed the reader to add own data to an already existing article, and to publish the "extended" version by including the authors of the previous article, after they had agreed. All images were displayed in a .pcx format. Although the peer reviewed and published articles had been cited worldwide, the application for inclusion into Reuter's citation index failed several times because a solely electronic distribution of scientific articles was not appreciated in the 1990s [21].

The Elec J Pathol Histol was redesigned in HTLM format and distributed on a CD in 2000. However, several advantages of the DOS environment had to be given up (executable programs, interactive publication), and the production of the Elec J Pathol Histol was terminated in July 2000. The front page of last edition is depicted in <Figure 2>.

The Elec J Pathol Histol has been brought back to live in 2006 once the internet had been established and acknowledged as a useful and flexible communication medium. The new journal was named Journal of Diagnostic Pathology (Diagn Pathol), <http://​www.​diagnosticpathol​ogy.​org>. The potential benefits of the internet embedding were analyzed, and corresponding mandatory formal changes were undertaken. These included:

Open access publication. The authors have to pay for the publication of their article. In compensation, they still inhabit the publication rights etc. of their article, and the readers have access to all published articles for free.

Authors who are working in developing countries can apply for waivers to publish their scientific results for free or at a remarkably reduced price.

The principle formal organization remains related to the conventional style of publication; i.e., Diagn Pathol still possesses its own "front page" and is not just a domain which might be embedded in another front page such as "medicine".

Access to and reading of the published articles is documented as well as the country of the reader, and statistically analyzed.

Reviewers are chosen from the editorial board and contemporary from authors who have published articles that are related to the content under request. These authors are chosen from articles cited in the National Institute of Health (NIH) library (PubMed).

The number of included color images, tables, or drawing etc. is not limited and free of additional charges.

All published articles are provided with an ID number (DOI) and indexed for interactive search of readers etc.

The introduction and establishment of Diagn Pathol lasted for about six months. The total number of published articles rose from 46 in 2006 to 83 in 2010, and will probably extent the level of 120 in 2011. The rejection rate increased from 30% in 2006 to > 50% in 2011. The journal possesses about 1,200 registered readers in 2011. Most articles are assessed by more than 500 readers within their first year of publication.

Diagn Pathol was included in the Citation Index (Reuter's) in 2010. It was calculated an impact factor of 1.39 for the year 2010.

Virtual slides (VS) are the digital representation of completely digitized glass slides [2, 13]. VS have been developed at the beginning of this century. Image acquisition machines (scanners) are commercially available from about 10 different companies (e.g. 3DHistech, Aperio, General Electric, Leica, Olympus, Phillips, and others). They are now in daily use for interdisciplinary conferences, teaching of medical students, and postgraduate education in most of the Medical Universities located in Western Europe and the United States. Some of the larger institutes of pathology are equipped with scanners that are embedded in the daily routine workflow of diagnosis, and several other institutes of pathology are investigating to replace the conventional workflow by virtual microscopy (VM) [2, 6, 10].

VS are comparable with conventional glass slides in image quality that has been demonstrated by several investigations [2, 6, 10]. The specific features of VM are summarized in <Figure 3>. In addition to the features of conventional microscopy they include contemporary viewing of different regions of interest (ROI) or stains, interactive labeling, automated scoring, and other electronic assistance.

Including VS in an open access peer reviewed scientific journal includes several advantages for the authors, readers, and the publisher <Figure 4>:

The authors can demonstrate the morphological findings of whole microscopic image. The ROI can be analyzed independent from the authors' view.

The readers are assured for the originality of the findings, and can proof their own strategy for ROI selection.

The publisher increases the attraction of the included articles, and opens the journal for additional perspectives such as implementation of a repository or specific case collection, etc.

Constraints can be seen in the mandatory logistics:

The authors have to submit the original glass slides to a selected institution that scans the glass slides, because an official VS standard does not exist at present. Most of the companies have developed own specific viewers. A conversation of their own image format into a more general and open standard such as jp2000 is difficult if not impossible without knowing the individual image structure [22‐27].

VS are large images of usually 2 - 3 GB in size. They usually require a specific viewer and a related image data bank that can handle these images. Thus, VS have to be organized in an electronic system that has to be separated from that of the published articles. In addition, the publication of still images should not be affected by VS.

After several trials, we decided to provide each article with dummy links that are kept empty if VS are not included in the article. These links are activated and connected to the corresponding images if VS are included.

The chosen solution possesses logistic and content related advantages: The publication procedure of an article is not affected or delayed by VS as it is completely disconnected from the VS production, and VS do not require a DOI number (Digital Object Identifyer for article identification). Furthermore, VS can be included into each article at any time; even years after its publication by replacing the dummy link with an active one, and the reader has access to VS separately from included still images. The logistic frame is depicted in <Figure 5>. The reviewers are not informed about potential inclusion of VS in order to completely separate this issue from scientific considerations. Thus, VS do not involve, delay or promote the articles' publication.

The preparations of the described procedure lasted for about 6 months. The authors of newly submitted articles were informed via email about the VS technology and asked to submit the corresponding glass slides for digitalization. A summary of the authors' responses, number of included VS, and published articles is shown in <Figure 6 >. About 50% of requested authors submitted glass slides for VS publication, most of them are working in Asia. There was no time delay in peer reviewing and production process in comparison to articles that do not contain VS. VS seem to promote the interest of readers and the journals reputation as the number of submissions increased remarkably after publication of VS. However, also other reasons might play a significant role in this manner, such as fast review process, the citation index, or increased number of subscribers. The quality of all published VS was judged good or even very good. The access to and display of VS naturally depends on the network and included servers of the readers internet. The navigation and VM control is fast and reliable as only the viewing partitions of VS have to be downloaded.

There are practical and theoretical reasons: One aim in electronic publication is to maintain a short review/production time which could be delayed in combining VS and review. In addition, the selection of published still images depends upon the author. Any errors in selecting additional correct ROI can be judged by the reader if VS are published too.

From the theoretical point of view, the diagnostic judgment in tissue based diagnosis can be distinguished in a) selection of the ROI, and b) the diagnostic statement [12, 28‐31]. In surgical pathology both algorithms are usually performed in only one combined step. Young colleagues are trained to perform the final diagnosis by viewing pre-selected areas (ROI) which are displayed in textbooks or atlases [2]. Publishing VS and still images independently from the review procedure separates these two steps, which is of additional educational significance.

The main constraint of the described method still remains in logistic "hardware" problems, especially in submission of requested glass slides. The mailing costs have to be covered by the authors. The glass slides have to be sent to an "image acquisition center" which controls the image quality. It is responsible to accurately document and handle VS. After several different trials to accurately link included still images with VS we decided to separately publish still images and VS. The separation of the "conventional" publication procedure from VS inclusion was favored by the production team. It could maintain the production speed. VS can be included at any time after publication. No intervention is needed. The provided links can be transferred to different servers at any time without involving the publication procedure of text, included images, and references. Thus, the journal diagnostic pathology has been "opened" to a new innovative distribution of image content information that offers new regions for additional applications.

Most of published VS are derived from case reports which usually describe rare diseases displaying educational significance. They are arranged in an appropriate data bank which will be transformed in an adequate repository in the next step. Certainly, the authors have to agree if the images will be included in such an electronic image retrieval system. Such a system will offer images of rare and educational significance combined with extensive clinical data such as the patient's history and development of the disease.

An additional potential application is related to quantitative object related measurements of VS [7, 15]. They can be performed on a broad variety of stains including conventional and fluorescent immunohistochemistry using open access systems such as EAMUS™ or equivalent systems [7, 15]. The reader can use these systems to select ROI from VS and to perform content related measurement.

In summary, the publication of VS in an open access peer reviewed scientific journal has been accepted by a broad community of pathologists and colleagues working in related medical and scientific fields shortly after its implementation. It significantly increases the reader's tools that can only be provided in an electronic communication environment such as individual selection of areas of interest, interactive measurements of specific objects, or image content information related selection of case reports for tissue based diagnosis and training.