The Digital Pathologist’s View

Can you tell us why there is a need for standardization?

About seven years ago, digital scanners were appearing and pathologists were taking an interest in what they could offer. At the same time the EURO-TELEPATH initiative was investigating how we could automate many of the manual processes in the laboratory (see sidebar, “Telepathology Networking” ). We concluded that digital imaging was something that needed developing as it could offer pathologists the opportunity to perform a complete scan of a digital slide or scans of sets of slides. However, we couldn’t recommend digital imaging without a standard. So, we set about collaborating with various international standardization organizations – DICOM, IHE and SNOMED CT – to address the need, which would hopefully and ultimately encourage the use of digital imaging.

Two main standards have been developed. DICOM Supplement 145, defines how we archive and store the very large image files. The second comes from the IHE initiative, which defines how different systems communicate with each other.

What about validation?

In the USA, the Food and Drug Administration (FDA) insists on validation studies before pathologists can use digital images for diagnosis. We don’t have this problem in Europe because we have sufficient validation studies. Here, it is the norm to perform validation studies once a lab buys a scanner and before using it for the first time. All the validation studies that I am aware of have demonstrated that these technologies are suitable for performing diagnoses.

However, we do need to perform further studies on efficiency; for example, we need to know that we are working efficiently with DICOM standard images. We also need to know the impact of using standards within our pathology workflows and that is something we are working on with the AIDPATH initiative (see sidebar “Collaborative Clout”).

How can we address the patent problem?

Patent problems are slowly being resolved. Various manufacturers were concerned about whether they could use technologies covered by DICOM because of patent concerns. Leica, for example, is allowing DICOM certified companies to use its technology, which is a step in the right direction.

I bought my first scanner about seven years ago, which used a specific file format and a specific viewer. Since then I have used several different file formats that are not interchangeable. Really, you should be able to use any format; if I can see any slide through my microscope then the same should be possible with viewing digitized images. So that’s why we need technology companies to adopt a uniform standard, which should be possible now that patent problems are being addressed. Standardized technology will enable pathologists to scan images with the confidence that in 10 years’ time they will be able to read those images without any problems.

Could you tell us about the technological hurdles?

Five years ago, we were debating whether or not this technology would be quick enough and produce high quality images. We’ve proven that it does and that it is suitable for performing on-screen diagnoses. However, the lack of standardized technology and the difficulty in implementing and using it needs to be addressed, if we expect to see wider adoption. In addition, we need to be able to provide adequate standardized training to users.

Scanning speed has reached a plateau and I doubt that it will get much faster before new technology emerges that captures the slide in one go. Feedback from the pathology community will play a big part in bringing forward the technological improvements we seek.

The AIDPATH initiative, for example, is enabling us to work with industry to drive such improvements. We are able to test new equipment and tell the manufacturers what they need to change. Also, the industry is working with university groups to improve imaging algorithms. Any successful work will be patented and distributed so that everyone will have access to standard algorithms.

What about file compatibility?

As long as we have different technologies that use proprietary file formats and specific viewing software, we will always have compatibility problems. I don’t think we should compress images using proprietary file formats because we may need to refer to them in the future. It’s so important that new technologies take backwards compatibility into account.

Are there any issues with LIS products?

Yes. The LIS companies are not evolving their products quickly enough to keep up with other areas of the market. Picture archiving and communication systems (PACS) and scanner companies are moving ahead but LIS vendors are trailing behind. It takes about five years to develop a LIS to ensure that a product is robust and reliable; the vendors need to speed up development. That said, I don’t think we need other companies to get involved and flood the market with competing products – that’s not the way to go.

We do try to involve LIS companies in AIDPATH by defining a project in which they can help. For example, seven years ago, we told the LIS vendors that they ought to produce web-based systems – at that time, there were no web-based solutions in Spain. We have learned that vendors only tend to respond when you ask them for something new; we need them to be more proactive.

Do you think that digital technology integration is moving too slowly?

On one hand, it is slower than I’d like it to be, but on the other, I think it is going as fast as the technology allows it to. For example, until about three years ago, scanners were slow and there were too many out of focus areas in an image. That has changed, but scanners remain difficult to implement, the images are very large and a PACS struggles with storing everything we’d like to archive.

Therefore, we need to get everything right with the technology so that we aren’t worrying about it and so that we can concentrate on doing our jobs as pathologists.

What has changed since you began using digital pathology?

We are now much more selective about scanning. We select cases that could be problematic, need quantitative analysis and image analysis to measure the width of infiltration of a melanoma, for example. We also use it for making second opinions. This amounts to about 15 to 20 percent of my cases. Only when we reach the point that everything is standardized will I decide it is time to digitize everything.

Despite the challenges, do you feel digital pathology has benefitted your practice?

Yes. I actually think my digital pathology images are becoming as popular as X-ray film!

The main benefit is that I am more confident with image analysis and measuring. For example, I can use it to measure Ki-67 protein levels or the size of melanomas. It also reduces the time my colleagues need to wait for a second opinion – I might get the request first thing, and I can email my assessment to them in the same morning.

In addition, I am able to share my images with the other medical specialists in the hospital – I really feel this has made pathology more respected by other departments. You can see it in the faces of the dermatologist and hematologists, for example; they are impressed when you show them digital slides. They really appreciate being able to see whether their markers are correctly excised in a tumor or whether morphological features in a leukemia case correspond with those in their own specimens, for example. Collaborating with other specialists is much easier with digital pathology.

Where do you see your lab in 10 years?

In 10 years’ time, I think I’ll be fully digital. That’s not to say I’ll be getting rid of my microscope! I’ll still

What is the general attitude to digital pathology in Europe?

I would say that the attitude is changing. Pathologists now view digital pathology more positively. They see that the technology is much more accessible and affordable; however, it needs to be easier to use and the pathology workflow it has introduced needs simplifying. We’ve been working on some of these issues at the hospital and have completely remodeled our processes, which has improved our workflow and made it easier to manage.

And in the next decade?

I think it will take longer than 10 years to see digital pathology fully embraced across Europe. Pathologists in private labs may not see the need, for example. However, I expect more hospitals to be using it. Also, as time passes, the pathology community will have access to the results of numerous validation studies, which should encourage wider adoption.

What do you say to the cynics?

All patients need the same opportunity and no one should get in the way of progress. Doing it your own way may not be the best way and, in the future, the best diagnosis will come with advances in digital pathology. You have to think of the patient and how to provide the best service and outcome for th