(Bio) Banking on Pathology’s Future

Curating Pathology’s Future

Biobanks are vital to biomedical research and clinical diagnostics, but we have a great deal of work to do before we can realize their true potential

The word “biobank” first began to take off in print in the mid-2000s. More than 15 years ago, an Internet search would have returned almost nothing; today, there are over a million results. It’s a very short existence for a concept that I believe is vital to modern pathology – both in research and in the clinic.

The biobanking initiative first came from the Organization for Economic Cooperation and Development (OECD), which not only started advocating for the importance of biobanks, but also insisted on the need to have an accreditation system. In the years following the proposal, the governments of various countries began funding research infrastructures for biobank operations. One such country was France, where I began my own career in the very first autonomous biobank to get ISO certification in 2005. And by 2008, the French government had developed and begun applying a national certification standard for biobanks. It’s approximately equivalent to ISO 9001: a basic quality management system, but nothing more. But professional biobanks – those whose sole purpose is sample collection, processing and management – should be held to a higher standard.

The preanalytical problem

The most important aspects of a biobank are consistency and quality. When researchers come to us and say, “I need 30 lung cancer samples,” we ask, “Okay, what kind of lung cancer? What kind of sample?” But most of them are not pathologists; they don’t know the different histological types or sample preservation options, so they just ask us for “lung cancer.” We have to educate basic and translational scientists to understand what they need in greater detail – because it’s difficult to provide a professional service when the clients can’t clearly articulate their requirements.

Sample characterization – clinical, pathological, immunohistochemical and preanalytical – is a large part of what we provide. Most of that may seem obvious but, until now, preanalytical characterization has been almost completely neglected despite its importance. We can’t just forget to take into account the potential impact of factors such as cold ischemia time, fixative type, or even storage temperature on the downstream results; these are all critical elements that professional biobanks should track – and, fortunately, most of them do. As a result, when asked for samples, the biobank can select them according to their suitability – and the researchers can then specify in their publications where their samples originated and how they were handled. Without that information, it’s easy to introduce invisible bias into the work – and then researchers are surprised when their findings cannot be reproduced!

Along with the Biospecimen Science Working Group at the International Society for Biological and Environmental Repositories (ISBER), we have developed a tool called SPREC – the Standard PREanalytical Code (1) – an evolving seven-element code that summarizes the nature of the sample and its history. For instance, the seven elements of a tissue sample SPREC are:

• specimen type,
• collection type,
• warm ischemia time,
• cold ischemia time,
• fixation/stabilization type,
• fixation time, and
• storage conditions.

So your specimen might carry the code TIS-BPS-N-E-NBF-G-P. That would make it a solid tissue specimen (TIS), collected via biopsy (BPS), with warm ischemia time not applicable (N), cold ischemia time of 30–60 minutes (E), fixed in neutral-buffered formalin (NBF) for 48–72 hours (G) and stored at room temperature in a paraffin block (P). Don’t have time to code all your samples by hand? A publicly available tool, the SPRECALC, will automatically generate the codes – and there’s even a second tool to convert them into barcodes for labeling.

Controlling quality

One major source of error in biobanking is poor annotation. Most clinical and pathological annotations come from medical records that lack standardized language and, on top of that, it’s not uncommon for them to be transcribed inaccurately. The other significant error source is the quality of the samples themselves; either the preanalytics aren’t accurately documented or quality control tests haven’t been run – or both.

Almost all of our existing samples suffer from the first problem. If you went into the average biobank today and tried to annotate its samples with SPREC, 90 percent of the time, you would simply write TIS-SRG-X-X-NBF-X-P, because some information was never recorded. Unfortunately, there’s no way to fix that; all we can do is ensure that protocols are documented going forward. But we can solve the second problem – even if you don’t know how samples were collected or processed, you can still apply quality control tests to them, or to their derivatives, and use that information to stratify them into quality categories. For example, you might extract DNA, perform a multiplex PCR, and see to what extent the genetic material is still amplifiable. Of course, that brings us to a further need: the development and validation of such quality control assays – but, in my opinion, that is the only solution that can allow us to use with confidence the millions of legacy samples stored in biobanks and pathology labs around the world (2).

Teaching and training

We are constantly involved in spreading the word about biobanking – why it’s necessary, who can benefit, how it’s done... When I worked at a university hospital in France, we organized training for clinician-researchers; now that I’m in Luxembourg, the work continues. We have developed a university certificate on biobanking that is targeted more at biobankers themselves, but we often see researchers and clinicians signing up because they want to learn more. We organize seminars at hospitals and research institutes to educate the faculty, and they are always very surprised when we explain to them, “You ask us for lung tissue – did you know that there are different histological types? Did you know that a sample with 10 percent tumor content will give you completely different results in your analyses than one with 80 percent tumor content?” It’s a revelation to them. Clearly, there’s a lot of work to be done!

For many years, I have been saying that professional biobankers need to submit abstracts to scientific society meetings. After all, our work is applicable to every area of biomedical science: immunology, cardiology, oncology, infectious diseases, hematology, and the list goes on! So any biobanker can assemble an abstract that addresses a few key questions:

• What are biobanks?
• What kind of work do they do?
• Why are they important?
• How can they help with your field of study?

We don’t do nearly enough of this kind of outreach work. In my opinion, we should be at all of the major scientific meetings. We need to make the research community aware of our services and help them to understand why they need us – and we need them.

Enabling access

The biggest obstacle to bringing researchers and biobanks together is the question of supply versus demand. If you are a researcher who needs samples and associated data and you try to request them from a biobank, you will almost never find what you are looking for. Why? Because the needs of each research project are so specific that often, even big biobanks won’t have what you need. In fact, this is a subject of much discussion in the biobanking community: what is the best way to operate? Should we operate on stock and try to build a huge library of samples so that we can provide as many different options as possible? Or should we operate on project-based demand? At the moment, most biobanks follow the first model – but experience shows that it is neither the best nor the most efficient method. Much of the time, researchers don’t have a use for what we have in stock, whereas we cannot provide them with what they do need.

I think the best approach is to switch to prospective, project-driven collections – but of course, for this you need professional biobanks with all of the necessary infrastructure in place to begin collecting immediately. If you have to wait a year while you assemble an ethics committee and establish everything you need from an administrative point of view, your clients will go elsewhere – or won’t be able to conduct their research at all. Professional biobanks already have the administrative and the quality management systems required. You send them your request; they begin collecting in a consistent and controlled manner; and after only a few months, they deliver exactly what you need.

I don’t know of any biobanks that currently work to this model, but it is something we are trying to develop. The first step is networking. You need to be in small, bottom-up networks to provide samples efficiently; if you don’t have what a client needs, it’s possible that another biobank does, which prevents the need to start from scratch. This kind of functional networking already exists in a few countries – in Spain and the United States, for example, and there’s a government initiative to establish something in Germany as well – but it’s lacking in most places. Even in those that claim to have such networks, it’s often more like a list or catalog of existing biobanks, rather than a true relationship between them.

At IBBL, we believe in “trusted biobank networks” – but it will take time to build them. In the interim, we advise potential users to locate biobanks that are serious and professional to help them get the samples they need.

It’s clear that our work as professional biobankers is just beginning – not only in our sample procurement and preservation work itself, but also as documenters, educators and promoters. The future of pathology lies in biobanking, and it’s up to us to step forward and make these services the best they can be.