Pathology 4.0 – inveox automates laboratory processes

Mix-ups, contamination and sample loss – most errors in histopathology happen when specimen are received. Countless samples arrive daily at the laboratory, while the sample entry process is very monotonous. As a result, the work is inefficient. The start-up company inveox has now developed a system that automates the processes in the pathology laboratory, thus making them more efficient.

In this MEDICA-tradefair.com interview, Maria Driesel describes the challenges of pathology laboratory processes, explains how an automated system can simplify these procedures and defines the role machine learning should play in all this.

Ms. Driesel, what is the current pathology laboratory setup?

Maria Driesel: The majority of the processes in a histopathology laboratory are handled manually. Typically, a simple container with a screw lid accompanied by a handwritten test request is used to transport the biopsies. That means, the ordering physician first collects the tissue sample, places it in the screw cap container that contains formalin – a stabilizing solution – and identifies the patient and analysis request on a handwritten form. All this is then sent to the lab. The problem is that labs receive hundreds or even thousands of samples each day. All of them must be registered and prepared. To do this, the data must first be collected and typed up. Subsequently, the samples are sorted, assigned and repackaged into a so-called biopsy cassette, which in turn requires another lettering. The samples cannot be processed in the laboratory until these steps have been completed. This is a very monotonous and time-consuming process due to the vast number of samples. Oftentimes, this can result in mix-ups, contamination or even the loss of samples. According to experts, these types of irregularities occur in one to fifteen percent of cases and – in the worst-case scenario – can lead to misdiagnosis or the wrong medical treatment.

What's more, thanks to dynamic medical advancements, diagnostics is becoming increasingly important, while more and more tests are being introduced. Consequently, more samples are being collected. Meanwhile, there is an acute talent shortage in this field. That’s why laboratories desperately need more efficient solutions to manage the increasing amount of work and ensure proper diagnostic testing.

You have developed a system that promises to deliver this type of efficiency. How does it work?

Driesel: Our system automates the lab sample entry process in histopathology laboratories. It is made up of three components. The first component is a novel sample container for biopsy samples that combines the transport container and biopsy cassette all-in-one. This avoids the frequent and error-prone repackaging process. The next component is the automation platform for the lab sample entry and the device that is ultimately located in the laboratory. It manages the complete "goods receiving procedure" of the histopathology laboratory. This includes sample identification, quality management, repackaging and labeling processes etc. The third component is a web-based software that connects the ordering physician with the pathologist.

Our technology achieves efficiency in a variety of ways. For example, thanks to a barcode, we can safely and accurately identify the sample at any time. What's more, we take pictures of the sample, thus setting a new quality management standard. By using the computer vision module and image recognition, it also opens the door to future opportunities as it pertains to machine learning for example. Currently, people largely appreciate our automation tool for being able to efficiently process a large number of samples.

How do the automation platform and database collaborate?

Driesel: This is done via a digital examination request. The database transmits the information provided by the ordering physician to the laboratory. The moment we identify the sample and read the code, the tool sends the information to our database, indicating that the sample has arrived safely. Via our database, we then electronically transfer the data we previously received from the doctor directly to the laboratory's information system with the matching case number – all of it in real-time. Our tool also delivers the image data – the so-called visual ID – by taking photos of the sample.

How does your system change the way a tissue sample is examined?

Driesel: At this juncture, it is important to point out how the pre-analytical phase in a histopathology laboratory actually works. It is made up of several steps. Our system covers the first and most critical portion. Once the sample arrives at the lab, it is first repackaged and registered. Then it is dehydrated, embedded into paraffin wax and cut. This thinly cut section is then placed on a slide, stained and covered, at which point the pathologist is able to examine the tissue under the microscope.

Our system automates the processes involved in sample registration and preparation. Essentially, the analysis itself doesn't take place in our automation device , only the preparation portion. We ensure that the data matches the sample. The tool collects the data, removes the formalin, while is also labels and takes photos of the sample. In doing so, the pathologist obtains additional information to make a diagnosis. Next to a more detailed macroscopic examination, which delivers accurate information pertaining to the sample's size, shape, and surface structure, we also utilize machine learning and study the possibilities of artificial intelligence.

You just mentioned machine learning. In what way will it someday replace manual labor and human beings?

Driesel: In our view, artificial intelligence still has a lot of untapped potential and will become increasingly important in the coming years. As we all know, human-machine interaction is already playing a vital role. We believe it is crucial to use machines where they make sense. They can take on monotonous and even dangerous tasks and ease human workload. Of course, occupational health and safety is also an important aspect in all this. We have to keep in mind that medical laboratory technicians handle formalin throughout the day and are thus exposed to potentially harmful and unhealthy substances. When it comes to these types of process steps, the use of a machine makes sense and frees up time laboratory technicians can then dedicate to entirely different crucial tasks that require unique human skill and knowledge. The ultimate goal is holistic, connected diagnostics where humans and machines can collaborate in perfect harmony.