“We believe the ability to work with just two drops of blood, plus the fast time-to-results and the reduction in operator overhead provided by our FDA-cleared moderate-complexity analyzer, will be highly attractive to providers.”
—Yossi Pollak, CEO, Sight Diagnostics, Ltd.
CEO SUMMARY: For 70 years, the Coulter Principle has been a bedrock technology in hematology. Now this seven-decades-old technology has a challenger. An Israeli company obtained clearance from the Food and Drug Administration (FDA) to market a new CBC instrument as a moderately-complex CLIA device. The system needs only two drops of blood placed on a specially-designed cartridge. The cartridge is inserted in a small analyzer, which uses machine imaging and artificial intelligence to produce a five-part CBC with 19 parameters and signature flagging capabilities.
EDITOR’S NOTE: Digital imaging, digital image analysis, and artificial intelligence/machine learning are technologies expected to disrupt anatomic pathology and its century-long reliance on the light microscope. Now Sight Diagnostics, Ltd., a young diagnostic company headquartered in Tel Aviv, Israel, is adapting these technologies to clinical laboratory testing—specifically hematology and complete blood counts (CBC). The company manufactures a diagnostic solution that has the potential to be disruptive in multiple ways. The following interview with Yossi Pollack, Co-Founder and CEO of SightDx, was conducted by Robert L. Michel and Donna Marie Pocius of The Dark Report.
EDITOR: Yossi, let’s set the scene for our readers by explaining that, in November, 2019, the federal Food and Drug Administration (FDA) issued a 501(k) clearance to your Sight OLO CBC analyzer for moderately-complex testing in CLIA-compliant facilities. This test produces five-part CBC results with 19 parameters and signature flagging capabilities. What differentiates the Sight CBC testing system is that it is engineered around a radically different mix of technologies, compared to most hematology systems used in clinical laboratories today. Could you explain more about why your OLO system is different from most hematology systems used by clinical laboratories today?
POLLAK: In simplest terms, the system takes two drops of blood—collected either by fingerstick or venipuncture—and turns that sample into a set of high-resolution digital images. The images are analyzed automatically using our algorithms to accurately perform a complete blood count (CBC)—to count the patient’s red and white blood cells, platelets, etc., including a white blood cell differential. Our OLO system can analyze 19 different blood parameters in minutes.
EDITOR: How large is your OLO system?
POLLAK: The system is about the size of a toaster oven—less than one foot per side—weighs 22 pounds, and can be placed on a bench, as well as in many other near-patient settings.
EDITOR: It is important for lab administrators and clinical pathologists to also understand that it is not just the small size that sets your CBC instrument apart from existing hematology systems. It does not count blood cells as they move through a gate, for example. Rather, it uses digital imaging and artificial intelligence (AI) in novel ways, correct?
POLLAK: Yes. Our OLO CBC analyzer essentially digitizes blood samples. We have a sample preparation method that allows us to automatically create a monolayer of cells within a self-contained cartridge. Next, this specimen cartridge is inserted into the analyzer and over 1,000 images are captured within 10 minutes. Then, in real time, the instrument analyzes these 1,000 images to produce the CBC results.
EDITOR: It this the first diagnostic for clinical purposes that you created?
POLLAK: Sight previously developed a test for malaria, which we launched in 2015. As laboratory scientists know, there is no malaria testing solution that is sufficiently inexpensive, rapid, and highly accurate. Yet it is a huge market globally: data from the World Health Organization (WHO) shows that about 500 million malaria tests are performed each year. These diagnostic tests are primarily microscopy and rapid diagnostic tests (RDTs). Our team saw this as an opportunity to apply new AI technologies that were being developed in domains such as self-driving cars and put them to use in clinical diagnostics.
EDITOR: That would seem to be quite a leap—from self-driving cars to a diagnostic test for malaria. How did this happen?
POLLAK: It started back in 2010. Some of us had worked at Mobileye, an Israeli-based company using digital images, digital image analysis, and artificial intelligence to guide self-driving cars. We looked outside this field to see how these same technologies could be used in novel ways. We saw a way to use machine vision—already capable of recognizing pedestrians, crosswalks, and traffic signals—and train it to recognize biomarkers in the blood.
EDITOR: What caused you to see a way to link technology that supports self-driving cars with some type of diagnostic test for use in patient care?
POLLAK: Imagine, for a moment, how much training a human needs to drive a car. Typically, an individual can become an expert after maybe 20 hours of actual driving. By contrast, it requires years of training for someone to achieve the skills needed for clinical-grade microscopy. That caused us to think that we might be able to adapt these technologies in malaria testing specifically to replace the human eye for the analyses. There was the potential to greatly reduce the cost of a malaria test while at the same time producing a fast time-to-result that would enable treatment to begin immediately at remote clinics in developing nations. With this type of solution, even I could be “behind the microscope” and use this system to produce an accurate answer.
EDITOR: When did you introduce this malaria test?
POLLAK: That happened in 2015, after five years of development. The test is marketed as the Parasight Malaria Detection Platform.
EDITOR: In researching your company for this interview, I noticed that just one year later, in 2016, your company entered into an agreement with Becton Dickinson for BD to distribute this test system in India.
POLLAK: Yes. The market numbers in India are huge, because almost 900,000 cases of malaria are diagnosed in India each year and 128 million tests are performed. BD’s interest in distributing this test system in India was an important confirmation of our malaria test’s accuracy and reliability at that time. In fact, we’ve now sold over one million of our Parasite malaria tests worldwide.
EDITOR: This is a credible track record for your digital imaging and artificial intelligence (AI) technologies in their use to test for malaria.
POLLAK: All this work also helped us better appreciate diagnostic accuracy because the sensitivity and specificity of a malaria test needs to be at very high levels to avoid the cost of following up on false positives, as well as the negative impact on patient outcomes from false negatives.
EDITOR: What came next after the introduction of the malaria test?
POLLAK: In 2016, we started to consider other types of blood tests. Our technology is a platform technology—meaning it can be adapted for a wide range of clinical and biological testing. And remember, for malaria, we were already creating a monolayer of human blood cells, digitizing the images, and analyzing those images with AI technologies.
EDITOR: So, you looked at the types of clinical laboratory tests that use whole blood and microscopy. Is that a good guess?
POLLAK: That’s on target. Before the days of Coulter counters, labs would perform blood counts by creating a slide that presents the blood cells for microscopy, and even today, 5% to 20% of CBC samples are sent to reflex testing that is based on microscopy. Our technology platform was already doing all of these actions for the malaria test.
EDITOR: However, in hematology, there are many reputable, highly-automated systems already delivering speedy, accurate CBC results at a low cost. Why did Sight Diagnostics believe it could successfully compete in this market?
POLLAK: Certainly there are many robust hematology systems in use globally. These systems have something in common. The highest-throughput instruments are generally large, complex, and most either utilize resistive pulse sensing—the Coulter Principle—or flow cytometry to assess the blood specimen. The Coulter Principle dates back to the early 1950s and is now 70 years old. We saw the opportunity to bring new technology to this market and also to address an underserved segment of the clinical market for blood testing.
EDITOR: Please explain, particularly about what you see as the underserved segment of the clinical market.
POLLAK: These hematology systems are commonly used in larger clinical laboratories. If your lab runs hundreds or thousands of CBC tests per day, then these hematology systems are efficient and cost-effective. But this model of CBC testing comes with delays in reporting results because specimens must be transported to the large central laboratories that operate big hematology systems. We saw the opportunity to move CBC testing closer to the patient to produce faster results.
EDITOR: Does this mean tailoring your OSO CBC analyzer to meet the needs of smaller labs?
POLLAK: Yes. We identified a huge gap in small and medium test volume settings—a need for on-site CBC testing that was closer to the patients and where the specimens didn’t need to be sent to one central lab.
EDITOR: How did you identify this unmet need?
POLLAK: We conducted interviews with doctors. We recognized that if technology was available that provided them with critical information during the patient visit, or otherwise within minutes of ordering the test, treatment could start faster. Armed with those insights, we worked from 2015 to 2019 to modify our platform to perform complete blood counts. It was in December, 2019, that the FDA issued clearance for this device for moderately-complex settings.
EDITOR: What you are talking about is point-of-care testing done in the doctor’s office and other near-patient settings.
POLLAK: We are targeting various near-patient settings today, and we plan to enter the point-of-care and doctors’ offices down the road, subject to additional regulatory approval. It is our strong belief that the key to these settings is to avoid any compromise in the quality of the result of running the CBC on OLO locally—as compared to sending it to a large laboratory. From day one we said, “we are going to replicate the best in class.” This is reflected in OLO’s design and in our choice of predicate device for the studies that we submitted to the FDA.
EDITOR: Could you explain your strategy for use of a two-blood-drop specimen to perform a CBC?
POLLAK: Not only is the Sight OLO both robust and compact, but it also allows—for the first time—a CBC test directly from a fingerpick of blood (as well as from a venipuncture). We expect this to prove to be an easier and more welcomed collection process for the patients.
EDITOR: Can the option of using either a finger-stick specimen or a venipuncture specimen be a benefit to the provider?
POLLAK: We think so. The ability to use fingerprick samples can avoid the time needed for scheduling and performing phlebotomy, permitting the process to be faster and more convenient. This can be of value, for example, in various hospital departments, including in the emergency department where speed is of the essence. Similarly, a fast CBC result would benefit patients showing up at an oncology center for treatment. We believe that the ability to work with just two drops of blood, plus the faster time-to-results and the reduction in operator overhead provided by our FDA-cleared moderate-complexity analyzer, will be highly attractive to providers.
EDITOR: Yossi, your platform has the potential to disrupt hematology testing in several ways. But let’s stay with near-patient settings and physicians’ offices capable of doing moderately-complex CLIA testing, which you’ve identified as the market segment you want to serve with the OLO CBC analyzer. Why are you confident this solution will gain traction?
POLLAK: The advantages offered by the speed of results and the ability to use finger-stick samples seem to be resonating across a number of moderately-complex settings, including hospital satellite facilities, oncology centers, and urgent-care centers. However, a less obvious advantage stems from OLO’s impact on operational efficiency and cost structure in low- and medium-volume settings: analyzers designed for large labs require a good deal of overhead in the form of washouts, calibrations, and frequent quality control (QC) runs. While these are effectively amortized across the large number of samples in larger facilities, they result in significant impact on test cost and workflow in the low- and medium-volume settings.
We believe that the ability to work with just two drops of blood, plus the faster time-to-results and the reduction in operator overhead provided by our FDA-cleared moderate-complexity analyzer, will be highly attractive to providers.
EDITOR: Explain these benefits, please.
POLLAK: Some of our customers have been surprised to learn just how much of an impact these factors have on their cost structures with legacy analyzers. In contrast, OLO’s use of disposable test kits means it does not require washouts; it is calibrated at the factory, so it does not require repeated calibration, and we enable our users to put in place independent QC plans (IQCPs) to reduce the number and cost of QC runs. We are building a collection of case studies that illustrate the resulting advantages in terms of total cost of ownership.
EDITOR: Finger-stick sampling is a patient-friendly feature, since many patients are uncomfortable with venipunctures. How was this part of your strategy?
POLLAK: Finger-stick collection has several equally important benefits. It eliminates the need for a phlebotomist or nurse to perform the venipuncture. That reduces the cost of collecting a specimen. It also gives the provider more flexibility in staffing, while preserving the ability to collect specimens and perform a CBC test with the moderately-complex device.
EDITOR: What about the benefits of time-to-result?
POLLAK: That is perhaps the most important source of value. The ability to speed up how a patient moves through the process of specimen collection, diagnosis based on test results, and decision on how to treat can have major positive consequences for providers. For example, during the COVID-19 pandemic, we are finding that hospitals are interested in the OLO CBC analyzer because it can help move patients faster through the appropriate care pathways in the emergency department. Hospitals want to triage patients as fast as possible to determine if they are positive for SARS-CoV-2. This moderately-complex testing solution can help cut the time to answer.
EDITOR: The Sysmex XN-2000 Hematology Analyzer is your predicate device and this instrument has FDA clearance as a moderate-complexity CLIA system. The Sight OLO CBC Analyzer also has FDA clearance as a moderate-complexity CLIA system. Do you have a timeline with the FDA at this point to obtain clearance as a waived CLIA test? And are you open to sites that might want to be study sites?
POLLAK: Hopefully, we will soon advance OLO by applying for a CLIA Waiver. The COVID-19 pandemic has unfortunately slowed our recruitment of additional clinical study sites, but we are back on track now. We have had clinical studies with Columbia University Medical Center, with Boston Children’s Hospital, and with TriCore Reference Laboratories in Albuquerque, N.M., and we plan to do more studies. We are interested in finding additional sites to take part in studies.
EDITOR: In coming to the United States, who is your perfect buyer and user? And who is an early adopter and user of the system?
We report 19 parameters in total, including red and white blood cells enumerated per microliter of blood as well as absolute and relative counts for the WBC differential.
POLLAK: We provide a lot of value to pediatrics, oncology centers, emergency rooms, hospital satellite facilities, and to urgent-care facilities that need to provide a speedy clinical service. Because of the excitement we see, we’ve accelerated expanding our sales distribution network. Our U.S headquarters is in Brooklyn, New York. This is where we warehouse devices, as well as recruit and train our U.S. sales professionals.
EDITOR: Some clinical insights are of interest to help our clients and regular readers who tend to be the first-mover types in labs in the U.S. They want to do the right thing for patients and be ahead of the curve with the kind of technology they think is ready for clinical service. How is information provided on the number of different blood cell types? Is it the same as you would obtain with a CBC with the cells enumerated per a given volume?
POLLAK: Yes, we report CBC results in a similar fashion to existing analyzers. We report 19 parameters in total, including red and white blood cells enumerated per microliter of blood as well as absolute and relative counts for the WBC differential. OLO also has a flagging system, which flags blasts, immature granulocytes and nucleated RBCs, and raises a number of messages. Results are displayed right on the instruments, optionally printed by an attached printer and transmitted to the laboratory information system.
EDITOR: Does OLO provide information about the function of the blood cells, most notably platelet function?
POLLAK: At this point, we are not going beyond the original CBC. However, we are collecting image and information to support ongoing research as to how we can adapt our platform for other types of diagnostic tests.
EDITOR: I understand you have six gigabytes of raw data coming out of a single assay. Is that being sent to an LIS?
POLLAK: It’s true: our method for “digitizing blood” results in six GB of image data per sample. However, we don’t currently store the images on the LIS, only the test results. Of course, when we do store image data, it will be HIPAA compliant and anonymized. We already have an extensive database of blood images—about half a petabyte worth. We are working with it to find trends, specifically with a few studies around stroke.
EDITOR: This is the future of laboratory medicine—to combine lab test results with other relevant clinical and demographic data to provide a more detailed picture of the patient that helps physicians make a more accurate diagnosis and select the most appropriate therapies. Thank you, Yossi, for taking the time to explain how Sight Diagnostics is using new technologies in the field of hematology testing.
POLLAK: As you can see, we are excited about the different ways that this new testing platform can help improve patient care. Thank you for the opportunity to share this information.
Contact Yossi Pollak at firstname.lastname@example.org; Judy Boniface-Chang at email@example.com.
Sight Diagnostics’ OLO 5-Part CBC Test Needs Just Two Drops of Blood to Produce Results
ONE WAY TO LOOK AT THE NEW INSTRUMENT SYSTEM MANUFACTURED by Sight Diagnostics of Tel Aviv, Israel, is that it brings new value to clinical laboratory testing by combining several rapidly-changing technologies to create a novel diagnostic testing solution. Sight’s OLO CBC analyzer utilizes machine imaging and artificial intelligence to deliver a 5-part CBC result in about 10 minutes, using two drops of blood.
Shown above are the three steps that Sight Diagnostics says are needed to produce a CBC test result using its OLO analyzer in moderately-complex CLIA labs. One drop of blood is placed in each of the two wells on the cartridge, which uses lateral flow technology to move the specimen through the required steps. After inserting the cartridge into the instrument, results are available within 10 minutes.
One point of competitive differentiation between the design and function of the OLO CBC analyzer and most existing hematology instruments is the absence of complex tubing, pumps, and moving parts. The manufacturer expects the OLO CBC analyzer to be simpler to operate and to require less maintenance when compared to existing, high-throughput hematology systems. At the same time, the OLO CBC analyzer’s cost per test will not be as low as the high-volume CBC analyzers used in the large central laboratories around the United States.