Advances make hematology more useful

The advent of digital image-based cell counting and application of AI will add more value to patient care, but microscopic examination by experienced technicians is still the gold standard in diagnosis of hematological abnormalities.

Over the past decades, hematology analyzers have experienced great technical advancements, featuring low turnaround time, enhanced precision, and accuracy. The challenges associated with manual methods like immature cells, reliability on results, distribution error, and statistical error are being replaced with precise, safe, and dependable automated hematology systems.

The hematology market in India is highly dependent on imports, the major suppliers being China, Europe, Japan, and the USA. The current fluctuation in the forex rates is one of the major concerns in the Indian hematology segment.

Out of the existing 110,000 laboratories in India, more than 40,000 laboratories are in the Tier-II and Tier-III cities, covering the rural and semi-urban population with limited access to majority of IVD technologies. As per statistics, more than 70 percent of the Indian population lives in the rural areas, and the benefits of IVD technologies remain underserved to the patients in these rural areas. One of the major changes in the hematology segment is the introduction of highly compact 3-part and 5-part hematology systems. These compact systems are well suited for rural India, and are expected to bring in changes in the IVD industry. The penetration of CBC in rural areas will help to assess the prevalence of various anemia among the rural population, and its control, which is an objective of the NRHM, and an affordable and accurate 3-part hematology analyzer will go a long way to achieve this objective.

With COVID-19 pandemic refusing to relent, most laboratories in the country are operating at very low output, resulting in minimal purchases of instruments. The procurement by government labs also is low. With the recent approval of COVID-19 package of Rs 15,000 crore to build on health infrastructure till March 2024, some expenditure is expected on IgG ELISA assay kits used for the qualitative detection of novel coronavirus-infected pneumonia cases, suspected clustering cases, and other new coronaviruses in serum samples through measurement of the COVID-19 IgG antibody. The PCR tests currently used to directly detect the presence of an antigen can be very labor intensive, with several stages at which errors may occur between sampling and analysis. The differential hematology analyzers may help in providing an accurate and precise evaluation and enumeration of the presence of reactive lymphocytes, which play a critical role in management of COVID-19 patients.

Global market
The global hematology analyzers and reagents market is expected to reach USD 9.6 billion by 2023 from USD 7.2 billion in 2019, at a CAGR of 7.4 percent, predicts MarketsandMarkets. The ability of a hematology test to effectively measure several blood components has made it an essential screening tool to diagnose a wide range of blood-related disorders, including anemia, blood cancer, hemorrhagic conditions, and blood infections, among others that affect millions of people each year across all age groups. Increasing incidences of hematology disorders and growing awareness related to the availability of a wide range of diagnostic options to treat these targeted disorders are the primary factors driving the growth of the hematology testing market.

Automation, being the primary focus of the hematology testing, the market has been witnessing considerable technological advancements with respect to products that offer effective, accurate, and fast testing results. The availability of automated hematology analyzers has further led to a reduced administrative error, making the process more effective and efficient in disease diagnosis, as compared to manual hematology testing. Furthermore, factors such as technological advancements in hematology analyzers and reagents and integration of flow cytometry techniques with hematology analyzers are expected to support the growth of this market during 2020-2023.

AI and hematology
The evidence-based literature on healthcare is currently expanding exponentially. The opportunities provided by the advancement in artificial intelligence (AI) tools such as machine learning (ML) are appealing in tackling many of the current healthcare challenges. As of 2020, radio-diagnostics is the field where AI has made the most progress to date; 90 percent of the practicing radiologists anticipate that AI will be incorporated in their future practices.

Recently, AI has also been employed in the analysis of hematopathology data for better-informed diagnosis, prognosis, and treatment planning. The foundation knowledge related to benign and malignant hematology is also evolving with the application of AI in hematology. Even the most skilled hematology specialist can overlook the patterns, deviations, and relations among the vast number of blood parameter measures by modern analyzers. In contrast, the AI algorithm can easily analyze hundreds of attributes (parameters) simultaneously with correlation, and recognize a pattern or fingerprint of certain hematological abnormalities.

The AI algorithms of modern hematology analyzers are trained on sufficiently large datasets of clinical cases that include laboratory blood tests, performed and confirmed by hematology specialist using various confirmatory tests.

The ML algorithms can uncover the disease-related patterns in a blood test that may be beyond medical knowledge, resulting in higher diagnostic accuracy as compared to traditional quantitative interpretation, based only over reference ranges. Also, it can lead to a fundamental change in differential diagnosis and result in the modifications in the presently accepted guidelines.

Currently, normal leukocytes can be differentiated, using automated hematology analyzers equipped with optical sensors and mathematical computer-based algorithms. However, because the morphology of dysplastic blood cells in patients with hematological disorders is much more elaborate than that of normal cells, manual microscopic examinations remain the mainstay of diagnosis, which are time-consuming, demanding, and subjective. Thus, many industrial and academic researchers have sought to develop efficient and accurate automated diagnostic systems. Current advances in computer technology have been used to derive automated diagnostic systems for leukemia.

Over the past decade, more than 20 studies have attempted to diagnose hematological malignancies, mainly acute lymphoblastic leukemia (ALL), using various mathematical algorithms to recognize and classify cell images. This process requires several complex steps like preprocessing, segmentation, feature extraction, and classification. Recently, convolutional neural networks (CNNs), advanced forms of deep learning, have been used to optimize the parameters automatically, without the need for mathematical algorithms. CNNs classify cell images more accurately than conventional neural networks or ML systems.

Researchers have successfully developed and tested a deep learning-based algorithm–CNN. This system is applied for differentiating myelodysplastic syndrome from aplastic anemia, with 96.2 percent of sensitivity and 100 percent specificity. This is the first such system for myelodysplastic syndrome (MDS), using peripheral blood smears; it can further be developed for automated diagnosis of various hematological abnormalities. This system recognizes over 100 patterns in cell size and cytoplasmic morphological features for detection and classification of myeloid malignant cells including MDS. Further training of this AI algorithm can improve the accuracy and reproducibility, which may surpass the human eye.

An understanding of AI’s basics will need to become a part of the physician’s statistical literacy in the very near future. As more widespread implementation of clinical AI nears, attention has also turned to the effects this will have on other areas in medicine. AI offers many promising tools to clinicians broadly, and specifically in the practice of hematology. Ongoing research into its various applications will likely result in an increasing utilization of AI by a broader swath of clinicians.

Recent developments
Today, the highest level of automation is scalable and configurable, starting from piercing the sample tubes to conducting variety of determinations, CBC, 6-part, WBC differential, NRBC, RET and immature retic fraction (IRF), automated immature platelet fraction (IPF), automated smear preparation, and staining. These tests are standardized assays that meet performance goals, decrease technologist hands-on time, eliminate batch testing, and provide results faster to physicians. Testing together with hematology-specific middleware product, laboratories are reporting up to 85 percent of their test volume without any operator intervention.

In addition to these, hematology automation platforms offer more than CBC testing from a single EDTA sample. Laboratories that have incorporated HbA1c testing on high-speed hematology lines are performing >90 percent of assays with minimal technologist intervention. Auto validation of HbA1c results can run as high as 90 percent. Further process improvements are coming to the forefront of hematology, such as automation of digital smear review. Now, these newly formed work areas can manage traditional hematology testing as well as the HbA1c, traditionally tested in the chemistry department.

For reagent preparation, hematology technologists frequently required to change 20-liter cubes of the diluent. Recent additions to automated hematology lines address this concern by including units that utilize concentrated reagent, which is diluted using an in-lab water supply. This approach minimizes the time and effort required to prep the analyzers prior to the highest volume run of the day, and enables laboratorians to avoid interruptions in testing owing to the need to frequently change the diluent in high-volume testing settings. Added to these innovations, the work in the direction of pre- and post-analytical sample sorting and archiving, effectively helps in minimizing errors in pre- and post-analytical work.

Other recent technological advances have led to the development and commercialization of innovative automated image analysis systems, which are suited for automation and can hence be directly connected (in series) with hematologic analyzers. These innovative platforms scan the slides (usually at a picture of ×100 objective), and store digitized images of blood smears at high magnification. The images are analyzed by artificial neural networks, based on a preexisting database of blood elements (thus including RBC), which can be locally customized or updated by the users. The images can be transmitted to, and displayed on, computer screens, which can be even placed at long distances from the scanner, for analysis and potential reclassification of blood elements. The operator can also increase the size of the images, or expand single sections of the scan, so obtaining a more accurate view. The operator can then accept and conserve the automatic classification or can move elements from one cell category to another, thus improving the final reclassification.

Although these automated image analysis systems have been originally developed for analysis of white blood cells (WBC), specific information can also be garnered on erythrocyte morphology, thus including the presence of anysocytosis, hypochromia, microcytosis or macrocytosis, spherocytosis, elliptocytosis, ovalocytosis, stomatocytosis, acanthocytosis, echinocytosis, polychromasia, poikilocytosis, and abnormal erythrocytes.

The recent data shows that diagnostic sensitivity of these systems for identifying some critical categories of abnormal erythrocytes is excellent, typically higher than 80 percent, thus making the use of digital image analysis a highly valuable, and probably more accurate and reproducible alternative to optical microscopy.

Notably, the use of these systems may also enable an efficient recognition of parasitoid infections like malaria, as well as the reliable identification of intravascular and spurious hemolysis, which would be otherwise undetectable on whole-blood specimens. Interestingly, most of these automated image analysis systems are also capable of optimizing the identification of rare RBC abnormalities, since morphological erythrocyte alterations can be more efficiently visualized on the computer screen. Finally, the creation of a large personalized database of images of suggestive RBC abnormalities represents a valuable resource for education and training of students and laboratory professionals.

Outlook
Hematology analyzers are the real work horses of clinical laboratories. The advent of digital image-based cell counting, and application of AI will add more value to patient care, but microscopic examination by experienced personnel is still the gold standard in diagnosis of hematological abnormalities. The detection of abnormal cells on the basis of complex morphology, especially the size and shape of abnormal cells like immature white blood cells or diseased cells. It is important, though, to accurately detect the presence of immature white blood cells since they are often associated with conditions like leukemia, infection, inflammation, or tissue injury. Accordingly, manufacturers are working to decrease the number of manual reviews that are required, through improvements in analyzer accuracy in differential count.

Opportunities in these areas are enormous and this segment will surely go a long way in coming years. 

Industry Speak

Second Opinion