Published in: Am J Hematol. 2022;97:514–518
Translated by: Ye Xiangjun
Since 2001, under the auspices of the World Health Organization (WHO) and its affiliated International Agency for Research on Cancer (IARC), the classification of hematologic tumors has been standardized as part of the WHO classification series commonly referred to as the “WHO Blue Book.” Before 2001, the classification of lymphomas, leukemias, and chronic myeloid diseases followed various different and often controversial paths. For lymphomas, pathologists initially adopted limited single-expert or regional classifications, such as those proposed by Rappaport, Lennert (Kiel), and Lukes and Collins; attempts to create a common language between classifications, known as working classifications, effectively became their own classifications. Although Rappaport included leukemias and chronic myeloid diseases in his 1966 Armed Forces Institute of Pathology volume, the widely accepted classifications of myeloma and leukemia were primarily proposed by hematologists, including the French-American-British (FAB) group and the Polycythemia Vera Study Group, among others. The criteria for these classifications varied and were based on different combinations of clinical features, cellular morphology, and cytochemical tests; in some cases, limited immunophenotypic analyses were also included, often with little (if any) assessment of prognostic significance. Despite these limitations, the various classifications provided urgently needed standards for the diagnosis of various hematologic tumors, allowing for further study and refinement. However, none of these classifications represented an international consensus and did not widely incorporate the opinions of experts in hematology, oncology, genetics, and pathology.
In 1994, the International Lymphoma Study Group (ILSG), composed of international lymphoma pathologists, proposed the Revised European-American Lymphoma (REAL) classification, attempting to define biological lymphoma entities based on clinical, morphological, immunophenotypic, and genetic findings. After the publication of the REAL classification, Les Sobin, co-editor of the WHO/IARC third edition Blue Book series (along with Paul Kleihues), contacted Elaine Jaffe, a co-author of the REAL classification and then-chair of the American Society of Hematopathology, to develop a similar classification for the WHO/IARC Blue Book series, which had not been widely used for the classification of hematologic diseases before. In fact, the second edition WHO series did not include hematologic tumors. Jaffe and the Executive Committee of the American Society of Hematopathology suggested that the work for the WHO third edition (now published by IARC) be supervised by two major hematopathology societies, namely the Society of Hematopathology (SH) and the European Association for Haematopathology (EAHP), convening a Clinical Advisory Committee (CAC) composed of leading international pathologists, oncologists, hematologists, and geneticists to provide input for the development of such classifications. They also suggested that the third edition classification not be limited to lymphomas but should include myeloid tumors and acute leukemias. The first CAC meeting was held in 1997 at Arlie House in Virginia, ultimately resulting in the third edition WHO classification of hematopoietic and lymphoid tissue tumors in 2001, thanks to the efforts of 75 invited authors from around the globe. Similar CAC meetings were held in Chicago in 2007 and 2014, chaired by James Vardiman and Michelle Le Beau, resulting in the fourth edition and the revised fourth edition WHO/IARC publications in 2008 and 2017, respectively. The revised fourth edition had over 200 contributors from 24 countries. The minutes of the CAC meetings were published by the leaders of the CAC before the official WHO/IARC books.
To gain deeper insights into the clinical input needed for the classification of myeloid tumors and acute leukemias and the impact of such classifications on the international community, the organizers of the international consensus meeting on myeloid and acute leukemia sought additional perspectives from leaders across four continents on the classification of myeloid and lymphoid tumors.
Professor Andrew Roberts, head of cancer research and treatment at the Walter and Eliza Hall Institute of Medical Research, Metcalf Chair of Leukemia Research at the University of Melbourne, and clinical hematologist at the Royal Melbourne Hospital and Peter MacCallum Cancer Centre, Australia
Never before have treatments for newly diagnosed patients with hematologic malignancies been so frequent. With a stronger ability than ever to explore the differences and commonalities of genetic abnormalities between patients, and equipped with an increasing number of targeted therapeutic agents, each with its own more selective activity range, physicians and patients now face more decisions than ever. These advancements bring the prospect of precision medicine closer to reality, but inevitably create greater uncertainty. For physicians, this uncertainty becomes apparent when we ask ourselves some critical questions, such as: How certain am I that this is the correct diagnosis for this disease? Have I ordered all necessary tests and interpreted them correctly? Does this test result seem inconsistent with the diagnosis, is that a concern? Does this diagnosis determine the best treatment approach? If I did not make the correct diagnosis, will the same treatment still work? For patients seeking transparent information and advice from their doctors, this uncertainty can become a barrier to accepting recommendations or increase sensitivity to misinformation spread through social media.
At the core of achieving rational and evidence-based decision-making in clinical practice is the correct diagnosis. Hematology has progressed because our predecessors have integrated insights into the pathophysiology of diseases into the clinical and morphological algorithms that generate our current comprehensive diagnostic classification system. As our knowledge grows, our diagnostic classification systems must evolve. In light of this, it is time to update the classification of myeloid tumors and acute leukemias. We need it to further integrate robust patterns of morphology, immunophenotype, cytogenetics, and molecular aberrations into rational diagnostic categories. In doing so, we need to maintain a constant focus on the performance characteristics of new molecular tests. However, the classification system should lead hematology forward and should not be hindered by variable issues in test implementation. When both physicians and patients believe that their significance for diagnosis is established and demand it, the latter will receive the fastest resolution.
In Australia, a coalition of professional and consumer non-government organizations recently assessed our healthcare system’s performance in caring for patients with hematologic malignancies. Although many patients have achieved good outcomes, there are still gaps that can be improved by applying existing knowledge. A loud voice in the advocacy is for equal access for all patients to accurate diagnoses. Patients know that accurate diagnoses help them understand their prognosis and receive the best treatment. They want medicine to assess their disease through appropriate and precise testing and confidently classify it (i.e., with as little uncertainty as possible). The evolving classification system for hematologic malignancies serves us well in this regard. Looking forward to the next iteration of the international classification, I urge all stakeholders—pathologists, physicians, manufacturers, and consumers—to work together to ensure its implementation, thus safeguarding the interests of patients.
Dr. Vikram Mathews, Professor of Hematology at Christian Medical College, Vellore, India
While it is often said that “our understanding of the biology of myeloid tumors has made revolutionary advances recently,” this is largely relevant to one-fifth of the world’s population living in developed countries. Myeloid tumors such as AML are among the most expensive to treat, and in developing countries, this is further complicated by high out-of-pocket costs, susceptibility to fungal and multidrug-resistant bacterial infections, and limited access to centers of excellence. Furthermore, the recently approved blockbuster drugs for treating myeloid tumors are often unavailable or affordable only to a small fraction of patients. An exception is the treatment of chronic myeloid leukemia (CML), where, despite repeated and coordinated efforts to undermine them, the use of generic tyrosine kinase inhibitors (TKIs) has indeed changed treatment outcomes for patients in these countries. The data claimed to be clinically relevant to the high level of evidence (Phase III RCT) generated in highly developed countries are unlikely to be widely applicable in India. It is also important to recognize that in India, due to high out-of-pocket expenses associated with obtaining healthcare, an estimated 39 million people live below the poverty line each year.
Many of the observations above relate to the relevance of general malignancy classification, and more specifically, the relevance of myeloid malignancy classification in low- and middle-high income countries (L/UMIC) can be considered under the following main headings: (1) While the anticipated derivatives of encouraging molecular classification and better-targeted therapies with minimal off-target side effects must be pursued, it is equally important to maintain the relevance of basic classification tools (such as morphology, cytochemistry, and immunophenotypic analysis), which are sufficient in most cases to make appropriate treatment decisions if done well. (2) Rare entities need to be clearly designated as rare. There is a trend to overemphasize rare and exotic cases while neglecting common and curable malignancies. In an environment where knowledge rather than actual practice is disproportionately emphasized and rewarded, this focus exacerbates a self-fulfilling prophecy dictated by the existing conditions in L/UMIC. (3) There is a need to strive to highlight highly curable/easily manageable myeloid tumors and incorporate the best care standards for such diseases into classifications. (4) Similarly, strategies for rapid and accurate diagnosis of very poor prognostic markers must be emphasized, which, where possible, is highly relevant to L/UMIC, as this data will rapidly change decisions about optimal supportive care rather than continuing to use expensive treatments that may impoverish patients’ families.
Dr. Anne Stidsholt Roug, Consultant and Associate Professor at Aarhus University Hospital, Denmark
In Europe, over the past few decades, significant energy (and generous funding) has been invested at both the community level and in clinical trials for molecular testing. It is recognized that molecular phenotyping holds prognostic information and is increasingly being used as a decision-making tool for targeted drug therapies and allogeneic hematopoietic stem cell transplantation. Consequently, many centers in Europe have incorporated extensive molecular profiling as part of routine checks for patients with acute leukemia suitable for intensive chemotherapy, and this approach is increasingly being applied to elderly or frail patients; however, access to such analyses in Europe remains inequitable, and platforms are not well standardized.
After more than a decade of molecular analysis of myeloid tumors and acute leukemias, it can be said that the list of drugs inspired by genomics remains relatively small, and the pathways for developing other such targeted therapies are limited. Moreover, in diagnosis and risk stratification, the data from molecular profiling has yet to be translated into non-overlapping disease entities. Additionally, the molecular associations of diseases are often confounded by age, and the classification of low-risk and high-risk categories is not always accurate. In some European countries, national and government-funded whole-genome sequencing programs are being developed. However, most of the data from molecular testing may be redundant when applied to current classification and risk stratification systems.
In the (2016) WHO classification of myeloid tumors and acute leukemias, the eight subtypes of AML are defined by specific cytogenetic abnormalities and certain mutations. The European Leukemia Network (ELN) initiative adopted a similar approach, stratifying AML risk based on 12 cytogenetic and 6 mutation aberrations. However, both these classification and risk stratification systems represent an oversimplification of the more complex underlying biology, thus requiring ongoing elaboration of genetic-driven diagnostic approaches in the context of developing knowledge bases to ensure uniform interpretation of genetic findings, using artificial intelligence to automate the development of decision support tools, integrate clinical and genetic data for risk scoring systems, and reliable studies combining clinical and molecular data to optimize and personalize patient treatment. Given these issues, classifying hematologic tumors as disease entities with clinically relevant and biologically homogeneous definitions requires further improvement and integration of all current disease diagnostic tools, continually supported by new molecular information. In Europe, collective efforts such as the ELN, which includes 220 participating centers from 44 countries, provide influential research papers, guidelines, and updated management recommendations, representing a formal cooperative start to this work.
Dr. Charles A. Schiffer, Emeritus Professor of Oncology at Wayne State University School of Medicine and Karmanos Cancer Institute, USA
The International Consensus Classification (ICC) of hematologic malignancies provides a significant update to previous classification systems by integrating updated molecular and clinical information, adding updated entities, and clarifying diagnoses that were marked as “provisional” in earlier efforts. Despite the pandemic, expert subspecialty hematopathologists and clinicians collaborated in this effort, using a combination of virtual meetings (less fun) and actual face-to-face meetings, aided by the protective benefits of broader vaccination.
Why? Beyond calling for clinical and diagnostic laboratories to focus on a few diagnoses with unique mutations that benefit from true targeted therapies, standardization of diagnoses is crucial to ensure comparable patient populations within and between clinical trials, so that outcomes using different therapies can be compared. Parallel work to standardize definitions of treatment response and toxicity also ensures that we are all speaking the same language. While more precise pathological diagnoses often guide treatment, there are usually few additional insights into the pathophysiological mechanisms controlling treatment choices. For example, why are certain morphological subtypes more sensitive to treatments with high cure rates? Consider CHOP therapy for diffuse large B-cell lymphoma and its close relative follicular lymphoma—the former has a significantly higher cure rate, while the latter responds well but has a lower cure rate. Why does hairy cell leukemia have much lower activity to purine analogs compared to other lymphoproliferative cancers at similar stages of differentiation? My group published an early large study on AML cytogenetics and, like others, noted that CBF AML patients (t(8;21) and inv16) have considerable cure rates under standard anthracycline and cytarabine regimens. I naively expected explanations might come soon, but I am now older and still waiting for clues that might apply to other leukemias. Similarly, in general, when we typically associate aneuploidy with more malignant phenotypes, why does extreme hyperdiploidy confer sensitivity? There are many other such examples that could be cited.
Recognizing that it is generally more straightforward to experimentally study resistance than sensitivity, we can draw important biological lessons from the hematologic malignancies we cure. Given the heterogeneity of curable hematologic malignancies, achieving extreme chemotherapeutic sensitivity seems unlikely to be solely a result of a lack of resistance mechanisms. The hypothesis that restoring immune surveillance plays a role is attractive, as reversing immune tolerance allows the growth of initially transformed cells, as this may be conducive to drug manipulation. A classification system that combines the goal of recognizing biological similarities with the heterogeneity of curable cancers would be intriguing and could become the next scientific focus for groups such as the ICC.
International Consensus Classification
The above perspectives highlight the highly impactful nature of changes in the classification of myeloid tumors and acute leukemias, and how these changes clearly necessitate extensive CAC work prior to any new classification. The organizers of the international consensus meeting agreed and deemed it necessary to elucidate the steps leading to this process to the hematology and pathology communities.
Despite multiple communications between the leaders and clinical leaders of SH and EAHP with Ian Cree, editor of the WHO/IARC fifth edition Blue Book series, including a formal request in February 2020 from 33 leaders in pathology, hematology, oncology, and genetics, consensus has yet to be reached on the necessity of convening a Clinical Advisory Committee (CAC) prior to initiating the fifth edition classification. Therefore, knowing that its conclusions might not inform the upcoming WHO classification, it was still decided to advance the CAC. This decision was not made lightly; it involved the opinions of former WHO editors and senior advisors, as well as leaders from SH and EAHP and top clinicians from around the globe. To organize this International Consensus Conference (ICC) on the classification of myeloid and lymphoid tumors, clinical and pathological co-chairs were selected for the lymphoma and myeloid/acute leukemia sections, respectively. The co-chairs for myeloid/acute leukemia pathology, Daniel Arber, Robert Hasserjian, and Attilio Orazi, along with clinical co-chairs Mario Cazzola, Hartmut Döhner, and Ayalew Tefferi, aimed to include diverse and geographically broad leadership among the participants across various disease categories. Once participants were identified, considerable preparatory work was needed to prepare for a productive two-day CAC meeting. Invited participants were assigned to groups, each of which pre-reviewed the existing classifications and proposed potential changes and key issues to be addressed in the CAC. Each group held a series of virtual meetings and shared key issues across disease groups, such as potential changes in primary cell counts or exclusion criteria.
On September 20-21, participants gathered at the Rubinstein Forum on the University of Chicago campus, working hard to present their respective statements to the entire CAC. The joint CAC for lymphomas and myeloid/acute leukemias included 138 participants from 23 countries across 5 continents (42 in-person, 98 remote). Participants discussed and debated potential classification changes and identified new questions and unresolved issues. After the meeting, groups reconvened to address key issues and propose the final classification for myeloid tumors and acute leukemias. For several contentious issues, the entire CAC myeloid/acute leukemia participants were surveyed for further opinions to ensure consensus on significant changes. The result became the International Consensus Classification of Myeloid and Lymphoid Neoplasms (ICC-MLN), which will be reported in a peer-reviewed manuscript with a publication target for mid-2022. The new name ICC-MLN was proposed by the pathology co-chairs, reflecting broad international input from authoritative pathologists and clinicians, along with updated genetic integrations. ICC-MLN will reflect the consensus among all parties involved in the diagnosis and treatment of hematologic malignancies regarding disease entities, terminology, and diagnostic standards.
The fundamental principle of the new ICC classification is to rely on input from pathology experts and clinical input to develop a rational classification based on genetic integration and clinical pathology.
Conclusion
The process of the Clinical Advisory Committee is unique for the classification of hematologic tumors, especially for myeloid tumors and acute leukemias. The available disease classifications are not merely a list of disease names proposed by one person or a small group, regardless of their expertise. Establishing consensus on the detailed criteria that define disease categories is crucial for pathologists to make diagnoses that are reproducible worldwide. Without well-defined standards, diagnostic ambiguity will arise, affecting both the treatment and outcomes of individual patients and the ability of clinical trials to provide meaningful results. For this reason, we deemed it necessary to convene an international expert CAC in 2021 to revise the existing classifications for guiding patient management of myeloid tumors and acute leukemias.
