Hairy Cell Leukemia Pathogenesis, Diagnosis, and Prognosis

Publication
Article
Special IssuesFocus on Pathways: Historical Perspective and Current Considerations in Hairy Cell Leukemia

The pathogenesis, diagnostic criteria, staging, and prognosis of hairy cell leukemia, an uncommon indolent B-cell lymphoid neoplasm, is examined.

Hairy cell leukemia (HCL) is an uncommon indolent B-cell lymphoid neoplasm.1 The first article in this 2-part series explored the discovery and characterization of HCL, its epidemiology, and the clinical presentation of the disease. The current article will examine the pathogenesis, diagnostic criteria, staging, and prognosis of HCL.

Pathogenesis of Hairy Cell Leukemia

Cell of Origin

The cellular origin of HCL has long been debated.2 Phenotypically and morphologically, HCL tumor cells do not resemble any stages of normal B-cell maturation and development.1 Because of morphologic and functional similarities between hairy cells and monocytes/macrophages, the former were thought to have derived from a tumor cell from the reticuloendothelium.1 The discovery of rearrangement of the B-cell receptor immunoglobulin genes in patients with the disease demonstrated that HCL has a B-cell origin.1,3

Studies conducted with monoclonal antibodies have confirmed that hairy cells are mature B cells.4 Hairy cells express B-cell markers (eg, CD19, CD20, and CD22)5,6 and possess light-chain—restricted immunoglobulins.4 These cells are unable to differentiate into antibody-secreting cells, despite the fact that they express the plasma cell—associated antigen PCA1. Other antigens expressed by malignant hairy cells include FMC7, CD22, CD25, CD72, and CD40L, which are activation antigens associated with B cells, as well as with other types of cells. Hairy cells do not express CD23, which distinguishes them from chronic lymphocytic leukemia cells.4 The cell-surface antigens expressed by hairy cells are summarized in Table 1.4

The discovery of these cell-surface markers led to the recognition of mature B cells as the origin of hairy cells. In addition, these antigens are used in immunophenotyping, when establishing the diagnosis of HCL.7 These antigens on hairy cells may serve as a therapeutic target for the disease. For example, cytotoxic drug conjugates of anti-CD22 antibodies are currently in development for HCL.8 Another cytotoxin with high affinity and selectivity for CD22 has recently been approved for the treatment of patients with relapsed or refractory HCL.8-10

BRAF V600E Mutation

The BRAF V600E mutation is a causal genetic event in the development of HCL. It is present in virtually all patients with classic HCL.11,12 BRAF V600E is a point mutation in which thymine is substituted with adenine at exon 15, position 1799, resulting in a change from valine (V) to glutamate (E) for amino acid 600.11 This amino acid substitution initiates constitutive activation of the RAS-RAF-mitogen—activated protein kinase (MAPK) extracellular signal-regulated kinase (ERK) signaling pathway (Figure 1),13 leading to enhanced cell survival and proliferation.2,11,12

Accumulating data suggest that the presence of BRAF V600E is the disease-causing event in HCL. This mutation is found in almost all cases of HCL, both at initial diagnosis and at relapse, regardless of the clinical presentation of the disease.11 In addition, the BRAF V600E mutation is detected in the entire tumor clone11 and is responsible for the hairy morphology, enhanced viability, and gene expression profile characteristic of HCL.14 In vitro data suggest that silencing the RAF-MAPK/ERK kinase (MEK) pathway using BRAF or MEK inhibitors leads to loss of the HCL-specific gene expression profile signature, reverses the morphology of the cells from “hairy” to “smooth,” and induces apoptosis.13,14 Hence, inhibiting the effects of the BRAF V600E mutation appears to reverse the morphology and characteristics of hairy cells.

In contrast to the classic form of HCL, BRAF V600E is not present in hairy cell leukemia variant (HCLv)—a rarer form of B-cell neoplasm with a clinical course and response to therapy that are distinct from those of HCL (see Differential Diagnosis section for further discussion).7

Tissue Infiltration Patterns of HCL

Hairy cells express the activated integrin receptors α4β1 and αVβ3 on their cell surface.2 Ligands of these integrin receptors—vascular cell adhesion molecule 1 (VCAM1) and vitronectin—home the leukemic cells to the bone marrow stroma; the splenic red pulp; and the splenic, hepatic, and bone marrow sinusoids. Hairy cells also overexpress inhibitors of prometastatic matrix metalloproteinases (MMPs), leading to the propensity of HCL cells to be confined in the peripheral blood and the aforementioned blood-related compartments. In addition, HCL cells upregulate annexin-1, which minimizes leukocyte extravasation during inflammatory responses. Downregulation of chemokine receptors and adhesion proteins, essential for lymph node homing, also contributes to the lack of consistent lymph node involvement in the disease.2 The profile of receptor and protein expression in HCL contributes to its distinct tissue infiltration patterns and to mostly peripheral blood, bone marrow, and vascular compartments within the spleen and liver.2

Hairy cells modify the infiltrated tissues by forming vascular lakes, or “pseudosinuses,” in the splenic red pulp, as well as fibrosis in the bone marrow.4 In the bone marrow, hairy cells interact with hyaluronan through the adhesion receptor CD44, resulting in fibroblast growth factor production, which in turn stimulates the synthesis of fibronectin by the tumor cells.4 From this process, the synthesis of hairy cells and the assembly of a fibronectin matrix lead to the characteristic bone marrow fibrosis of the disease.15 In the spleen, hairy cells modify and replace the splenic red pulp.16 Through α4β1 expressed on the cell surface, the infiltrating hairy cells interact with VCAM on endothelial cells, forming “pseudosinuses,” which then lead to red cell pooling in the spleen and the anemia that is associated with HCL.16

The tissue infiltration process by hairy cells leads to bone marrow infiltration and hypersplenism, which contribute to the severe cytopenia reported in patients with HCL.1

Diagnosis and Staging

Initial Workup and Diagnosis

Fatigue and infections are common among patients with HCL.7 Although splenomegaly was a common presentation in the past, incidental findings of pancytopenia have now been detected at earlier stages of the disease and are much more common.7 The initial workup in a patient with suspected HCL is depicted in Table 2.1,7,17

Close evaluation of the peripheral blood smear, including a differential count, should be a part of the initial workup.7 Although leukemic cells are often rare, monocytopenia is a rather sensitive and specific finding in HCL detection.7 Hairy cells have an abundant pale blue cytoplasm, an oblong nucleus, a serrated cytoplasmic border, and hair-like/threadlike cytoplasmic extensions.7,18

Immunophenotyping of the leukemic cells is essential for establishing a diagnosis of HCL.7 Detection of monotypic B cells is an important step in identifying mature B-cell neoplasms, as these cells express only a single immunoglobulin species.19 Peripheral blood mononuclear cells in HCL exhibit light-chain restriction of either kappa- or lambda-expressing B cells.7 Table 319,20 provides additional information on kappa and lambda light chains, along with cluster of differentiation nomenclature.

Immunophenotyping by flow cytometry of peripheral blood or bone marrow aspirate is an important part of the diagnostic workup.1 Diagnosis of the classical form of HCL is confirmed by bright positivity for CD11c, CD20, and CD22, and positivity for CD19, CD103, and CD123.7,8 The neoplastic cells are negatively stained for CD27 antigen but intensely stained for CD200 expression.7 Although no single marker can distinguish HCL from other B-cell leukemias, the use of a panel of immunophenotypic markers is valuable in differentiating among other types of lymphoid neoplasm.1,21

Bone marrow aspirates may yield a “dry tap,” which is due to bone marrow fibrosis and loss of hematopoietic cells in patients with HCL.1 The extent of bone marrow infiltration can be ascertained via bone marrow trephine biopsy and aspirate.7 Adequate trephine biopsies should be ≥20 mm in length after processing22 and are needed in patients with bone marrow fibrosis or hypocellular bone marrow.22

Bone marrow specimens in HCL appear patchy, with a diffuse infiltrate of small lymphoid cells that resemble fried eggs and widely spaced nuclei because of the abundant cytoplasm.1,23 The extent of bone marrow involvement is determined by immunohistochemical stains for CD20, annexin-1A, tartrate-resistant acid phosphatase (TRAP), DBA.44, and VE1 (a stain for BRAF V600E).7,8 Presence of the BRAF V600E mutation differentiates classic HCL from other lymphoid neoplasms and thus has therapeutic implications. Therefore, highly sensitive molecular assays for the mutation should be used (eg, next-generation sequencing or allele-specific polymerase chain reaction), rather than less sensitive methods (eg, melting curve analysis, pyrosequencing, or Sanger sequencing).7 These highly sensitive methods may detect the few leukemic cells present in the peripheral blood or bone marrow aspirates from a dry tap. If these techniques are not available, however, then the application of VE1 immunohistochemical stain to the bone marrow biopsy may allow for detection of the BRAF V600E mutation.7

Differential Diagnosis

HCL and other lymphoid neoplasms should be distinguished from each other, in order to attain the correct diagnosis and treatment. Immunophenotyping is a useful step in distinguishing among the different lymphoproliferative neoplasms.21

A rare, aggressive variant of HCL, known as HCLv, does not respond well to classic HCL treatments.1 In 2008, the World Health Organization recognized HCLv as a separate disease.7 In contrast to HCL, HCLv lacks expression of CD25, annexin-A1, TRAP, and BRAF V600E.8 In addition, patients with HCLv often have increased white blood cell counts (>10 x 109/L), less severe cytopenias, and more lymphocytosis compared with those with the classic form of HCL.1,8 In HCLv, the hairy cells are atypical and have prolymphocytic features.1 In addition, HCLv expresses dim or negative CD123,1,5 whereas classic HCL expresses bright homogeneous CD123.5 Moreover, the BRAF V600E mutation is absent in HCLv. Mitogen-activated protein kinase (MAP)2K1 mutations are detected in 50% of HCLv cases, and 17p (tumor protein p53) deletions are detected in 30% of cases.11 Negativity for CD25, annexin-A1, and BRAF V600E are common features of splenic diffuse red pulp small B-cell lymphoma (SDRPBCL) and HCLv.11 Currently, the genomic profile of SDRPBCL has not been clarified, and it is still considered a provisional entity.13

Even though it is associated with more pronounced peripheral blood involvement, splenic marginal zone lymphoma (SMZL) does exhibit some clinical and morphologic similarities to HCL.1 Unlike HCL, however, SMZL lacks TRAP expression. It also has a different immunophenotype, because it lacks CD25 and CD103 expression.1,21 CD123 and annexin-A1 are typically negative, and BRAF V600E is absent, in patients with SMZL.1

Staging

An initial staging system was proposed by Jansen and Hermans based on a retrospective analysis of clinical outcomes in 391 patients.24 The investigators reported that 4 factors contributed significantly to survival among patients with HCL: (1) duration of symptoms, (2) spleen size, (3) hemoglobin level, and (4) number of hairy cells.24 Only the degree of splenomegaly and hemoglobin level were used in a survival prediction model, however, because the number of hairy cells and symptom duration reports were deemed too subjective.24 A clinical staging system that comprised 3 prognosis groups was proposed, based on hemoglobin levels and spleen size24:

Stage I:

Hemoglobin >12.0 g/dL and spleen <10 cm under costal margin (UCM) or Hemoglobin >8.5 g/dL and spleen <4 cm UCM

Stage II:

Hemoglobin >12.0 g/dL and spleen >10 cm UCM or Hemoglobin 8.5 to 12.0 g/dL and spleen 4 to 10 cm UCM or Hemoglobin <8.5 g/dL and spleen <4 cm UCM

Stage III:

Hemoglobin 8.5 to 12.0 g/dL and spleen >10 cm UCM or Hemoglobin <8.5 g/dL and spleen >4 cm UCM

Use of this proposed staging system has not been confirmed in the most recent consensus guidelines for the diagnosis and management of HCL, however, which have been published by the International Hairy Cell Leukemia Research Foundation and by the National Comprehensive Cancer Network.7,17 In place of a staging system, both guidelines recommended that indications for treatment include the following7,17:

  • Systemic symptoms
  • Splenic discomfort
  • Recurrent infection
  • Excessive fatigue
  • Symptomatic organomegaly
  • Unexplained weight loss of >10% within the previous 6 months
  • Progressive lymphadenopathy or lymphocytosis
  • Hemoglobin <11 g/dL
  • Platelets <100 x 109/dL
  • Absolute neutrophil count <1000 /mL

Prognosis

The first-line treatment for patients with HCL should comprise a standard regimen of a purine nucleoside analog—either pentostatin or cladribine.7 Purine analogs are highly effective initial therapies for HCL. The overall response rates following a single course of cladribine range from 75% to 100%, with complete response rates of 72% to 98%.1 The efficacy of pentostatin is similar, with overall response rates ranging from 79% to 100% and complete response rates of 44% to 89%.1 Data from one study suggest that at 10 years and 15 years after diagnosis, relapse rates with pentostatin were 42% and 47%, respectively, compared with 42% and 48%, respectively, with cladribine.25 Patients with complete responses who had hemoglobin levels >100 g/dL and platelet counts >100 x 109/L experienced the longest relapse-free survival. In contrast, those with a partial response who had hemoglobin levels <100 g/dL and/or platelet counts <100 x 109/L experienced the shortest relapse-free survival.25

Data from the 1973 to 2003 US National Cancer Institute’s Surveillance, Epidemiology, and End Results Program estimate that the 5-year survival rate among patients with HCL is about 87%.26 More recent data are available from a population-based analysis that was conducted in The Netherlands.27 Relative survival in patients with HCL, as defined by overall survival corrected for life expectancy of the general population, improved gradually during 3 time periods that were examined: 1989 to 1993, 1994 to 2000, and 2001 to 2015.27 Improvement among patients 60 to 69 years of age was statistically significant.27 The 10-year relative survival rate in patients <70 years of age was between 95% and 97% in the most recent data series (2001 to 2015).27 For this group of patients, relative survival decreased in the first 2 years following diagnosis and plateaued thereafter.27 Low mortality can be attributed to the high rates of complete remission with purine analog therapies, improvements in infection prevention and control, and other supportive measures.27 In this Dutch study, outcomes in older patients were inferior. The 10-year relative survival rate in patients ≥70 years of age was 83% in 2001 to 2015 time period.27 In this subgroup, excess mortality existed for ≤4 years after diagnosis.27 Reasons for increased mortality with increasing age may include concomitant comorbidities, which are associated with a reluctance to use purine analog therapies, and the risk for experiencing infection while undergoing purine analog treatments.27 In summary, these findings suggest that the majority of patients, especially those <70 years of age, may enjoy a relatively normal life expectancy, even with an HCL diagnosis.

Despite durable responses achieved in most patients with the use of purine nucleoside analogs, approximately 40% of patients with HCL experience a relapse 10 years after treatment.1 Few predictors of response have been reported, but leukocytosis (hairy cells >10 x 109/L), splenomegaly (spleen size >10 cm), and the presence of unmutated immunoglobulin heavy chain variable region genes all have been associated with a shorter event-free survival (defined as the time from first treatment to disease progression, treatment for relapse, or death due to HCL) following cladribine treatment.28

Minimal residual disease (MRD), which is present in 40% to 60% of patients with HCL who are in complete remission after initial treatment, may be predictive of relapses.29,30 The elimination of MRD leading to a longer relapse-free interval, however, has not been uniformly established.1 MRD prognostic implications may be confounded by the sensitivity of techniques that are used to assess MRD, the lack of uniform criteria to define MRD, and the small sample sizes used in published studies.29 Some patients with HCL may have MRD or even morphologically evident disease following their initial treatment, but they do not experience overt clinical relapses.1 Despite the debate, accumulating data suggest that the persistence of MRD after initial therapy with a purine nucleoside analogs is predictive of eventual relapse.

Wheaton and colleagues reported that during a 3-year follow up, 50% of patients with MRD, detected by immunohistochemistry relapsed, compared with 4% of those without MRD (P = .016).30 López-Rubio and coworkers also reported that patients with MRD following treatment with pentostatin or cladribine had a shorter treatment-free interval than did those without MRD (97 months vs not reached, respectively; P <.049).31

The prognostic implication of MRD may evolve with improvements in detection technologies. Advanced flow cytometry techniques have been developed to detect MRD in patients with HCL.32 Flow cytometry that has the ability to detect a level of HCL cells of <0.01% is considered the most sensitive assay.8 Other tests that use sequence-specific probes and primers are in development, but are considered investigational. 8

At relapse, additional courses of nucleoside analogs can be used. Complete response rates, however, are lower with subsequent courses of purine analogs. In addition, pentostatin and cladribine can cause immunosuppression for a considerable period of time.1 Hence, concerns exist regarding the use of multiple courses of these agents. Although rituximab can be used and has the advantage of sparing T lymphocytes, the overall response rate with this agent is lower (26% to 80%).1 Complete response rates with the use of rituximab in patients with HCL range from 10% to 66%.1

Novel therapeutic agents are necessary to improve the treatment of patients with relapsed HCL. One of these drugs—a high-affinity anti-CD22 immunotoxin&mdash;has recently been approved.9,10 Other agents, including BRAF inhibitors and MAPK inhibitors, are currently in development.29

Conclusions and Future Directions

Advances regarding the pathogenesis of HCL have helped to elucidate the significance of the BRAF V600E mutation in the cycle of disease. Therapeutically, purine nucleoside analogs have shown to be effective agents. And yet, despite high 5-year survival rates, some patients relapse as far out as 10 years following treatment. Therefore, effective treatments for those with relapsed HCL are needed. Novel therapeutic agents in development capitalize on the knowledge of the unique biologic and immunophenotypic characteristics of hairy cells, including their surface antigen expressions (eg, CD22), BRAF V600E mutations, and downstream signaling pathways.

References

  1. Ravandi F. Diagnosis and treatment of diffuse large B-cell lymphoma (DLBCL) and Burkitt lymphoma (BL). In: Hoffman R, Benz EJ, Silberstein LE, et al, eds. Hematology: Basic Principles and Practice. 7th ed. Philadelphia:PA: Elsevier; 2018:1265-1276.
  2. Tiacci E, Liso A, Piris M, Falini B. Evolving concepts in the pathogenesis of hairy-cell leukaemia. Nat Rev Cancer. 2006;6(6):437-448.
  3. Korsmeyer SJ, Greene WC, Cossman J, et al. Rearrangement and expression of immunoglobulin genes and expression of Tac antigen in hairy cell leukemia. Proc Natl Acad Sci U S A. 1983;80(14):4522-4526.
  4. Cawley JC. The pathophysiology of the hairy cell. Hematol Oncol Clin North Am. 2006;20(5):1011-1021.
  5. Shao H, Calvo KR, Grönborg M, et al. Distinguishing hairy cell leukemia variant from hairy cell leukemia: development and validation of diagnostic criteria. Leuk Res. 2013;37(4):401-409.
  6. Robbins BA, Ellison DJ, Spinosa JC, et al. Diagnostic application of two-color flow cytometry in 161 cases of hairy cell leukemia. Blood. 1993;82(4):1277-1287.
  7. Grever MR, Abdel-Wahab O, Andritsos LA, et al. Consensus guidelines for the diagnosis and management of patients with classic hairy cell leukemia. Blood. 2017;129(5):553-560.
  8. Kreitman RJ, Arons E. Update on hairy cell leukemia. Clin Adv Hematol Oncol. 2018;16(3):205-215.
  9. AstraZeneca. US FDA approves Lumoxiti (moxetumomab pasudotox-tdfk) for certain patients with relapsed or refractory hairy cell leukaemia [news release]. September 14, 2018. astrazeneca.com/media-centre/press-releases/2018/us-fda-approves-lumoxiti-moxetumomab-pasudotox-tdfk-for-certain-patients-with-relapsed-or-refractory-hairy-cell-leukaemia.html. Accessed April 19, 201
  10. Lumoxiti [prescribing information]. Wilmington, DE: AstraZeneca Pharmaceuticals LP; 2018. https://www.accessdata.fda.gov/drugsatfda_docs/label/2018/761104s000lbl.pdf. Accessed May 14, 2019.
  11. Falini B, Martelli MP, Tiacci E. BRAF V600E mutation in hairy cell leukemia: from bench to bedside. Blood. 2016;128(15):1918-1927.
  12. Tiacci E, Trifonov V, Schiavoni G, et al. BRAF mutations in hairy-cell leukemia. N Engl J Med. 2011;364(24):2305-2315.
  13. 1Roider T, Falini B, Dietrich S. Recent advances in understanding and managing hairy cell leukemia. F1000Res. 2018;7. pii: F1000 Faculty Rev-509. doi: 10.12688/f1000research.13265.1.
  14. Pettirossi V, Santi A, Imperi E, et al. BRAF inhibitors reverse the unique molecular signature and phenotype of hairy cell leukemia and exert potent antileukemic activity. Blood. 2015;125(8):1207-1216.
  15. Burthem J, Cawley JC. The bone marrow fibrosis of hairy-cell leukemia is caused by the synthesis and assembly of a fibronectin matrix by the hairy cells. Blood. 1994;83(2):497-504.
  16. Cawley JC, Hawkins SF. The biology of hairy-cell leukaemia. Curr Opin Hematol. 2010;17(4):341-349.
  17. National Comprehensive Cancer Network (NCCN). NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines): Hairy Cell Leukemia. Version 3.2019 — January 31, 2019. NCCN Website. https://www.nccn.org/professionals/physician_gls/pdf/hairy_cell.pdf. Accessed May 15, 2019.
  18. Shastri A. Hairy cell leukemia. American Society of Hematology website. imagebank.hematology.org/image/14395/hairy-cell-leukemia. Published October 18, 2012. Accessed April 19, 2019.
  19. Seegmiller AC, Hsi ED, Craig FE. The current role of clinical flow cytometry in the evaluation of mature B-cell neoplasms. Cytometry B Clin Cytom. 2019;96(1):20-29.
  20. Engel P, Boumsell L, Balderas R, et al. CD nomenclature 2015: Human Leukocyte Differentiation Antigen Workshops as a driving force in immunology. J Immunol. 2015;195(10):4555-4563.
  21. Quest GR, Johnston JB. Clinical features and diagnosis of hairy cell leukemia. Best Pract Res Clin Haematol. 2015;28(4):180-192.
  22. Bain BJ. Bone marrow trephine biopsy. J Clin Pathol. 2001;54(10):737-742.
  23. Shastri A. Bone marrow biopsy in hairy cell leukemia: the fried egg appearance. American Society of Hematology website. imagebank.hematology.org/image/14396/bone-marrow-biopsy-in-hairy-cell-leukemia--the-fried-egg-appearance?type=upload. Published October 18, 2012. Accessed April 19, 2019.
  24. Jansen J, Hermans J. Clinical staging system for hairy-cell leukemia. Blood. 1982;60(3):571-577.
  25. Else M, Dearden CE, Matutes E, et al. Long-term follow-up of 233 patients with hairy cell leukaemia, treated initially with pentostatin or cladribine, at a median of 16 years from diagnosis. Br J Haematol. 2009;145(6):733-740.
  26. Han X, Kilfoy B, Zheng T, et al. Lymphoma survival patterns by WHO subtype in the United States, 1973—2003. Cancer Causes Control. 2008;19(8):841-858. &#8233;Dinmohamed AG, Posthuma EFM, Visser O, et al. Relative survival reaches a plateau in hairy cell leukemia: a population-based analysis in The Netherlands. Blood. 2018;131(12):1380-1383.
  27. Forconi F, Sozzi E, Cencini E, et al. Hairy cell leukemias with unmutated IGHV genes define the minor subset refractory to single-agent cladribine and with more aggressive behavior. Blood. 2009;114(21):4696-4702.
  28. López-Rubio M, Garcia-Marco JA. Current and emerging treatment options for hairy cell leukemia. Onco Targets Ther. 2015;8:2147-2156.
  29. Wheaton S, Tallman MS, Hakimian D, Peterson L. Minimal residual disease may predict bone marrow relapse in patients with hairy cell leukemia treated with 2-chlorodeoxyadenosine. Blood. 1996;87(4):1556-1560.
  30. López Rubio M, Da Silva C, Loscertales J, et al. Hairy cell leukemia treated initially with purine analogs: a retrospective study of 107 patients from the Spanish Cooperative Group on Chronic Lymphocytic Leukemia (GELLC). Leuk Lymphoma. 2014;55(5):1007-1012.
  31. Garnache Ottou F, Chandesris MO, Lhermitte L, et al. Peripheral blood 8 colour flow cytometry monitoring of hairy cell leukaemia allows detection of high-risk patients. Br J Haematol. 2014;166(1):50-59.
Related Videos
Corey Cutler, MD, MPH, and Hana Safah, MD, experts on GvHD
Guenther Koehne, MD, PhD
Lori A. Leslie, MD, an expert on lymphoma
Lori A. Leslie, MD, an expert on lymphoma
A panel of 4 experts on MDS
Elias Jabbour, MD
Corey Cutler, MD, MPH, and Hana Safah, MD, experts on GvHD
Corey Cutler, MD, MPH, and Hana Safah, MD, experts on GvHD