International evidence-based consensus diagnostic and treatment guidelines for unicentric Castleman disease

Castleman disease (CD) comprises a heterogeneous group of disorders involving a broad lymph node histopathological spectrum and is classified as unicentric (UCD) or multicentric CD (MCD). Classically, MCD presents with lymphadenopathy affecting multiple lymph node stations and is associated with systemic symptoms, such as fever, weight loss, and fatigue, driven by interleukin-6 (IL-6) and other cytokines. MCD has been subclassified into human herpesvirus-8 (HHV-8)–associated MCD, POEMS-associated MCD, and HHV-8⁻/idiopathic MCD (iMCD).

UCD, which involves a single enlarged lymph node or multiple enlarged lymph nodes within a single lymph node station, was first described by Castleman and Towne¹  in a patient with a mediastinal mass in 1954 and soon thereafter by Castleman et al²  in a series of 12 cases in 1956. The disease typically behaves in an indolent fashion, and enlargement of lymph nodes is gradual. Patients can be diagnosed at any age (range, 2-84 years), with a median age of diagnosis of 34 years and possibly a mild preponderance of female patients.³  There are no known epidemiological factors that predispose to the development of UCD. The incidence of UCD has been estimated at 16 to 19 per million in the US population, which translates to 5000 to 6000 patients per year.⁴  UCD is virtually always HHV-8⁻, but rare positive cases have been reported, and these should be categorized and managed as localized HHV-8–associated CD.^5-8  Frequently, the diagnosis of UCD is incidental, and the lymphadenopathy is often asymptomatic. However, some patients present with symptoms resulting from compression of vital structures (eg, airways, neurovascular bundles, or ureters), whereas others will experience iMCD-like inflammatory syndromes. In a large series of UCD cases, the most common sites of involvement were the mediastinum (29%), neck (23%), abdomen (21%), and retroperitoneum (17%).⁵  UCD can also occur in the axilla and inguinal regions as well as orbits, nasopharynx, and small bowel.^9-12 

Histopathological features of lymph nodes in UCD range from what is considered to be hyaline vascular (HV) histopathology to plasmacytic (PC) histopathology, with a mixed subgroup in between that exhibits both features. HV histopathological features are most often described in UCD, and these patients do not typically have constitutional symptomatology or excess cytokine secretion.^13-16  The lymph nodes show follicular hyperplasia with abnormal atretic germinal centers, increased vascularity with hyalinization, prominent and/or dysplastic follicular dendritic cells (FDCs), and expanded mantle zones in an onion-skin appearance. Often there are radially penetrating blood vessels causing a lollipop appearance, multiple germinal centers within the same mantle zone referred to as twinning or budding, tight aggregates of plasmacytoid dendritic cells, and architectural disruption with obliterated sinuses. No interfollicular plasmacytosis is present in HV UCD. An estimated 10% to 20% of UCD cases are considered to have PC histopathology, involving variably sized germinal centers, interfollicular plasmacytosis, and occasionally systemic inflammatory symptoms. Patients with PC UCD or mixed UCD often demonstrate iMCD-like symptoms, such as night sweats, fever, anorexia, or weight loss, as well as abnormal laboratory markers, such as anemia, elevated erythrocyte sedimentation rate (ESR) or C-reactive protein (CRP), hypoalbuminemia, or hypergammaglobulinemia.^13,15-19  Occasionally, AA amyloidosis as well as non–amyloidosis-related renal disease can occur with UCD, mainly in the PC subtype.^20-23  Life expectancy is usually not changed after the diagnosis of UCD. However, patients with UCD are at increased risk of developing paraneoplastic pemphigus (PNP), bronchiolitis obliterans (BO), AA amyloidosis, vascular neoplasms (eg, FDC sarcoma [FDCS]), and possibly lymphomas.^23-26 

The pathogenesis of UCD is not well understood, but recent evidence suggests that it may be caused in some patients by a clonal expansion of lymph node stromal cells, including FDCs.²⁷  FDCs are located in the germinal center of secondary follicles and interact with B cells during germinal center reaction through antigen presentation as well as chemokine and cytokine secretion. The atrophic B cell–depleted germinal centers in UCD may reflect FDC dysfunction.^27-29  Recently, whole-exome sequencing revealed gain-of-function somatic mutations in the kinase domain of the PDGFRB gene in 7 of 41 UCD patients with HV histopathology; the mutations were localized to CD45⁻ cells, likely representing stromal cells.³⁰  No clear associations have been made between UCD and active viral infections, such as HIV or HHV-8.³¹  Further research is under way to investigate the etiology and pathogenesis of UCD.

The Castleman Disease Collaborative Network (CDCN) has established international consensus treatment guidelines for the treatment of iMCD. Herein, we report our recommendations for the management of UCD.

In 2016, the CDCN embarked on the establishment of treatment guidelines for UCD and iMCD based on expert opinion and review of the available literature by the Patient Care and Therapy subcommittee of its scientific advisory board. This international group of 42 participants from the United States, Japan, Hong Kong, France, the United Kingdom, Germany, Italy, Norway, and New Zealand comprises experts in hematology/oncology, hematology, hematopathology, infectious diseases, immunology, and surgery, as well as an iMCD patient. The working group first met in December 2016 in San Diego, California, with a follow-up meeting in December 2017 in Atlanta, Georgia, both during America Society of Hematology annual conferences. Three additional Web-based teleconferences were held in August 2017, November 2017, and March 2018. The iMCD guidelines were first published in 2018. Further discussion and refinement of the UCD guidelines occurred from 2018 to 2020.

All relevant English-language literature was identified through PubMed using the search term unicentric Castleman disease. Included were case series and retrospective cohort studies involving ≥5 patients in which treatments and responses were reported and a clear distinction between UCD and MCD was apparent. Literature from before 2012 was captured in the large review by Talat et al.⁵  Systematic literature reviews after 2012 were excluded out of concern of possible overreporting on the same patients. Articles primarily focusing on treatment of UCD complications, such as PNP or BO, were also excluded. Table 1 summarizes published data and data from the ongoing ACCELERATE natural history registry³²  from the CDCN and University of Pennsylvania.^12,19,33-45  The consensus focused on 2 main topics: treatment of resectable UCD and treatment of unresectable UCD. Levels of evidence were modeled after National Comprehensive Cancer Network guidelines.⁴⁶ 

UCD is diagnosed after a lymph node biopsy to investigate solitary lymphadenopathy, usually on clinical suspicion of lymphoma. Figure 1 displays the spectrum of changes seen in UCD, and Figure 2 provides representative images.

In addition to identifying histopathological features consistent with UCD, a thorough workup for conditions with overlapping histopathology is required (Table 2). It is important that a complete excisional lymph node biopsy be performed rather than a fine-needle aspirate or core biopsy. Unlike MCD with its many overlapping conditions, few diseases other than UCD and lymphomas present with a solitary enlarged lymph node with CD-like histopathology. Thymomas, progressive transformation of germinal centers, unusual morphological variants of follicular lymphoma (FL), and lymphoproliferations with regressive germinal centers, such as angioimmunoblastic T-cell lymphoma, can sometimes show histopathological features reminiscent of HV UCD. PC UCD–like histopathological features may be seen in many other conditions, such as infections, autoimmune diseases, primary or acquired immunodeficiencies, such as advanced phases of HIV-related lymphadenopathy, and malignancies, including Hodgkin lymphoma, but the lymphadenopathy is usually not unicentric.^2,47-49 

A series of laboratory and radiological investigations should be performed (Table 3). First, computed tomography (CT) of the neck, chest, abdomen, and pelvis is recommended to establish the number of lymph node stations involved and differentiate UCD from MCD. UCD should be suspected in the setting of a single persistently enlarged lymph node or multiple enlarged lymph nodes within a single lymph node station associated with moderate to intense postcontrast enhancement on CT. UCD can present in any lymph node in the body, and the size of the enlarged lymph is often more bulky than the enlarged lymph nodes seen in MCD. If >1 lymph node is enlarged, those lymph nodes, by definition, must be confined to a single lymph node region to be considered to represent UCD.

CT–positron emission tomography scanning may provide additional information regarding the metabolic activity of the affected lymph nodes.⁵⁰  High specific uptake values should raise the suspicion of lymphoma, which is an important differential diagnosis. If suspected, immunostaining for markers of lymphoma and clonality testing should also be performed. Immunostaining for HHV-8 should also be performed to identify rare cases of HHV-8⁺ UCD. Clinical evaluation includes assessment for constitutional symptoms (night sweats, fever >38°C, weight loss, fatigue) and features such as fluid accumulation (peripheral edema, pleural effusion, ascites). Recommended laboratory testing comprises screening for anemia, elevated CRP and/or ESR, hypoalbuminemia, hypergammaglobulinemia, and other markers of a cytokine-induced inflammatory response. It should be noted that most cases of UCD are asymptomatic, and often no laboratory abnormalities are present. HIV serology is typically negative. As clinically indicated, further workup may require assessment for PNP, BO, and POEMS syndrome. Rare cases of UCD have light chain restriction of plasma cells, which should prompt a workup for a clonal plasma cell dyscrasia with associated POEMS syndrome.

Important considerations in the management of UCD are: the location of the enlarged lymph node, the resectability of the enlarged lymph node, and the degree to which the enlarged lymph node causes symptoms by compressing neighboring anatomy or inducing a systemic inflammatory syndrome. Further consideration should also be given to cases that are currently asymptomatic but likely to become symptomatic through compression of neighboring structures if the mass continues to grow. Resectability is a subjective decision that should weigh the risks and benefits of surgery and be made in consultation between the surgeon, clinician, and patient. Multidisciplinary review is encouraged and should include surgical specialists, radiologists, pathologists, and oncologists to discuss treatment options to optimize patient outcomes.

We recommend that all patients first be evaluated to determine whether the mass is amenable to complete surgical extirpation (Figure 3). Complete surgical excision will usually eliminate any systemic symptomatology and laboratory abnormalities, if present. A review of the literature and real-world data from the ACCELERATE study found a high response rate to surgical excision (Table 1). In their first descriptions, Castleman et al^1,2  reported that surgery was an effective treatment modality, and several early series confirmed that resection is the preferred method of first-line treatment.^{15,16,48,51-53}  Dispenzieri et al³⁴  reported a 5-year overall survival (OS) of 91% in a series of 53 patients surgically treated at the Mayo Clinic and University of Nebraska. Talat et al⁵  reported a similar OS of >90% in a retrospective literature study comprising 278 UCD patients treated with surgery, with a follow-up of up to 10 years. In the Talat et al series, outcome was better when surgery was performed for peripheral lymphadenopathy rather than central lymphadenopathy located in the chest, abdomen, or pelvis. Death resulting from disease occurred in only 1 (1.5%) of 68 patients with peripheral lymphadenopathy vs 9 (6.1%) of 148 patients with central lymphadenopathy (P < .05), presumably because resection was virtually always feasible in those with peripheral lymphadenopathy. Inability to resect the involved lymph node proved the only factor predictive of a fatal outcome. In the absence of persistent systemic symptoms or concern for neighboring anatomic structures, residual small satellite lymph nodes after resection can be observed, with spontaneous involution often seen, after resection of the principal enlarged lymph node.⁵⁴  Symptoms resulting from AA amyloidosis as well as non–amyloidosis-related renal disease typically improve over the ensuing years after removal of unicentric disease.^20-22 

UCD should be managed in the first-line setting with surgery in both children and adults. Parez et al⁵⁵  reported 72 pediatric cases who fared well with surgery (n = 70) or radiotherapy (n = 2). Taken together, these and other reports provide level 2A evidence establishing surgical resection as the treatment of choice for UCD.^{19,37,45,56-59}  Surgical resection is the most appropriate first-line intervention, according to uniform panel consensus. No high-quality randomized controlled trials exist to substantiate a level 1 grade recommendation. Recurrences after complete resection are uncommon, although they have occasionally been described.^{12,13,45,53,60,61}  Recurrent UCD should be managed according to the algorithm described in Figure 3 and surgically excised, if feasible. Follow-up with yearly CT scan imaging studies for 5 years is therefore recommended. Thereafter, imaging should be clinically driven based upon suspicion of disease recurrence.^50,62 

The average size of the lymphadenopathy in UCD is larger than that in MCD (5.5 vs 3.8 cm).⁵  These larger lymph nodes can be more challenging to remove, not only because of their size but also because of the increased vascularity of UCD, which can cause severe perioperative bleeding.^63-65  Typically, dominant feeding vessels are present, and several investigators have used preoperative embolization to render surgery safer.^66-76  We recommend, based on level 2B evidence, that consideration be given to presurgical embolic devascularization of large UCD lymph nodes with high risk of bleeding.

Occasionally, a UCD lymph node may be unresectable because of size or location. This is frequently encountered in the setting of a mediastinal mass that is close to a main bronchus or major blood vessel. If surgery is not possible or too morbid, further treatment is dependent upon the presence of symptoms resulting from compression of neighboring structures or systemic inflammation (Figure 3). When possible, expert advice should be sought to identify the most appropriate therapy for a given patient.

The management of unresectable UCD requires a judicial approach. The lymphadenopathy in patients with UCD can remain stable or grow slowly over time. There is consensus that a careful watch-and-wait approach can be adopted in patients with normal laboratory values who are currently asymptomatic and unlikely to develop future symptoms resulting from compression of neighboring structures if the mass continues to grow (level 2B evidence).^17,19,77  Selected patients who are currently asymptomatic but likely to develop symptoms if the mass continues to grow may need to be treated more aggressively, as if they have compression-related symptomatic unresectable UCD. We recommend regular CT scan monitoring at 12-month intervals as well as annual physical examination and laboratory investigations.

Several therapeutic options are available if the lymphadenopathy is causing compressive symptoms but cannot be surgically removed. Partial resection to debulk the disease without achieving complete removal can yield good results, but follow-up is required to monitor for growth, with the frequency of follow-up individualized by the location, degree of partial resection, and degree of compression of neighboring structures.^15,48,53,62  There have been case reports in which UCD lymph nodes have been rendered amenable to surgery through medical therapy with rituximab with or without steroids. In selected cases, there may be a role for embolization as adjunctive or alternative therapy.

There have been no systematic studies to evaluate the most effective approach in these patients (Table 1).^13,69,78,79  In a recent series, 21 patients with unresectable UCD received a variety of therapies, including steroids, chemotherapy, rituximab, and tocilizumab; 7 went on to have successful surgery.¹⁹  Overall, the available evidence and consensus opinion (level 2B) support the use of rituximab with or without steroids to reduce the size of compression-related symptomatic unresectable UCD. Any persisting lymphadenopathy should be surgically resected, if feasible. Patients who have unresectable disease after medical therapy but are rendered asymptomatic can be observed. Those remaining symptomatic are candidates for radiotherapy (as described in “Persistent symptomatic unresectable UCD”).

We consider any symptomatic UCD with inflammatory symptoms (night sweats, fevers, anorexia, weight loss) and/or laboratory abnormalities (elevated ESR, elevated CRP, or anemia) to be inflammation related. IL-6 is the well-established driver of symptomatology in most iMCD patients and rare UCD patients with inflammation-related symptoms and laboratory abnormalities. Early studies correlated local production of IL-6 with the systemic manifestations seen in a minority of patients with UCD, because lymph node excision resulted in relief of symptoms along with a decrease in serum IL-6 levels.^80-82  However, levels of IL-6 and other inflammatory mediators have not been systematically studied in a large number of UCD cases. Anti–IL-6 agents siltuximab and tocilizumab have been thoroughly studied in iMCD and found to be effective. There is consensus among the panel that it is appropriate to treat inflammation-related symptomatic unresectable UCD with siltuximab or tocilizumab, if elevated CRP/ESR or anemia is present, before considering extensive surgery or radiotherapy (level 2B).⁸³  However, it should be pointed out that limited data exist regarding the effectiveness of this approach.^19,84  Any persisting lymphadenopathy should be surgically resected, if feasible. As with compression-related unresectable UCD, patients who continue to have unresectable UCD lymphadenopathy after treatment with anti–IL-6 monoclonal antibody therapy but have become otherwise asymptomatic can be observed. Patients with ongoing troublesome symptomatology should be considered for radiotherapy (as described in “Persistent symptomatic unresectable UCD”), with rituximab with or without steroids being an alternative therapeutic option.

Among unresectable UCD patients with persistent symptoms and lymphadenopathy despite therapies directed at the compression- or inflammation-related symptoms, radiotherapy is an alternative treatment modality (Table 1). Several early case reports attest to the potential utility of radiotherapy in UCD.^53,85-90  Chronowski et al⁹¹  reported on a small series of 4 patients, 3 of whom achieved complete resolution after irradiation with doses in the order of 4000 cGy. Chan et al⁹²  identified 17 cases of unresectable UCD in the literature treated with similar doses of radiotherapy.^{53,85,87,88,90,91,93-98}  Thirteen achieved disease control, with 7 patients attaining a partial response and 6 a complete response. At a median follow-up of 20 months, only 1 had died as a result of progressive disease. In general, radiotherapy was well tolerated, although 1 patient was reported to have developed PNP.⁹⁵  However, there is no uniform consensus among the panel members that therapeutic irradiation is desirable (level 3). Radiotherapy does carry the concern of late sequelae, which temper enthusiasm for its routine application in a benign disorder, especially in younger patients.

An alternative approach to radiotherapy for unresectable UCD patients with persistent symptoms and lymphadenopathy is treatment with immunomodulators/immunosuppressants, such as corticosteroids, cyclosporine A, and sirolimus, which are more commonly used for iMCD. However, limited data exist on these treatment approaches (level 3).

After completion of the initially planned treatment, patients should be evaluated to determine the disease response to treatment and should be followed longitudinally for relapse and complications. One to 3 months after the initial therapy, the response to treatment should be documented by history, physical examination, laboratory studies, and imaging.

After excision, patients are followed annually with CT and laboratory studies, which include complete blood count, lactate dehydrogenase, and chemistries with liver and renal function and electrolytes, albumin, CRP, and quantitative immunoglobulins. Annual imaging may be discontinued after 5 years if the patient remains disease free.

Occasionally, patients continue to report symptoms after complete surgical extirpation that include persisting fatigue, lethargy, and other constitutional symptoms. These patients should be reevaluated to ensure that an alternative diagnosis, such as iMCD, autoimmune condition, endocrinopathy, depression, chronic fatigue syndrome, or fibromyalgia, was not missed, which may be responsible for the symptoms. Appropriate testing (eg, antinuclear antibody, ESR, CRP, cytokine panel) can help to underpin the presence of an unrecognized disorder giving rise to inflammatory symptomatology. Absent an alternative etiology to explain these symptoms, the systemic therapy approaches described in “Inflammation-related symptomatic unresectable UCD” could be considered, although additional research is needed to assess their use in this setting.

There has been recent recognition that patients can present with a clinical picture that seems to fall between UCD (1 enlarged lymph node or 1 station of enlarged lymph nodes) and iMCD (>1 station of enlarged lymph nodes, and often >5 stations, with cytokine-driven systemic inflammation). Specifically, these patients often have enlarged lymph nodes in 2 to 3 adjacent lymph node stations, but they lack sufficient clinical and laboratory abnormalities to meet the iMCD diagnostic criteria. These cases have been tentatively described as regional or oligocentric CD. More research is needed, but we felt it was important to recognize this entity in this report, because it may require different treatments than iMCD. In fact, the clinical and pathological features of regional CD resemble those of UCD closely.⁸³  Based on the current limited experiences of the panel and lack of available peer-reviewed evidence, we believe that surgical removal or debulking should be considered (level 2B) over extensive morbid surgeries and systemic therapies, particularly cytotoxic chemotherapies, if possible.^19,83  This more conservative approach avoids long-term adverse effects and is prudent given the often indolent clinical course. However, regional CD with inflammatory symptoms may benefit from treatment according to the iMCD treatment algorithm, particularly when exhibiting the PC histopathological subtype.

A number of disorders can present in conjunction with UCD or arise after UCD. These include PNP, Hodgkin disease (HD) or NHL, and FDCS. A survey of the French registry of CD showed that 18% of UCD patients had a complicating disorder.¹⁹ 

UCD associated with PNP deserves special mention, because it is life threatening and often accompanied by progressive BO.⁹⁹  The frequency of this associated disorder seems to be particularly high in the Asian population.¹⁰⁰  PNP and BO typically occur in the setting of the HV histopathologic subtype of UCD, sometimes associated with stroma-rich features.⁴¹  Complete surgical removal of UCD should be performed in patients with UCD and PNP; this has been reported to often halt or reverse the PNP.^100,101  BO has a high fatality rate, despite treatment with a variety of agents alone or in combination, including corticosteroids, rituximab, cyclosporine A, and cyclophosphamide.^{25,99,102,103}  The expert panel agrees that therapy with an anti–IL-6 monoclonal antibody has limited anecdotal efficacy in PNP.

FDCS is a rare form of sarcoma arising from FDCs that can cooccur with the HV histopathologic subtype of UCD.^104-106  In a large single-institution series of 66 patients with FDCS, 6 were found to have UCD, and in 2, the diagnosis of UCD preceded that of FDCS. The preferred therapy for localized FDCS is resection, and additional consolidative radiotherapy should be considered for local control.^107,108  Chemotherapy is recommended for metastatic FDCS.

Distinguishing lymphoma from UCD can be challenging. HD has been reported in patients with UCD; however, the UCD plasmacytic-like histopathological changes may simply represent reactive changes secondary to the HD. UCD and HD are most frequently observed in the same lymph node, emphasizing the need for complete surgical excision.¹⁰⁹  A few cases of UCD preceding HD have been described.^109-113  A review of 23 cases reported in the literature revealed that 21 cases were described as having PC histopathology, and 2 had mixed histopathology.¹⁰⁹  A majority of the HD patients had B symptoms. PC UCD–like lymph node changes may be due to IL-6 and other cytokines produced by Reed-Sternberg cells. Occasionally, CD-like changes are found in a distant lymph node, which can delay the diagnosis of HD.

In contrast, NHL is more often associated with UCD with HV histopathology. In a series reporting NHL and UCD cooccurrence, NHL was less commonly present in the same lymph node as UCD (37%) and was instead found in different anatomical areas (72%).¹¹⁴  Both B- and T-cell lymphomas have been described to cooccur with UCD. FL with features reminiscent of HV UCD may be present in some patients as an unusual morphological variant of FL that may create diagnostic difficulties.¹¹⁴  Taken together, the precise relationship between UCD and both HD and NHL remains unclear. Regardless, the risk of having an HD or NHL diagnosis concurrently or subsequent to a UCD diagnosis seems to be low.

Therapy for UCD and NHL or HD should primarily be dictated by the lymphoma and may include excision, radiotherapy, and chemotherapy.

The present consensus guidelines for UCD complement the previous recommendations for iMCD. It is important to accurately diagnose UCD by ruling out conditions with overlapping histopathology, distinguishing UCD from iMCD, and evaluating the resectability and symptomatology associated with each case. We recommend complete surgical resection as the preferred intervention for UCD whenever possible. Recurrence of UCD after complete surgical removal is rare. Occasionally, the distinction between UCD and iMCD is not clearly delineated, with some patients exhibiting more regional lymphadenopathy. These patients typically have a more indolent clinical course and may benefit from UCD-directed rather than iMCD-directed treatment. The most challenging scenario in UCD is that of unresectable UCD. These cases require a judicial approach, with some benefiting from medical cytoreduction followed by surgery and others better served with surgical debulking followed by careful observation. Unresectable asymptomatic UCD patients may also be managed with surveillance alone. In patients with unresectable UCD and inflammation-related symptomatology, anti–IL-6 therapy may be effective. Radiotherapy is an acceptable alternative for unresectable UCD but should probably be avoided in younger patients when possible. Further research is urgently needed into the most effective approaches for unresectable UCD, regional CD, and anti–IL-6 refractory iMCD patients. The CDCN international ACCELERATE natural history registry³²  collects data regarding the clinical characteristics, therapies, and outcomes of all forms of CD to advance knowledge regarding this complex and heterogeneous disorder. The University of Pennsylvania and CDCN are also able to consent, coordinate, and receive excess lymph node tissue from clinically warranted lymph node biopsies for research purposes.¹¹⁵  The ACCELERATE registry and lymph node biospecimens support ongoing clinical and translational research efforts, with the twin goals of strengthening the evidence base to help physicians in selecting optimal treatment approaches and identifying new rational therapeutic modalities, particularly for patients with unresectable UCD.

The CDCN coordinated the meetings. The authors of the guidelines had full responsibility for the consensus: building process/methods, data interpretations, treatment recommendations, and writing of the report.

Contribution: D.C.F., F.v.R., S.K.P., and S.P. wrote the paper, analyzed data, and designed research; and F.v.R., E.O., G.S., P.V., M.L., A.D., M.I., S.P., S.S., M.S., R.W., D.W., I.M., J.B., N.M., W.B., K.S.E.-J., A.F., M.J.L., S.C., S.K.P., A.G., S.N., K.Y., R.K., T.S.U., C.C., A.C., and D.C.F. wrote the paper.

Conflict-of-interest disclosure: A.D. receives research support from Celgene, Takeda, Alnylam, Caelum, and Pfizer and serves in an advisory capacity for Janssen, Prothena, and Akcea. D.C.F. has received grant funding from EUSA Pharma and Janssen Pharmaceuticals for the ACCELERATE natural history registry (registered at www.clinicaltrials.gov as #NCT02817997) and has received study drug from Pfizer for a clinical trial of sirolimus (registered at www.clinicaltrials.gov as #NCT03933904). R.K. receives research funding from Genentech, Merck Serono, Pfizer, Boehringer Ingelheim, TopAlliance, Takeda, Incyte, Debiopharm, Medimmune, Sequenom, Foundation Medicine, Konica Minolta, Grifols, Omniseq, and Guardant; receives consultant and/or speaker fees and/or serves on an advisory board for X-Biotech, Neomed, Pfizer, Actuate Therapeutics, Roche, Turning Point Therapeutics, TD2/Volastra, and Bicara Therapeutics, Inc.; has an equity interest in IDbyDNA and CureMatch, Inc.; serves on the boards of CureMatch, Inc., and CureMetrix; and is a cofounder of CureMatch, Inc. E.O. receives consulting fees and serves on an advisory board for EUSA Pharma. S.S. serves on the European Advisory Board for EUSA. G.S. receives consultant and/or speaker fees and/or serves on an advisory board for EUSA Pharma, Janssen, and Takeda. T.S.U. receives research support from Celgene/Bristol-Myers Squibb, Roche, and Merck and serves as a consultant for Seattle Genetics. F.v.R. receives research support from Janssen Pharmaceuticals and consultant fees from EUSA Pharma. R.W. has received research funding from Janssen Pharmaceuticals and served on advisory boards for Janssen Pharmaceuticals. D.W., on behalf of the University of Washington, received research support from Janssen Pharmaceuticals for a role in the central pathology review for the 2014 van Rhee et al Lancet Oncology work. The remaining authors declare no competing financial interests.

Correspondence: Frits van Rhee, University of Arkansas for Medical Sciences, Myeloma Center, 4301 W Markham #816, Little Rock, AR 72205; e-mail: vanrheefrits@uams.edu.