Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Analysis of a cohort of 279 patients with hairy-cell leukemia (HCL): 10 years of follow-up

Analysis of a cohort of 279 patients with hairy-cell leukemia (HCL): 10 years of follow-up In total, 279 patients with hairy-cell leukemia (HCL) were analyzed, with a median follow-up of 10 years. Data were collected up to June 2018. We analyzed responses to treatment, relapses, survival, and the occurrence of second malignancies during follow-up. The median age was 59 years. In total, 208 patients (75%) were treated with purine analogs (PNAs), either cladribine (159) or pentosatin (49), as the first-line therapy. After a median follow-up of 127 months, the median overall survival was 27 years, and the median relapse-free survival (RFS) was 11 years. The cumulative 10-year relapse incidence was 39%. In patients receiving second-line therapy, the median RFS was 7 years. For the second-line therapy, using the same or another PNA was equivalent. We identified 68 second malignancies in 59 patients: 49 solid cancers and 19 hematological malignancies. The 10-year cumulative incidences of cancers, solid tumors, and hematological malignancies were 15%, 11%, and 5.0%, respectively, and the standardized incidence ratios were 2.22, 1.81, and 6.67, respectively. In multivariate analysis, PNA was not a risk factor for second malignancies. HCL patients have a good long-term prognosis. PNAs are the first-line treatment. HCL patients require long-term follow-up because of their relatively increased risk of second malignancies. Introduction survival (OS) of 90% with PNA treatment . However, the Hairy cell leukemia (HCL) is a rare B-cell chronic management of HCL patients remains under investiga- lymphoproliferative disorder characterized by atypical tion, particularly the care of patients with relapsed/ lymphoid cells with hairy projections in the peripheral refractory disease and the evaluation of the risk of second 1,2 blood, bone marrow, spleen, and/or liver . HCL is malignancies. For the first relapse, using the same PNA or 3 4 9 responsible for 2% of leukemias . Splenectomy and switching to another PNA may be effective . Anti-CD20 5 10 interferon-alpha (IFNα) were the first-line treatments, monoclonal antibodies (rituximab), alone or associated 6 11 and purine analogs (PNAs), either cladribine or pentos- with PNA , can also be alternative treatments for relap- tatin , were subsequently introduced. HCL prognosis ses. With the recent identification of the BRAF V600E improved consistently over time, with a 10-year overall mutation in most classic HCL (HCLc) , BRAF inhibitors, namely, vemurafenib or dabrafenib, could be indi- 13,14 15 cated . MEK inhibitors (trametinib) , BCR pathway Correspondence: Xavier Troussard (troussard-x@chu-caen.fr) inhibitors (ibrutinib) , and anti-CD22 immunotoxins Service des Maladies du Sang, CHU d’Angers, Angers, France (moxetumomab pasudotox) are the newest therapeutic Laboratoire d’Hematologie Biologique, CHU de Caen, Caen, France Full list of author information is available at the end of the article © The Author(s) 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to theCreativeCommons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Blood Cancer Journal 1234567890():,; 1234567890():,; 1234567890():,; 1234567890():,; Paillassa et al. Blood Cancer Journal (2020) 10:62 Page 2 of 12 alternatives. The risk of second malignancies occurring single treatment modality if not associated with adjuvant during follow-up in HCL patients is controversial, with drug therapy. Multiple-agent therapies were defined by some studies describing a higher risk of cancer than in the the use of more than one drug within a period of 18–20 general population and others describing no 6 months. Responses were defined according to the 21–24 27 increased risk . The reasons for these discrepancies Consensus Resolution . CR required the morphologic might be the variability in the methods used to define absence of hairy cells in peripheral blood and bone mar- second malignancies with pooling of second cancers row aspiration or biopsy specimens and the normalization occurring before and after HCL diagnosis in some of any organomegaly and cytopenia. Immunophenotypic studies . analysis of peripheral blood or bone marrow biopsy was To answer this question, we investigated a large cohort not required. PR was defined as the normalization of the of 279 HCL patients, with a 10-year median follow-up peripheral counts associated with at least a 50% reduction period, and we analyzed the treatments, responses, sur- in organomegaly and bone marrow hairy cells and <5% vival, relapses, and occurrence of second cancers. circulating hairy cells. All other outcomes were con- sidered non-responses. Relapse was defined as any dete- Subjects and methods rioration in blood counts related to the detection of hairy Patients cells in the peripheral blood and bone marrow. The eligibility criteria for the HCL diagnosis were established according to the WHO 2008 and 2016 clas- Statistical analyses sifications, including morphological and flow cytometric Survival curves were drawn according to the analyses of blood, bone marrow or tissue specimens. Each Kaplan–Meier method. OS and RFS were compared using patient signed an informed consent form. The study was the log-rank test. The cumulative incidence of relapse performed in accordance with the Declaration of Helsinki. (CIR) and cumulative incidence of second cancer were evaluated considering death as a competing risk. The Study design and data collection CIRs were compared according to the first-line treatment In the first analysis, we collected data up to 2012 .We with Gray’s test. Multivariate analyses of OS and RFS were then updated the data up to June 2018. A questionnaire performed using the Cox regression model, and multi- was sent to the physicians, who were members of the variate analyses for the cumulative incidence of second French National Society of Hematology (SFH), with malignancies were performed using the Fine and Gray requests for the following data: date of last observation; regression model, considering death as a competing risk. last disease status: complete response (CR), partial An excess of second malignancies was expressed by the response (PR), or progressive disease (PD); date of relapse standardized incidence ratio (SIR), which was defined as (s); treatments and responses (CR, PR, failure); treatment the ratio between the number of observed and expected start and end dates; second solid cancers (date, histology); cases from the general population in France. The expec- hematological malignancies (date, WHO 2016 classifica- ted malignancy rates were calculated from the age- tion); death (date, cause); and other complications. The indexed (in 5-year categories) cancer incidence in 29,30 second cancers were defined as either synchronous can- France . These incidence rates were multiplied by the cers or metachronous cancers. Synchronous cancers were observed person-years at risk in each age category to cancers that occurred at the same time as the diagnosis of calculate the expected frequency of second malignancies. HCL and those occurring within two months, as recom- Confidence intervals (CIs) of the SIR were obtained by mended by the Surveillance, Epidemiology, and End assuming a Poisson distribution for the observed num- Results (SEER) Program. Metachronous cancers were bers. The SIR was calculated for second cancers, second defined by cancers occurring more than two months after solid cancers, and second hematological malignancies. As the diagnosis of HCL. Overall survival (OS) was defined as the French registries did not include nonmelanoma skin the time from the date of HCL diagnosis until death from cancers or monoclonal gammopathies of undetermined any cause or the date of the last observation. Relapse-free significance/monoclonal gammopathies of clinical sig- survival (RFS) was defined as the time from the start of nificance (MGUS/MGCS), we did not take into account treatment until relapse or death, and patients who these malignancies in the “observed cases” to calculate the ® ® remained free from disease were censored at the date of SIR. SPSS (version 16.0) and R (version 3.5.3) were used ® ® the last observation. Excel and FileMaker software were for the statistical analyses. used for data collection. Results Treatments and evaluation of outcomes Patient characteristics Single-agent therapies included cladribine, pentostatin, Two hundred and seventy-nine patients from 19 French IFNα, and rituximab. Splenectomy was also considered a centers were analyzed, with a median follow-up of Blood Cancer Journal Paillassa et al. Blood Cancer Journal (2020) 10:62 Page 3 of 12 Table 1 Patient characteristics at baseline. PNAs are the treatment of choice for first-line treatment and treatment of the first relapse Age at HCL diagnosis (years), median [range] 59 [29–88] The median number of lines of treatment was 1 (range 0–7). The treatments received in each line of therapy are Hemoglobin (g/dL), median [range] 12 [3–16.7] shown in Table 2. Platelet count (×10 /L), median [range] 93.5 [7.4–503] As a first-line therapy, 208 patients (75%) received PNA, White blood cell count (×10 /L), median [range] 2.68 [0.5–107.2] either cladribine (159 patients: 57%) or pentostatin (49 patients: 18%). Fifty-nine patients received various other Neutrophil count (×10 /L), median [range] 0.99 [0.027–12.96] treatments (21%), with 40 patients treated with IFNα Hairy cells (%), median [range] 3 [0–94] (14%) (Table 2). Twelve patients (4%) never received any Flow cytometry analysis, n (%) 170 (61) treatment. Only the year of HCL diagnosis significantly CD25 , n (%) 118 (69) influenced the first-line treatment (more patients were treated with PNA after 2000 (p < 0.001)), which was not CD103 , n (%) 123 (72) the case for age, infection, Hb level, platelet count, or CD11c , n (%) 109 (64) neutrophil count at diagnosis. Infectious disease at diagnosis, n (%) 58 (21) With regard to second-line therapies, 77 patients (69%) were re-treated with PNAs: 59 patients with cladribine (53%) and 18 with pentostatin (16%). Twenty-five patients (22%) received other treatments, with 11 patients receiv- 127 months (range 2–413). All the patients had HCL ing IFNα (10%). Ten patients (9%) did not receive treat- diagnosed between 1980 and 2011. The characteristics of ment for a relapse. the 279 patients are listed in Table 1. The median age at The subsequent lines of treatment were heterogeneous HCL diagnosis was 59 years (range 29–88). Twenty-one (Table 2). Forty-four patients (16%) received 3 lines of percent of patients presented had an infectious disease at treatment, 16 patients (6%) received 4 lines, 9 patients the time of the diagnosis of HCL. The median hemoglobin (3%) received 5 lines, 4 patients (1%) received 6 lines, and (Hb) level was 12 g/dL, the median platelet count was 4 patients (1%) received 7 lines of treatment. Only one 93.5 × 10 /L, the median white blood cell count was patient received 4 cycles of vemurafenib (960 mg bid 2.68 × 10 /L, and the median neutrophil count was 0.99 × 21 days/28) as a third-line therapy, which was stopped after achieving a CR and developing invasive pulmonary 10 /L. The median percentage of bone marrow hairy cells evaluated in the bone marrow aspirations was 3% (range aspergillosis. Two years after stopping vemurafenib, the 0–94). Sixty-one percent of patients underwent flow patient was still in CR. cytometry analysis at diagnosis. There were personal histories of cancer in 31 patients, representing 11% of the PNA induced the highest CR rate and duration of patients (Supplementary Information 1). The median time response (DOR) between the personal history of cancer and the diagnosis The overall response rates (ORRs), including CR and PR of HCL was 4 years (range 0–44). There were 85 familial after first-line treatment, were 99% for all patients, 100% histories of malignancy (69 solid cancers and 16 hema- for patients treated with PNA, 99% for those treated with tological malignancies) in 63 patients, representing 23% of cladribine, 100% for those treated with pentostatin, 96% the patients (Supplementary Information 2). In total, 45, for those treated with other treatments (including IFNα), 15, and 3 patients had 1, 2, and 3 family members with and 90% for those treated exclusively with IFNα. The CR histories of malignancy, respectively. rates for the patients were 78%, 83%, 83%, 84%, 58%, and 50%, respectively. The CR rate was significantly lower for New events patients who received other treatments or IFNα (p < New events were defined as new relapses, death or 0.001) (Supplementary Information 4). second cancers occurring since the first analysis. In total, The DOR (median (min; max)) was 91 months (3; 335) 99/279 patients (36%) experienced at least one new event. for all patients who received first-line therapy, 94 months We observed 130 new events: 60 relapses (1 new relapse (4; 260) for those who received PNA, 94 months (8; 260) in 54 patients (19% of patients), 2 new relapses in 6 for those who received cladribine, 88 months (4; 251) for patients (2%)), 25 solid second cancers, 12 s hematological those who received pentostatin, and only 52 months (3; malignancies, and 33 new deaths. At the last follow-up, 335) for those who received other treatments. The dif- 229 patients were still alive: 193 were in CR (84%), 19 ference in the median DOR among patients receiving were in PR (8%), 10 had PD (5%), and 7 had an unknown cladribine, pentostatin and other treatments was statisti- disease status (3%) (Supplementary Information 3). cally significant (p = 0.02). The median DOR also sig- nificantly decreased with the line of treatment (p < 0.001) Blood Cancer Journal Paillassa et al. Blood Cancer Journal (2020) 10:62 Page 4 of 12 (Supplementary Information 5), which remained sig- Table 2 Treatments according to the line of treatment. nificant when we only considered treatment with PNAs 1st line treatment n = 279 (100%) (p < 0.001). PNA 208 (75) PNA provided the best RFS and the lowest CIR Cladribine 159 (57) Fifty patients died (Supplementary Information 1). Pentostatin 49 (18) Among the 33 known causes of death, second malig- nancies accounted for the most deaths (33%). The median Other 59 (21) OS was 27 years (328 months; 95% CI: 299; 357). IFN 40 After excluding the 12 patients who never received any IFN then cladribine 7 treatment and the patient for whom the first-line treat- Splenectomy 3 ment was unknown, the median OS was not reached for patients who received cladribine or pentostatin, whereas it Cladribine + rituximab 2 was 328 months (95% CI: 300–356) for patients treated Pentostatin then cladribine 2 with other treatments and 321 months (95% CI: 298–344) Splenectomy then pentostatin 1 for patients treated with IFNα (Fig. 1). Patients receiving IFN then pentostatin 1 cladribine had a significantly better OS than patients receiving pentostatin (log rank test, p = 0.039). Indeed, R-CHOP 1 the 5-year OS was 97% in patients receiving cladribine Pentostatin + rituximab 1 versus 86% in patients receiving pentostatin (Table 3). In Unknown 1 multivariate analysis (Supplementary Information 6), only No treatment 12 (4) age at diagnosis was a predictor of an inferior OS (p < 0.001, HR = 1.082, 95% CI: 1.044–1.121). 2nd line treatment n = 112 (100%) The median RFS after first-line treatment was 11 years (136 months (95% CI: 109; 163)): 163, 159, 55, and PNA 77 (69) 50 months after treatment with cladribine, pentostatin, Cladribine 59 (53) other treatments and IFNα, respectively (Supplementary Information 7). Patients treated with cladribine or pen- Pentostatin 18 (16) tostatin as a first-line therapy had a significantly better Other 25 (22) RFS than patients treated with other treatments or IFNα IFN 11 (log rank test, p < 0.001). However, the difference in RFS Rituximab 7 between patients treated with cladribine and those treated with pentostatin was not statistically significant. In mul- Cladribine + rituximab 3 tivariate analysis (Supplementary Information 8), the Pentostatin + rituximab 1 percentage of hairy cells at diagnosis as a continuous IFN + rituximab 1 variable (p = 0.006, HR 1.010, 95% CI: 1.003; 1.018) and IFN then pentostatin 1 using other first-line treatments (p = 0.004, HR 2.533, 95% CI: 1.349; 4.754) were both predictors of a worse RFS, Splenectomy 1 whereas achieving CR1 (p < 0.001, HR 0.355, 95% CI: No treatment 10 (9) 0.228; 0.555) was a predictor of a longer RFS. The proportions of patients with relapses after the first, 3rd line treatment n= 49 (100%) second, third, fourth, fifth, and sixth lines of treatment were 106/257 (41%), 49/102 (48%), 16/44 (36%), 9/16 PNA 29 (59) (56%), 5/9 (56%), and 4/4 (100%), respectively. With Cladribine 12 (24) regard to the first-line treatment, the proportions of Pentostatin 17 (35) patients with relapses after treatment with PNA, cla- Other 15 (31) dribine, pentostatin, and other treatments were 31%, 31%, 29%, and 80% (90% for IFNα), respectively. For all patients Rituximab 5 who responded to the first-line treatment, considering Cladribine + rituximab 4 death as a competing risk, the CIR increased with time IFN and was 3.5% (95% CI: 1.7; 6.3), 5.8% (95% CI: 3.4; 9.2), 21.1% (95% CI: 16.4; 26.4), and 39.4% (95% CI: 32.8; 45.9) at 1, 2, 5, and 10 years, respectively. Patients who received other treatments had a statistically significant higher CIR Blood Cancer Journal Paillassa et al. Blood Cancer Journal (2020) 10:62 Page 5 of 12 Table 2 continued than patients who received cladribine or pentostatin (Gray’s test , subdistribution hazard ratio (sdHR) = 44.6, 3rd line treatment n= 49 (100%) p < 0.001) (Fig. 2). The median RFS after second-line therapy (called RFS2) was 86 months (95% CI: 70; 102) for all patients. According to the second-line therapy used, the median Cladribine + IFN 1 RFS2 was 79, 116, and 40 months after cladribine, pen- tostatin, and other treatments, respectively. Patients who Rituximab + bendamustine then vemurafenib 1 received pentostatin had a significantly longer RFS2 than No treatment 5 (10) patients who received other treatments (log-rank test, p = 0.050) (Supplementary Information 9). To test the effect 4th line treatment n= 16 (100%) of switching one PNA for another, we analyzed patients who received PNA in the first and second lines (n = 44, PNA 9 (56) Supplementary Information 10). Among them, 31 patients Cladribine 6 (38) were treated with the same PNA in both lines (no Pentostatin 3 (18) switching: 30 patients cladribine then cladribine; 1 patient Other 7 (44) pentostatin then pentostatin), whereas 13 patients swit- ched PNAs (switching: 8 patients cladribine then pen- IFN 3 tostatin; 5 patients pentostatin then cladribine). Overall, Rituximab 2 5/44 patients (11%) experienced a relapse (4 patients who Pentostatin + rituximab 1 had received cladribine then cladribine and 1 patient who IFN + rituximab 1 had received pentostatin then cladribine). The median DOR was 48 months (range 4–120). There was no sta- 5th line treatment n = 9 tistically significant difference in terms of the median DOR between patients who switched (52 months) and PNA 5 those who did not (36 months). The median RFS was Cladribine 3 71 months. There was no statistically significant differ- ence in terms of the median RFS between patients who Pentostatin 2 switched (55 months) and those who did not (71 months) Other 4 (Supplementary Information 11). Thus, for the second- IFN 2 line treatment, we found that pentostatin had an advan- Fludarabine + rituximab 1 tage. We did not observe a significant difference in out- comes between patients who switched and those who R-CHOP 1 were re-treated with the same PNA according to these results based on the DOR and RFS. 6th line treatment n = 5 Table 3 summarizes the results after first-line treatment. PNA 2 Risk of infectious or immune complications Cladribine 1 Forty-three of 279 patients (15%) experienced at least Pentostatin 1 one infection during follow-up. Lower and upper Other 2 respiratory tract infections were the most frequent com- IFN 1 plications. Three patients experienced invasive pulmonary aspergillosis, including the patient treated with vemur- Splenectomy 1 afenib. A total of 7/279 patients (2.5%) had at least one No treatment 1 immune complication. We observed the following com- plications: 3 with vasculitis, 1 with anti-MAG neuropathy 7th line treatment n = 4 with cryoglobulin, 1 with sarcoidosis, 1 with acute poly- arthritis, 1 with rheumatoid arthritis, 1 with immune PNA 1 thrombocytopenia, and 1 with glomerulopathy with IgA Cladribine 1 mesangial deposits. Other 3 Risk of second cancers IFN 2 Twenty-one percent of patients (59/279) experienced at R-DHAX then HSCT 1 least 1 second cancer (68 s cancers), 17% (46/279) Blood Cancer Journal Paillassa et al. Blood Cancer Journal (2020) 10:62 Page 6 of 12 Fig. 1 Overall survival. a all 279 HCL patients, b in line with the first-line treatment: cladribine, pentostatin, and other treatments including IFNα, c in line with the first-line treatment: cladribine, pentostatin, and IFNα. Kaplan–Meier method. experienced at least 1 solid cancer (three of them had two death as a competing risk, the 10-year cumulative inci- successive solid cancers), and 6.8% (19/279) experienced a dences of all second cancers, solid cancers, and hemato- hematological malignancy. The most prevalent solid logical malignancies were 15% (95% CI: 11; 19), 11% (95% tumors were prostate and nonmelanoma skin cancers. CI: 7.2; 15), and 5.0% (95% CI: 2.8; 8.2), respectively (Fig. The most prevalent hematological malignancies were 3). Then, we performed univariate and multivariate ana- MGUS/MGCS (Table 4). The patient treated with lyses (Supplementary Information 12) using the Fine and vemurafenib did not experience any second malignancy. Gray regression model, considering death as a com- The median times between the diagnosis of HCL and all peting risk, and including age at HCL diagnosis, a familial second cancers, solid cancers and hematological malig- history of cancer, a personal history of cancer, treatment nancies were 81 months (range 0–374), 99 months (range with cladribine (regardless of the line of treatment), 0–374), and 78 months (range 2–262), respectively. The treatment with pentostatin, and treatment with IFNα as median ages at the diagnosis of all second cancers, solid covariates. In multivariate analysis, IFNα was a protective cancers or hematological malignancies were 70, 69, and factor against second cancers (p = 0.038, sdHR 0.529, 95% 77 years, respectively. When comparing the occurrence of CI: 0.290; 0.966), a familial history of cancer was a risk all second cancers, solid cancers and hematological factor for solid cancers (p = 0.017, sdHR 2.117, 95% CI: malignancies according to first-line treatment, the dif- 1.146; 3.910), and a personal history of cancer was a risk ferences were not statistically significant. Considering factor for hematological malignancies (p = 0.028, sdHR Blood Cancer Journal Paillassa et al. Blood Cancer Journal (2020) 10:62 Page 7 of 12 Table 3 Results after first-line treatment. Cladribine Pentostatin “Other” treatments All treated patients Follow-up (months), median (range) 116 (21–293) 126 (2–292) 221 (3–413) 128 (2–413) ORR, n (%) 157 (99) 45 (100) 52 (96) 255 (99) CR, n (%) 131 (83) 38 (84) 32 (58) 201 (78) Median RFS (months) 163 159 55 136 CIR 1 year (%) 2.5 4.4 5.5 3.5 5 years (%) 14 13 49 21 10 years (%) 33 30 66 39 Median OS (months) NR NR 328 321 OS, evaluable patients 1 year, n (%) 100 (158) 96 (47) 98 (57) 99 (263) 5 years, n (%) 97 (150) 86 (42) 95 (54) 95 (246) 10 years, n (%) 94 (77) 81 (25) 89 (43) 90 (145) 20 years, n (%) 69 (8) 71 (3) 82 (26) 76 (37) Fig. 2 Cumulative incidence of relapse (CIR): a all patients who responded to first-line treatment, b according to first-line treatment. 3.473, 95% CI: 1.144; 10.550). Only 1/5 of the patients patients had an excess of cancers (SIR: 2.22; 95% CI: with a personal history of hematological malignancies 1.61–2.83), solid cancers (SIR: 1.81; 95% CI: 1.24–2.38), developed a second hematological cancer during and hematological malignancies (SIR: 6.67; 95% CI: follow-up. Compared to the French population, our 3.04–10.30). Blood Cancer Journal Paillassa et al. Blood Cancer Journal (2020) 10:62 Page 8 of 12 Table 4 Second solid cancers (left) and second 50%, and the median DORs were 2.7, 2.5, 2.2, 1.6 and 1.3 hematological malignancies (right) observed during the years, respectively. follow-up period. Our study confirmed the good long-term prognosis of HCL. The median OS was 27 years, which is in line with Solid cancers n = 49 Hematological n= 19 duration reported in the literature (Supplementary malignancies 8,18,20,22,24,32–42 38 Information 13) . In a study by Else et al. including 233 patients with long-term follow-up, the 15- Prostate 11 MGUS/MGCS 6 year OS was 78%. In another study including 44 patients Nonmelanoma 11 MDS 4 treated with cladribine, the 12-year OS was 79% . Strik- skin cancer ingly, in our cohort, we found that patients had a better Lung 7 NHL 3 OS with cladribine than with pentostatin, which was not Colorectal 7 MDS/MPN 2 the case in our first analysis or in the literature. However, we did not find any difference in OS between first-line Kidney 4 MPN 1 treatments in multivariate analysis. One explanation Pancreas 2 AML 1 could be that patients who received cladribine were in Esophagus 1 MM 1 better condition than the patients treated with pentosta- Pleural 1 CLL 1 tin. We compared the characteristics at baseline of patients stratified by the first-line therapy they received, Unknown 1 but we did not find any statistically significant differences Kaposi 1 in terms of age or infectious disease at diagnosis between Biliary tract 1 the treatment groups. However, because of the retro- Bladder 1 spective nature of this study, with data obtained via responses to a questionnaire, we had no information Breast 1 about the performance status, fitness or comorbidities of 1 DLBCL, 1 FL, 1 SMZL. the patients. Only age at HCL diagnosis was a predictor of AML acute myeloid leukemia, MM multiple myeloma, CLL chronic lymphocytic leukemia, NHL non-Hodgkin’s lymphoma, DLBCL diffuse large B-cell lymphoma, OS in multivariate analysis. FL follicular lymphoma, SMZL splenic marginal zone lymphoma. The median RFS was 11 years, which was in line with the RFS observed in previous cohorts (Supplementary 8,18,20,22,24,32–42 Information 13) Discussion . In a review of retro- We analyzed 279 HCL patients, with a median follow- spective studies with a long-term follow-up including up of 10 years. HCL patients treated with PNA have a patients treated with PNA, the median RFS was 13–16 good long-term prognosis. However, relapses and second years . In our cohort, we found that first-line treatment malignancies are common. with PNAs, CR1 achievement, and a low percentage of In our cohort, PNAs remained the first choice for first- hairy cells in the bone marrow at diagnosis were pre- line treatment and treatment of the first relapse. PNAs dictors of a better RFS in multivariate analysis. In a ret- gave the best CR rate, DOR, and RFS, and were associated rospective study comparing three first-line therapies in 71 with the lowest relapse rate (Supplementary Information patients (31 cladribine, 19 IFNα, 16 splenectomy), the 8,18,20,22,24,32–42 8 13) . Madanat et al. analyzed the data of 61 progression-free survival (PFS) time was significantly patients with cladribine as the first-line treatment. The longer and the relapse rate was lower for patients treated ORR and CR rates were 97% and 78%, respectively. with cladribine than for those undergoing the other two Nineteen patients relapsed, 12 of whom received second- treatments . In a Spanish retrospective cohort of 107 line treatment with cladribine, resulting in an ORR of patients, the median treatment-free interval (TFI) was 8 37 36 83% . Hacioglu et al. described a cohort of 94 HCL shorter for patients in PR than for those in CR .We patients, most of whom received first-line treatment with confirmed that relapses remain an issue, with a 10-year cladribine. For the patients treated with cladribine, the CIR of 39% after first-line treatment. This highlights the ORR, CR and relapse rates were 97%, 81%, and 17%, issue of late relapses occurring several years after treat- respectively. Cladribine was the first treatment choice in ment. Moreover, PNAs were again found to be the best the second and third lines, with CR rates of 68% and 67%, choice, resulting in a lower CIR than other treatments. respectively . In our cohort, IFNα remained an important option; it In our cohort, the responses were shorter at each was used in 40/279 patients (14%) as the first-line therapy relapse. Zinzani et al. evaluated the long-term outcomes and in 11/112 patients (10%) as the second-line therapy. of 121 patients: the CR rates after the first, second, third, The outcomes after treatment with IFNα were inferior to fourth and fifth lines of treatment were 77, 74, 71, 65, and those obtained after treatment with PNAs in terms of the CR, RFS and relapse rates. However, there was no Blood Cancer Journal Paillassa et al. Blood Cancer Journal (2020) 10:62 Page 9 of 12 Fig. 3 Cumulative incidence of a second cancer, b second solid cancers, c second hematological malignancies. statistically significant difference in ORR or OS. In our primary cause of death. In some studies, the risk of second opinion, IFNα should be the preferred option for patients cancer was especially high when the cancer was a hema- with an active infection at diagnosis and those for whom tological malignancy, and this was also observed in our 48,49 PNAs are not an option. study . We found that a personal history of cancer was We did not find any difference in the outcomes of a risk factor for hematological malignancies. Our patients who switched or did not switch PNAs between hypothesis is that these patients have received che- the first and second lines of treatment. In their review, motherapy/radiotherapy, favoring the development of Else et al. described no difference in CR rates between hematological malignancies. patients who did and did not switch PNAs. In a long-term In other studies, HCL patients did not have a higher risk 21–24,40,43,50 follow-up of 233 patients with HCL initially treated with of developing a second cancer (Table 5) . Some pentostatin or cladribine, the CR rates, relapse rates, PFS, studies found an excess of risk only for certain cancers 51–53 and RFS were not significantly different between patients (Table 5) . who did and did not switch treatments after relapsing . In our study, IFNα was a protective factor against During follow-up, 15% of the patients experienced at least cancer in multivariate analysis. However, cladribine and one infection. This was probably underestimated because pentostatin were not risk factors for cancers, even if the infections were underdeclared. In a retrospective study, results of our multivariate analysis should be interpreted Damaj et al. included 73 patients, most of whom were with caution because most patients received PNAs. Of treated with IFNα as a first-line treatment. With a median note, ten patients with a second cancer did not receive follow-up of 13 years, 37% of the patients experienced at PNAs. Therefore, it seems that the risk of second malig- least one infection, and 15% experienced a severe infec- nancies might be related to HCL itself rather than to the tion . In our study, one patient developed tuberculosis, and treatments, which is in line with the findings of other three patients experienced pulmonary invasive aspergillosis. studies . According to the literature, HCL patients have a higher risk Our study has several strengths. First, the sample size was of mycobacterial infections, but invasive fungal diseases large considering the rarity of HCL. In addition, the follow- 33,45–47 (IFDs) have rarely been described . up was long, with a median of 10 years. Moreover, this Our study also showed that patients with HCL had a study showed the “real-life” experience of treatments for risk of developing second cancers. This risk was higher HCL, with the inclusion of non-selected patients. Finally, it than in the general population: the SIRs for all cancers, was a collaborative and multicentric study, including solid tumors, and hematological malignancies were 2.22, patients from 19 centers. However, the study has several 1.81, and 6.67, respectively. Several studies also found that limitations. First, it was a retrospective study with data HCL patients have a higher risk of developing second obtained via responses to a questionnaire. Therefore, there cancer than the general population (SIR 1.24–4.33, Table was a risk of information bias. In addition, there was also a 5). These studies also found relatively increased cancer- declaration bias regarding infections and autoimmune dis- 18–20,48,49 related mortality rates among HCL patients .In eases. Indeed, the incidence of autoimmune complications our cohort, we found that second cancers were the was ten times lower than that reported in the literature Blood Cancer Journal Paillassa et al. Blood Cancer Journal (2020) 10:62 Page 10 of 12 Table 5 (Part 1). Other HCL cohorts with cancer incidence during follow-up. (Part 2). Other HCL cohorts with cancer incidence during follow-up. Study Type n Treatments Follow-up Cumulative incidence of cancers SIR or observed/expected ratio (95% CI) Hisada et al. Retrospective 3104 6.5 years 32% at 25 years 1.24 (1.11; 1.37) Kampmeier et al. Prospective 69 IFNα 91 months 19% 4.33 Goodman et al. Retrospective 209 Cladribine ≥7 years 23% 2.03 (1.49; 2.71) Saven et al. Retrospective 358 Cladribine 58 months 8% 1.88 (1.24; 2.74) Au et al. Retrospective 117 Cladribine 68 months 24% 2.60 (1.82; 3.61) pentostatin splenectomy IFNα Paltiel et al. Retrospective 181 Cladribine 80 months 11% 1.3 (0.68; 2.28) for all cancers 3.23 (1.39; 6.36) for urogenital cancers Kurzrock et al. Retrospective 350 Cladribine 6 years 7.40% 1.34 for all cancers pentostatin 13.04 for myelomas IFNα 8.7 for lymphomas Federico et al. Retrospective 1022 14% at 15 years 1.01 (0.74; 1.33) for all cancers 5.3 (1.9; 11.5) for NHL Else et al. Retrospective 233 Pentostatin 16 years 12% (excluding nonmelanoma skin No data cladribine cancers) Flinn et al. Prospective 241 Pentostatin 9.3 years 16% 1.26 (0.86; 1.77) +/−IFNα Pawson et al. Retrospective 200 Cladribine 65 months 4% 1.29 (0.60; 2.65) pentostatin IFNα Maloisel et al. Retrospective 238 Pentostatin 63.5 months 7.60% 0.95 (0.5; 1.92) Rosenberg et al. Retrospective 88 Cladribine 21 years 9.10% 1.60 (0.80; 2.89) Watts et al. Retrospective 267 Cladribine 11% at 10 years (melanoma and non- 1.30 (0.78; 2.03) (for pentostatin melanoma skin cancers only) melanoma only) Troussard et al. Retrospective 107 IFNα 102 months 9.5% at 10 years 1.24 (0.54; 2.45) Getta et al. Retrospective 331 Cladribine 69 months Age ≤ 40: 21% at 10 years No data splenectomy Age > 40: 29% at 10 years IFNα Damaj et al. Retrospective 73 Cladribine 13 years 27+/−6% at 13 years No data pentostatin splenectomy IFNα Excess of second malignancies. Excess for some cancers only. No excess of second malignancies. No data about the excess of risk compared to the general population. 54–56 (2.5% versus 25%) . Moreover, we had no information outcomes, second malignancies and infections. Regarding about the proportion of HCLv patients, who have a worse the evaluation of response, there was a lack of standardi- prognosis than HCL patients. Regarding treatments, the zation among the different centers in terms of the type and number of cycles of pentostatin has rarely been reported. It timing of the evaluation, which led to a measurement bias. is unknown whether the number of cycles has an impact on In addition, there was a lack of statistical power in our Blood Cancer Journal Paillassa et al. Blood Cancer Journal (2020) 10:62 Page 11 of 12 25 26 multivariate analyses for the cumulative incidences of solid France. UFR Sante, Universite d’Angers, Angers, France. CRCINA, INSERM, Universite de Nantes, Universite d’Angers, Angers, France cancers and hematological malignancies due to the lack of events. Then, many cancers were nonmelanoma skin can- Conflict of interest cers and MGUS/MGRS. Some authors did not take into The authors declare that they have no conflict of interest. account these types of cancer . As the French cancer registries do not include these cancers in their databases, we Publisher’s note did not include them in the calculation of the SIR to avoid Springer Nature remains neutral with regard to jurisdictional claims in overestimation. Finally, we did not collect data in order to published maps and institutional affiliations. deal with analytic epidemiology. Some previous studies Supplementary Information accompanies this paper at (https://doi.org/ assessed for potential risk factors of developing HCL and 10.1038/s41408-020-0328-z). identified some of them: farming, pesticide exposure, diesel, petrol, and ionizing radiations. Surprisingly, tobacco seems Received: 15 March 2020 Revised: 2 May 2020 Accepted: 12 May 2020 to be a protective factor in several of these studies, even if 57–60 this should be confimed . In this 10-year follow-up analysis, we confirmed the References favorable prognosis of HCL. PNAs are the best choice of 1. Bouroncle, B. A.,Wiseman,B. K.& Doan, C.A.Leukemic reticuloendotheliosis. treatment and result in the best CR rate, DOR, RFS, and Blood 13,609–630 (2016). CIR. Relapses are still an issue, with a 10-year CIR of 39%. 2. Schrek,R.&Donnelly,W.J. ‘Hairy’ cells in blood in lymphoreticular neoplastic disease and ‘flagellated’ cells of normal lymph nodes. Blood 27,199–211 In our cohort, there was no difference between patients (1966). who switched or did not switch between PNAs from the 3. Tadmor, T. & Polliack, A. Epidemiology and environmental risk in hairy cell first to the second line. Interestingly, we found a relatively leukemia. Best Pract. Res. Clin. Haematol. 28,175–179 (2015). 4. Mintz, U. & Golomb, H. M. Splenectomy as initial therapy in twenty-six patients higher risk of solid cancers and hematological malig- with leukemic reticuloendotheliosis (hairy cell leukemia). Cancer Res. 39, nancies in HCL patients: it seems that PNAs are not a risk 2366–2370 (1979). factor for second cancers. Few prospective studies exist 5. Quesada, J. R.,Reuben, J.,Manning,J.T., Hersh, E. M. &Gutterman,J.U.Alpha interferon for induction of remissioninhairy-cellleukemia. N. Engl.J.Med. 310, for HCL patients . Therefore, we plan to perform a 15–18 (2000). prospective national cohort study that will include 6. Piro,L.D., Carrera,C.J., Carson,D.A.&Beutler, E. Lasting remissions in hairy-cell patients with HCLc, HCLv, and splenic diffuse red pulp leukemia induced by a single infusion of 2-chlorodeoxyadenosine. N. Engl. J. Med. 322,1117–1121 (1990). small B-cell lymphoma (SDRPL). 7. Grever, M. et al. Randomized comparison of pentostatin versus interferon alfa- 2a in previously untreated patients with hairy cell leukemia: an intergroup Acknowledgements study. J. Clin. Oncol. 13,974–982 (1995). We would like to thank the participating centers for their responses and J.R., V. 8. Madanat, Y. F. et al. Long-term outcomes of hairy cell leukemia treated with L., and M.B. for their help with statistics. purine analogs: a comparison with the general population. Clin. Lymphoma Myeloma Leuk. 17,857–862 (2017). Author details 9. Troussard, X. & Cornet, E. Hairy cell leukemia 2018: update on diagnosis, risk- 1 2 Service des Maladies du Sang, CHU d’Angers, Angers, France. Laboratoire stratification, and treatment. Am.J.Hematol. 92, 1382–1390 (2017). d’Hematologie Biologique, CHU de Caen, Caen, France. Service d’Oncologie 10. Lauria, F. et al. Efficacy of anti-CD20 monoclonal antibodies (Mabthera) in Hematologique et Therapie Cellulaire, CHU de Poitiers, et CIC Inserm U1402, patients with progressed hairy cell leukemia. Haematologica 86,1046–1050 Poitiers, France. Inserm U1245 et Service d’Hematologie, Centre Henri (1997). Becquerel et Normandie Univ UNIROUEN, Rouen, France. Service 11. Chihara,D.etal. Long-termdurable remission by cladribine followed by d’Hematologie Clinique, CHU Henri Mondor, Assistance Publique des Hopitaux rituximab in patients with hairy cell leukaemia: update of a phase II trial. Br. J. de Paris, Creteil, France. Service d’Hematologie Clinique, CH Sud Francilien, Haematol. 174,760–766 (2016). Corbeil Essonnes, France. Service d’Hematologie, CHU de Grenoble, Grenoble, 12. Tiacci,E.etal. BRAF mutationsin hairy-cellleukemia. N. Engl. J. Med. 364, France. Service Hemato-Oncologie, Hopital Saint Louis, Assistance Publique 2305–2315 (2011). des Hopitaux de Paris, Paris, France. Unite d’Investigation Clinique, 13. Tiacci, E. et al. Targeting mutant BRAF in relapsed or refractory hairy-cell Departement de Medecine Oncologique, Institut Curie, Paris, France. Service leukemia. N. Engl. J. Med. 373,1733–1747 (2015). d’Hematologie Clinique, Institut Curie, Paris, France. Service d’Hematologie 14. Tiacci, E. et al. Vemurafenib plus rituximab in hairy cell leukemia: a promising Clinique, Groupe Hospitalier Regional de Mulhouse, Mulhouse, France. chemotherapy-free regimen for relapsed or refractory patients. Blood 128, Service d’Hematologie Clinique et Therapie Cellulaire, CHU de Limoges, 1214 (2016). Limoges, France. Service de Medecine Interne et Immunologie Clinique, 15. Kreitman, R. J. et al. Treatment with combination of dabrafenib and trametinib Hopital Bicêtre, Assistance Publique des Hopitaux de Paris, Paris, France. in patients with recurrent/refractory BRAF V600E-mutated hairy cell leukemia Service d’Hematologie Adulte, Hopital Necker-Enfants Malades, Assistance (HCL). Blood 132, 391 (2018). Publique des Hopitaux de Paris, Paris, France. Service d’Hematologie Clinique, 16. Jones,J.A. et al. Safety andefficacy of the bruton tyrosine kinase inhibitor CHU de Nice, Nice, France. Service d’Hematologie, CHRU de Nancy, Nancy, ibrutinib in patients with hairy cell leukemia: interim results of a phase 2 study. France. Service de Medecine Interne, Maladies Infectieuses, Oncologie et Blood 100, 313 (2016). Hematologie, CH d’Auch, Auch, France. Service de Medecine Interne, CHI 17. Kreitman, R. J. et al. Moxetumomab pasudotox in relapsed/refractory hairy cell Frejus Saint Raphaël, Frejus, France. Laboratoire d’Hematologie Biologique, leukemia. Leukemia 32, 1768–1777 (2018). CHU de Brest, Brest, France. Service d’Hematologie Oncologie, CH de 18. Goodman, G. R., Burian, C., Koziol, J. A. & Saven, A. Extended follow-up of Beauvais, Beauvais, France. Unite de Recherche Clinique, Hopital Avicenne, patients with hairy cell leukemia after treatment with cladribine. J. Clin. Oncol. Assistance Publique des Hopitaux de Paris, Bobigny, France. MINT UMR 21,891–896 (2003). INSERM 1066, CNRS 6021, Universite d’Angers, Angers, France. Institut 19. Au, W. Y. et al. Second malignancies in patients with hairy cell leukemia in d’Hematologie de Basse-Normandie, CHU de Caen, Caen, France. Federation British Columbia: a 20-year experience. Blood 92,1160–1164 (1998). Hospitalo-Universitaire ‘Grand Ouest Against Leukemia’ (FHU GOAL), Angers, Blood Cancer Journal Paillassa et al. Blood Cancer Journal (2020) 10:62 Page 12 of 12 20. Saven,A., Burian,C., Koziol,J.A. & Piro, L.D.Long-term follow-up of patients 40. Rosenberg,J.D., Burian,C., Waalen,J.&Saven, A. Clinical characteristics and with hairy cell leukemia after cladribine treatment. Blood 92,1918–1926 (1998). long-term outcome of young hairy cell leukemia patients treated with cla- 21. Pawson, R., A’Hern,R.&Catovsky,D.Secondmalignancyinhairy cell leu- dribine: a single-institution series. Blood 123,177–183 (2014). kaemia: no evidence of increased incidence after treatment with interferon 41. Johnston, J. B. et al. Long-term outcome following treatment of hairy cell alpha. Leuk. Lymphoma 22,103–106 (1996). leukemia with pentostatin (Nipent): A National Cancer Institute of Canada 22. Maloisel, F. et al. Long-term outcome with pentostatin treatment in hairy cell study. Semin. Oncol. 27,32–36 (2000). leukemia patients. A French retrospective study of 238 patients. Leukemia 17, 42. Getta, B. M. et al. Treatment outcomes and secondary cancer incidence in 45–51 (2003). young patients with hairy cell leukaemia. Br. J. Haematol. 175, 402–409 23. Troussard,X., Henry-Amar,M. & Flandrin,G.Secondcancerriskafter interferon (2016). therapy? Blood 84, 3242–3244 (1994). 43. Else, M., Dearden, C. E. & Catovsky, D. Long-term follow-up after purine ana- 24. Flinn, I. W. et al. Long-term follow-up of remission duration, mortality, and logue therapy in hairy cell leukaemia. Best Pract. Res. Clin. Haematol. 28, second malignancies in hairy cell leukemia patients treated with pentostatin. 217–229 (2015). Blood 96, 2981–2986 (2000). 44. Damaj, G. et al. Risk factors for severe infection in patients with hairy cell 25. Nielsen, B., Braide, I. & Hasselbalch, H. Evidence for an association between leukemia: A long-term study of 73 patients. Eur. J. Haematol. 83,246–250 hairy cell leukemia and renal cell and colorectal carcinoma. Cancer 70, (2009). 2087–2090 (1992). 45. Golomb, H. M. & Hadad, L. J. Infectious complications in 127 patients with 26. Cornet, E. et al. Long-term follow-up and second malignancies in 487 patients hairy cell leukemia. Am.J.Hematol. 16,393–401 (1984). with hairy cell leukaemia. Br.J.Haematol. 166,390–400 (2014). 46. Green,L., Coumbe,A., Sawicka, E. & DeLord, C. Mycobacterium kansasii in a 27. Catovsky, D., Quesada, J. R., Golomb, H. M. & Golde, D. N. Consensus resolution: patient with hairy cell leukaemia. Br. J. Haematol. 144,2 (2009). proposed criteria for evaluation of response to treatment in hairy cell leuke- 47. Ramasamy, C. et al. Atypical mycobacterial infectioninhairy cell leu- mia. Leukemia 1, 405 (2003). kemia treated with cladribine. Indian J. Hematol. Blood Transfus 30, 28. Fine, J. P. & Gray, R. J. A proportional hazards model for the subdistribution of a 59–61 (2014). competing risk. J. Am. Stat. Assoc. 94,496–509 (1999). 48. Hisada, M.,Chen, B. E.,Jaffe,E.S.& Travis,L.B.Secondcancerincidence and 29. Binder-Foucard, F. et al. Cancer incidence and mortality in France over the cause-specificmortalityamong 3104 patients with hairycellleukemia: a 1980–2012 period: solid tumors. Rev. Epidemiol. Sante Publique 62,95–108 population-based study. J. Natl Cancer Inst. 99,215–222 (2007). (2014). 49. Kampmeier, P. et al. Increased incidence of second neoplasms in patients 30. Le Guyader-Peyrou, S. et al. Cancer incidence in France over the 1980–2012 treated with interferon alpha 2b for hairy cell leukemia: a clinicopathologic period: hematological malignancies. Rev. Epidemiol. Sante Publique 64, assessment. Blood 83, 2931–2938 (1994). 103–112 (2016). 50. Watts,J.M.etal. Melanoma andnon-melanoma skin cancers in hairy cell 31. Gray, R. J. A class of K-sample tests for comparing the cumulative incidence of leukaemia: a surveillance, epidemiology and end results population analysis a competing risk. Ann. Stat. 16, 1141–1154 (1988). and the 30-year experience at memorial sloan kettering cancer center. Br. J. 32. Bastie, J. N. et al. Five years follow-up after 2-chloro deoxyadenosine treatment Haematol. 171,84–90 (2015). in thirty patients with hairy cell leukemia: evaluation of minimal residual dis- 51. Paltiel, O., Adler, B., Barchana, M. & Dann, E. J. A population-based study of ease and CD4+ lymphocytopenia after treatment. Leuk. Lymphoma 35, hairy cell leukemia in Israel. Eur. J. Haematol. 77,372–377 (2006). 555–565 (1999). 52. Kurzrock, R. et al. Second cancer risk in hairy cell leukemia: analysis of 350 33. Öngören, Ş. et al. Retrospective evaluation of hairy cell leukemia patients patients. J. Clin. Oncol. 15, 1803–1810 (1997). treated with three different first-line treatment modalities in the last two 53. Federico, M. et al. Risk of second cancer in patients with hairy cell leukemia: decades: a single-center experience. Turk.J.Haematol. 34,291–299 (2017). Long-term follow-up. J. Clin. Oncol. 20, 638–646 (2002). 34. Jehn, U., Bartl, R., Dietzfelbinger, H., Haferlach, T. & Heinemann, V. An update: 54. Tadmor, T. & Polliack, A. Hairy cell leukemia: uncommon clinical features, 12-year follow-up of patients with hairy cell leukemia following treatment with unusual sites of involvement and some rare associations. Best Pract. Res. Clin. 2-chlorodeoxyadenosine. Leukemia 18,1476–1481 (2004). Haematol. 28,193–199 (2015). 35. Găman, A. M. Hairy cell leukemia-a rare type of leukemia. A retrospective study 55. Anderson, L. A. & Engels, E. A. Autoimmune conditions and hairy cell leukemia: on 39 patients. Rom. J. Morphol. Embryol. 54,575–579 (2018). an exploratory case–control study. J. Hematol. Oncol. 3, 35 (2010). 36. López Rubio, M. et al. Hairy cell leukemia treated initially with purine analogs: a 56. Dasanu, C. A., Van den Bergh, M., Pepito, D. & Alvarez Argote, J. Autoimmune retrospective study of 107 patients from the Spanish Cooperative Group on disorders in patients with hairy cell leukemia: are they more common than Chronic Lymphocytic Leukemia (GELLC). Leuk. Lymphoma 55, 1007–1012 previously thought? Curr.Med.Res. Opin. 31,17–23 (2014). (2013). 57. Hardell, L., Eriksson, M. & Nordstrom, M. Exposure to pesticides as risk factor for 37. Hacioglu, S. et al. Multicenter retrospective analysis regarding the clinical non-Hodgkin’s lymphoma and hairy cell leukemia: pooled analysis of two manifestations and treatment results in patients with hairy cell leukemia: Swedish case-control studies. Leuk. Lymphoma 43,1043–1049 (2002). twenty-four year Turkish experience in cladribine therapy. Hematol. Oncol. 33, 58. Monnereau, A. et al. Medical history, lifestyle, and occupational risk factors for 192–198 (2014). hairy cell leukemia: the interlymph non-Hodgkin lymphoma subtypes project. 38. Else, M. et al. Long-term follow-up of 233 patients with hairy cell leukaemia, J. Natl Cancer Inst. Monogr. 2014,115–124 (2014). treated initially with pentostatin or cladribine, at a median of 16 years from 59. Clavel, J. et al. Hairy cell leukaemia, occupation, and smoking. Br.J.Haematol. diagnosis. Br.J.Haematol. 145,733–740 (2009). 91,154–161 (1995). 39. Zinzani, P. L. et al. Hairy cell leukemia: evaluation of the long-term outcome in 60. Clavel, J. et al. Farming, pesticide use and hairy-cell leukemia. Scand. J. Work 121 patients. Cancer 116, 4788–4792 (2010). Environ. Health 22,285–293 (1996). Blood Cancer Journal http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Blood Cancer Journal Springer Journals

Loading next page...
 
/lp/springer-journals/analysis-of-a-cohort-of-279-patients-with-hairy-cell-leukemia-hcl-10-JvzPVy0lh4

References (60)

  • JN Bastie (1999)

    555

    Leuk. Lymphoma, 35

  • BM Getta (2016)

    402

    Br. J. Haematol., 175

  • YF Madanat (2017)

    857

    Clin. Lymphoma Myeloma Leuk., 17

  • GR Goodman (2003)

    891

    J. Clin. Oncol., 21

  • R Schrek (1966)

    199

    Blood, 27

  • E Cornet (2014)

    390

    Br. J. Haematol., 166

  • RJ Kreitman (2018)

    391

    Blood, 132

  • A Saven (1998)

    1918

    Blood, 92

  • O Paltiel (2006)

    372

    Eur. J. Haematol., 77

  • JR Quesada (2000)

    15

    N. Engl. J. Med., 310

  • U Mintz (1979)

    2366

    Cancer Res., 39

  • JP Fine (1999)

    496

    J. Am. Stat. Assoc., 94

  • T Tadmor (2015)

    175

    Best Pract. Res. Clin. Haematol., 28

  • PL Zinzani (2010)

    4788

    Cancer, 116

  • E Tiacci (2016)

    1214

    Blood, 128

  • F Maloisel (2003)

    45

    Leukemia, 17

  • J Clavel (1995)

    154

    Br. J. Haematol., 91

  • JM Watts (2015)

    84

    Br. J. Haematol., 171

  • JB Johnston (2000)

    32

    Semin. Oncol., 27

  • WY Au (1998)

    1160

    Blood, 92

  • C Ramasamy (2014)

    59

    Indian J. Hematol. Blood Transfus, 30

  • F Lauria (1997)

    1046

    Haematologica, 86

  • AM Găman (2018)

    575

    Rom. J. Morphol. Embryol., 54

  • RJ Kreitman (2018)

    1768

    Leukemia, 32

  • B Nielsen (1992)

    2087

    Cancer, 70

  • G Damaj (2009)

    246

    Eur. J. Haematol., 83

  • L Hardell (2002)

    1043

    Leuk. Lymphoma, 43

  • M López Rubio (2013)

    1007

    Leuk. Lymphoma, 55

  • M Hisada (2007)

    215

    J. Natl Cancer Inst., 99

  • E Tiacci (2011)

    2305

    N. Engl. J. Med., 364

  • D Chihara (2016)

    760

    Br. J. Haematol., 174

  • RJ Gray (1988)

    1141

    Ann. Stat., 16

  • P Kampmeier (1994)

    2931

    Blood, 83

  • M Grever (1995)

    974

    J. Clin. Oncol., 13

  • JD Rosenberg (2014)

    177

    Blood, 123

  • X Troussard (1994)

    3242

    Blood, 84

  • IW Flinn (2000)

    2981

    Blood, 96

  • M Federico (2002)

    638

    J. Clin. Oncol., 20

  • T Tadmor (2015)

    193

    Best Pract. Res. Clin. Haematol., 28

  • J Clavel (1996)

    285

    Scand. J. Work Environ. Health, 22

  • BA Bouroncle (2016)

    609

    Blood, 13

  • E Tiacci (2015)

    1733

    N. Engl. J. Med., 373

  • U Jehn (2004)

    1476

    Leukemia, 18

  • X Troussard (2017)

    1382

    Am. J. Hematol., 92

  • F Binder-Foucard (2014)

    95

    Rev. Epidemiol. Sante Publique, 62

  • D Catovsky (2003)

    405

    Leukemia, 1

  • LA Anderson (2010)

    35

    J. Hematol. Oncol., 3

  • CA Dasanu (2014)

    17

    Curr. Med. Res. Opin., 31

  • Ş Öngören (2017)

    291

    Turk. J. Haematol., 34

  • S Hacioglu (2014)

    192

    Hematol. Oncol., 33

  • LD Piro (1990)

    1117

    N. Engl. J. Med., 322

  • JA Jones (2016)

    313

    Blood, 100

  • A Monnereau (2014)

    115

    J. Natl Cancer Inst. Monogr., 2014

  • M Else (2015)

    217

    Best Pract. Res. Clin. Haematol., 28

  • R Pawson (1996)

    103

    Leuk. Lymphoma, 22

  • S Le Guyader-Peyrou (2016)

    103

    Rev. Epidemiol. Sante Publique, 64

  • HM Golomb (1984)

    393

    Am. J. Hematol., 16

  • M Else (2009)

    733

    Br. J. Haematol., 145

  • R Kurzrock (1997)

    1803

    J. Clin. Oncol., 15

  • L Green (2009)

    2

    Br. J. Haematol., 144

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2020
eISSN
2044-5385
DOI
10.1038/s41408-020-0328-z
Publisher site
See Article on Publisher Site

Abstract

In total, 279 patients with hairy-cell leukemia (HCL) were analyzed, with a median follow-up of 10 years. Data were collected up to June 2018. We analyzed responses to treatment, relapses, survival, and the occurrence of second malignancies during follow-up. The median age was 59 years. In total, 208 patients (75%) were treated with purine analogs (PNAs), either cladribine (159) or pentosatin (49), as the first-line therapy. After a median follow-up of 127 months, the median overall survival was 27 years, and the median relapse-free survival (RFS) was 11 years. The cumulative 10-year relapse incidence was 39%. In patients receiving second-line therapy, the median RFS was 7 years. For the second-line therapy, using the same or another PNA was equivalent. We identified 68 second malignancies in 59 patients: 49 solid cancers and 19 hematological malignancies. The 10-year cumulative incidences of cancers, solid tumors, and hematological malignancies were 15%, 11%, and 5.0%, respectively, and the standardized incidence ratios were 2.22, 1.81, and 6.67, respectively. In multivariate analysis, PNA was not a risk factor for second malignancies. HCL patients have a good long-term prognosis. PNAs are the first-line treatment. HCL patients require long-term follow-up because of their relatively increased risk of second malignancies. Introduction survival (OS) of 90% with PNA treatment . However, the Hairy cell leukemia (HCL) is a rare B-cell chronic management of HCL patients remains under investiga- lymphoproliferative disorder characterized by atypical tion, particularly the care of patients with relapsed/ lymphoid cells with hairy projections in the peripheral refractory disease and the evaluation of the risk of second 1,2 blood, bone marrow, spleen, and/or liver . HCL is malignancies. For the first relapse, using the same PNA or 3 4 9 responsible for 2% of leukemias . Splenectomy and switching to another PNA may be effective . Anti-CD20 5 10 interferon-alpha (IFNα) were the first-line treatments, monoclonal antibodies (rituximab), alone or associated 6 11 and purine analogs (PNAs), either cladribine or pentos- with PNA , can also be alternative treatments for relap- tatin , were subsequently introduced. HCL prognosis ses. With the recent identification of the BRAF V600E improved consistently over time, with a 10-year overall mutation in most classic HCL (HCLc) , BRAF inhibitors, namely, vemurafenib or dabrafenib, could be indi- 13,14 15 cated . MEK inhibitors (trametinib) , BCR pathway Correspondence: Xavier Troussard (troussard-x@chu-caen.fr) inhibitors (ibrutinib) , and anti-CD22 immunotoxins Service des Maladies du Sang, CHU d’Angers, Angers, France (moxetumomab pasudotox) are the newest therapeutic Laboratoire d’Hematologie Biologique, CHU de Caen, Caen, France Full list of author information is available at the end of the article © The Author(s) 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to theCreativeCommons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Blood Cancer Journal 1234567890():,; 1234567890():,; 1234567890():,; 1234567890():,; Paillassa et al. Blood Cancer Journal (2020) 10:62 Page 2 of 12 alternatives. The risk of second malignancies occurring single treatment modality if not associated with adjuvant during follow-up in HCL patients is controversial, with drug therapy. Multiple-agent therapies were defined by some studies describing a higher risk of cancer than in the the use of more than one drug within a period of 18–20 general population and others describing no 6 months. Responses were defined according to the 21–24 27 increased risk . The reasons for these discrepancies Consensus Resolution . CR required the morphologic might be the variability in the methods used to define absence of hairy cells in peripheral blood and bone mar- second malignancies with pooling of second cancers row aspiration or biopsy specimens and the normalization occurring before and after HCL diagnosis in some of any organomegaly and cytopenia. Immunophenotypic studies . analysis of peripheral blood or bone marrow biopsy was To answer this question, we investigated a large cohort not required. PR was defined as the normalization of the of 279 HCL patients, with a 10-year median follow-up peripheral counts associated with at least a 50% reduction period, and we analyzed the treatments, responses, sur- in organomegaly and bone marrow hairy cells and <5% vival, relapses, and occurrence of second cancers. circulating hairy cells. All other outcomes were con- sidered non-responses. Relapse was defined as any dete- Subjects and methods rioration in blood counts related to the detection of hairy Patients cells in the peripheral blood and bone marrow. The eligibility criteria for the HCL diagnosis were established according to the WHO 2008 and 2016 clas- Statistical analyses sifications, including morphological and flow cytometric Survival curves were drawn according to the analyses of blood, bone marrow or tissue specimens. Each Kaplan–Meier method. OS and RFS were compared using patient signed an informed consent form. The study was the log-rank test. The cumulative incidence of relapse performed in accordance with the Declaration of Helsinki. (CIR) and cumulative incidence of second cancer were evaluated considering death as a competing risk. The Study design and data collection CIRs were compared according to the first-line treatment In the first analysis, we collected data up to 2012 .We with Gray’s test. Multivariate analyses of OS and RFS were then updated the data up to June 2018. A questionnaire performed using the Cox regression model, and multi- was sent to the physicians, who were members of the variate analyses for the cumulative incidence of second French National Society of Hematology (SFH), with malignancies were performed using the Fine and Gray requests for the following data: date of last observation; regression model, considering death as a competing risk. last disease status: complete response (CR), partial An excess of second malignancies was expressed by the response (PR), or progressive disease (PD); date of relapse standardized incidence ratio (SIR), which was defined as (s); treatments and responses (CR, PR, failure); treatment the ratio between the number of observed and expected start and end dates; second solid cancers (date, histology); cases from the general population in France. The expec- hematological malignancies (date, WHO 2016 classifica- ted malignancy rates were calculated from the age- tion); death (date, cause); and other complications. The indexed (in 5-year categories) cancer incidence in 29,30 second cancers were defined as either synchronous can- France . These incidence rates were multiplied by the cers or metachronous cancers. Synchronous cancers were observed person-years at risk in each age category to cancers that occurred at the same time as the diagnosis of calculate the expected frequency of second malignancies. HCL and those occurring within two months, as recom- Confidence intervals (CIs) of the SIR were obtained by mended by the Surveillance, Epidemiology, and End assuming a Poisson distribution for the observed num- Results (SEER) Program. Metachronous cancers were bers. The SIR was calculated for second cancers, second defined by cancers occurring more than two months after solid cancers, and second hematological malignancies. As the diagnosis of HCL. Overall survival (OS) was defined as the French registries did not include nonmelanoma skin the time from the date of HCL diagnosis until death from cancers or monoclonal gammopathies of undetermined any cause or the date of the last observation. Relapse-free significance/monoclonal gammopathies of clinical sig- survival (RFS) was defined as the time from the start of nificance (MGUS/MGCS), we did not take into account treatment until relapse or death, and patients who these malignancies in the “observed cases” to calculate the ® ® remained free from disease were censored at the date of SIR. SPSS (version 16.0) and R (version 3.5.3) were used ® ® the last observation. Excel and FileMaker software were for the statistical analyses. used for data collection. Results Treatments and evaluation of outcomes Patient characteristics Single-agent therapies included cladribine, pentostatin, Two hundred and seventy-nine patients from 19 French IFNα, and rituximab. Splenectomy was also considered a centers were analyzed, with a median follow-up of Blood Cancer Journal Paillassa et al. Blood Cancer Journal (2020) 10:62 Page 3 of 12 Table 1 Patient characteristics at baseline. PNAs are the treatment of choice for first-line treatment and treatment of the first relapse Age at HCL diagnosis (years), median [range] 59 [29–88] The median number of lines of treatment was 1 (range 0–7). The treatments received in each line of therapy are Hemoglobin (g/dL), median [range] 12 [3–16.7] shown in Table 2. Platelet count (×10 /L), median [range] 93.5 [7.4–503] As a first-line therapy, 208 patients (75%) received PNA, White blood cell count (×10 /L), median [range] 2.68 [0.5–107.2] either cladribine (159 patients: 57%) or pentostatin (49 patients: 18%). Fifty-nine patients received various other Neutrophil count (×10 /L), median [range] 0.99 [0.027–12.96] treatments (21%), with 40 patients treated with IFNα Hairy cells (%), median [range] 3 [0–94] (14%) (Table 2). Twelve patients (4%) never received any Flow cytometry analysis, n (%) 170 (61) treatment. Only the year of HCL diagnosis significantly CD25 , n (%) 118 (69) influenced the first-line treatment (more patients were treated with PNA after 2000 (p < 0.001)), which was not CD103 , n (%) 123 (72) the case for age, infection, Hb level, platelet count, or CD11c , n (%) 109 (64) neutrophil count at diagnosis. Infectious disease at diagnosis, n (%) 58 (21) With regard to second-line therapies, 77 patients (69%) were re-treated with PNAs: 59 patients with cladribine (53%) and 18 with pentostatin (16%). Twenty-five patients (22%) received other treatments, with 11 patients receiv- 127 months (range 2–413). All the patients had HCL ing IFNα (10%). Ten patients (9%) did not receive treat- diagnosed between 1980 and 2011. The characteristics of ment for a relapse. the 279 patients are listed in Table 1. The median age at The subsequent lines of treatment were heterogeneous HCL diagnosis was 59 years (range 29–88). Twenty-one (Table 2). Forty-four patients (16%) received 3 lines of percent of patients presented had an infectious disease at treatment, 16 patients (6%) received 4 lines, 9 patients the time of the diagnosis of HCL. The median hemoglobin (3%) received 5 lines, 4 patients (1%) received 6 lines, and (Hb) level was 12 g/dL, the median platelet count was 4 patients (1%) received 7 lines of treatment. Only one 93.5 × 10 /L, the median white blood cell count was patient received 4 cycles of vemurafenib (960 mg bid 2.68 × 10 /L, and the median neutrophil count was 0.99 × 21 days/28) as a third-line therapy, which was stopped after achieving a CR and developing invasive pulmonary 10 /L. The median percentage of bone marrow hairy cells evaluated in the bone marrow aspirations was 3% (range aspergillosis. Two years after stopping vemurafenib, the 0–94). Sixty-one percent of patients underwent flow patient was still in CR. cytometry analysis at diagnosis. There were personal histories of cancer in 31 patients, representing 11% of the PNA induced the highest CR rate and duration of patients (Supplementary Information 1). The median time response (DOR) between the personal history of cancer and the diagnosis The overall response rates (ORRs), including CR and PR of HCL was 4 years (range 0–44). There were 85 familial after first-line treatment, were 99% for all patients, 100% histories of malignancy (69 solid cancers and 16 hema- for patients treated with PNA, 99% for those treated with tological malignancies) in 63 patients, representing 23% of cladribine, 100% for those treated with pentostatin, 96% the patients (Supplementary Information 2). In total, 45, for those treated with other treatments (including IFNα), 15, and 3 patients had 1, 2, and 3 family members with and 90% for those treated exclusively with IFNα. The CR histories of malignancy, respectively. rates for the patients were 78%, 83%, 83%, 84%, 58%, and 50%, respectively. The CR rate was significantly lower for New events patients who received other treatments or IFNα (p < New events were defined as new relapses, death or 0.001) (Supplementary Information 4). second cancers occurring since the first analysis. In total, The DOR (median (min; max)) was 91 months (3; 335) 99/279 patients (36%) experienced at least one new event. for all patients who received first-line therapy, 94 months We observed 130 new events: 60 relapses (1 new relapse (4; 260) for those who received PNA, 94 months (8; 260) in 54 patients (19% of patients), 2 new relapses in 6 for those who received cladribine, 88 months (4; 251) for patients (2%)), 25 solid second cancers, 12 s hematological those who received pentostatin, and only 52 months (3; malignancies, and 33 new deaths. At the last follow-up, 335) for those who received other treatments. The dif- 229 patients were still alive: 193 were in CR (84%), 19 ference in the median DOR among patients receiving were in PR (8%), 10 had PD (5%), and 7 had an unknown cladribine, pentostatin and other treatments was statisti- disease status (3%) (Supplementary Information 3). cally significant (p = 0.02). The median DOR also sig- nificantly decreased with the line of treatment (p < 0.001) Blood Cancer Journal Paillassa et al. Blood Cancer Journal (2020) 10:62 Page 4 of 12 (Supplementary Information 5), which remained sig- Table 2 Treatments according to the line of treatment. nificant when we only considered treatment with PNAs 1st line treatment n = 279 (100%) (p < 0.001). PNA 208 (75) PNA provided the best RFS and the lowest CIR Cladribine 159 (57) Fifty patients died (Supplementary Information 1). Pentostatin 49 (18) Among the 33 known causes of death, second malig- nancies accounted for the most deaths (33%). The median Other 59 (21) OS was 27 years (328 months; 95% CI: 299; 357). IFN 40 After excluding the 12 patients who never received any IFN then cladribine 7 treatment and the patient for whom the first-line treat- Splenectomy 3 ment was unknown, the median OS was not reached for patients who received cladribine or pentostatin, whereas it Cladribine + rituximab 2 was 328 months (95% CI: 300–356) for patients treated Pentostatin then cladribine 2 with other treatments and 321 months (95% CI: 298–344) Splenectomy then pentostatin 1 for patients treated with IFNα (Fig. 1). Patients receiving IFN then pentostatin 1 cladribine had a significantly better OS than patients receiving pentostatin (log rank test, p = 0.039). Indeed, R-CHOP 1 the 5-year OS was 97% in patients receiving cladribine Pentostatin + rituximab 1 versus 86% in patients receiving pentostatin (Table 3). In Unknown 1 multivariate analysis (Supplementary Information 6), only No treatment 12 (4) age at diagnosis was a predictor of an inferior OS (p < 0.001, HR = 1.082, 95% CI: 1.044–1.121). 2nd line treatment n = 112 (100%) The median RFS after first-line treatment was 11 years (136 months (95% CI: 109; 163)): 163, 159, 55, and PNA 77 (69) 50 months after treatment with cladribine, pentostatin, Cladribine 59 (53) other treatments and IFNα, respectively (Supplementary Information 7). Patients treated with cladribine or pen- Pentostatin 18 (16) tostatin as a first-line therapy had a significantly better Other 25 (22) RFS than patients treated with other treatments or IFNα IFN 11 (log rank test, p < 0.001). However, the difference in RFS Rituximab 7 between patients treated with cladribine and those treated with pentostatin was not statistically significant. In mul- Cladribine + rituximab 3 tivariate analysis (Supplementary Information 8), the Pentostatin + rituximab 1 percentage of hairy cells at diagnosis as a continuous IFN + rituximab 1 variable (p = 0.006, HR 1.010, 95% CI: 1.003; 1.018) and IFN then pentostatin 1 using other first-line treatments (p = 0.004, HR 2.533, 95% CI: 1.349; 4.754) were both predictors of a worse RFS, Splenectomy 1 whereas achieving CR1 (p < 0.001, HR 0.355, 95% CI: No treatment 10 (9) 0.228; 0.555) was a predictor of a longer RFS. The proportions of patients with relapses after the first, 3rd line treatment n= 49 (100%) second, third, fourth, fifth, and sixth lines of treatment were 106/257 (41%), 49/102 (48%), 16/44 (36%), 9/16 PNA 29 (59) (56%), 5/9 (56%), and 4/4 (100%), respectively. With Cladribine 12 (24) regard to the first-line treatment, the proportions of Pentostatin 17 (35) patients with relapses after treatment with PNA, cla- Other 15 (31) dribine, pentostatin, and other treatments were 31%, 31%, 29%, and 80% (90% for IFNα), respectively. For all patients Rituximab 5 who responded to the first-line treatment, considering Cladribine + rituximab 4 death as a competing risk, the CIR increased with time IFN and was 3.5% (95% CI: 1.7; 6.3), 5.8% (95% CI: 3.4; 9.2), 21.1% (95% CI: 16.4; 26.4), and 39.4% (95% CI: 32.8; 45.9) at 1, 2, 5, and 10 years, respectively. Patients who received other treatments had a statistically significant higher CIR Blood Cancer Journal Paillassa et al. Blood Cancer Journal (2020) 10:62 Page 5 of 12 Table 2 continued than patients who received cladribine or pentostatin (Gray’s test , subdistribution hazard ratio (sdHR) = 44.6, 3rd line treatment n= 49 (100%) p < 0.001) (Fig. 2). The median RFS after second-line therapy (called RFS2) was 86 months (95% CI: 70; 102) for all patients. According to the second-line therapy used, the median Cladribine + IFN 1 RFS2 was 79, 116, and 40 months after cladribine, pen- tostatin, and other treatments, respectively. Patients who Rituximab + bendamustine then vemurafenib 1 received pentostatin had a significantly longer RFS2 than No treatment 5 (10) patients who received other treatments (log-rank test, p = 0.050) (Supplementary Information 9). To test the effect 4th line treatment n= 16 (100%) of switching one PNA for another, we analyzed patients who received PNA in the first and second lines (n = 44, PNA 9 (56) Supplementary Information 10). Among them, 31 patients Cladribine 6 (38) were treated with the same PNA in both lines (no Pentostatin 3 (18) switching: 30 patients cladribine then cladribine; 1 patient Other 7 (44) pentostatin then pentostatin), whereas 13 patients swit- ched PNAs (switching: 8 patients cladribine then pen- IFN 3 tostatin; 5 patients pentostatin then cladribine). Overall, Rituximab 2 5/44 patients (11%) experienced a relapse (4 patients who Pentostatin + rituximab 1 had received cladribine then cladribine and 1 patient who IFN + rituximab 1 had received pentostatin then cladribine). The median DOR was 48 months (range 4–120). There was no sta- 5th line treatment n = 9 tistically significant difference in terms of the median DOR between patients who switched (52 months) and PNA 5 those who did not (36 months). The median RFS was Cladribine 3 71 months. There was no statistically significant differ- ence in terms of the median RFS between patients who Pentostatin 2 switched (55 months) and those who did not (71 months) Other 4 (Supplementary Information 11). Thus, for the second- IFN 2 line treatment, we found that pentostatin had an advan- Fludarabine + rituximab 1 tage. We did not observe a significant difference in out- comes between patients who switched and those who R-CHOP 1 were re-treated with the same PNA according to these results based on the DOR and RFS. 6th line treatment n = 5 Table 3 summarizes the results after first-line treatment. PNA 2 Risk of infectious or immune complications Cladribine 1 Forty-three of 279 patients (15%) experienced at least Pentostatin 1 one infection during follow-up. Lower and upper Other 2 respiratory tract infections were the most frequent com- IFN 1 plications. Three patients experienced invasive pulmonary aspergillosis, including the patient treated with vemur- Splenectomy 1 afenib. A total of 7/279 patients (2.5%) had at least one No treatment 1 immune complication. We observed the following com- plications: 3 with vasculitis, 1 with anti-MAG neuropathy 7th line treatment n = 4 with cryoglobulin, 1 with sarcoidosis, 1 with acute poly- arthritis, 1 with rheumatoid arthritis, 1 with immune PNA 1 thrombocytopenia, and 1 with glomerulopathy with IgA Cladribine 1 mesangial deposits. Other 3 Risk of second cancers IFN 2 Twenty-one percent of patients (59/279) experienced at R-DHAX then HSCT 1 least 1 second cancer (68 s cancers), 17% (46/279) Blood Cancer Journal Paillassa et al. Blood Cancer Journal (2020) 10:62 Page 6 of 12 Fig. 1 Overall survival. a all 279 HCL patients, b in line with the first-line treatment: cladribine, pentostatin, and other treatments including IFNα, c in line with the first-line treatment: cladribine, pentostatin, and IFNα. Kaplan–Meier method. experienced at least 1 solid cancer (three of them had two death as a competing risk, the 10-year cumulative inci- successive solid cancers), and 6.8% (19/279) experienced a dences of all second cancers, solid cancers, and hemato- hematological malignancy. The most prevalent solid logical malignancies were 15% (95% CI: 11; 19), 11% (95% tumors were prostate and nonmelanoma skin cancers. CI: 7.2; 15), and 5.0% (95% CI: 2.8; 8.2), respectively (Fig. The most prevalent hematological malignancies were 3). Then, we performed univariate and multivariate ana- MGUS/MGCS (Table 4). The patient treated with lyses (Supplementary Information 12) using the Fine and vemurafenib did not experience any second malignancy. Gray regression model, considering death as a com- The median times between the diagnosis of HCL and all peting risk, and including age at HCL diagnosis, a familial second cancers, solid cancers and hematological malig- history of cancer, a personal history of cancer, treatment nancies were 81 months (range 0–374), 99 months (range with cladribine (regardless of the line of treatment), 0–374), and 78 months (range 2–262), respectively. The treatment with pentostatin, and treatment with IFNα as median ages at the diagnosis of all second cancers, solid covariates. In multivariate analysis, IFNα was a protective cancers or hematological malignancies were 70, 69, and factor against second cancers (p = 0.038, sdHR 0.529, 95% 77 years, respectively. When comparing the occurrence of CI: 0.290; 0.966), a familial history of cancer was a risk all second cancers, solid cancers and hematological factor for solid cancers (p = 0.017, sdHR 2.117, 95% CI: malignancies according to first-line treatment, the dif- 1.146; 3.910), and a personal history of cancer was a risk ferences were not statistically significant. Considering factor for hematological malignancies (p = 0.028, sdHR Blood Cancer Journal Paillassa et al. Blood Cancer Journal (2020) 10:62 Page 7 of 12 Table 3 Results after first-line treatment. Cladribine Pentostatin “Other” treatments All treated patients Follow-up (months), median (range) 116 (21–293) 126 (2–292) 221 (3–413) 128 (2–413) ORR, n (%) 157 (99) 45 (100) 52 (96) 255 (99) CR, n (%) 131 (83) 38 (84) 32 (58) 201 (78) Median RFS (months) 163 159 55 136 CIR 1 year (%) 2.5 4.4 5.5 3.5 5 years (%) 14 13 49 21 10 years (%) 33 30 66 39 Median OS (months) NR NR 328 321 OS, evaluable patients 1 year, n (%) 100 (158) 96 (47) 98 (57) 99 (263) 5 years, n (%) 97 (150) 86 (42) 95 (54) 95 (246) 10 years, n (%) 94 (77) 81 (25) 89 (43) 90 (145) 20 years, n (%) 69 (8) 71 (3) 82 (26) 76 (37) Fig. 2 Cumulative incidence of relapse (CIR): a all patients who responded to first-line treatment, b according to first-line treatment. 3.473, 95% CI: 1.144; 10.550). Only 1/5 of the patients patients had an excess of cancers (SIR: 2.22; 95% CI: with a personal history of hematological malignancies 1.61–2.83), solid cancers (SIR: 1.81; 95% CI: 1.24–2.38), developed a second hematological cancer during and hematological malignancies (SIR: 6.67; 95% CI: follow-up. Compared to the French population, our 3.04–10.30). Blood Cancer Journal Paillassa et al. Blood Cancer Journal (2020) 10:62 Page 8 of 12 Table 4 Second solid cancers (left) and second 50%, and the median DORs were 2.7, 2.5, 2.2, 1.6 and 1.3 hematological malignancies (right) observed during the years, respectively. follow-up period. Our study confirmed the good long-term prognosis of HCL. The median OS was 27 years, which is in line with Solid cancers n = 49 Hematological n= 19 duration reported in the literature (Supplementary malignancies 8,18,20,22,24,32–42 38 Information 13) . In a study by Else et al. including 233 patients with long-term follow-up, the 15- Prostate 11 MGUS/MGCS 6 year OS was 78%. In another study including 44 patients Nonmelanoma 11 MDS 4 treated with cladribine, the 12-year OS was 79% . Strik- skin cancer ingly, in our cohort, we found that patients had a better Lung 7 NHL 3 OS with cladribine than with pentostatin, which was not Colorectal 7 MDS/MPN 2 the case in our first analysis or in the literature. However, we did not find any difference in OS between first-line Kidney 4 MPN 1 treatments in multivariate analysis. One explanation Pancreas 2 AML 1 could be that patients who received cladribine were in Esophagus 1 MM 1 better condition than the patients treated with pentosta- Pleural 1 CLL 1 tin. We compared the characteristics at baseline of patients stratified by the first-line therapy they received, Unknown 1 but we did not find any statistically significant differences Kaposi 1 in terms of age or infectious disease at diagnosis between Biliary tract 1 the treatment groups. However, because of the retro- Bladder 1 spective nature of this study, with data obtained via responses to a questionnaire, we had no information Breast 1 about the performance status, fitness or comorbidities of 1 DLBCL, 1 FL, 1 SMZL. the patients. Only age at HCL diagnosis was a predictor of AML acute myeloid leukemia, MM multiple myeloma, CLL chronic lymphocytic leukemia, NHL non-Hodgkin’s lymphoma, DLBCL diffuse large B-cell lymphoma, OS in multivariate analysis. FL follicular lymphoma, SMZL splenic marginal zone lymphoma. The median RFS was 11 years, which was in line with the RFS observed in previous cohorts (Supplementary 8,18,20,22,24,32–42 Information 13) Discussion . In a review of retro- We analyzed 279 HCL patients, with a median follow- spective studies with a long-term follow-up including up of 10 years. HCL patients treated with PNA have a patients treated with PNA, the median RFS was 13–16 good long-term prognosis. However, relapses and second years . In our cohort, we found that first-line treatment malignancies are common. with PNAs, CR1 achievement, and a low percentage of In our cohort, PNAs remained the first choice for first- hairy cells in the bone marrow at diagnosis were pre- line treatment and treatment of the first relapse. PNAs dictors of a better RFS in multivariate analysis. In a ret- gave the best CR rate, DOR, and RFS, and were associated rospective study comparing three first-line therapies in 71 with the lowest relapse rate (Supplementary Information patients (31 cladribine, 19 IFNα, 16 splenectomy), the 8,18,20,22,24,32–42 8 13) . Madanat et al. analyzed the data of 61 progression-free survival (PFS) time was significantly patients with cladribine as the first-line treatment. The longer and the relapse rate was lower for patients treated ORR and CR rates were 97% and 78%, respectively. with cladribine than for those undergoing the other two Nineteen patients relapsed, 12 of whom received second- treatments . In a Spanish retrospective cohort of 107 line treatment with cladribine, resulting in an ORR of patients, the median treatment-free interval (TFI) was 8 37 36 83% . Hacioglu et al. described a cohort of 94 HCL shorter for patients in PR than for those in CR .We patients, most of whom received first-line treatment with confirmed that relapses remain an issue, with a 10-year cladribine. For the patients treated with cladribine, the CIR of 39% after first-line treatment. This highlights the ORR, CR and relapse rates were 97%, 81%, and 17%, issue of late relapses occurring several years after treat- respectively. Cladribine was the first treatment choice in ment. Moreover, PNAs were again found to be the best the second and third lines, with CR rates of 68% and 67%, choice, resulting in a lower CIR than other treatments. respectively . In our cohort, IFNα remained an important option; it In our cohort, the responses were shorter at each was used in 40/279 patients (14%) as the first-line therapy relapse. Zinzani et al. evaluated the long-term outcomes and in 11/112 patients (10%) as the second-line therapy. of 121 patients: the CR rates after the first, second, third, The outcomes after treatment with IFNα were inferior to fourth and fifth lines of treatment were 77, 74, 71, 65, and those obtained after treatment with PNAs in terms of the CR, RFS and relapse rates. However, there was no Blood Cancer Journal Paillassa et al. Blood Cancer Journal (2020) 10:62 Page 9 of 12 Fig. 3 Cumulative incidence of a second cancer, b second solid cancers, c second hematological malignancies. statistically significant difference in ORR or OS. In our primary cause of death. In some studies, the risk of second opinion, IFNα should be the preferred option for patients cancer was especially high when the cancer was a hema- with an active infection at diagnosis and those for whom tological malignancy, and this was also observed in our 48,49 PNAs are not an option. study . We found that a personal history of cancer was We did not find any difference in the outcomes of a risk factor for hematological malignancies. Our patients who switched or did not switch PNAs between hypothesis is that these patients have received che- the first and second lines of treatment. In their review, motherapy/radiotherapy, favoring the development of Else et al. described no difference in CR rates between hematological malignancies. patients who did and did not switch PNAs. In a long-term In other studies, HCL patients did not have a higher risk 21–24,40,43,50 follow-up of 233 patients with HCL initially treated with of developing a second cancer (Table 5) . Some pentostatin or cladribine, the CR rates, relapse rates, PFS, studies found an excess of risk only for certain cancers 51–53 and RFS were not significantly different between patients (Table 5) . who did and did not switch treatments after relapsing . In our study, IFNα was a protective factor against During follow-up, 15% of the patients experienced at least cancer in multivariate analysis. However, cladribine and one infection. This was probably underestimated because pentostatin were not risk factors for cancers, even if the infections were underdeclared. In a retrospective study, results of our multivariate analysis should be interpreted Damaj et al. included 73 patients, most of whom were with caution because most patients received PNAs. Of treated with IFNα as a first-line treatment. With a median note, ten patients with a second cancer did not receive follow-up of 13 years, 37% of the patients experienced at PNAs. Therefore, it seems that the risk of second malig- least one infection, and 15% experienced a severe infec- nancies might be related to HCL itself rather than to the tion . In our study, one patient developed tuberculosis, and treatments, which is in line with the findings of other three patients experienced pulmonary invasive aspergillosis. studies . According to the literature, HCL patients have a higher risk Our study has several strengths. First, the sample size was of mycobacterial infections, but invasive fungal diseases large considering the rarity of HCL. In addition, the follow- 33,45–47 (IFDs) have rarely been described . up was long, with a median of 10 years. Moreover, this Our study also showed that patients with HCL had a study showed the “real-life” experience of treatments for risk of developing second cancers. This risk was higher HCL, with the inclusion of non-selected patients. Finally, it than in the general population: the SIRs for all cancers, was a collaborative and multicentric study, including solid tumors, and hematological malignancies were 2.22, patients from 19 centers. However, the study has several 1.81, and 6.67, respectively. Several studies also found that limitations. First, it was a retrospective study with data HCL patients have a higher risk of developing second obtained via responses to a questionnaire. Therefore, there cancer than the general population (SIR 1.24–4.33, Table was a risk of information bias. In addition, there was also a 5). These studies also found relatively increased cancer- declaration bias regarding infections and autoimmune dis- 18–20,48,49 related mortality rates among HCL patients .In eases. Indeed, the incidence of autoimmune complications our cohort, we found that second cancers were the was ten times lower than that reported in the literature Blood Cancer Journal Paillassa et al. Blood Cancer Journal (2020) 10:62 Page 10 of 12 Table 5 (Part 1). Other HCL cohorts with cancer incidence during follow-up. (Part 2). Other HCL cohorts with cancer incidence during follow-up. Study Type n Treatments Follow-up Cumulative incidence of cancers SIR or observed/expected ratio (95% CI) Hisada et al. Retrospective 3104 6.5 years 32% at 25 years 1.24 (1.11; 1.37) Kampmeier et al. Prospective 69 IFNα 91 months 19% 4.33 Goodman et al. Retrospective 209 Cladribine ≥7 years 23% 2.03 (1.49; 2.71) Saven et al. Retrospective 358 Cladribine 58 months 8% 1.88 (1.24; 2.74) Au et al. Retrospective 117 Cladribine 68 months 24% 2.60 (1.82; 3.61) pentostatin splenectomy IFNα Paltiel et al. Retrospective 181 Cladribine 80 months 11% 1.3 (0.68; 2.28) for all cancers 3.23 (1.39; 6.36) for urogenital cancers Kurzrock et al. Retrospective 350 Cladribine 6 years 7.40% 1.34 for all cancers pentostatin 13.04 for myelomas IFNα 8.7 for lymphomas Federico et al. Retrospective 1022 14% at 15 years 1.01 (0.74; 1.33) for all cancers 5.3 (1.9; 11.5) for NHL Else et al. Retrospective 233 Pentostatin 16 years 12% (excluding nonmelanoma skin No data cladribine cancers) Flinn et al. Prospective 241 Pentostatin 9.3 years 16% 1.26 (0.86; 1.77) +/−IFNα Pawson et al. Retrospective 200 Cladribine 65 months 4% 1.29 (0.60; 2.65) pentostatin IFNα Maloisel et al. Retrospective 238 Pentostatin 63.5 months 7.60% 0.95 (0.5; 1.92) Rosenberg et al. Retrospective 88 Cladribine 21 years 9.10% 1.60 (0.80; 2.89) Watts et al. Retrospective 267 Cladribine 11% at 10 years (melanoma and non- 1.30 (0.78; 2.03) (for pentostatin melanoma skin cancers only) melanoma only) Troussard et al. Retrospective 107 IFNα 102 months 9.5% at 10 years 1.24 (0.54; 2.45) Getta et al. Retrospective 331 Cladribine 69 months Age ≤ 40: 21% at 10 years No data splenectomy Age > 40: 29% at 10 years IFNα Damaj et al. Retrospective 73 Cladribine 13 years 27+/−6% at 13 years No data pentostatin splenectomy IFNα Excess of second malignancies. Excess for some cancers only. No excess of second malignancies. No data about the excess of risk compared to the general population. 54–56 (2.5% versus 25%) . Moreover, we had no information outcomes, second malignancies and infections. Regarding about the proportion of HCLv patients, who have a worse the evaluation of response, there was a lack of standardi- prognosis than HCL patients. Regarding treatments, the zation among the different centers in terms of the type and number of cycles of pentostatin has rarely been reported. It timing of the evaluation, which led to a measurement bias. is unknown whether the number of cycles has an impact on In addition, there was a lack of statistical power in our Blood Cancer Journal Paillassa et al. Blood Cancer Journal (2020) 10:62 Page 11 of 12 25 26 multivariate analyses for the cumulative incidences of solid France. UFR Sante, Universite d’Angers, Angers, France. CRCINA, INSERM, Universite de Nantes, Universite d’Angers, Angers, France cancers and hematological malignancies due to the lack of events. Then, many cancers were nonmelanoma skin can- Conflict of interest cers and MGUS/MGRS. Some authors did not take into The authors declare that they have no conflict of interest. account these types of cancer . As the French cancer registries do not include these cancers in their databases, we Publisher’s note did not include them in the calculation of the SIR to avoid Springer Nature remains neutral with regard to jurisdictional claims in overestimation. Finally, we did not collect data in order to published maps and institutional affiliations. deal with analytic epidemiology. Some previous studies Supplementary Information accompanies this paper at (https://doi.org/ assessed for potential risk factors of developing HCL and 10.1038/s41408-020-0328-z). identified some of them: farming, pesticide exposure, diesel, petrol, and ionizing radiations. Surprisingly, tobacco seems Received: 15 March 2020 Revised: 2 May 2020 Accepted: 12 May 2020 to be a protective factor in several of these studies, even if 57–60 this should be confimed . In this 10-year follow-up analysis, we confirmed the References favorable prognosis of HCL. PNAs are the best choice of 1. Bouroncle, B. A.,Wiseman,B. K.& Doan, C.A.Leukemic reticuloendotheliosis. treatment and result in the best CR rate, DOR, RFS, and Blood 13,609–630 (2016). CIR. Relapses are still an issue, with a 10-year CIR of 39%. 2. Schrek,R.&Donnelly,W.J. ‘Hairy’ cells in blood in lymphoreticular neoplastic disease and ‘flagellated’ cells of normal lymph nodes. Blood 27,199–211 In our cohort, there was no difference between patients (1966). who switched or did not switch between PNAs from the 3. Tadmor, T. & Polliack, A. Epidemiology and environmental risk in hairy cell first to the second line. Interestingly, we found a relatively leukemia. Best Pract. Res. Clin. Haematol. 28,175–179 (2015). 4. Mintz, U. & Golomb, H. M. Splenectomy as initial therapy in twenty-six patients higher risk of solid cancers and hematological malig- with leukemic reticuloendotheliosis (hairy cell leukemia). Cancer Res. 39, nancies in HCL patients: it seems that PNAs are not a risk 2366–2370 (1979). factor for second cancers. Few prospective studies exist 5. Quesada, J. R.,Reuben, J.,Manning,J.T., Hersh, E. M. &Gutterman,J.U.Alpha interferon for induction of remissioninhairy-cellleukemia. N. Engl.J.Med. 310, for HCL patients . Therefore, we plan to perform a 15–18 (2000). prospective national cohort study that will include 6. Piro,L.D., Carrera,C.J., Carson,D.A.&Beutler, E. Lasting remissions in hairy-cell patients with HCLc, HCLv, and splenic diffuse red pulp leukemia induced by a single infusion of 2-chlorodeoxyadenosine. N. Engl. J. Med. 322,1117–1121 (1990). small B-cell lymphoma (SDRPL). 7. Grever, M. et al. Randomized comparison of pentostatin versus interferon alfa- 2a in previously untreated patients with hairy cell leukemia: an intergroup Acknowledgements study. J. Clin. Oncol. 13,974–982 (1995). We would like to thank the participating centers for their responses and J.R., V. 8. Madanat, Y. F. et al. Long-term outcomes of hairy cell leukemia treated with L., and M.B. for their help with statistics. purine analogs: a comparison with the general population. Clin. Lymphoma Myeloma Leuk. 17,857–862 (2017). Author details 9. Troussard, X. & Cornet, E. Hairy cell leukemia 2018: update on diagnosis, risk- 1 2 Service des Maladies du Sang, CHU d’Angers, Angers, France. Laboratoire stratification, and treatment. Am.J.Hematol. 92, 1382–1390 (2017). d’Hematologie Biologique, CHU de Caen, Caen, France. Service d’Oncologie 10. Lauria, F. et al. Efficacy of anti-CD20 monoclonal antibodies (Mabthera) in Hematologique et Therapie Cellulaire, CHU de Poitiers, et CIC Inserm U1402, patients with progressed hairy cell leukemia. Haematologica 86,1046–1050 Poitiers, France. Inserm U1245 et Service d’Hematologie, Centre Henri (1997). Becquerel et Normandie Univ UNIROUEN, Rouen, France. Service 11. Chihara,D.etal. Long-termdurable remission by cladribine followed by d’Hematologie Clinique, CHU Henri Mondor, Assistance Publique des Hopitaux rituximab in patients with hairy cell leukaemia: update of a phase II trial. Br. J. de Paris, Creteil, France. Service d’Hematologie Clinique, CH Sud Francilien, Haematol. 174,760–766 (2016). Corbeil Essonnes, France. Service d’Hematologie, CHU de Grenoble, Grenoble, 12. Tiacci,E.etal. BRAF mutationsin hairy-cellleukemia. N. Engl. J. Med. 364, France. Service Hemato-Oncologie, Hopital Saint Louis, Assistance Publique 2305–2315 (2011). des Hopitaux de Paris, Paris, France. Unite d’Investigation Clinique, 13. Tiacci, E. et al. Targeting mutant BRAF in relapsed or refractory hairy-cell Departement de Medecine Oncologique, Institut Curie, Paris, France. Service leukemia. N. Engl. J. Med. 373,1733–1747 (2015). d’Hematologie Clinique, Institut Curie, Paris, France. Service d’Hematologie 14. Tiacci, E. et al. Vemurafenib plus rituximab in hairy cell leukemia: a promising Clinique, Groupe Hospitalier Regional de Mulhouse, Mulhouse, France. chemotherapy-free regimen for relapsed or refractory patients. Blood 128, Service d’Hematologie Clinique et Therapie Cellulaire, CHU de Limoges, 1214 (2016). Limoges, France. Service de Medecine Interne et Immunologie Clinique, 15. Kreitman, R. J. et al. Treatment with combination of dabrafenib and trametinib Hopital Bicêtre, Assistance Publique des Hopitaux de Paris, Paris, France. in patients with recurrent/refractory BRAF V600E-mutated hairy cell leukemia Service d’Hematologie Adulte, Hopital Necker-Enfants Malades, Assistance (HCL). Blood 132, 391 (2018). Publique des Hopitaux de Paris, Paris, France. Service d’Hematologie Clinique, 16. Jones,J.A. et al. Safety andefficacy of the bruton tyrosine kinase inhibitor CHU de Nice, Nice, France. Service d’Hematologie, CHRU de Nancy, Nancy, ibrutinib in patients with hairy cell leukemia: interim results of a phase 2 study. France. Service de Medecine Interne, Maladies Infectieuses, Oncologie et Blood 100, 313 (2016). Hematologie, CH d’Auch, Auch, France. Service de Medecine Interne, CHI 17. Kreitman, R. J. et al. Moxetumomab pasudotox in relapsed/refractory hairy cell Frejus Saint Raphaël, Frejus, France. Laboratoire d’Hematologie Biologique, leukemia. Leukemia 32, 1768–1777 (2018). CHU de Brest, Brest, France. Service d’Hematologie Oncologie, CH de 18. Goodman, G. R., Burian, C., Koziol, J. A. & Saven, A. Extended follow-up of Beauvais, Beauvais, France. Unite de Recherche Clinique, Hopital Avicenne, patients with hairy cell leukemia after treatment with cladribine. J. Clin. Oncol. Assistance Publique des Hopitaux de Paris, Bobigny, France. MINT UMR 21,891–896 (2003). INSERM 1066, CNRS 6021, Universite d’Angers, Angers, France. Institut 19. Au, W. Y. et al. Second malignancies in patients with hairy cell leukemia in d’Hematologie de Basse-Normandie, CHU de Caen, Caen, France. Federation British Columbia: a 20-year experience. Blood 92,1160–1164 (1998). Hospitalo-Universitaire ‘Grand Ouest Against Leukemia’ (FHU GOAL), Angers, Blood Cancer Journal Paillassa et al. Blood Cancer Journal (2020) 10:62 Page 12 of 12 20. Saven,A., Burian,C., Koziol,J.A. & Piro, L.D.Long-term follow-up of patients 40. Rosenberg,J.D., Burian,C., Waalen,J.&Saven, A. Clinical characteristics and with hairy cell leukemia after cladribine treatment. Blood 92,1918–1926 (1998). long-term outcome of young hairy cell leukemia patients treated with cla- 21. Pawson, R., A’Hern,R.&Catovsky,D.Secondmalignancyinhairy cell leu- dribine: a single-institution series. Blood 123,177–183 (2014). kaemia: no evidence of increased incidence after treatment with interferon 41. Johnston, J. B. et al. Long-term outcome following treatment of hairy cell alpha. Leuk. Lymphoma 22,103–106 (1996). leukemia with pentostatin (Nipent): A National Cancer Institute of Canada 22. Maloisel, F. et al. Long-term outcome with pentostatin treatment in hairy cell study. Semin. Oncol. 27,32–36 (2000). leukemia patients. A French retrospective study of 238 patients. Leukemia 17, 42. Getta, B. M. et al. Treatment outcomes and secondary cancer incidence in 45–51 (2003). young patients with hairy cell leukaemia. Br. J. Haematol. 175, 402–409 23. Troussard,X., Henry-Amar,M. & Flandrin,G.Secondcancerriskafter interferon (2016). therapy? Blood 84, 3242–3244 (1994). 43. Else, M., Dearden, C. E. & Catovsky, D. Long-term follow-up after purine ana- 24. Flinn, I. W. et al. Long-term follow-up of remission duration, mortality, and logue therapy in hairy cell leukaemia. Best Pract. Res. Clin. Haematol. 28, second malignancies in hairy cell leukemia patients treated with pentostatin. 217–229 (2015). Blood 96, 2981–2986 (2000). 44. Damaj, G. et al. Risk factors for severe infection in patients with hairy cell 25. Nielsen, B., Braide, I. & Hasselbalch, H. Evidence for an association between leukemia: A long-term study of 73 patients. Eur. J. Haematol. 83,246–250 hairy cell leukemia and renal cell and colorectal carcinoma. Cancer 70, (2009). 2087–2090 (1992). 45. Golomb, H. M. & Hadad, L. J. Infectious complications in 127 patients with 26. Cornet, E. et al. Long-term follow-up and second malignancies in 487 patients hairy cell leukemia. Am.J.Hematol. 16,393–401 (1984). with hairy cell leukaemia. Br.J.Haematol. 166,390–400 (2014). 46. Green,L., Coumbe,A., Sawicka, E. & DeLord, C. Mycobacterium kansasii in a 27. Catovsky, D., Quesada, J. R., Golomb, H. M. & Golde, D. N. Consensus resolution: patient with hairy cell leukaemia. Br. J. Haematol. 144,2 (2009). proposed criteria for evaluation of response to treatment in hairy cell leuke- 47. Ramasamy, C. et al. Atypical mycobacterial infectioninhairy cell leu- mia. Leukemia 1, 405 (2003). kemia treated with cladribine. Indian J. Hematol. Blood Transfus 30, 28. Fine, J. P. & Gray, R. J. A proportional hazards model for the subdistribution of a 59–61 (2014). competing risk. J. Am. Stat. Assoc. 94,496–509 (1999). 48. Hisada, M.,Chen, B. E.,Jaffe,E.S.& Travis,L.B.Secondcancerincidence and 29. Binder-Foucard, F. et al. Cancer incidence and mortality in France over the cause-specificmortalityamong 3104 patients with hairycellleukemia: a 1980–2012 period: solid tumors. Rev. Epidemiol. Sante Publique 62,95–108 population-based study. J. Natl Cancer Inst. 99,215–222 (2007). (2014). 49. Kampmeier, P. et al. Increased incidence of second neoplasms in patients 30. Le Guyader-Peyrou, S. et al. Cancer incidence in France over the 1980–2012 treated with interferon alpha 2b for hairy cell leukemia: a clinicopathologic period: hematological malignancies. Rev. Epidemiol. Sante Publique 64, assessment. Blood 83, 2931–2938 (1994). 103–112 (2016). 50. Watts,J.M.etal. Melanoma andnon-melanoma skin cancers in hairy cell 31. Gray, R. J. A class of K-sample tests for comparing the cumulative incidence of leukaemia: a surveillance, epidemiology and end results population analysis a competing risk. Ann. Stat. 16, 1141–1154 (1988). and the 30-year experience at memorial sloan kettering cancer center. Br. J. 32. Bastie, J. N. et al. Five years follow-up after 2-chloro deoxyadenosine treatment Haematol. 171,84–90 (2015). in thirty patients with hairy cell leukemia: evaluation of minimal residual dis- 51. Paltiel, O., Adler, B., Barchana, M. & Dann, E. J. A population-based study of ease and CD4+ lymphocytopenia after treatment. Leuk. Lymphoma 35, hairy cell leukemia in Israel. Eur. J. Haematol. 77,372–377 (2006). 555–565 (1999). 52. Kurzrock, R. et al. Second cancer risk in hairy cell leukemia: analysis of 350 33. Öngören, Ş. et al. Retrospective evaluation of hairy cell leukemia patients patients. J. Clin. Oncol. 15, 1803–1810 (1997). treated with three different first-line treatment modalities in the last two 53. Federico, M. et al. Risk of second cancer in patients with hairy cell leukemia: decades: a single-center experience. Turk.J.Haematol. 34,291–299 (2017). Long-term follow-up. J. Clin. Oncol. 20, 638–646 (2002). 34. Jehn, U., Bartl, R., Dietzfelbinger, H., Haferlach, T. & Heinemann, V. An update: 54. Tadmor, T. & Polliack, A. Hairy cell leukemia: uncommon clinical features, 12-year follow-up of patients with hairy cell leukemia following treatment with unusual sites of involvement and some rare associations. Best Pract. Res. Clin. 2-chlorodeoxyadenosine. Leukemia 18,1476–1481 (2004). Haematol. 28,193–199 (2015). 35. Găman, A. M. Hairy cell leukemia-a rare type of leukemia. A retrospective study 55. Anderson, L. A. & Engels, E. A. Autoimmune conditions and hairy cell leukemia: on 39 patients. Rom. J. Morphol. Embryol. 54,575–579 (2018). an exploratory case–control study. J. Hematol. Oncol. 3, 35 (2010). 36. López Rubio, M. et al. Hairy cell leukemia treated initially with purine analogs: a 56. Dasanu, C. A., Van den Bergh, M., Pepito, D. & Alvarez Argote, J. Autoimmune retrospective study of 107 patients from the Spanish Cooperative Group on disorders in patients with hairy cell leukemia: are they more common than Chronic Lymphocytic Leukemia (GELLC). Leuk. Lymphoma 55, 1007–1012 previously thought? Curr.Med.Res. Opin. 31,17–23 (2014). (2013). 57. Hardell, L., Eriksson, M. & Nordstrom, M. Exposure to pesticides as risk factor for 37. Hacioglu, S. et al. Multicenter retrospective analysis regarding the clinical non-Hodgkin’s lymphoma and hairy cell leukemia: pooled analysis of two manifestations and treatment results in patients with hairy cell leukemia: Swedish case-control studies. Leuk. Lymphoma 43,1043–1049 (2002). twenty-four year Turkish experience in cladribine therapy. Hematol. Oncol. 33, 58. Monnereau, A. et al. Medical history, lifestyle, and occupational risk factors for 192–198 (2014). hairy cell leukemia: the interlymph non-Hodgkin lymphoma subtypes project. 38. Else, M. et al. Long-term follow-up of 233 patients with hairy cell leukaemia, J. Natl Cancer Inst. Monogr. 2014,115–124 (2014). treated initially with pentostatin or cladribine, at a median of 16 years from 59. Clavel, J. et al. Hairy cell leukaemia, occupation, and smoking. Br.J.Haematol. diagnosis. Br.J.Haematol. 145,733–740 (2009). 91,154–161 (1995). 39. Zinzani, P. L. et al. Hairy cell leukemia: evaluation of the long-term outcome in 60. Clavel, J. et al. Farming, pesticide use and hairy-cell leukemia. Scand. J. Work 121 patients. Cancer 116, 4788–4792 (2010). Environ. Health 22,285–293 (1996). Blood Cancer Journal

Journal

Blood Cancer JournalSpringer Journals

Published: May 27, 2020

There are no references for this article.