short report Prognostic heterogeneity of adult B-cell precursor acute lymphoblastic leukaemia patients with t(1;19)(q23;p13)/ TCF3-PBX1 treated with measurable residual disease-oriented protocols Jordi Ribera,1 Isabel Granada,1 Summary Mireia Morgades, Teresa Gonzalez,2 1 Juana Ciudad,3,4 Esperanza Such,5 The prognosis of t(1;19)(q23;p13)/transcription factor 3-pre-B-cell leukaemia home- Marıa-Jose Calasanz,6 obox 1 (TCF3-PBX1) in adolescent and adult patients with acute lymphoblastic leu- Santiago Mercadal,7 Rosa Coll,8 kaemia (ALL) treated with measurable residual disease (MRD)-oriented trials Jose Gonzalez-Campos, Mar Tormo,10 9 remains controversial. In the present study, we analysed the outcome of adolescent Irene Garcıa-Cadenas,11 Cristina Gil,12 and adult patients with t(1;19)(q23;p13) enrolled in paediatric-inspired trials. The Marta Cervera,13 Pere Barba,14 patients with TCF3-PBX1 showed similar MRD clearance and did not have different Dolors Costa,15 Rosa Ayala,16 survival compared with other B-cell precursor ALL patients. However, patients with Arancha Berm udez,17 TCF3-PBX1 had a significantly higher cumulative incidence of relapse, especially Alberto Orfao 3,4 and among patients aged ≥35 years carrying additional cytogenetic alterations. These Josep-Maria Ribera,1 on behalf of the patients might benefit from additional/intensified therapy (e.g. immunotherapy in Programa para el Tratamiento de first complete remission with or without subsequent haematopoietic stem cell trans- Hemopatias Malignas (PETHEMA) plantation). Group (Spanish Society of Hematology, Keywords: acute lymphoblastic leukaemia, adults, t(1;19)(q23;p13)/TCF3- SEHH) 1 Josep Carreras Leukaemia Research PBX1, prognosis, cytogenetic alterations. Institute, ICO-Hospital Germans Trias i Pujol, Universitat Autonoma de Barcelona, Badalona, 2Hospital Universitario de Salamanca, Universidad de Salamanca, IBMCC (CSIC/USAL), IBSAL and CIBERONC, 3Cytometry Service (NUCLEUS) and Department of Medicine, Cancer Research Center (IBMCC-CSIC/ USAL-IBSAL), University of Salamanca, Salamanca, 4Centro de Investigacion Biomedica en Red de Cancer (CIBER- ONC) CB16/12/00400, Instituto de Salud Carlos III, Madrid, 5Hematology Department, Hospital Universitari Politecnic La Fe, Valencia, 6Centro de Investigacion Clınica Aplicada (CIMA), Universidad de Navarra, Pamplona, 7 Hematology Department, ICO-Hospital Duran i Reynals, Hospitalet de Llobregat, 8 Hematology Department, ICO-Hospital Josep Trueta, Girona, 9Hematology Department, Hospital Universitario Virgen del Rocıo, Sevilla, 10Hematology Department, Hospital Clınico Universitario, Valencia, 11Hematology Department, First published online 21 September 2021 ª 2021 The Authors. British Journal of Haematology published by British Society for doi: 10.1111/bjh.17844 Haematology and John Wiley & Sons Ltd. British Journal of Haematology, 2022, 196, 670–675 This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. Short Report Hospital de Sant Pau, Josep Carreras Leukaemia Research Institute, Barcelona, 12 Hematology Department, Hospital General de Alicante, Alicante, 13 Hematology Department, ICO-Hospital Joan XXIII, Tarragona, 14Hematology Department, Hospital Universitari Vall d’Hebron, 15Haematopathology Section, Department of Pathology, Hospital Clınic, Barcelona, 16Hematology Department, Hospital Universitario Doce de Octubre, Madrid, and 17Hematology Department, Hospital Universitario Marques de Valdecilla, Santander, Spain Received 2 July 2021; accepted for publication 6 September 2021 Correspondence: Jordi Ribera, Josep Carreras Leukaemia Research Institute, ICO-Hospital Germans Trias i Pujol, Universitat Aut onoma de Barcelona, Carretera de Canyet s/n, Camı de les Escoles s/n, 08916 Badalona, Spain. E-mail: jribera@carrerasresearch.org When possible, patients with normal karyotype or no meta- Introduction phases were screened for the most recurrent BCP ALL subtypes, The t(1;19)(q23;p13)/transcription factor 3-pre-B-cell leukae- including TCF3-PBX1, by fluorescence in situ hybridisation mia homeobox 1 (TCF3-PBX1) is a subtype of B-cell precursor (FISH) and by reverse transcription polymerase chain reaction acute lymphoblastic leukaemia (BCP ALL) with an incidence of to a lesser extent. Among patients with Ph-negative ALL, high- around 5–7%.1 In paediatric ALL, therapy intensification has risk was defined by age of >30 years, white blood cell (WBC) improved the outcome, moving from poor to intermediate– count of >30 9 109/l or the presence of 11q23/Lysine Methyl- good prognosis. In adults the scenario is less clear, with reports transferase 2A (KMT2A) rearrangements. Therapeutic deci- showing favourable,2 intermediate3–5 and poor prognosis.6–8 sions were based on MRD levels in the two high-risk protocols. The low incidence of this alteration limits the prognosis assess- A good MRD response was considered when the MRD level ment in homogeneously treated patients, and differences was <01% at the end of induction and <001% at the end of within chemotherapy schedules and in the indication of allo- consolidation. Patients with end-consolidation MRD levels of geneic haematopoietic stem cell transplantation (alloHSCT) ≥001% received alloHSCT, while patients with good MRD may explain these controversial results. However, the prognos- clearance received delayed consolidation and maintenance tic significance of TCF3-PBX1 in paediatric-inspired, measur- therapy. MRD was centrally assessed by multiparametric flow able residual disease (MRD)-based trials has not been cytometry in a EuroFlow laboratory. Comparisons between extensively evaluated in adolescents and adults with ALL. groups were performed with the chi-square test, Fisher’s exact test and the median test as appropriate. Overall survival (OS) was estimated using the Kaplan–Meier method and curves were Patients and methods compared by the log-rank test. The cumulative incidence of The outcomes of adolescent and adults (aged 15–60 years) with relapse (CIR) was calculated using cumulative incidence func- Philadelphia (Ph) chromosome-negative BCP ALL and TCF3- tions considering non-relapse mortality (NRM) as a competing PBX1 rearrangement diagnosed between 2003 and 2017 and risk and was compared by Gray’s test. Two-sided P < 005 were treated with MRD-oriented protocols from the Programa considered statistically significant. Espa~nol de Tratamientos en Hematologıa (PETHEMA) Group (high-risk protocols ALL AR03 and ALL HR11, and standard- Results risk protocol ALL RE08) were retrospectively analysed.9–11 Patients were re-classified into cytogenetic subtypes according Of 539 patients with BCP ALL, 154 were excluded for unin- to the World Health Organization (WHO) 2017 classification. formative cytogenetics. From 385 patients with Ph-negative ª 2021 The Authors. British Journal of Haematology published by British Society for 671 Haematology and John Wiley & Sons Ltd. British Journal of Haematology, 2022, 196, 670–675 Short Report BCP ALL with informative cytogenetics, 26 (7%) were classi- higher female gender, more frequent pre-B phenotype and a fied as TCF3-PBX1. Of them, 22 had t(1;19)(q23;p13) or der trend towards higher WBC counts. There were no significant (19)t(1;19) by G-banding and four cases with normal kary- differences in age and WBC count between patients with bal- otype (two) or no metaphases (two) showed TCF3-PBX1 anced translocation, der(19)t(1;19), isolated t(1;19)/der(19)t rearrangement by FISH. Patients with t(1;19) were classified (1;19) and t(1;19)/der(19)t(1;19) with ACA. All the patients according to the type of rearrangement [balanced transloca- with TCF3-PBX1 achieved complete remission (CR). End- tion versus der(19)t(1;19)] and the karyotype complexity induction MRD response (available for 23/26 TCF3-PBX1 [isolated t(1;19)/der(19)t(1;19) versus t(1;19)/der(19)t(1;19) and 277/359 patients) was not significantly different between with additional chromosomal alterations (ACA)]. In all, 10/ the two groups of patients [MRD <001% TCF3-PBX1: 16/23 22 (45%) patients with TCF3-PBX1 had t(1;19) balanced (70%) vs. 169/277 (61%) respectively, P = 0417]. There were rearrangements, while 12/22 (55%) had der(19)t(1;19). In no significant differences according to good MRD response turn, nine of the 22 (41%) cases had a t(1;19) or der(19)t (MRD <001%) between patients with isolated t(1;19)/der (1;19) as an isolated aberration, while 13/22 (59%) had ACA (19)t(1;19) and those with t(1;19)/der(19)t(1;19) with ACA, [median (range) 2 (1–10)]. The most recurrent ACA were as well as between patients with balanced rearrangements trisomy 1/1q (four), 6/del(6q) [MYB Proto-Oncogene (C- and der(19)t(1;19). Only two patients with TCF3-PBX1 MYB); three], del(9p) [cyclin-dependent kinase inhibitor 2A/ received an alloHSCT due to an end-consolidation MRD B (CDKN2A/B), paired box 5 (PAX5); five] and 13/del level of 005% and to poor response to the first induction (13q) [retinoblastoma 1 (RB1); four]. Although der(19)t cycle respectively. One of them is alive in third CR after chi- (1;19) was preferentially seen in association with ACA, there meric antigen receptor T cells targeting CD19 (CART19) was no statistically significant relationship between both clas- therapy and the second died due to a concomitant mela- sifications (balanced/derivative versus isolated/ACA). Four noma, with ALL in CR. The remaining patients with TCF3- patients (15%) were treated within the ALL RE08 protocol, PBX1 completed delayed consolidation and maintenance (16 while 10 and 12 patients were treated according to the ALL patients), relapsed during consolidation (five), died in con- AR03 and ALL HR11 trials respectively. solidation (one), withdrew from the protocol (one) or were The main clinical and biological characteristics of the 26 under consolidation (one). patients with TCF3-PBX1 were compared with those of the The median follow-up of alive patients using the reverse 359 patients with Ph-negative BCP ALL with informative Kaplan–Meier method was 359 years [95% confidence cytogenetics (Table I). The patients with TCF3-PBX1 showed interval (CI) 210–580, range 020–725] for the patients Table I. Comparison of the main characteristics of patients with TCF3-PBX1 and patients with other B-cell precursor acute lymphoblastic leukae- mia with evaluable cytogenetics. TCF3-PBX1 (n = 26) Remaining B-cell precursor ALL (n = 359) P Age, years Median (range) 36 (17–60) 35 (15–60) 0675 Gender, n (%) Male 7 (27) 197 (55) 0006 Female 19 (73) 162 (45) ECOG PS score, n/N (%) 0 5/22 (23) 109/339 (32) 0623 1 14/22 (64) 177/339 (52) ≥2 3/22 (14) 53/339 (16) WBC count, 9109/l Median (range) 162 (0–199.5) 12 (02–638) 0144 Mediastinal mass, n/N (%) No 22/22 (100) 339/346 (98) 1000 Yes 0 7/346 (2) CNS involvement, n/N (%) No 21/24 (88) 311/336 (93) 0417 Yes 3/24 (12) 25/336 (7) Phenotype, n/N (%) Pro-B 0 67 (19) <0001 Common B 9/24 (37) 231 (64) Pre-B 15/24 (63) 61 (17) ALL, acute lymphoblastic leukaemia; CNS, central nervous system; ECOG PS, Eastern Cooperative Oncology Group Performance Status; TCF3- PBX1, transcription factor 3-pre-B-cell leukaemia homeobox 1; WBC, white blood cell. 672 ª 2021 The Authors. British Journal of Haematology published by British Society for Haematology and John Wiley & Sons Ltd. British Journal of Haematology, 2022, 196, 670–675 Short Report (A) (B) (C) Fig 1. (A) Probability of overall survival for patients with transcription factor 3-pre-B-cell leukaemia homeobox 1 (TCF3-PBX1) rearrangement and the other B-cell precursor acute lymphoblastic leukaemia (ALL) patients with evaluable cytogenetics. (B) Cumulative incidence of relapse for patients with TCF3-PBX1 rearrangement and the other B-cell precursor ALL patients with evaluable cytogenetics. (C) Cumulative incidence of relapse among patients with TCF3-PBX1 rearrangement depending on the presence versus absence of additional cytogenetic alterations (ACA). [Colour figure can be viewed at wileyonlinelibrary.com] with TCF3-PBX1 and 371 (95% CI 327–420, range 013– Patients with TCF3-PBX1 had a significantly higher CIR 947) for the other patients with BCP ALL. The probability than the remaining BCP ALL cases [5-year CIR of 60% (95% of OS for patients with TCF3-PBX1 was similar among the CI 36–77%) for TCF3-PBX1 vs. 42% (95% CI 35–48%) three protocols and was not significantly different from that respectively; P = 0023) (Fig 1B). The higher relapse propen- of the remaining patients with BCP ALL [5-year OS 41% sity seen in patients with TCF3-PBX1 was mainly observed in (95% CI 19–63%) vs. 44% (95% CI 38–50%), P = 0562] those with ACA (5-year CIR 69% [33–89%] vs. 28% (Fig 1A). The OS was better for patients with isolated t [3–63%], P = 0137] (Fig 1C). In addition, younger patients (1;19)/der(19)t(1;19) than for those with t(1;19)/der(19)t had a trend for lower CIR than those aged ≥35 years [2-year (1;19) with ACA [5-year OS 59% (95% CI 23–95%) vs. CIR 40% (95% CI 11–69%) vs. 70% (95% CI 31–90%), 31% (95% CI 0–62%), P = 0558]. Patients aged <35 years P = 0073]. Patients with both adverse prognostic factors had a significantly better OS than those ≥35 years [3-years (aged ≥35 years with ACA, nine patients) showed a trend for OS 77% (95% CI 49–100%) vs. 25% (95% CI 1–49%), higher CIR compared with the remaining 13 patients [2-year P = 0002]. CIR 67% (95% CI 20–90%] vs. 36% (95% CI 10–63%), ª 2021 The Authors. British Journal of Haematology published by British Society for 673 Haematology and John Wiley & Sons Ltd. British Journal of Haematology, 2022, 196, 670–675 Short Report P = 0089]. All patients aged ≥35 years with ACA relapsed, inhibitors and/or phosphatidylinositol-3 kinase (PI3K)/pro- which correlated with poorer OS for this subgroup [2-year tein kinase B (Akt) inhibitors may also be of special interest OS 22% (95% CI 0–49%) vs. 74% (95% CI 48–100%), for this subset of patients.15 P = 0020]. The Fine–Gray model showed that age was the only predictor for CIR in patients with TCF3-PBX1 (hazard Acknowledgements ratio 4433, 95% CI 1014–19380; P = 0048). The outcome of the 14 relapsing patients was poor, This work was supported in part by CERCA/Generalitat de despite six of them being rescued with blinatumomab (three) Catalunya SGR 2017 288 (GRC), a restricted grant from ‘La or inotuzumab (three), followed by CART19 in two of them. Caixa’ and Healthcare Alliance for Resourceful Medicine Currently, four patients are alive. The OS of the patients Offensive against Neoplasms (HARMONY). We are very with TCF3-PBX1 after first relapse was not different from grateful to the following authors for their contribution to that of the other patients with BCP ALL (data not shown). this multicentric study: Lurdes Zamora (Josep Carreras Disease progression was the main cause of death among Leukaemia Research Institute, ICO-Hospital Germans Trias i patients with TCF3-PBX1 (10/13, 77%), while three of 13 Pujol, Universitat Aut onoma de Barcelona), Anna Torrent patients (23%) died from NRM. In the remaining BCP ALL, (Josep Carreras Leukaemia Research Institute, ICO-Hospital death by disease progression was observed in 94/160 (59%) GermansTrias i Pujol, Universitat Aut onoma de Barcelona), and 66/160 (41%) died from NRM (transplant-related mor- Susana Barrena (Cancer Research Center (IBMCC-CSIC/ tality in 13 of them). USAL-IBSAL), Cytometry Service (NUCLEUS) and Depart- ment of Medicine, University of Salamanca), Beatriz Soriano (Cancer Research Center (IBMCC-CSIC/USAL-IBSAL), Discussion Cytometry Service (NUCLEUS) and Department of Medi- The present analysis of adolescent and adult patients treated cine, University of Salamanca), Pau Montesinos (Hospital with MRD-oriented protocols shows that patients with TCF3- Universitari i Politecnic La Fe), Eulalia Genesca (Josep Car- PBX1 have a high probability of relapse, especially those aged reras Leukaemia Research Institute, ICO-Hospital Germans >35 years with ACA, despite excellent response to treatment Trias i Pujol, Universitat Aut onoma de Barcelona), Celia and scarce need of alloHSCT according to the MRD status. Gonzalez-Gil (Josep Carreras Leukaemia Research Institute, The relapse-associated mortality in these patients might coun- ICO-Hospital Germans Trias i Pujol, Universitat Aut onoma teract the higher NRM registered in the non-TCF3-PBX1 de Barcelona), Francisco Fuster-Tormo (Josep Carreras Leu- group, explaining the similar OS of these groups. kaemia Research Institute, ICO-Hospital Germans Trias i Age was a prognostic factor for patients with TCF3-PBX1, Pujol, Universitat Aut onoma de Barcelona), Clara Maluquer as shown in any ALL subtype. The median age of the patients (ICO-Hospital Duran i Reynals), Laura Blanco (Josep Car- with TCF3-PBX1 in our present series (36 years) was older reras Leukaemia Research Institute, Hospital de Sant Pau), than that of other similar studies2,3,5 and the frequency of Josep Nomdedeu (Josep Carreras Leukaemia Research Insti- alloHSCT was lower,4 as alloHSCT indication was exclusively tute, Hospital de Sant Pau), Margarita Ortega (Hospital based on MRD-status after induction and consolidation. Universitari Vall d’Hebr on), Jordi Esteve (Haematology The presence of ACA showed a trend for higher relapse, Department, Hospital Clınic Barcelona, IDIBAPS, Josep Car- as has also been shown in patients with Ph+-ALL treated reras Leukaemia Research Institute), Marina Dıaz-Beya within PETHEMA trials.12 Unfortunately, detailed genomic (Haematology Department, Hospital Clınic Barcelona, data are not available, although the ACA could indicate the IDIBAPS, Josep Carreras Leukaemia Research Institute), presence of recurrent abnormalities with potential poor out- Pilar Martınez-Sanchez (Hospital Universitario Doce de come (i.e. MYB, CDKN2A/B and RB1).13,14 Besides MRD, Octubre), Andres Novo (Hospital Universitari Son Espases), genetic differences at the molecular level might contribute, at Mª Paz Queipo de Llano (Hospital Clınico Virgen de la Vic- least in part, to the different outcomes observed among toria), Ana Batlle (Hospital Universitario Marques de Valde- patients with TCF3-PBX1. cilla), Juan Bergua (Hospital San Pedro Alcantara), Mª In addition to the lack of genomic data, the main limita- Teresa Olave (Hospital Clınico Universitario Lozano Blesa), tion of the present study, and others showing similar fre- Beatriz de Rueda (Hospital Universitario Miguel Servet), quency, is the low number of patients, despite including Pilar Herrera(Hospital Universitario de Donostia), Mª Ter- FISH analysis to ensure full detection of patients with TCF3- esa Artola (Hospital Universitario de Donostia), Jes us Mª PBX1 when available. Given the low incidence of this ALL Hernandez-Rivas (Hospital Universitario de Salamanca, subtype and the conflicting data in the literature, a pooling Universidad de Salamanca, IBMCC (CSIC/USAL), IBSAL of data from several studies or a meta-analysis might help to and CIBERONC). give more reliable results. From our present results, patients with TCF3-PBX1 aged >35 years with ACA might benefit Conflicts of interest from consolidation with immunotherapy followed or not by alloHSCT. In addition, targeted therapies with kinase The authors have no competing interests. 674 ª 2021 The Authors. British Journal of Haematology published by British Society for Haematology and John Wiley & Sons Ltd. British Journal of Haematology, 2022, 196, 670–675 Short Report References in adults with standard-risk acute lymphoblastic leukemia in first complete remission. Biol Blood Marrow Transplant. 2014;20:1314–21. 1. Paietta E, Roberts KG, Wang V, Gu Z, Buck G, Pei D, et al. Molecular 9. Ribera J-M, Oriol A, Morgades M, Montesinos P, Sarra J, Gonzalez- classification improves risk assessment in adult BCR-ABL1-negative B- Campos J, et al. Treatment of high-risk Philadelphia chromosome- ALL. Blood. 2021. https://doi.org/10.1182/blood.2020010144. negative acute lymphoblastic leukemia in adolescents and adults according 2. Yilmaz M, Kantarjian HM, Toruner G, Yin CC, Kanagal-Shamanna R, to early cytologic response and minimal residual disease after consolida- Cortes JE, et al. Translocation t(1;19)(q23;p13) in adult acute lymphoblas- tion assessed by flow cytometry: final results of the PETHEMA ALL-AR- tic leukemia - a distinct subtype with favorable prognosis. Leuk Lym- 03 trial. J Clin Oncol. 2014;32:1595–604. phoma. 2021;62:224–8. 10. Ribera JM, Morgades M, Ciudad J, Montesinos P, Esteve J, Genesca E, 3. Burmeister T, Gokbuget N, Schwartz S, Fischer L, Hubert D, Sindram A, et al. Chemotherapy or allogeneic transplantation in high-risk Philadelphia et al. Clinical features and prognostic implications of TCF3-PBX1 and chromosome-negative adult lymphoblastic leukemia. Blood. 2021;137: ETV6-RUNX1 in adult acute lymphoblastic leukemia. Haematologica. 1879–94. 2010;95:241–6. 11. Ribera J-M, Morgades M, Montesinos P, Tormo M, Martınez-Carballeira 4. Lafage-Pochitaloff M, Baranger L, Hunault M, Cuccuini W, Lefebvre C, D, Gonzalez-Campos J, et al. A pediatric regimen for adolescents and Bidet A, et al. Impact of cytogenetic abnormalities in adults with Ph- young adults with Philadelphia chromosome-negative acute lymphoblastic negative B-cell precursor acute lymphoblastic leukemia. Blood. leukemia: results of the ALLRE08 PETHEMA trial. Cancer Med. 2017;130:1832–44. 2020;9:2317–29. 5. Moorman AV, Harrison CJ, Buck GAN, Richards SM, Secker-Walker LM, 12. Motll o C, Ribera JM, Morgades M, Granada I, Montesinos P, Mercadal S, Martineau M, et al. Karyotype is an independent prognostic factor in et al. Frequency and prognostic significance of additional cytogenetic adult acute lymphoblastic leukemia (ALL): analysis of cytogenetic data abnormalities to the Philadelphia chromosome in young and older adults from patients treated on the Medical Research Council (MRC) UKALL- with acute lymphoblastic leukemia. Leuk Lymphoma. 2018;59:146–54. XII/Eastern Cooperative Oncology Group (ECOG) 2993 trial. Blood. 13. Gonzalez-Gil C, Ribera J, Ribera JM, Genesca E. The Yin and Yang-like 2007;109:3189–97. clinical implications of the CDKN2A/ARF/CDKN2B gene cluster in acute 6. Mancini M, Scappaticci D, Cimino G, Nanni M, Derme V, Elia L, et al. A lymphoblastic leukemia. Genes (Basel). 2021;12:79. comprehensive genetic classification of adult acute lymphoblastic leukemia 14. Messina M, Chiaretti S, Fedullo AL, Piciocchi A, Puzzolo MC, Lauretti A, (ALL): analysis of the GIMEMA 0496 protocol. Blood. 2005;105:3434–41. et al. Clinical significance of recurrent copy number aberrations in B- 7. Pullarkat V, Slovak ML, Kopecky KJ, Forman SJ, Appelbaum FR. Impact of lineage acute lymphoblastic leukaemia without recurrent fusion genes cytogenetics on the outcome of adult acute lymphoblastic leukemia: results across age cohorts. Br J Haematol. 2017;178:583–7. of Southwest Oncology Group 9400 study. Blood. 2008;111:2563–72. 15. Eldfors S, Kuusanm€aki H, Kontro M, Majumder MM, Parsons A, Edgren 8. Yan CH, Jiang Q, Wang J, Xu LP, Liu DH, Jiang H, et al. Superior sur- H, et al. Idelalisib sensitivity and mechanisms of disease progression in vival of unmanipulated haploidentical hematopoietic stem cell transplanta- relapsed TCF3-PBX1 acute lymphoblastic leukemia. Leukemia. 2017; tion compared with chemotherapy alone used as post-remission therapy 31:51–7. ª 2021 The Authors. British Journal of Haematology published by British Society for 675 Haematology and John Wiley & Sons Ltd. British Journal of Haematology, 2022, 196, 670–675