Letters to the Editor 2254 treated with vitamin E, which reduced ROS by approximately 3Abteilung Haematologie/Onkologie, Universitaetsklinikum twofold (Figure 1a).9,11 Jena, Jena, Germany and Human CD45 þ cells collected from bone marrows and spleens 4Department of Immunology, Medical University of Warsaw, of NSG mice formed colonies in vitro, and vitamin E treatment did Warsaw, Poland not affect the engraftment (Figure 1b). -resistant clones E-mail: [email protected] carrying either E255K or T315I BCR-ABL1 kinase were detected in three out of five untreated xenografts, but in none of the vitamin E-treated samples (Table 1). REFERENCES In conclusion, we postulate that anti-oxidants such as vitamin E 1 Melo JV, Barnes DJ. Chronic myeloid leukaemia as a model of disease evolution in may be applicable in prevention of TKI-resistance, in particular of human cancer. Nat Rev Cancer 2007; 7: 441–453. that driven by BCR-ABL1 kinase mutations. Moreover, as imatinib- 2 Eiring AM, Khorashad JS, Morley K, Deininger MW. Advances in the treatment of treated CML-CP LSCs and LPCs continue to accumulate high levels chronic myeloid . BMC Med 2011; 9:99. of ROS resulting in TKI-resistant mutations,9,10 anti-oxidant treatment 3 Willis SG, Lange T, Demehri S, Otto S, Crossman L, Niederweiser D et al. High- could be combined with TKIs to extend/improve the therapeutic sensitivity detection of BCR-ABL kinase domain mutations in imatinib-naive effects of ABL1 kinase inhibitors. This speculation is reinforced by the patients: correlation with clonal cytogenetic evolution but not response to observation that anti-oxidants vitamin E and N-acetyl-cysteine therapy. Blood 2005; 106: 2128–2137. 4 Hochhaus A, La Rosee P. Imatinib therapy in chronic myelogenous reduced the percentage of the resistant clones emerging in vitro 11 leukemia: strategies to avoid and overcome resistance. Leukemia 2004; 18: from imatinib-treated BCR-ABL1–positive cells. 1321–1331. 5 Eide CA, Adrian LT, Tyner JW, MacPartlin M, Anderson DJ, Wise SC et al. The ABL switch control inhibitor DCC-2036 is active against the chronic myeloid leukemia CONFLICT OF INTEREST mutant BCR-ABLT315I and exhibits a narrow resistance profile. Cancer Res 2011; 71: 3189–3195. The authors declare no conflict of interest. 6 Khorashad JS, Kelley TW, Szankasi P, Mason CC, Soverini S, Adrian LT et al. BCR-ABL1 compound mutations in inhibitor-resistant CML: frequency and clonal relationships. Blood 2013; 121: 489–498. ACKNOWLEDGEMENTS 7 Skorski T. BCR-ABL1 kinase: hunting an elusive target with new weapons. Chem MN-S and GH share equal contribution. This work was supported by NIH/NCI Biol 2011; 18: 1352–1353. CA123014 and CA134458 to T Skorski, N401 039037 from the Polish Ministry of 8 Gutteridge JM, Halliwell B. Free radicals and antioxidants in the year 2000. A Education and Science to G Hoser and EU program: FP7-REGPOT-2012-CT2012- historical look to the future. Ann NY Acad Sci 2000; 899: 136–147. 316254-BASTION to T Stoklosa. 9 Nieborowska-Skorska M, Kopinski PK, Ray R, Hoser G, Ngaba D, Flis K et al. Rac2-MRC-cIII-generated ROS cause genomic instability in chronic myeloid M Nieborowska-Skorska1, G Hoser2, A Hochhaus3, leukemia stem cells and primitive progenitors. Blood 2012; 119: 4253–4263. T Stoklosa4 and T Skorski1 10 Bolton-Gillespie E, Schemionek M, Klein HU, Flis S, Hoser G, Lange T et al. Genomic 1 instability may originate from imatinib-refractory chronic myeloid leukemia stem Department of Microbiology and Immunology, cells. Blood 2013 (in press). School of Medicine, Temple University, Philadelphia, PA, USA; 11 Koptyra M, Falinski R, Nowicki MO, Stoklosa T, Majsterek I, Nieborowska-Skorska M 2 Department of Clinical Cytology, Medical Center for et al. BCR/ABL kinase induces self-mutagenesis via reactive oxygen species to Postgraduate Education, Warsaw, Poland; encode imatinib resistance. Blood 2006; 108: 319–327.

Long-term follow-up of treatment with imatinib in -associated myeloid/lymphoid neoplasms with PDGFR rearrangements in blast phase

Leukemia (2013) 27, 2254–2256; doi:10.1038/leu.2013.129 Thirteen patients were FIP1L1-PDGFRA-positive (myeloid BP, n ¼ 8; chloroma, n ¼ 2; lymphoid BP, n ¼ 3). In all three patients investigated, FIP1L1-PDGFRA could be identified contempora- neously in myeloid cells derived from BM and CD3-positive Several cooperative study groups and single institutions lymphoblasts in LN biopsies. Four patients (myeloid BP, n ¼ 2; have reported on high rates of rapid and durable complete lymphoid BP, n ¼ 2) were positive for a rearrangement of PDGFRB hematologic (CHR) and complete molecular remissions (CMR) (partner : ETV6, n ¼ 1; SART3, n ¼ 1; unknown partner , 10 on imatinib in eosinophilia-associated myeloid/lymphoid n ¼ 2). All patients were male, median age was 46 years (range 9 9 neoplasms (MLN-Eo) with rearrangements of PDGFR (PDGFRA 37–66). Significant eosinophilia 41.5 Â 10 /l (median 7.7 Â 10 /l, and PDGFRB) in chronic phase (CP). These responses have range 1.3–20.7) was present in 12 of 13 (92%) FIP1L1-PDGFRA- 9 translated into vastly improved progression-free and overall positive patients and 2 of 4 (50%) patients (median 1.2 Â 10 /l, survival (OS).1–9 In this report, we present data from the range 0.1–33.0) with rearrangements of PDGFRB. In addition to the ‘German Registry on Disorders of and Mast Cells’ on specific morphologic features of PB, BM and LN, the most long-term follow-up of 17 patients primarily diagnosed in blast common clinical characteristics included splenomegaly (17/17, phase (BP). BP was defined as (i) myeloid BP (420% blasts in 100%), involvement of lung (4/17, 24%; pleural effusion, n ¼ 4), peripheral blood (PB) and/or bone marrow (BM)), (ii) chloroma skin (4/17, 24%) and heart (2/17, 12%; endo-/myocarditis, n ¼ 1; (extramedullary tissue infiltration by myeloid blasts) or (iii) intracardial thrombus, n ¼ 1). extramedullary lymphoid BP (lymph node (LN) infiltration by Without initial knowledge of the underlying , nine lymphoblasts of T-cell origin). patients received primary according to protocols

Accepted article preview online 25 April 2013; advance online publication, 21 May 2013

Leukemia (2013) 2242 – 2267 & 2013 Macmillan Publishers Limited Letters to the Editor 2255 100 after SCT. Overall, only 2 of 17 (12%) patients have died after a median observation time of 65 months (range 7–106), which is not 80 different as compared to 46 patients in CP (median 51 months, range 1–103; Figure 1a and b). 60 Besides the excellent and sustained responses to imatinib in PDGFR fusion gene-positive MLN-eo in BP, our data also under- 40 score the failure of standard diagnostic work-up at initial diagnosis to identify pathogenetically relevant tyrosine kinase (TK) fusion genes, for example, PDGFRA, PDGFRB or FGFR1, in patients with Overall survival (%) Overall 20 eosinophilia and supposed ‘ (AML)’ or ‘T-cell ’. It is noteworthy that the frequency of so called 0 1224 36 48 60 72 84 96 108 120 ‘T-cell ’ is even more frequent in patients with FGFR1 fusion genes as consequence of reciprocal translocations with imatinib (months) involvement of 8p11, particularly in t(8;13)(p11;q13) with a ZNF198-FGFR1 fusion gene.12 Careful attention should therefore 100 be paid to the potential presence of these TK fusion genes in primarily diagnosed myeloid and lymphoid malignancies through 80 distinctive clinical (for example, splenomegaly), biochemical (for example, elevated serum levels of tryptase or vitamin B12), 60 morphological (for example, leukocytosis with left-shift and dysplasia, eosinophilia, fibrosis and/or increased numbers of 40 loosely scattered mast cells in BM) and molecular genetic (for example, FIP1L1-PDGFRA, reciprocal translocations with

Overall survival (%) Overall 20 involvement of 4q12-PDGFRA, 5q33-PDGFRB or 8p11-FGFR1) features of an eosinophilia-associated and TK fusion-positive 0 stem cell disorder with involvement of the myeloid and 0 1224 36 48 60 72 84 96 108 120 lymphoid lineage. Having established the presence of a fusion imatinib (months) gene, we consider that the diagnosis of these cases should be revised to ‘myeloid BP’ and ‘extramedullary lymphoid BP’ of MLN- Figure 1. (a) Overall survival of 49 patients with a FIP1L1-PDGFRA by eo, respectively, the latter reflecting the fact that these are stem FIP1L1-PDGFRA fusion gene (upper graph) and 14 patients with cell disorders in which affected TK fusion gene-positive diverse PDGFRB fusion genes (lower graph) in chronic and blast T-lymphoblasts predominantly reside in the LN.11 phase (P ¼ nonsignificant (NS)). (b) Overall survival of 63 patients (middle graph) with a FIP1L1-PDGFRA fusion gene (n ¼ 49) or diverse In conclusion, our data highlight several new aspects for PDGFRB fusion genes (n ¼ 14). Patients in chronic phase (FIP1L1- diagnosis and treatment of MLN-eo with PDGFR fusion genes: (i) PDGFRA, n ¼ 36; X-PDGFRB, n ¼ 10) are shown in the top graph while mimicking de novo AML and T-cell lymphoma, myeloid and patients in blast phase (FIP1L1-PDGFRA, n ¼ 13; X-PDGFRB, n ¼ 4; lymphoid BP of MLN-eo with an underlying imatinib-sensitive P ¼ NS) are shown in the lower graph. PDGFR fusion gene is frequently not identified, (ii) in the absence of CBF fusion genes, patients with eosinophilia-associated AML but for acute leukemia or lymphoma.11 Despite clear signs of also patients with eosinophilia-associated T-cell lymphoma should remission, for example, clearance of blasts or partial/complete be actively screened for PDGFR fusion genes by PCR, cytogenetics clinical and/or hematologic remission, eosinophilia persisted at and/or fluorescence in situ hybridization and (iii) the excellent various levels in all nine patients. The underlying involvement of long-term data for patients with MLN-eo in BP suggest that PDGFRA (FIP1L1-PDGFRA, n ¼ 6) or PDGFRB (n ¼ 3) was therefore monotherapy with imatinib should be initiated early because only identified weeks or even months after initial morphological durable remissions may only rarely be achieved through intensive diagnosis by PCR, cytogenetics and/or fluorescence in situ chemotherapy or allogeneic SCT. hybridization and imatinib was initiated in 7 of 9 (78%) patients. In 8 of 17 patients (47%), the PDGFR rearrangement (FIP1L1- PDGFRA, n ¼ 7; PDGFRB, n ¼ 1) was already identified at diagnosis CONFLICT OF INTEREST and imatinib was initiated directly. Overall, CHR was achieved GM, AH, NCPC, WKH and AR received honoraria and travel support from Novartis in all 15 imatinib-treated patients (starting dose 400 mg/day, Pharma. n ¼ 6; 100 mg/day, n ¼ 9) after median 1 month (0.1–14). CMR was detected in all 12 FIP1L1-PDGFRA-positive patients after median 5 months (3–32). One patient died due to ACKNOWLEDGEMENTS a cerebral hemorrhage while in CMR for 9 months. The This work was supported by the ‘Deutsche Jose´ Carreras Leuka¨mie Stiftung e.V.’ remaining 11 patients are in sustained CMR for median 65 (R09/29f and H11/03), Germany. months (7–103). The three patients with a PDGFRB rearrangement are in sustained CHR for median 56 months (24–56). Two patients G Metzgeroth1, J Schwaab1, D Gosenca1, A Fabarius1, (FIP1L1-PDGFRA, n ¼ 1; ETV6-PDGFRB plus complex karyotype, C Haferlach2, A Hochhaus3, NCP Cross4,5, n ¼ 1) received an allogeneic stem cell transplantation (SCT) W-K Hofmann1 and A Reiter1 within 6 months (related donor, n ¼ 1; unrelated donor, n ¼ 1) 1III. Medizinische Klinik, Universita¨tsmedizin Mannheim, after diagnosis of myeloid BP and lack of complete remission after Mannheim, Germany; intensive chemotherapy. Both patients relapsed early (o3 2Mu¨nchner Leuka¨mie Labor, Mu¨nchen, Germany; months). Following relapse, the FIP1L1-PDGFRA-positive patient 3Abteilung Ha¨matologie/Onkologie, Klinik fu¨r Innere Medizin II, was treated with imatinib (100 mg/day) and is in sustained CMR Universita¨tsklinikum Jena, Jena, Germany; þ 9 months after SCT. The ETV6-PDGFRB-positive patient died due 4Wessex Regional Genetics Laboratory, Salisbury, UK and to progressive leptomeningeal involvement 9 months after SCT 5Faculty of Medicine, University of Southampton, and 12 months after diagnosis of myeloid BP despite achievement Southampton, UK of CHR in PB and BM on imatinib, which was initiated þ 1 month E-mail: [email protected]

& 2013 Macmillan Publishers Limited Leukemia (2013) 2242 – 2267 Letters to the Editor 2256 REFERENCES 7 Helbig G, Moskwa A, Hus M, Piszcz J, Swiderska A, Urbanowicz A et al. Durable 1 Apperley JF, Gardembas M, Melo JV, Russell-Jones R, Bain BJ, Baxter EJ et al. remission after treatment with very low doses of imatinib for FIP1L1-PDGFRalpha- Response to imatinib mesylate in patients with chronic myeloproliferative positive chronic eosinophilic leukaemia. Cancer Chemother Pharmacol 2011; 67: diseases with rearrangements of the platelet-derived growth factor beta. 967–969. N Engl J Med 2002; 347: 481–487. 8 Metzgeroth G, Walz C, Erben P, Popp H, Schmitt-Graeff A, Haferlach C et al. Safety 2 Bain BJ. Myeloid and lymphoid neoplasms with eosinophilia and abnormalities of and efficacy of imatinib in chronic eosinophilic leukaemia and hypereosinophilic PDGFRA, PDGFRB or FGFR1. Haematologica 2010; 95: 696–698. syndrome: a phase-II study. Br J Haematol 2008; 143: 707–715. 3 Cools J, DeAngelo DJ, Gotlib J, Stover EH, Legare RD, Cortes J et al. A tyrosine 9 Pardanani A, D’Souza A, Knudson RA, Hanson CA, Ketterling RP, Tefferi A. Long- kinase created by fusion of the PDGFRA and FIP1L1 genes as a therapeutic target term follow-up of FIP1L1-PDGFRA-mutated patients with eosinophilia: survival of imatinib in idiopathic hypereosinophilic syndrome. N Engl J Med 2003; 348: and clinical outcome. Leukemia 2012; 26: 2439–2441. 1201–1214. 10 Erben P, Gosenca D, Mu¨ller MC, Reinhard J, Score J, Del Valle F et al. Screening for 4 Baccarani M, Cilloni D, Rondoni M, Ottaviani E, Messa F, Merante S et al. The efficacy diverse PDGFRA or PDGFRB fusion genes is facilitated by generic quantitative of imatinib mesylate in patients with FIP1L1-PDGFRalpha-positive hyper- reverse transcriptase polymerase chain reaction analysis. Haematologica 2010; 95: eosinophilic syndrome. Results of a multicenter prospective study. Haematologica 738–744. 2007; 92: 1173–1179. 11 Metzgeroth G, Walz C, Score J, Siebert R, Schnittger S, Haferlach C et al. 5 David M, Cross NC, Burgstaller S, Chase A, Curtis C, Dang R et al. Durable Recurrent finding of the FIP1L1-PDGFRA fusion gene in eosinophilia-associated responses to imatinib in patients with PDGFRB fusion gene-positive and BCR-ABL- acute myeloid leukemia and lymphoblastic T-cell lymphoma. Leukemia 2007; 21: negative chronic myeloproliferative disorders. Blood 2007; 109: 61–64. 1183–1188. 6 Gotlib J, Cools J. Five years since the discovery of FIP1L1-PDGFRA: what we have 12 Cross NC, Reiter A. Fibroblast growth factor receptor and platelet-derived growth learned about the fusion and other molecularly defined . Leukemia factor receptor abnormalities in eosinophilic myeloproliferative disorders. 2008; 22: 1999–2010. Acta Haematol 2008; 119: 199–206.

Long-term follow-up of ETV6–RUNX1 ALL reveals that NCI risk, rather than secondary genetic abnormalities, is the key risk factor

Leukemia (2013) 27, 2256–2259; doi:10.1038/leu.2013.136 two delayed intensification blocks and continuing therapy for a total of 2 (girls) or 3 (boys) years. ETV6–RUNX1 was determined by fluorescent in situ hybridization (FISH) using the TEL-AML1 ES probe (Abbott Diagnostics, Maiden- 9 The ETV6–RUNX1 fusion gene is present in 25% of children head, UK) or by reverse transcription-PCR. Secondary abnormalities diagnosed with B-cell precursor acute lymphoblastic leukaemia affecting ETV6 or RUNX1 were determined in a representative cohort (BCP-ALL) and is associated with an excellent outcome.1 of 247 (67%) patients (Table 1) by FISH using two dual-colour break- Although ETV6–RUNX1-positive patients have a low relapse apart probes: ETV6 (Dako Ltd, Ely, UK) and home-grown RUNX1 9 rate and good outcome after relapse,2 averting relapse in (RP11-272A03/RP11-396G11, Sanger Institute, UK). Deletions of this prevalent subgroup warrants investigation. Previous IKZF1, CDKN2A/B, PAX5, EBF1, ETV6, BTG1 and RB1 and the P2RY8– studies investigating risk factors in this subgroup have produced CRLF2 fusion were determined by multiplex ligation-dependent inconsistent results and have been hampered by small probe amplification (MLPA) in a representative cohort of 114 (31%) patient cohorts and/or treatment heterogeneity.3–6 ETV6–RUNX1 patients (Table 1) using the SALSA P335 kit (MRC Holland, alone is insufficient to cause overt leukaemia and numerous Amsterdam, The Netherlands).9 Event-free survival (EFS) and cooperating mutations have been described.7 Therefore, we overall survival (OS) were calculated from the start of treatment to sought to assess the prognostic relevance of secondary relapse/death and death, respectively. Patients without an event of abnormalities targeting the ETV6 and RUNX1 genes as well as interest were censored at the date of last contact. Survival estimates other abnormalities (for example, IKZF1 ) in a large cohort were calculated using the Kaplan–Meier method and compared of patients. using Cox proportional hazard regression models, which conformed A total of 368 children with ETV6–RUNX1 BCP-ALL were treated to the proportional hazards assumption. on MRC ALL97/99.8 Both phases, ALL97 and ALL99, included a The clinical features and outcome of ETV6–RUNX1 patients by steroid and thiopurine randomization in induction and age group, WCC group, phase of trial, NCI risk group and key maintenance. In ALL97, patients received a 3-drug induction, 2/3 secondary abnormalities are detailed in Table 1. The majority of intensification blocks, central nervous system-directed treatment ETV6–RUNX1 patients were NCI-SR and the proportion did not and continuing therapy for a total of 2 years. High-risk patients, vary between ALL97 and ALL99: 135/175 (77%) versus 148/193 identified by the Oxford Hazard Score or cytogenetics, were (77%), respectively. After a median follow-up time of 9.2 years, transferred to a more intensive protocol. In ALL99, children were 47 (13%) relapses, 54 (15%) events and 20 (5%) deaths had been stratified according to National Cancer Institute (NCI) risk to recorded (Table 1). Only two patients failed to achieve a regimen A (standard risk (SR); o10 years and white cell count complete remission (CR), both of whom died within a month. (WCC) o50 Â 109 /L) or regimen B (high risk (HR); others). Patients Among the ALL99 cohort, nine (5%) patients were SER; two died received a 3/4-drug induction (regimen A/B) and were deemed to in remission, two relapsed and five achieved sustained CR. be slow early responders (SERs) if the day 15/8 marrow contained In agreement with previous observations,10–12 the temporal X25% blasts. Patients who were SERs or had high-risk pattern of relapses was later than observed for other subtypes of cytogenetics were transferred to regimen C. After induction, ALL and 490% occurred after the end of treatment (EOT): 3 (6%) patients received consolidation, two interim maintenance blocks, very early (o18 months after diagnosis), 4 (9%) early (418 months

Accepted article preview online 2 May 2013; advance online publication, 21 May 2013

Leukemia (2013) 2242 – 2267 & 2013 Macmillan Publishers Limited