Page 58 - 2021_06-Haematologica-web
P. 58

S. Chiaretti et al.
that Ph-like patients have a lower complete remission rate, event-free survival and disease-free survival, as well as a greater MRD persistence also in a pediatric-oriented and MRD-driven adult ALL protocol, thus rein- forcing that the early recognition of Ph-like ALL patients at diagnosis is crucial to refine risk-stratification and to optimize therapeutic strategies. Clinicaltrials gov. Identifier: 02067143.
Introduction
Philadelphia-like (Ph-like) acute lymphoblastic leukemia (ALL) accounts for 15-30% of B-lineage ALL, with an increasing incidence starting from adolescence. The grow- ing interest in this subgroup of ALL arises from the distinc- tive gene expression profile - that resembles that of the true Ph-positive cases - and by the unfavorable clinical outcome.1,2 In-depth and large-scale genetic characteriza- tion has shown that the majority of Ph-like ALL cases carry fusion genes involving tyrosine kinases (i.e., ABL- class and JAK2 rearrangements), or cytokine receptor rearrangements (i.e., P2RY8/CRLF2 and IGH/CRLF2), fre- quently associated with mutations of the JAK/STAT path- way genes.3-5 Among the other co-operating events, a rele- vant role is played by IKZF1 deletions present in about 70% of cases.4-7 The possibility of recognizing these cases at diagnosis has important prognostic implications and would also pave the way to testing tyrosine kinase inhibitors (TKI) and other targeted therapeutic approaches that have proven successful in pre-clinical models and in vivo in a few relapsed patients.3,8-12 So far, several strate- gies13-15 have been reported in an attempt to identify Ph- like cases, but none of them is deemed as the gold stan- dard for the diagnostic work-up of these patients. To this end, our group recently reported a predictive tool called “BCR/ABL1-like predictor” based on the levels of expres- sion of nine genes together with CRLF2 transcript quan- tification.7 From a clinical standpoint, Ph-like patients are characterized by a worse outcome which is due to an infe- rior response to induction therapy, a higher incidence of relapses and lower survival.1,2,4 Since minimal residual dis- ease (MRD) is considered today the most important prog- nostic factor in ALL, the role of the Ph-like status has been investigated in the context of MRD-driven protocols, with contradicting results. Roberts and colleagues reported in a pediatric cohort that Ph-like patients, though displaying higher MRD levels at the end of induction, had a survival probability similar to that of non-Ph-like childhood ALL when treated with intensive therapies.16 Opposite results were obtained by Heatley et al.14 who demonstrated that, despite a risk-adjusted treatment approach, a high rate of relapse was recorded among children who were retrospec- tively identified as Ph-like. In adolescents and young adults, the results of the CALGB10403 trial, based on a pediatric inspired regimen, have shown that parameters associated with inferior survival rates were indeed repre- sented by the Ph-like signature and obesity.17 In adult cohorts, all reported studies so far agree on a shorter sur- vival likelihood for Ph-like ALL compared to non-Ph-like patients.5-7,18,19 However, the data are still insufficient to elucidate whether intensive treatments are capable of abolishing the negative impact of the Ph-like status on prognosis: conflicting results have been reported in the studies by Jain et al.20 and Herold et al.6 Likewise, the role of the Ph-like status in the context of MRD-driven clinical trials is still unclear, since the data produced by the
German study group were derived from a small cohort of patients.6
In order to clarify these aspects, we hereby evaluated the incidence and clinical-biologial features of Ph-like cases - identified using the BCR/ABL1-like predictor7 - and the prognostic role of the Ph-like profile in terms of com- plete remission (CR) achievement, MRD persistence and survival in a cohort of adult ALL patients homogeneously and intensively treated in the pediatric-oriented, MRD- driven LAL1913 GIMEMA front-line protocol for adult Ph- negative ALL.
Methods
Study population and experimental strategy
This study included B-lineage ALL patients negative for major molecular aberrations (BCR/ABL1, KT2MA and TCF3/PBX1, B- NEG) enrolled in the GIMEMA LAL1913 front-line clinical trial (clinicaltrials gov. Identifier: 02067143; Online Supplementary Figure S1) - designed for Ph-negative ALL patients aged 18-65 years - based on a pediatric-oriented backbone, in which Peg- asparaginase was administered instead of asparaginase, and on a MRD-driven transplant allocation;20 MRD time-points and MRD analysis are detailed in the Online Supplementary Materials and Methods. The EC study number approval is 5629.
Diagnostic bone marrow samples were available from 105 patients (median age 38.7 years, range, 18.2-64.7). Baseline patients’ characteristics are summarized in the Online Supplementary Table S1; there were no differences in clinical- biologial features between our cohort and the remaining popu- lation enrolled in the protocol (Online Supplementary Table S2). All cases underwent centralized molecular screening: i) the “BCR/ABL1-like predictor” assay, ii) sequencing of the JAK/STAT and RAS cascades by next-generation sequencing (NGS), iii) Multiplex Ligation-dependent Probe Amplification (MLPA), iv) targeted RNA sequencing. In 17 cases, the BCR/ABL1-like pre- dictor was not feasible due to lack of RNA (Online Supplementary Table S3; Online Supplementary Figure S2).
BCR/ABL1-like predictor
In order to detect the Ph-like cases, we applied the
“BCR/ABL1-like predictor”7 to 88 patients (Online Supplementary Materials and Methods).
Screening of recurrent mutations and deletions
The members of the JAK/STAT (JAK1, JAK2, JAK3, IL7R and CRLF2) and RAS (FLT3, NRAS, KRAS and PTPN11) pathways (181 amplicons) were sequenced by NGS (Online Supplementary Materials and Methods).
NGS experiments were performed in 91 cases (74 in common with the BCR/ABL1-like predictor analysis - 24 Ph-like and 50 non-Ph-like ALL cases -, Online Supplementary Materials and Methods and Table 3). Variants recognized as single nucleotide polymorphisms (SNP) were excluded, unless of prognostic value or previously reported in Ph-like ALL.21
Recurrent deletions (IKZF1, CDKN2A/2B, PAX5, EBF1, BTG1,
1560
haematologica | 2021; 106(6)


































































































   56   57   58   59   60