Clinical utility of fluorescence in situ hybridization‐based diagnosis of BCR‐ABL1 like ( P hiladelphia chromosome like) B ‐acute lymphoblastic leukemia

This is the authors' abstract. We don't add key points for this paper.

To the Editor: The outcomes of patients with B acute lymphoblastic leukemia (B-ALL) have improved over the past few decades, mainly due to the adaptation of therapy based on risk that incorporates recurrent primary cytogenetic subtypes along with clinical risk factors for relapse. However, long term overall survival (OS) rates of B-ALL in adults remain low, with 5-year OS 20-40%,1 whereas 5-year OS is over 85% in children.2 Note, BCR-ABL1-like ALL (Ph-like ALL) is a recently characterized high risk subtype, recognized as a provisional entity in the World Health Organization (WHO), that primarily affects children and young adults.3 It was initially defined by a gene expression profile similar to that of Ph (+) ALL that activates tyrosine kinase (TK) or cytokine receptor signaling. The incidence of Ph-like ALL varies between 20-35% in different studies. Several studies have described inferior outcomes with conventional chemotherapy and long term survival rates are 10-35%.3 There are two different groups of genetic alterations described in Ph-like ALL; ABL-class fusions that phenocopy BCR-ABL1, involving ABL1, ABL2, CSF1R and PDGFRB and gene rearrangements that activate JAK/STAT signaling, including alterations of CRFL2, JAK2 and EPOR.4 In prior studies, approximately half of the cases with Ph-like ALL had a CRLF2 rearrangement and over half of those had concomitant JAK1 or JAK2 mutations.4, 5 Other common alterations were non-BCR ABL-gene fusions (9.8%), JAK2 or EPOR rearrangements (12.4%), JAK/STAT sequence mutations (7.2%), other kinase alterations (4.1%) and RAS pathway mutations (3.6%). Importantly, the majority of these alterations were potentially targetable with TKIs, JAK inhibitors or inhibitors of ALK.5 Thus, despite the complexity of multiple genetic alterations in Ph-like ALL, most of those can be identified with standard molecular and cytogenetic approaches and have therapeutic potential. We have designed a fluorescence in situ hybridization (FISH) panel with probes targeting most therapeutically relevant rearrangements seen in Ph-like ALL. The purpose of designing an in-house FISH panel is to provide a rapid test to detect the Ph-like abnormalities in patients that could potentially benefit from the addition of TKIs or JAK-inhibitors to conventional chemotherapy. In this study, we aimed to assess the utility of our in-house FISH panel in a cohort of Ph (−) B-ALL patients. We retrospectively identified all patients with Ph (−) B-ALL in a clinically annotated database of patients with B-ALL treated at Mayo Clinic Rochester from 1995-2017. Both adult and pediatric patients were included. Adolescent and young adults (AYA) were defined as patients 15-39 years of age. With FISH, a targeted Ph-like ALL panel was retrospectively performed in all cases, with available fixed frozen bone marrow cell pellets at the time of diagnosis (n = 112), in the Mayo Clinic Cytogenetics Core Laboratory. This FISH panel was developed by our Clinical Genomics Laboratory in order to identify Ph-like arrangements and included the following probes: 1) BCR-ABL1 for ABL1 abnormalities. If additional ABL1 signals not fused with BCR were identified, an ABL1 break apart probe was utilized to detect a potential ABL1 rearrangement, 2) t(1q25;var) for ABL2 rearrangements, 3) t(5q33;var) for PDGFRB or CSF1R rearrangements (Figure S1), 4) t(9p24.1;var) for JAK2 rearrangements, 5) t(Xp22.3;var) or t(Yp11.32;var) for CRLF2 rearrangements (Cytocell, NY). Reflex probes for CRLF2 rearrangements included a double fusion strategy for: 1) t(X;14)(p22.3;q32) or t(Y;14)(p11.32;q32) for CRLF2/IGH fusion, or 2) t(Xp22.3;var) or t(Yp11.32;var) for P2RY8 rearrangements (Cytocell, NY). Mate pair sequencing (MPseq) was performed on three specimens (patients two, seven, and eight) to further clarify two CRLF2/P2RY8 abnormalities and an atypical JAK2 abnormality. The OS was defined as time from diagnosis until death or last follow-up. Treatment-Free Survival (TFS) was defined as time from treatment to relapse or last hematology visit, and calculated accounting for competing risk of death using the Fine Gray method. A two-sided P value of <.05 was considered statistically significant. All statistical analyses were performed using SAS 9.4 (SAS Institute, Cary, NC). Between 1995 and 2017, 270 B-ALL cases were treated at Mayo Clinic Rochester;144 adults, 72 AYA and 54 children. Among these, 204 were negative for the Ph chromosome and 112 of these Ph (−) cases had available frozen bone marrow fixed cell pellets. Our FISH testing strategy identified a Ph-like abnormality in 10 cases (8.9%). The baseline characteristics of all 112 patients are summarized in Table S1. Two patients were children (patients one and two), three were AYA (patients 3-5) and five were adults (patients 6-10) with median ages 4.4, (range 3.1-5.8) 34.4, (range 30.1-35.5) and 59.2 years, (range 40.2-63.6) respectively. The clinical outcomes and cytogenetics of these 10 patients are summarized in Table 1. Both children achieved a complete response (CR) with chemotherapy and are disease-free, with a median follow up of 6.4 (range 6-6.7) years. All three AYA died of refractory ALL. Patient three had primary refractory disease, whereas patients four and five relapsed 15 and 121 months after achieving a CR. Patient five had a double umbilical cord blood allogeneic stem cell transplant (SCT), however, he suffered a relapse 3 years later. Four out of five adult patients achieved a CR with induction, while one patient treated on the ECOG trial, E1910, died of complications related to treatment. Among the four patients that achieved a CR, 3 patients proceeded with allogeneic SCT and are currently disease-free (patients six, seven and 10). The fourth patient (patient nine) also remains in remission at last follow up. Application of our Ph-like FISH panel identified a loss of 5′ CRLF2 in 9 cases (patients 1-4, 6-10). Reflex testing demonstrated CRFL2/IGH fusions in three cases (patients three, seven, 10), and showed CRLF2/P2RY8 fusion in six cases (patients one, two, four, six, eight, nine). Finally, an atypical (2F1R) FISH signal pattern indicating two intact copies of the JAK2 probe, and an additional copy of 3′ portion of the JAK2 probe was identified in one case (patient five) (Figure S2). MPseq identified a rearrangement that involved the 3′ region of the JAK2 probe, which appeared to fuse 2 genes, CD274 (PDL1) with DLGAP1, on chromosome 18p11 in patient 5 (Figure S3). This is not a known rearrangement and the significance is uncertain, thus highlighting the importance of MPseq in detection of partner genes in patients with atypical FISH results to document the presence or absence of gene fusions. At a median follow up of 3.8 (range 0.2-9.4) years, OS of patients with Ph-like B-ALL (n = 10) was 50% (five of 10 patients alive). Median OS was 2 (IQR 0.2-NR) years for patients that did not undergo SCT (n = 6) versus not reached (NR) (IQR 1.2-NR) for patients that underwent SCT (n = 4). Causes of death were disease relapse (n = 3), SCT complications (n = 1) and infectious complications (n = 1). In comparison, OS of patients with Ph (−) B-ALL and available pellets (n = 102) was 53.9% (55 of 102 patients alive). Median TFS of Ph-like B-ALL patients was NR (IQR 0.5-NR) and median TFS of Ph (−) cases was NR (IQR 9.5-NR) (Figure S4). Although preclinical data and data from small series demonstrate that Ph-like ALL is responsive to targeted therapies, early diagnosis, allowing for the prompt administration of TKIs, remains a challenge.4-6 Genome wide assays, such as RNA sequencing (RNAseq) and array based techniques, are able to identify unique gene expression profiles and gene fusions, however, they are labor intense and not incorporated into timely clinical management. FISH, on the other hand, can be performed on readily accessible material. However, there is currently no commercially available Ph-like FISH panel. In comparison to genome wide assays, FISH may underestimate the true prevalence of Ph-like ALL, however it detects most therapeutically relevant abnormalities, with the exception of EPOR rearrangements (seen in 4% of cases and technically challenging for detection with FISH based strategies). Finally, RNAseq identifies chimeric RNA fusions but not upregulation events, such as the CRLF2-IGH rearrangement, which leads to CRLF2 overexpression by putting CRLF2 in proximity of the IGH enhancer. On the other hand, chromosomal microarray does not detect balanced rearrangements, but detects unbalanced rearrangements, such as the microdeletion mechanisms involving EPOR or resulting in CRLF2/P2RY8 fusion. There are no established guidelines on how to diagnose Ph-like cases and different groups are starting to build diagnostic algorithms ranging between targeted testing in patients with high clinical suspicion, to whole transcriptome sequencing in all patients. At Mayo Clinic, we have developed a tiered diagnostic approach for new B-ALL cases which incorporates conventional chromosome analysis, FISH, chromosomal microarray and MPseq and is the same regardless of the patient's age (Figure S5). Chromosome analysis and FISH are performed on all new cases, while microarray and MPseq are performed when there is a normal or an atypical pattern of rearrangements on chromosome and/or FISH testing, and/or a concern for either a masked hyplodiploid clone or a structural abnormality suggesting involvement of a kinase gene. In conclusion, our panel is a practical tool that can be used to identify Ph-like ALL cases in a timely manner, with potential to guide treatment decisions. We also demonstrate the importance of selective use of whole genome techniques, such as MPseq, in better defining gene rearrangements in select cases. Current publication is supported in part by grants from the “The Henry J. Predolin Foundation for Research in Leukemia, Mayo Clinic, Rochester, MN, USA and the Division of Hematology, Mayo Clinic, Rochester, MN, USA, Small Grants Program”. Dr. Patnaik has served on the advisory board for StemLine Pharmaceuticals. Figure S1 FISH results demonstrating abnormal signal patterns associated with a PDGFRB or CSF1R translocation. Figure S2 FISH results of the representative cases with Ph-like rearrangements. The numbers of the samples correspond to the number of the cases on Table 1. A. CRLF2/IGH fusion (patient 3) B. Rearrangement of CRLF2 and P2RY8 (at Xp22.33/Yp11.32) likely representing CRLF2/P2RY8 fusion (patient 4). C. Atypical JAK2 rearrangement (patient 5). FISH = fluorescence in situ hybridization, Ph = Philadelphia chromosome. Figure S3 Mate pair sequencing results demonstrating CRLF2/IGH fusion present on the derivative chromosome 14. Of note, deletions were observed within the IGH locus as well as the region just proximal to CRLF2 on the X chromosome. Figure S4 A. OS of Ph-like and Ph-negative B-ALL cases. B. TFS of Ph-like and Ph-negative B-ALL cases. OS = overall survival, TFS = treatment free survival. Figure S5 Diagnostic algorithm for the diagnosis of B-cell ALL. ALL = acute lymphoblastic leukemia, FISH = fluorescence in situ hybridization, BALLF = B-ALL FISH, CMAH = chromosomal microarray, MTRBM = mate pair targeted rearrangements bone marrow, PCR = polymerase chain reaction. Table S1 Baseline characteristics of Ph-like and Ph-negative B-ALL cases Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.