Page 59 - 2021_06-Haematologica-web

Ph-like ALL correlates with MRD+ and outcome
Table 1. Comparison between Philadelphia-like (Ph-like) and non-Ph-like clinical features.
Ph-like non-Ph-like P
N 28 60
Age, median (range)
Wbc x109/L, median (range) Hb g/dL, median (range) Plt x109/L, median (range) Sex, N (%)
M
F
Risk category, N (%)
Standard risk
No Standard risk
42.24 (18.18-64.53) 3.34 (0.23-347) 8.70 (3.70-13.00) 40 (1.23-399)
19 (67.9%)
9 (32.1%) 14 (56%)
11 (44%)
34.52 (18.23-64.59) ns 5.74 (1-75.5) ns
9.75 (5.00-15.70)
66.5 (7.5- 630) ns
34 (56.7%) ns
26 (43.3%)
34 (63%) ns
20 (37%)
0.034
Ph-like: Philadelphia-like; N: number; WBC: while blood cell; Plt: platelet; M. male; F: female; ns: not significant.
RB1, ETV6 and CRLF2) were screened in 87 samples (70 in com- mon with the BCR/ABL1-like predictor analysis - 22 Ph-like and 48 non-Ph-like ALL cases -, Online Supplementary Table S3), by the Salsa MLPA P335 ALL-IKZF1 kit (MRC-Holland, Amsterdam, the Netherlands) and analyzed according to the Coffalyser manual.22 P2RY8/CRLF2 was inferred when a deletion within the PAR1 region was documented. Samples were defined IKZF1+ CDKN2A/2B and/or PAX5 when IKZF1 deletion co- occurred with CDKN2A/2B and/or PAX5 deletions.23
Targeted RNA-sequencing and FISH analysis
In order to detect fusion genes, libraries were prepared using the TruSight RNA Pan-Cancer Panel (Illumina, San Diego, CA) kit, targeting 1385 cancer- genes (Online Supplementary Materials and Methods).
Double-color fluorescence in situ hybridization (FISH) studies were performed in 20 B-ALL, 13 Ph-like and seven non-Ph-like with high levels of CRLF2 expression (Online Supplementary Materials and Methods).
Overall, 85 cases were screened (25 Ph-like and 60 non-Ph-like ALL cases, Online Supplementary Table S3).
Statistical analyses
Patients’ characteristics were compared by chi-squared or Fisher’s exact test for categorical variables and Wilcoxon test for continuous data. Overall survival (OS), disease-free survival (DFS) and event-free survival (EFS) were estimated by the Kaplan-Meier product-limit and compared by log-rank test. OS was defined as the time between the date of diagnosis and death for any cause; patients still alive were censored at the time of the last follow-up. DFS was defined as the time between the evalu- ation of CR - after the induction phase - and relapse or death in CR; patients still alive in first CR, were censored at the time of the last follow-up. Finally, EFS was defined as the time between diagnosis and non-achievement of CR in the induction phase, relapse or death in CR, whichever occurred first; patients still alive, in first CR, were censored at the time of the last follow- up.
Multivariate analysis was performed with the Cox propor- tional hazards regression model to adjust the effect of BCR/ABL1-like predictor for clinically relevant parameters (age, white blood cell [WBC] count, hemoglobin [Hb] level, platelet count, sex and allogeneic transplant [HSCT] and for genetic aberrations impacting on prognosis [IKZF1+ CDKN2A/2B and/or PAX5, K/NRAS clonal mutations, JAK/STAT clonal muta- tions].21,22 All tests were 2-sided, accepting P<0.05 as statistically
significant. All analyses relied on the SAS v9.4 software. Study data were collected and managed using REDCap24 electronic data capture tools hosted at the GIMEMA Foundation.
Results
Incidence and clinical features of Ph-like acute lymphoblastic leukemia
We identified 28 (31.8%) Ph-like cases with a median score of 0.85 (range, -0.18 to 6.37); the remaining 60 cases had a median score equal to -1.24 (range, -1.7 to -0.33). Overall, the clinical features (age, sex, WBC and platelet counts) at diagnosis of Ph-like and of non-Ph-like cases were similar. Ph-like patients had lower Hb levels (P=0.016), as detailed in Table 1. The incidence of Ph-like ALL cases was slightly higher in adults (≥36 years) than in young adults (18-35 years), being 36.2% (17 of 47) and 26.8% (11 of 41), respectively. As per clinical protocol guidelines, only 45% of Ph-like cases were assigned to the high-risk category.
Genetic features of Ph-like acute lymphoblastic leukemia cases
The identified Ph-like cases were evaluated for the fol- lowing genetic features: CRLF2 expression levels (n=28), JAK/STAT and RAS pathways mutations (n=24), CNA aberrations (n=22) and fusion genes (n=23), the latter either by RNA-sequencing and/or FISH. A CRLF2 overex- pression, defined as ΔCt <8,25 was found in 10 of 28 Ph- like cases (35.7%). Among the CRLF2-high cases with a ΔCt value <4.5, we observed that three harbored a CRLF2 rearrangement, with one displaying a concomitant F232C CRLF2 mutation. Of the remaining seven CRLF2-high cases, three had a concomitant rearrangement (two ABL- class and one DDX3X/USP9X), one displayed a JAK1 and RAS mutation, and in two cases the mutational screening could not be performed due to lack of genomic material; finally, in one case no additional lesions were detected. Among the 24 Ph-like cases analyzed for the mutational status, we detected a total of 13 JAK/STAT pathway mutations - nine clonal and four subclonal - in eight cases (33.3%). Despite a high heterogeneity among samples, the most frequently mutated genes were JAK1 - affected by five mutations mainly targeting the hotspot V658 - and JAK2 - affected by three mutations focused in the
haematologica | 2021; 106(6)
1561