Page 172 - Haematologica May 2020

A. Lazaryan et al.
Selection of patients
Cytogenetic abnormalities included high hyperdiploidy (n=29), tetraploidy (n=9), near triploidy (n=6), low hypodiploidy (n=11), complex karyotype (n=51), monosomal karyotype (n=84), mono- somy 17 (n=21), i(17q) (n=5), del(17p) (n=6), t(1;19) (n=33), t(4;11) (n=95), t(8;14) (n=10), t(10;11) (n=8), t(11;19) (n=10), add(5q) (n=7), del(5q) (n=20), add(7p) (n=8), i(7q) (n=10), add(12p) (n=10), del(12p) (n=18), t(14;18) (n=6), del(6q) (n=48), del(7q) (n=7), mono- somy 7 (n=33), add(9p) (n=11), del(9p) (n=52), i(9q) (n=17), add(12p) (n=10), del(12p) (n=18), del(11q) (n=18), del(13q) (n=12), and trisomy 8 (n=35). Each cytogenetic abnormality was tested individually for its association with post-HCT relapse while adjusted for potential confounders. Statistically significant (P<0.05) clinical factors other than cytogenetics [conditioning reg- imen, remission status, donor type, and graft-versus-host disease (GvHD) prophylaxis among other potential confounders] were retained in the multivariable Cox proportional hazards model. Abnormalities with a hazard ratio (HR) ≥1.4 for relapse were sub- sequently grouped as adverse risk; abnormalities with a HR ≤0.6 for relapse were grouped as favorable, whereas all other abnor- malities, and normal cytogenetics, were grouped as intermediate risk. Relapse was used as the primary endpoint for evaluation of individual cytogenetic abnormalities and it was calculated as the cumulative incidence of ALL recurrence with treatment-related mortality as the competing risk. Leukemia-free survival was used as the primary endpoint for evaluation of previously established and study-derived cytogenetic risk classifications and was defined as the time to death or relapse with survivors in continuing com- plete remission censored at last follow-up. Adjusted probabilities of leukemia-free survival and relapse were calculated using multi- variable models, stratified by cytogenetic risk scheme and weight- ed by the pooled sample proportion value for each prognostic fac- tor.10,11 Overall survival was a secondary study endpoint and was defined as the time to death from any cause with surviving patients censored at last follow-up. Treatment failure (1 – leukemia-free survival) and overall mortality (1 – overall survival) were used to model all Cox regression HR estimates. SAS version 9.4 (SAS Institute, Cary, NC, USA) and GraphPad Prism version 7.04 were used for all data analysis and graphics.
Results
Study population and transplant characteristics
A description of the entire study population and the dis- tribution of the main study variables among patients with abnormal and normal cytogenetics are summarized in Table 2. The study cohort consisted predominantly of young (82% <45 years) male (63%) patients with B-pre- cursor ALL (69%). Patients with hyperleukocytosis (white blood cell count >30×109/L for B-ALL and >100×109/L for T-ALL) at the time of initial diagnosis accounted for 22% of the entire cohort and 57% of patients underwent allo- geneic HCT in CR1 with a median time to achieve CR1 of 6 weeks (range, 1-123).
Post-transplant outcomes classified by established cytogenetic schemes
Patients with abnormal cytogenetics had 5-year leukemia-free and overall survival rates of 40% and 42%, respectively, which were similar to those of patients with a normal karyotype (both P>0.6). The cytogenetic risk cat- egories defined by the MRC-ECOG, SWOG, NILG-ALL, North UK, and GIMEMA 0496 (Table 1) had no prognostic significance for leukemia-free survival, relapse, or overall survival (all P-values >0.15). However, the cytogenetic risk
The initial study population included 3,275 adults (age ≥16 years) with Ph– ALL in first or second complete remission (CR1 or CR2, corresponding to morphological remission with <5% bone marrow blasts) who underwent allogeneic HCT between 1995- 2011 and whose data were reported to the CIBMTR. Further restriction of the study population to the recipients of HLA- matched sibling and unrelated donor peripheral blood or bone marrow allografts (with consent to submit at least 100 days of post-transplant research reports) resulted in 2,903 eligible study participants. The CIBMTR data center requested original cytoge- netic reports for cases with reportedly abnormal or unknown cytogenetics at either the time of diagnosis or prior to allogeneic HCT. Cytogenetic reports were received from participating cen- ters for 1,013 cases, all of which were reviewed and validated by the study’s principal investigators (AL, MD). Data on cytogenetics from the existing CIBMTR records were used for 743 cases for which no original cytogenetic reports were received from the queried centers. For 342 cases (12%) with prior CIBMTR cytoge- netics status reported as “unknown” or “not tested” the original cytogenetic reports were requested, but not received from the transplant centers. Normal conventional cytogenetic results were confirmed with over 95% accuracy upon review of all original reports received and the remaining 805 cases with normal cytoge- netics were included in the final study sample of 1,099 patients with normal cytogenetics reported. Patients with abnormal con- ventional cytogenetics (n=632) were included in the study popula- tion after review of all available original cytogenetic reports. Thus, a total study population of 1,731 patients from 256 reporting cen- ters and 38 countries was analyzed.
Cytogenetics
Blood and marrow samples at the time of ALL diagnosis and prior to transplantation were cultured and evaluated for cytoge- netic abnormalities by G-banding according to the standard prac- tices of the participating centers. Original cytogenetic data report- ed to the CIBMTR conformed to the International System of Cytogenetic Nomenclature (ISCN).5 According to the ISCN, a clonal abnormality was defined as the presence of a gain of the same chromosome or the presence of the same structural abnor- mality in ≥2 cells or the loss of the same chromosome in ≥3 cells. A normal conventional cytogenetic result was defined as the absence of clonal abnormalities in at least 20 metaphase cells. Abnormal cytogenetics were classified according to previously established cytogenetic risk classifications for Ph– ALL (Table 1). Standard definitions for hypodiploid, hyperdiploid, complex, and monosomal karyotypes were based on the following modal chro- mosome numbers: (i) low hypodiploidy (30-39 chromosomes), (ii) high hypodiploidy (40-43), (iii) low hyperdiploidy (47-50), (iv) high hyperdiploidy (>50), (v) near triploidy (60-78), (vi) tetraploidy (>80), (vii) complex with ≥5 abnormalities6-8 (adopted here) in the absence of established translocations or ploidy abnormalities; or ≥3 abnormalities used exclusively by the Northern Italy Leukemia Group (NILG)9 (Table 1), and (viii) monosomal (≥2 autosomal monosomies or a single autosomal monosomy combined with a single structural abnormality). Fluorescence in situ hybridization (FISH) findings and/or other molecular data were available for the minority of patients and were, therefore, only used to validate cytogenetic reports when available.
Statistical analysis
Individual Ph– cytogenetic abnormalities were included in the analysis if they were detected in ≥20 patients or in <20 patients but with previously established prognostic significance in ALL.
1332
haematologica | 2020; 105(5)