Page 140 - Haematologica March 2020

J. Woo et al.
panel (ArcherDX, Boulder, CO, USA) following the manufactur- er’s instructions. We employed highly sensitive and specific target- ed sequencing methods, exploiting unique molecular barcodes by anchored multiplex polymerase chain reaction (PCR) chemistry to tag the starting DNA material before PCR amplification.33 These methods significantly enhance the sensitivity and specificity of variant detection by filtering out duplicate reads and PCR errors. We followed variant calling procedures as previously reported.34 Pairwise associations between mutations and clinical parameters were evaluated by the Fisher exact test and Pearson correlation, corrected for the testing of multiple hypotheses. Hierarchical clus- tering of associations was performed by in-house scripts in R/Bioconductor. The Euclidean distance between variables was calculated, and the final partition of variables was obtained by the hclust function in the R package ‘stat’ with complete linkage as the argument. Details are provided in the Online Supplementary Methods.
Statistical analysis
Survival was defined as the time from transplantation to death or date of last contact. Relapse-free survival was defined as the time from transplantation to relapse or death from causes other than relapse. Non-relapse mortality (NRM) was defined as death without prior relapse. Estimates of the probability of overall and relapse-free survival were obtained by the Kaplan-Meier method,
and estimates of the probability of relapse, NRM, and GvHD were summarized using cumulative incidence estimates. NRM was considered a competing risk for relapse, and death without chron- ic GvHD a competing risk for chronic GvHD. Associations of var- ious factors with the cause-specific hazards of failure for each of these endpoints were assessed using Cox regression. All P-values are two-sided and do not incorporate adjustment for multiple comparisons; however, because of the large number of individual mutations considered, we report only associations significant at the 0.01 level of significance, unless the mutation is considered of interest based on prior studies.
Results
Conventional risk classification
As shown in Table 1, according to WHO criteria, 21 patients (16%) had CMML-0, 31 (24%) had CMML-1, 38 (30%) had CMML-2, and 37 patients (29%) had CMML- T (progressed to a blast count ≥20% with previous history of CMML). In two patients, the staging was inconclusive. In 88 patients (68%), the white blood cell count was <13 × 109/L, thus qualifying as MD-CMML, and 40 patients (31%) had MP-CMML with a white cell count ≥13 × 109/L. In one patient, the white blood cell count from the pre-HCT period was not available. Among 124 patients with cytogenetic data, 72 (56%) were considered low risk, 22 (17%) intermediate risk, and 30 (23%) high risk accord- ing to the CMML-specific cytogenetic classification.28 More than 80% of patients were “higher” risk according to the CPSS classification (i.e., 30 were high risk and 74 inter- mediate-2 risk), while 50% of patients were “higher” risk according to the MDAPS (i.e., 15 were high risk and 46 intermediate-2 risk).
Engraftment and graft-versus-host disease
One hundred twenty patients (93%) achieved sustained engraftment, as defined by absolute neutrophil counts of ≥0.5 × 105/mL for three or more consecutive days. Seven of the remaining nine patients had donor cell engraftment, as determined by chimerism analysis, but died before day 100 without achieving an absolute neutrophil count ≥0.5 × 105/mL, while two patients died with recurrent CMML. Of the 126 patients evaluable for GvHD (surviving beyond day 28 with donor cell engraftment), acute GvHD of grades II-IV developed in 93 (74%) and grades III–IV in 32 (25%). Chronic GvHD occurred in 57 patients within 2
years, for a cumulative incidence of 45%.
Relapse
Relapse or disease progression occurred in 40 patients between 7 and 2,490 (median 154) days after transplanta- tion. The estimated probability of relapse or disease progres- sion was 28% at 3 years, and 32% at 10 years (Figure 1A).
Survival
With a median follow-up of 9.3 (range 0.4-25.2) years, 39 patients are alive while 90 patients have died. The overall survival rates at 3 and 10 years were 38% and 28%, respectively, while the relapse-free survival rates at the corresponding times were 37% and 29% (Figure 1A).
Clinical determinants of post-transplant outcomes
The results of the univariate analyses are summarized in Table 3. Relapse incidence was significantly increased
Table 2. Donor and transplant characteristics. Variable
Donor Related
HLA-matched
HLA-mismatched Unrelated
HLA-matched HLA-mismatched*
Cell source
Bone marrow Peripheral blood Cord blood
Conditioning regimen Myeloablative conditioning
BU (16 mg/kg) / CY (120 mg/kg) ± ATG FLU (120 mg/m2) / BU(16 mg/kg)
TBI (2 Gy) / I-131-antiCD45
BU (7 mg/kg) / TBI (12 Gy)
BU (7 mg/kg) / CY (50 mg/kg) / TBI (12Gy)
CY (120 mg/kg) / TBI (14.4 or 13.2 Gy)
TREO (42 g/m2) / FLU (150 mg/m2) / TBI (2 Gy)
Reduced-intensity conditioning FLU (90 mg/m2) / TBI (2-3 Gy) Others†
GvHD prophylaxis CSP/MTX CSP/MMF TAC/MTX TAC/MMF CSP/MMF/MTX CSP/others
N. % of patients
42 32.6 38 29.5 4 3.1 87 67.4 68 52.7 19 14.7
34 26.4 93 72.1 2 1.6
41 31.8 13 10.1 12 9.3 11 8.5 10 7.8 10 7.8 9 7.0
21 16.3 2 1.6
43 33.3 36 27.9 35 27.1 6 4.7 2 1.6 7 5.4
*Includes two cord blood transplants. †FLU (175 mg/m2) / CY (50 mg/kg) / TBI (3 Gy), FLU (125 mg/m2) / melphalan (140 mg/m2). HLA: human leukocyte antigen; BU: busul- fan;CY:cyclophosphamide;ATG:antithymocyteglobulin; FLU:fludarabine;TBI:total- body irradiation; TREO: treosulfan; I-131-antiCD45, iodine-131 monoclonal antibody BC8; GvHD, graft-versus-host disease; CSP, cyclosporine; MMF, mycophenolate mofetil; TAC, tacrolimus; MTX, methotrexate.
654
haematologica | 2020; 105(3)