Patient characteristics and correlation with clinical and biological parameters in the whole cohort
A total of 1472 AML patients underwent allo-HSCT during this period, with 1247 (85%) in CR1, 148 (10%) in CR2 and 77 (5%) in CR3. A total of 111 patients (8%) did not achieve CR at the time of pretransplant evaluation. The median age of our patients was 34 years (range 1–65 years), and 811 (55%) of the patients were male. All CSF samples were examined by CC and MFC, which were performed on 1472 samples. Before transplantation, 1428 (97%) patients were CNS negative (CNS-), while 44 (3%) were CNS positive (CNS+). Of these 44 patients, 10 (23%) were positive for CC, 6 (14%) were positive for both CC and MFC (CC+), and they were diagnosed with CNSL; the other 28 (64%) were only MFC positive (MFC+). All CSF-positive patients received intrathecal chemotherapy according to the CNSL. After intrathecal injection, CSF test results were negative for all 16 patients in the CC + group before transplantation. CSF MFC results were negative in 26 patients in the MFC + group but were still positive in 2 patients. The CNS + group had a greater white blood cell count at diagnosis than the CNS- group did (P < .001). The distributions of sex, age, CR status, and risk stratification were not significantly different between the two groups.
After PSM according to sex, age and WBC count at diagnosis before transplantation, a total of 175 CNS (control) and 44 CNS (28 MFC + and 16 CC+) patients were included in the analysis (Table 1).
In the CC + cohort, the median follow-up period was 65 months (IQR 6 to 106, range 3 to 128) for all 16 patients and 92 months (IQR 71 to 109, range 52 to 128) for the 10 surviving patients. In the MFC + group, the median follow-up period was 56 months (IQR 16 to 69, range 2 to 93) for all 28 patients and 61 months (IQR 54 to 73, range 51 to 93) for the 19 surviving patients. In the control group, the median follow-up period was 69 months (IQR 54 to 96, range 2 to 143) for all 175 patients and 75 months (IQR 60 to 101, range 49 to 143) for 145 surviving patients.
The effect of CNS abnormalities on engraftment and GVHD
All patients were successfully engrafted. In the CC+, MFC+, and control groups, the median durations of myeloid engraftment were 14 days (IQR, 11–16), 14 days (IQR, 12–17) and 13 days (IQR, 12–15) after transplantation, respectively, and there was no significant difference in the duration of myeloid engraftment among the three groups (P = .16). The rates of platelet engraftment were 99%, 100% and 96%, respectively. The median durations of platelet engraftment were 15 days (IQR, 11–22), 17 days (IQR, 11–32) and 16 days (IQR, 12–20) after transplantation. The difference was not statistically significant (P = .73) (Table 2). There was no significant difference in the incidence of Grade II-IV acute GVHD (13% vs. 21% vs. 15%, P = .67) or cGVHD (38% vs. 46% vs. 46%, P = .81) among the three groups.
Table 1. Patient and characteristics after PSM
Characteristics
|
CNS-(n=175)
|
CNS+(n=44)
|
P value
|
Age, Y, median (range)
|
31 (1, 62)
|
33 (2, 61)
|
0.760
|
Gender, no. (%)
|
|
|
0.610
|
male
|
104 (60%)
|
28 (64%)
|
|
female
|
71 (40%)
|
16 (36%)
|
|
French-American-British subtype, no. (%)
|
|
|
0.020
|
M1
|
19 (11%)
|
2 (5%)
|
|
M2
|
58 (33%)
|
17 (39%)
|
|
M3
|
0 (0%)
|
3 (7%)
|
|
M4
|
34 (19%)
|
6 (14%)
|
|
M5
|
56 (32%)
|
16 (36%)
|
|
M6
|
1 (1%)
|
0 (0%)
|
|
M7
|
3 (2%)
|
0 (0%)
|
|
Undetermined
|
4 (2%)
|
0 (0%)
|
|
WBC, × 109/L median (range)
|
105 (1, 356)
|
79 (1, 348)
|
0.662
|
Hemoglobin, g/L median (range)
|
91 (29, 164)
|
93 (41, 132)
|
0.606
|
Platelet, × 109/L median (range)
|
45 (6, 856)
|
55 (14, 176)
|
0.211
|
Blasts in BM, % median (range)
|
77 (21, 100)
|
80 (19, 99)
|
0.330
|
ELN risk group, no. (%)
|
|
|
0.167
|
Low risk,
|
26 (15%)
|
9 (21%)
|
|
Intermediate risk
|
118 (67%)
|
23 (52%)
|
|
High risk
|
31 (18%)
|
12 (27%)
|
|
CR status pretransplant, no. (%)
|
|
|
0.016
|
CR 1
|
156 (89%)
|
32 (73%)
|
|
CR 2
|
13 (7%)
|
7 (16%)
|
|
CR 3
|
6 (3%)
|
5 (11%)
|
|
Transplant type, no. (%)
|
|
|
0.724
|
Identical-sibling
|
36 (21%)
|
8 (18%)
|
|
Haploidentical
|
139 (79%)
|
36 (82%)
|
|
Abbreviations: propensity score matching (PSM), year (Y), central nervous system (CNS), white blood cell (WBC), blasts in bone marrow (BM), European LeukemiaNet (ELN), complete remission (CR)
Table 2
The effect of CNS abnormalities on outcomes.
Outcomes
|
CNS-
(n = 175)
|
CC+
(n = 16)
|
MFC+
(n = 28)
|
P value
|
CNS- vs. CC+
|
CNS- vs. MFC+
|
CC + vs. MFC+
|
Rates of myeloid engraftment (%)
|
100%
|
100%
|
100%
|
1.000
|
1.000
|
1.000
|
Rates of platelet engraftment (%)
|
99%
|
100%
|
96%
|
0.667
|
0.524
|
0.444
|
Time to myeloid engraftment, median (range)
|
13 (9, 45)
|
14 (10, 21)
|
14 (11, 21)
|
0.907
|
0.054
|
0.264
|
Time to platelet engraftment,median (range)
|
16 (7, 161)
|
15 (10, 137)
|
18 (6, 187)
|
0.734
|
0.408
|
0.521
|
Rates of grade Ⅱ~Ⅳ aGVHD, n, (%)
|
27 (15%)
|
2 (13%)
|
6 (21%)
|
0.755
|
0.669
|
0.460
|
Rates of cGVHD, n, (%)
|
80 (46%)
|
6 (38%)
|
13 (46%)
|
0.527
|
1.000
|
0.565
|
The 8-year cumulative incidence of relapse (%)
|
14.5 ± 0.1%
|
31.2 ± 1.4%
|
32.1 ± 0.8%
|
0.007
|
The 8-year cumulative incidence of non-relapse mortality (%)
|
6.3 ± 0.1%
|
6.3 ± 0.4%
|
3.6 ± 0.1%
|
0.861
|
The 8-year leukemia free survival (%)
|
79%±0.03%
|
63%±0.12%
|
64%±0.09%
|
0.048
|
0.023
|
0.900
|
The 8-year overall survival (%)
|
83%±0.03%
|
63%±0.12%
|
68%±0.09%
|
0.021
|
0.046
|
0.614
|
Abbreviations: central nervous system (CNS), conventional cytology (CC), multiparameter flow cytometry (MFC), graft-versus-host disease (GVHD) |
The effect of CNS abnormalities on relapse and NRM
In the CC + group, 5 patients relapsed at a median of 4 months posttransplant (IQR 3 to 5, range 3 to 6) (4 with haematological relapse and 1 with CNS leukemia relapse), and 6 died (relapse in 5 patients and NRM in 1 patient) at a median follow-up of 5 months posttransplant (IQR 4 to 9, range 3 to 13). The 8-year NRM and CIR were 6.3 ± 0.4% (Fig. 1A) and 31.2 ± 1.4% (Fig. 1B), respectively. In the MFC + group, 9 patients relapsed at a median of 5 months posttransplant (IQR 2 to 13, range 1 to 16) (6 with haematological relapse and 3 with CNS leukemia relapse), and 9 died (relapse in 8 patients and NRM in 1 patient) at a median follow-up of 11 months posttransplant (IQR 4 to 17, range 2 to 23). The 8-year NRM and CIR were 3.6 ± 0.1% (Fig. 1A) and 32.1 ± 0.8% (Fig. 1B), respectively. In the CNS- group, 23 patients experienced haematological relapse at a median of 10 months post-HSCT (IQR 5 to 25, range 2 to 79), and 30 died of relapse (n = 19) or transplant-related mortality (n = 11) at a median follow-up of 9 months (IQR 6 to 20, range 2 to 57). The 8-year NRM and CIR were 6.3 ± 0.1% and 14.5 ± 0.1%, respectively. The 8-year cumulative incidences of NRM were similar in the 3 groups (P = .816) (Fig. 1A), while the 8-year CIR was significantly lower in the control group than in the other 2 groups (P = .007) (Fig. 1B).
The effect of CNS abnormalities on survival
The LFS in the control group was 79%, which was significantly better than that in the CC+ (63%, P = .048) and MFC + groups (64%, P = .021) (Fig. 2A). The 8-year OS in the control group was 83%, which was significantly better than that in the CC+ (63%, P = .021) and MFC + groups (68%, P = .046) (Fig. 2B). The survival outcomes were not significantly comparable between the CC + and MFC + groups.
Multivariate analysis
The above outcomes indicated that both MFC + and CC + predicted poorer outcomes post allo-HSCT. To further explore the predictive role of MFC in CSF, the clinical impact of MFC positivity as an independent prognostic factor for OS was also assessed via multivariate Cox proportional hazards analysis applied to models including prognostic factors proven to be significant in univariate analysis (Table S1). Multivariate analysis confirmed that MFC positivity at the time of diagnosis was a factor independently affecting OS and LFS (Table 3).
Table 3
Results of multivariate analysis of outcomes.
Outcome
|
Hazard ratio (95%Confidence interval)
|
P value
|
the whole cohort (n = 292)
|
|
|
Disease free survival
|
|
|
Age
|
1.044 (1.021–1.067)
|
0.000
|
FAB classification
|
1.280 (1.050–1.561)
|
0.014
|
MFC positivity
|
2.435 (1.163–5.098)
|
0.018
|
Overall survival
|
|
|
Age
|
1.039 (1.014–1.064)
|
0.002
|
PLT count at diagnosis
|
1.004 (1.001–1.007)
|
0.016
|
ELN risk stratification
|
1.993 (1.175–3.380)
|
0.010
|
MFC positivity
|
2.303 (1.049–5.057)
|
0.038
|
Abbreviations: French-American-British subtype(FAB), multiparameter flow cytometry (MFC), Platelets (PLT), European LeukemiaNet (ELN) |