The Impact of Pretransplant Use of Tyrosine Kinase Inhibitors on Allogeneic Stem Cell Transplantation in Patients with Chronic Myeloid Leukemia - A Single-institution Retrospective Study -

Abstract

Objective

Chronic myeloid leukemia (CML) is a malignant hematological disorder, and allogeneic stem cell transplantation (allo-SCT) was its only curative treatment until the introduction of tyrosine kinase inhibitors (TKIs). Allo-SCT is still considered for CML patients who are resistant to TKIs and in an advanced phase. Currently, second- and third-generation (2/3G TKIs are typically incorporated into the first-line treatment of CML. However, the impact of 2/3G TKIs on subsequent allo-SCT remains unclear. We therefore evaluated the effect of 2/3G TKIs on allo-SCT.

Methods

We retrospectively evaluated the effect of pretransplant therapy with TKIs on the outcome of allo-SCT for CML using clinical data at our institution.

Patients

Thirty-two CML patients who received their first allo-SCT procedure at our institute from 2001 to 2020 were included. We divided the patients into three subgroups based on TKI treatment before allo-SCT. Patients receiving no TKIs, only imatinib (IM), and 2/3G TKIs were classified into the Non-TKI, IM, and 2/3G TKI groups, respectively.

Results

In a univariate analysis, the pretransplant use of 2/3G TKIs was significantly associated with a higher 5-year overall survival (91.7%) and relapse-free survival (75.0%) than the use of IM (37.5% and 12.5%) in patients presenting with or progressing to the advanced phase. In addition, pretransplant use of 2/3G TKIs did not increase the incidence of graft-versus-host disease (GVHD).

Conclusion

We demonstrated that the pretransplant use of 2/3G TKIs was safe and improved the outcome of CML patients who presented with or progressed to the advanced phase without increasing the frequency of GVHD.

Introduction

Chronic myeloid leukemia (CML) is a malignant hematological disease characterized by uncontrolled proliferation of myeloid cells in the bone marrow (1). CML is typically associated with a specific chromosomal abnormality known as the Philadelphia chromosome, which results from a reciprocal translocation between chromosomes 9 and 22. This translocation leads to the formation of a BCR-ABL fusion gene, which encodes a constitutively active tyrosine kinase. Consequently, if left untreated, CML acquires mutations over time, progressing from the initial phase, the chronic phase (CP), to the accelerated phase (AP), to the most advanced stage, the blastic phase (BP) (1).

Until the advent of the first tyrosine kinase inhibitor (TKI), imatinib (IM), the prognosis of patients with CML was dismal, and allogeneic stem cell transplantation (allo-SCT) was the only curative treatment. Following the successful introduction of IM, second- and third-generation (2/3G) TKIs, such as dasatinib, nilotinib, bosutinib, and ponatinib, were developed. These newer TKIs demonstrated a stronger treatment response than IM in not only newly diagnosed CML-CP (2,3) but also patients with IM resistance/intolerance (4-6) and AP/BP (7-9). As a result, the treatment landscape for CML patients has shifted, and 2/3G TKIs have become the preferred first-line treatment option.

Even with the introduction of these newer TKIs, however, allo-SCT remains an important treatment option for patients presenting with or progressing to BP, patients progressing to AP while on TKI treatment, and patients in the first chronic phase (CP1) with resistance or intolerance to multiple TKIs (10,11). Based on data from the Japanese Data Center for Hematopoietic Cell Transplantation (JDCHCT), after the introduction of TKIs, the number of allo-SCT procedures for CML in Japan continued to decline, reaching 96 in 2003 and 54 in 2007 (12); however, the number remained relatively unchanged thereafter, indicating that allo-SCT remains an essential treatment option for CML.

Currently, CML patients undergoing allo-SCT receive at least one 2/3G TKI prior to allo-SCT, and while their use prior to transplantation can indeed control disease, it may also increase the risk of lethal complications related to allo-SCT. In allo-SCT, recipient patients undergo intense chemotherapy and/or radiation, called conditioning regimens, and are then transplanted with donor cells. As a result, fatal complications, such as veno-occlusive disease (VOD; wherein hepatic sinusoidal vessels are damaged and the blood flow is obstructed) (13), graft-versus-host disease (GVHD), and opportunistic infections, can occur and cause mortality after allo-SCT. Therefore, it is crucial to understand the impact of 2/3G TKIs on allo-SCT.

Several studies have reported the effect of the pretransplant use of first- and second-generation TKIs on the outcome of allo-SCT (14-19) and suggested that their use prior to allo-SCT did not increase transplant-related toxicity. However, few studies have investigated the effects of the 3G TKI ponatinib. Therefore, the prognostic impact of TKI use before transplantation remains unclear.

To clarify this issue, we retrospectively analyzed the data of CML patients undergoing allo-SCT at our institute, focusing on the impact of the pretransplant use of TKIs on transplant outcomes.

Materials and Methods

Patients and definitions

This retrospective study analyzed the clinical records of patients with CML at Keio University Hospital (Tokyo, Japan). Thirty-two patients who underwent their first allo-SCT procedure at our institute between 2001 and 2020 were included in this study. To analyze the impact of pretransplant use of TKIs on the outcomes of patients with CML receiving allo-SCT, we divided the patients into three subgroups based on TKI treatment before allo-SCT: the Non-TKI group included patients who did not receive any TKI; the IM group included patients who received only IM, and the 2/3G TKI group included patients who received at least one 2G and/or 3G TKI.

We also divided the patients into three subgroups according to the disease phase at the diagnosis (20) and added another category (progressing phase) based on the trajectory changes of the disease phase between diagnosis and transplant. Patients transplanted in the second CP (CP2) were either diagnosed in the AP or BP or diagnosed initially as CP but progressed to more advanced disease and returned to CP by the time of allo-SCT (18,21), and progressing phase CML was defined as CP2 or AP or BP at the time of SCT. When patients presented with CML-BP, we performed neutrophil fluorescence in situ hybridization and ruled out the possibility of major BCR-ABL-positive acute lymphoblastic leukemia or mixed-phenotype acute leukemia.

This study was approved by the Ethics Committee of the Keio University School of Medicine.

Responses were defined using the European LeukemiaNet (ELN) criteria (10). Transplantation-related risk was assessed using the European Group for Blood and Marrow Transplantation (EBMT) score (22). The indications for allo-SCT for CML were determined based on the available National Comprehensive Cancer Network (NCCN) guidelines and the ELN recommendations of the era. Pretransplant TKI treatment was continued until the day before the start of the conditioning regimen. Relapse included molecular relapse, as previously reported (23-25), and was defined as follows: progression to AP/BP, hematologic relapse, cytogenetic relapse, or molecular relapse. Molecular relapse was defined as positivity for BCR::ABL transcripts in quantitative reverse transcription polymerase chain reaction (RT-PCR) with a BCR::ABL to ABL of ≥10^-5 in at least 2 measurements. We did not routinely perform prophylactic TKI treatment after allo-SCT and started post-transplant TKI treatment if patients receiving allo-SCT met the definition of relapse after transplantation. Acute and chronic GVHD was graded according to clinical consensus criteria (26,27). BCR::ABL was measured at our institute for clinical genetics. In terms of conditioning regimens in allo-SCT, myeloablative and reduced-intensity conditioning were defined according to the definition of the Center for International Blood and Marrow Transplant Research (28).

Statistical analyses

The differences between the groups were compared using Fisher's exact test or the Mann-Whitney U test, as appropriate. The probabilities of the overall survival (OS), relapse-free survival (RFS), and post-relapse survival (PRS) were calculated using the Kaplan-Meier method, with the PRS defined as the time from the first relapse until the last follow-up. Estimates of non-relapse mortality (NRM) and relapse/progression were calculated using the cumulative incidence rate to accommodate competing risks and were compared using Gray's test. The following variables were considered in the univariate analysis: age at SCT, sex, donor-recipient sex matching, duration from the diagnosis to transplant, disease status at the diagnosis and at transplant, use of TKIs before transplant, types and numbers of TKIs used before transplantation, type of donor, use of TKIs after transplant, and EBMT score. Significance was set at p<0.05. All statistical analyses were performed using EZR, a graphical user interface for R (29).

Results

Patients and pretransplant treatment characteristics

A total of 32 patients were included in the analysis. Patient characteristics are shown in Table 1. In the 2/3G TKI group, seven patients received not only IM but also a 2G TKI, one patient received a 2G TKI, and four patients received a 3G TKI. The year of SCT, disease status at the diagnosis, disease status at SCT, number of TKIs used before transplant, EBMT risk score, and pretransplant chemotherapy were significantly different among treatment groups. There were more patients with AP/BP at the diagnosis and CP2 at transplant in the 2/3G TKI group than in the IM and Non-TKI groups. In addition, significantly more patients achieved partial cytogenetic response (PCyR), complete cytogenetic response (CCyR), or molecular response (MR) in the 2/3G TKI group than in the IM and Non-TKI groups. The frequency of patients achieving a complete hematologic response (CHR) was not significantly different among the subgroups. In addition, the EBMT risk score, number of TKIs used before transplant, and number of patients receiving pretransplant chemotherapy were significantly higher in the 2/3G TKI group than in the IM and Non-TKI groups.

Table 1.

Patients and Pretransplant Treatment Characteristics in Entire Patients.

	Total	Non-TKI	IM	2/3G TKI	p value
No. of patients	32	9	9	14
Median age at SCT, years [range]	34 [24-54]	33 [25-49]	41 [24-45]	42 [24-54]	0.34
Age at transplantation, y
<40	14	6	3	5	0.33
≥40	18	3	6	9
Sex, n
Male	17	6	4	7	0.74
Female	15	3	5	7
Median months from diagnosis to SCT [range]	19 [2-97]	19 [7-71]	24 [3-47]	13 [2-97]	0.77
Year of SCT, n
2001-2010	21	9	9	3	<0.01
2011-2020	11	0	0	11
Disease status at diagnosis, n
CP	22	9	7	6	0.01
AP/BP	3/7	0	1/1	2/6
Disease status at SCT, n
CP1	11	8	1	2	<0.01
CP2	14	0	4	10
AP/BP	4/3	1/0	2/2	1/1
Disease status at SCT, n
AP/BP	4/3	1/0	2/2	1/1	<0.01
Non-CHR	4	4	0	0
CHR	5	4	1	0
PCyR	5	0	3	2
CCyR	9	0	1	8
MR	2	0	0	2
Achieving CHR at SCT, n
No	11	5	4	2	0.10
Yes	21	4	5	12
Achieving CCyR at SCT, n
No	21	9	8	4	<0.01
Yes	11	0	1	10
Type of donor and stem cell source, n
HLA-matched related BM/PBSC	3/5	0/2	2/1	1/2	0.30
HLA-mismatched related BM/PBSC	0	0	0	0
HLA-matched unrelated BM/PBSC	17/0	7/0	3/0	7/0
HLA-mismatched unrelated BM/PBSC	0	0	0	0
Cord blood from unrelated donor	7	0	3	4
EBMT risk score
0-1	1	0	1	0	<0.01
2-4	17	9	2	6
≥5	14	0	6	8
Numbers of TKIs used before transplant, n
0	9	9	0	0	<0.01
1	9	0	9	0
≥2	14	0	0	14
Pretransplant chemotherapy, n
No	15	8	3	4	0.01
Yes	17	1	6	10
Conditioning, n
MAC					0.18
TBI-based	28	7	7	14
Non-TBI based	4	2	2	0
RIC	0	0	0	0
TKI treatment after SCT, n
No	27	9	6	12	0.15
Yes	5	0	3	2
GVHD prophylaxis, n
Tacrolimus-based	15	4	3	8	0.61
Cyclosporin-based	17	5	6	6

SCT: stem cell transplantation, CP: chronic phase, AP: accelerated phase, BP: blastic phase, CP1: first chronic phase, CP2: second chronic phase, CHR: complete hematologic response, PCyR: partial cytogenetic response, CCyR: complete cytogenetic response, MR: molecular response, HLA: human leukocyte antigen, BM: bone marrow, PBSC: peripheral blood stem cell, MAC: myeloablative conditioning, TBI: total body irradiation, RIC: reduced intensity conditioning, GVHD: graft-versus-host disease, TKI: tyrosine kinase inhibitor, IM: imatinib

The transitions of disease status from the diagnosis to transplantation according to pre-SCT treatment are shown in Fig. 1. In the Non-TKI group, one of nine patients progressed to AP at transplant. In the IM group, six of seven patients progressed to AP/BP, but four patients had returned to CP2 by the time of transplant among patients with CP at the diagnosis. Patients in the IM group with AP/BP at the diagnosis could not achieve CP2 at transplantation. In the 2/3G TKI group, four out of six patients progressed to AP/BP, but two patients had returned to CP2 by the time of transplant among patients with CP at the diagnosis. In addition, all patients with AP/BP at the diagnosis achieved CP2 by the time of transplantation in the 2/3G TKI group.

Figure 1.

Transitions of the disease status from the diagnosis to transplantation. The disease status from the diagnosis to transplant for the (a) Non-TKI, (b) IM, and (c) 2/3G TKI groups. The number of patients in each state is indicated in parentheses, and the number of patients representing the transition of the disease status is indicated by the arrows. AP: accelerated phase, BP: blastic phase, CP1: first chronic phase, CP2: second chronic phase, CHR: complete hematologic response, PCyR: partial cytogenetic response, CCyR: complete cytogenetic response, MR: molecular response

Patient characteristics in the progressing phase category are shown in Table 2, and details of the treatment course in the progressing phase category are shown in Table 3. There were 1, 8, and 21 patients in the Non-TKI, IM, and 2/3G TKI group, respectively. The year of SCT, disease status at SCT, achievement of CCyR at SCT, and number of TKIs used before transplantation were significantly different among the treatment groups. There were more patients who achieved higher responses, such as CCyR or MR, in the 2/3G TKI group than in the IM and Non-TKI groups. The frequency of patients achieving CHR at allo-SCT did not differ among the subgroups. In the Non-TKI group, one patient did not achieve CP2 by the time of transplant but survived after allo-SCT. In the IM group, four of eight patients achieved CP2 at transplant, and three and two patients died due to relapse and NRM, respectively. In the 2/3G TKI group, 10 out of 12 patients achieved CP2 by the time of transplant, and only 1 patient died because of NRM (Table 3).

Table 2.

Patients and Pretransplant Treatment Characteristics in Progressing Phase Category.

	Total	Non-TKI	IM	2/3G TKI	p value
No. of patients	21	1	8	12
Median age at SCT, years [range]	41 [24-54]	33-	41 [24-45]	42 [25-54]	0.80
Age at transplantation, y
<40	8	1	2	5	0.51
≥40	13	0	6	7
Sex, n
Male	11	1	4	6	1.00
Female	10	0	4	6
Median months from diagnosis to SCT [range]	20 [2-97]	71-	26 [3-47]	6 [2-97]	0.40
Year of SCT, n
2001-2010	11	1	8	2	<0.01
2011-2020	10	0	0	10
Disease status at diagnosis, n
CP	11	1	6	4	0.12
AP/BP	3/7	0	1/1	2/6
Disease status at SCT, n
CP1	0	0	0	0	0.10
CP2	14	0	4	10
AP/BP	4/3	1/0	2/2	1/1
Disease status at SCT, n
AP/BP	4/3	1/0	2/2	1/1	0.04
Non-CHR	0	0	0	0
CHR	1	0	1	0
PCyR	3	0	2	1
CCyR	8	0	1	7
MR	2	0	0	2
Achieving CHR at SCT, n
No	7	1	4	2	0.10
Yes	14	0	4	10
Achieving CCyR at SCT, n
No	11	1	7	3	0.01
Yes	10	0	1	9
Type of donor and stem cell source, n
HLA-matched related BM/PBSC	1/3	0/0	1/1	0/2	0.82
HLA-mismatched related BM/PBSC	0	0	0	0
HLA-matched unrelated BM/PBSC	10/0	1/0	3/0	6/0
HLA-mismatched unrelated BM/PBSC	0	0	0	0
Cord blood from unrelated donor	7	0	3	4
EBMT risk score
0-1	0	0	0	0	0.44
2-4	7	1	2	4
≥5	14	0	6	8
Numbers of TKIs used before transplant, n
0	1	1	0	0	<0.01
1	8	0	8	0
≥2	12	0	0	12
Pretransplant chemotherapy, n
No	5	1	2	2	0.32
Yes	16	0	6	10
Conditioning, n
MAC					0.43
TBI-based	20	1	7	12
non-TBI based	1	0	1	0
RIC	0	0	0	0
TKI treatment after SCT, n
No	16	1	5	10	0.50
Yes	5	0	3	2
GVHD prophylaxis, n
Tacrolimus-based	11	1	3	7	0.51
Cyclosporin-based	10	0	5	5

SCT: stem cell transplantation, CP: chronic phase, AP: accelerated phase, BP: blastic phase, CP1: first chronic phase, CP2: second chronic phase, CHR: complete hematologic response, PCyR: partial cytogenetic response, CCyR: complete cytogenetic response, MR: molecular response, HLA: human leukocyte antigen, BM: bone marrow, PBSC: peripheral blood stem cell, MAC: myeloablative conditioning, TBI: total body irradiation, RIC: reduced intensity conditioning, GVHD: graft-versus-host disease, TKI: tyrosine kinase inhibitor, IM: imatinib

Table 3.

Patient and Their Treatment Characteristics in the Progressing Phase Category.

Non-TKI group
No.	Age/ Sex	Year of SCT	Time from diagnosis to SCT (mo)	TKI dose/duration (mg/mo) before SCT	Disease state and treatment from diagnosis to SCT	Details of diesease state at SCT	Type of BP crisis	CHT before SCT	Conditioning	Cell source	Relapse after allo-SCT	TKI dose/duration (mg/mo) after SCT	Cause of death	OS (mo)
1	33/M	2002	71	-	CP→ IFN→ AP	AP	-	-	TBI/CY	MU-BM	-	-	-	239
IM group
1	41/F	2001	24	IM (600/3.0)	CP→IFN→BP→ IM+CHT→ CP2	CHR	Myeloid	HDAraC	TBI/AraC/CY/GCSF	UCB	Molecular relapse	-	-	243
2	43/M	2002	37	IM (600/6.0)	CP→IFN→BP→ IM+CHT→ CP2	PCyR	Lymphoid	VP	TBI/AraC/GCSF	MU-BM	Molecular relapse	IM (400/111)	-	240
3	41/M	2002	47	IM (300/0.25)	CP→IFN→AP→IM→ AP	AP	-	-	TBI/AraC/CY	MR-PB	-	-	cGVHD (liver)	11
4	45/M	2002	27	IM (600/6.0)	CP→IFN→BP→ IM+CHT→CP2	CCyR	Lymphoid	VP	TBI/AraC/CY	MU-BM	-	-	Bacterial pneumonia	1
5	24/M	2002	20	IM (400/14)	CP→IFN→BP→ IM+CHT→BP	BP	Myeloid	HDAraC	Bu/CY	UCB	Progression to BP	IM (600/8.0)	Relapse	26
6	31/F	2003	28	IM (600/27)	CP→IFN→AP→IM→ CP2	PCyR	-	-	TBI/AraC/GCSF	MMU-BM	-	-	-	200
7	41/F	2002	12	IM (300/0.75)	AP→IFN→IM→AP	AP	-	-	TBI/CY	MR-BM	Progression to BP	-	Relapse	41
8	45/F	2006	3	IM (600/3.0)	BP→IM+CHT→CP2→ BP	BP	Lymphoid	VP	TBI/AraC/CY	UCB	Progression to BP	IM (600/2.5), DAS (140/4.5)	Relapse	35
2/3G TKI group
1	50/M	2008	92	IM (600/42), NIL (800/28)	CP→IM→BP→ IM+CHT→CP2→BP→ NIL+CHT→CP2→BP	BP	Myeloid	HDAraC	TBI/AraC/CY/GCSF	UCB	-	-	-	167
2	43/F	2009	88	IM (400/44), DAS (140/4.0)	CP→IM→AP→DAS→ AP	AP	-	-	TBI/AraC/CY/GCSF	UCB	Molecular relapse	DAS (50/12)	-	50
3	41/M	2012	19	IM (400/13), DAS (140/5.0)	CP→IM→AP→DAS→ CP2	CCyR	-	-	TBI/AraC/CY/GCSF	MU-BM	-	-	-	124
4	34/M	2012	97	IM (400/51), DAS (140/1.0)	CP→IM→BP→NIL→ BP→DAS+CHT→CP2	CCyR	Lymphoid	VP	TBI/AraC/CY/GCSF	MMU-BM	-	-	CMV colitis	1
5	26/F	2012	3	IM (600/0.5), DAS(140/1.0)	BP→IM+CHT→CNS relapse→DAS→CP2	CCyR	Lymphoid	JALSG Ph+ ALL202 induction regimen	TBI/AraC/CY/GCSF	UCB	-	-	-	113
6	33/F	2014	2	IM (600/0.5), DAS (140/1.0)	BP→IM→BP→DAS→ CP2	CCyR	Lymphoid	-	TBI/AraC/CY/GCSF/Flu/Plerixafor	MR-PB	Molecular relapse	NIL (600/5.5)	-	94
7	48/M	2016	26	DAS (140/13), NIL (800/10)	BP→DAS+CHT→ CP2→V299L mutation→NIL→ BP→NIL+CHT→CP2	CCyR	Biphenotypic	CALGB9111 induction regimen, AdVP, VP	TBI/AraC/CY/GCSF	MU-BM	-	-	-	66
8	54/M	2017	7	DAS (100/3.0), PON (45/0.33)	AP→DAS→BP→ PON+CHT→CP2	PCyR	Lymphoid	CALGB9111 induction regimen, VP	TBI/AraC/CY/GCSF	UCB	-	-	-	60
9	27/F	2017	4	IM (800/0.5), DAS (140/2.0)	BP→DAS+CHT→CP2	MoR	Lymphoid	VP	TBI/AraC/CY/GCSF	MU-BM	-	-	-	60
10	25/M	2018	5	DAS (100/3.0), PON (30/2.0)	AP→DAS→BP→ PON+CHT→CP2	CCyR	Lymphoid	VP	TBI/AraC/CY/GCSF	MR-PB	-	-	-	45
11	53/F	2018	5	DAS (140/4.0), PON (45/1.0)	BP→DAS→BP→ PON+CHT→CP2	MoR	Lymphoid	VP, Hyper-CVAD	Flu/Mel/TBI	MMU-BM	-	-	-	45
12	42/F	2019	5	DAS (140/3.0), PON (30/2.0)	BP→DAS→CP2→ DAS intolerance→ PON→CP2	CCyR	Lymphoid	-	Flu/Mel/TBI	MR-BM	-	-	-	32

Underlined words mean treatment between disease states. IM: imatinib, DAS: dasatinib, NIL: nilotinib, PON: ponatinib, IFN: interferon, CHT: chemotherapy, AP: accelerated phase, BP: blastic phase, CHR: complete hematologic response, PCyR: partial cytogenetic response, CCyR: complete cytogenetic response, MoR: molecular, response, CP2: second chronic phase, HDAraC: high-dose cytarabine, VP: vincristine and prednisolone, JALSG: Japan Adult Leukemia Study Group, CALGB: Cancer and Leukemia Group B, AdVP: doxorubicin: vincristine and prednisolone, TBI: total body irradiatin, CY: cyclophosphamide, AraC: cytarabine, GCSF: granulocyte colony-stimulating factor, Flu: fludarabine, Mel: melphalan, UCB: umbilical cord blood, MR: matched related, MU: matched unrelated, MMU: mismatched unrelated, BM: bone marrow, PB: peripheral blood, cGVHD: chronic graft-versus-host disease, CMV: cytomegalovirus

Transplant outcomes

All patients achieved engraftment after allo-SCT. Twenty-two of the 32 patients were alive at the last follow-up. The median follow-up period for survivors was 66 (range: 1-251) months. The cumulative incidence of acute (Grade >I) GVHD at day 100 was 56.2% [95% confidence interval (CI), 37.1-71.6%], and the cumulative incidence of chronic GVHD at 1 year was 58.7% (95% CI, 38.7-74.2%). The causes of NRM were bacterial pneumonia (n=2), chronic GVHD (n=2), acute GVHD (n=3), and cytomegalovirus (CMV) colitis (n=1). The cumulative incidence of NRM at 5 years was 21.9% (95% CI, 9.4-37.6%), and the cumulative incidence of disease relapse/progression at 5 years was 28.3% (95% CI, 13.8-44.7%). The 5-year OS rates were 68.8% (95% CI, 49.7-81.8%), and the 5-year RFS rates were 49.8% (95% CI, 31.7-65.5%).

Five patients received post-transplant TKIs, all of whom were in the progressing phase category. The details of post-transplant treatment with TKIs are shown in Table 3. Post-transplant TKIs were used in two patients for progression to BP in the IM group, one patient for molecular relapse in the IM group, and two patients for molecular relapse in the 2/3G TKI group. The frequency of post-transplant TKI use did not differ among the three subgroups in all patients and patients in the progressing category (Tables 1, 22).

Factors affecting transplant outcomes

The results of the univariate analysis of all patients are summarized in Table 4. Achieving CHR at transplant, type of TKIs before transplant, and number of TKIs used before transplant were factors associated with the OS. Achieving CHR at transplantation and the use of TKIs after transplantation were associated with the RFS. Achieving CHR at transplantation was the only factor associated with NRM. Use of TKIs before transplant, types of TKIs used before transplant, number of TKIs used before transplant, and use of TKIs after transplant were associated with the cumulative incidence of relapse (CIR).

Table 4.

Risk Factor Analyses for OS, RFS, NRM, and CIR after allo-SCT in Entire Patients.

Variables		n	5-y OS (95% CI)	p value	5-y RFS (95% CI)	p value
Age, years	<40	14	78.6 (47.2-92.5)	0.32	71.4 (40.6-88.2)	0.054
	≥40	18	61.1 (35.3-79.2)		32.4 (12.7-54.0)
Sex	Male	17	58.8 (32.5-77.8)	0.16	47.1 (23.0-68.0)	0.63
	Female	15	80.0 (50.0-93.1)		52.5 (25.2-74.0)
Donor recipient sex combination	Female donor, male recipient	3	33.3 (0.9-77.4)	0.09	33.3 (0.9-77.4)	0.53
	Others	29	72.4 (52.3-85.1)		55.2 (35.6-71.0)
Time from diagnosis to transplant, month	<12	13	76.9 (44.2-91.9)	0.41	69.2 (37.3-87.2)	0.09
	≥12	19	63.2 (37.9-80.4)		36.8 (16.5-57.5)
Treatment response at transplant	MR	2	100.0 (-)	<0.01	100.0 (-)	0.08
	CCyR	9	77.8 (36.5-93.9)		55.6 (20.4-80.5)
	PCyR	5	100.0 (-)		56.2 (27.2-77.6)
	CHR	5	100.0 (-)		80.0 (20.4-96.9)
	Non-CHR	4	NR		NR
	AP/BP	7	42.9 (9.8-73.4)		28.6 (4.1-61.2)
Disease status at transplant	AP/BP	7	42.9 (9.8-73.4)	0.21	28.6 (4.1-61.2)	0.49
	CP1/CP2	25	76.0 (54.2-88.4)		55.7 (34.4-72.6)
CHR at transplant	Yes	21	90.5 (67.0-97.5)	<0.01	66.3 (42.0-82.3)	0.03
	No	11	27.3 (6.5-53.9)		18.2 (2.9-44.2)
CCyR at transplant	Yes	11	81.8 (44.7-95.1)	0.35	63.6 (29.7-84.5)	0.52
	No	21	61.9 (38.1-78.8)		42.9 (21.9-62.3)
MR at transplant	Yes	2	100.0 (-)	0.37	100.0 (-)	0.22
	No	30	66.7 (46.9-80.5)		46.4 (28.1-62.9)
Disease status at transplant (AP/BP/CP2 vs. CP1)	AP/BP/CP2	21	71.4 (47.2-86.0)	0.57	52.4 (29.7-70.9)	0.80
	CP1	11	63.6 (29.7-84.5)		45.5 (16.7-70.7)
Use of TKI before transplant (Non-TKI vs. TKI)	Non-TKI	9	55.6 (20.4-80.5)	0.25	55.6 (20.4-80.5)	0.72
	TKI	23	73.9 (50.9-87.3)		47.4 (26.3-65.9)
Types of TKIs used before transplant	None	9	55.6 (20.4-80.5)	0.04	55.6 (20.4-80.5)	0.16
	IM	9	44.4 (13.6-71.9)		22.2 (3.4-51.3)
	2/3G TKI	14	92.9 (59.1-99.0)		63.5 (33.1-83.0)
Numbers of TKIs used before transplant	0	9	55.6 (20.4-80.5)	0.04	55.6 (20.4-80.5)	0.16
	1	9	44.4 (13.6-71.9)		22.2 (3.4-51.3)
	≥2	14	0.929 (0.591-0.990)		63.5 (33.1-83.0)
Type of donor	HLA-matched related	8	50.0 (5.2-77.5)	0.23	25.0 (3.7-55.8)	0.24
	Others	24	75.0 (52.6-87.9)		58.3 (36.4-75.0)
Use of TKI after tranplant	No	27	70.4 (49.4-0.83.9)	0.81	66.7 (45.7-81.1)	<0.01
	Yes	5	60.0 (12.6-88.2)		-
EBMT score	0-4	18	667.7 (40.4-83.4)	0.81	50.0 (25.9-70.1)	0.77
	≥5	14	714 (40.6-88.2)		50.0 (22.9-72.2)

Table 4.

Risk Factor Analyses for OS, RFS, NRM, and CIR after allo-SCT in Entire Patients. (Continued)

Variables		n	5-y NRM (95% CI)	p value	5-y CIR (95% CI)	p value
Age, years	<40	14	14.3 (2.1-37.6)	0.40	14.3 (2.1-37.6)	0.14
	≥40	18	27.8 (9.5-49.8)		39.8 (16.7-62.3)
Sex	Male	17	35.3 (13.7-57.9)	0.053	17.6 (4.0-39.3)	0.18
	Female	15	6.7 (0.4-27.0)		40.8 (15.7-64.9)
Donor recipient sex combination	Female donor, male recipient	3	66.7 (0.2-97.3)	0.055	0.0 (0.0-0.0)	0.30
	Others	29	17.2 (6.1-33.1)		27.6 (12.8-44.7)
Time from diagnosis to transplant, month	<12	13	15.4 (2.2-40.0)	0.44	15.4 (2.2-39.9)	0.22
	≥12	19	26.3 (9.1-47.6)		36.8 (15.7-58.4)
Disease status at transplant	AP/BP	7	14.3 (0.3-51.3)	0.53	57.1 (12.2-86.2)	0.09
	CP1/CP2	25	24.0 (9.5-42.2)		20.3 (7.1-38.2)
Treatment response at transplant	MR	2	0.0 (-)	<0.01	0.0 (-)	0.55
	CCyR	9	22.2 (3.4-51.3)		22.2 (3.4-51.3)
	PCyR	5	0.0 (-)		40.0 (5.2-75.3)
	CHR	5	0.0 (-)		20.0 (0.8-58.2)
	Non-CHR	4	NR		NR
	AP/BP	7	14.2 (0.7-46.5)		57.1 (17.2-83.7)
CHR at transplant	Yes	21	9.5 (1.5-26.7)	0.03	24.1 (8.4-44.2)	0.57
	No	11	45.5 (14.4-72.5)		36.4 (9.0-65.4)
CCyR at transplant	Yes	11	18.2 (2.5-45.5)	0.83	18.2 (2.4-45.9)	0.46
	No	21	23.8 (8.3-43.7)		33.3 (14.3-53.8)
MR at transplant	Yes	2	0.0 (-)	0.45	0.0 (-)	0.38
	No	30	23.3 (10.3-39.4)		30.2 (15.1-46.9)
Disease status at transplant (AP/BP/CP2 vs. CP1)	AP/BP/CP2	21	14.3 (3.4-32.7)	0.18	33.3 (14.3-53.7)	0.35
	CP1	11	36.4 (10.0-64.2)		18.2 (2.3-46.2)
Use of TKI before transplant (Non-TKI vs. TKI)	Non-TKI	9	44.4 (11.7-73.6)	0.07	0.0 (0.0-0.0)	0.03
	TKI	23	13.0 (3.1-30.3)		39.5 (19.4-59.1)
Types of TKIs used before transplant	None	9	44.4 (11.7-73.6)	0.13	0.0 (0.0-0.0)	0.04
	IM	9	22.2 (2.4-54.4)		55.6 (16.3-82.7)
	2/3G TKI	14	7.1 (0.4-28.5)		29.4 (8.3-54.7)
Numbers of TKIs used before transplant	0	9	44.4 (11.7-73.6)	0.13	0.0 (0.0-0.0)	0.04
	1	9	22.2 (2.4-54.4)		55.6 (16.3-82.7)
	≥2	14	7.1 (0.4-28.5)		29.4 (8.3-54.7)
Type of donor	HLA-matched related	8	37.5 (6.9-69.0)	0.28	37.5 (6.1-70.9)	0.60
	Others	24	16.7 (5.0-34.2)		25.0 (9.8-43.7)
Use of TKI after tranplant	No	27	25.9 (11.2-43.5)	0.27	11.1 (2.7-26.3)	<0.01
	Yes	5	-		-
EBMT score	0-4	18	27.8 (9.6-49.6)	0.45	22.2 (6.5-43.8)	0.37
	≥5	14	14.3 (2.1-37.6)		35.7 (12.0-60.6)

AP: accelerated phase, BP: blastic phase, CP1: first chronic phase, CP2: second chronic phase, CHR: complete hematologic response, PCyR: partial cytogenetic response, CCyR: complete cytogenetic response, MR: molecular response, HLA: human leukocyte antigen, NR: not reached, OS: overall survival, RFS: relapse-free survival, NRM: non-relapse mortality, CIR: cumulative incidence of relapse, allo-SCT: allogeneic stem cell transplantation, CI: confidence interval, TKI: tyrosine kinase inhibitor, IM: imatinib

To investigate the effect of pretransplant use of TKIs on the outcomes of patients with CML receiving allo-SCT, we analyzed the survival outcomes according to disease status at the diagnosis and transplant (Table 5). In patients with CP at the diagnosis, pretransplant TKI use did not correlate with transplantation outcomes. We could not perform a univariate analysis in patients with AP/BP at the diagnosis because there were only two and eight patients in the IM and 2/3G TKI groups, respectively. However, neither of the patients in the IM group were able to achieve CP2 by the time of transplant, relapsing as progression to BP and ultimately dying. In contrast, all eight patients in the 2/3G TKI group achieved CP2 by the time of transplant, and while one patient suffered molecular relapse, they survived with the administration of posttransplant TKIs (Table 3).

Table 5.

Risk Factor Analyses for OS, RFS, NRM, and CIR after allo-SCT according to Disease Stage.

Variables	Disease status	Subgroups	n		5-y OS (95% CI)	p value		5-y RFS (95% CI)	p value
Disease status atdiagnosis	CP	None	9		55.6 (20.4-80.5)	0.59		55.6 (20.4-80.5)	0.54
		IM	7		57.1 (17.2-83.7)			28.6 (4.1-61.2)
		2/3G TKI	6		83.3 (27.3-97.5)			33.3 (4.6-67.6)
Disease status at transplant	CP	None	8		50.0 (15.2-77.5)	0.14		50.0 (15.2-77.5)	0.42
		IM	5		80.0 (20.4-96.9)			40.0 (5.2-75.3)
		2/3G TKI	12		91.7 (53.9-98.8)			65.6 (32.0-85.6)
	Progressing	None	1		-	0.01		-	0.01
		IM	8		37.5 (8.7-67.4)			12.5 (0.7-42.3)
		2/3G TKI	12		91.7 (53.9-98.8)			75.0 (40.8-91.2)
Variables	Disease status	Subgroups	n		5-y NRM (%)	p value		5-y CIR (%)	95% CI
Disease status at diagnosis	CP	None	9		44.4 (11.7-73.6)	0.60		0.0 (0.0-0.0)	0.073
		IM	7		28.6 (2.5-65.5)			42.9 (7.2-76.2)
		2/3G TKI	6		16.7 (0.5-54.9)			50.0 (6.9-83.6)
Disease status at transplant	CP	None	8		50.0 (12.5-79.0)	0.15		0.0 (0.0-0.0)	0.19
		IM	5		20.0 (0.4-63.2)			40.0 (2.6-79.0)
		2/3G TKI	12		8.3 (0.4-32.3)			26.0 (5.5-53.7)
	Progressing	None	1		-	0.33		-	0.06
		IM	8		25.0 (2.3-60.0)			62.5 (15.5-88.8)
		2/3G TKI	12		8.3 (0.4-32.3)			16.7 (2.3-42.6)

CP: chronic phase, AP: accelerated phase, BP: blastic phase, IM: imatinib, OS: overall survival, RFS: relapse-free survival, NRM: non-relapse mortality, CIR: cumulative incidence of relapse, allo-SCT: allogeneic stem cell transplantation, CI: confidence interval, TKI: tyrosine kinase inhibitor, IM: imatinib

The median OS was 32 (range, 29-35) months in the IM group and 63 (range, 35-116) months in the 2/3G TKI group. The 5-year OS and RFS were 100.0% and 87.5% (95% CI, 38.7-98.1%), respectively, in the 2/3G TKI group. In patients with CP at transplant, pretransplant TKI use did not correlate with transplantation outcomes. In patients with AP/BP at transplant, we were unable to perform a univariate analysis because of the small sample size. In the progressing phase category, there were 1, 8, and 12 patients in the Non-TKI, IM, and 2/3G TKI groups, respectively (Table 2). Having only one patient in a group was not sufficient for performing Kaplan-Meier analyses, so we excluded the Non-TKI group.

The 5-year OS and RFS were significantly higher in the 2/3G TKI group [5-year OS: 91.7% (95% CI, 53.9-98.8%), 5-year RFS: 75.0% (95% CI, 40.8-91.2%)] than in the IM group [5-year OS: 37.5% (95% CI, 8.7-67.4%), 5-year RFS: 12.5% (95% CI, 7.0-42.3%)] (Fig. 2a, b, Table 5). In addition, the 5-year CIR was considerably lower in the 2/3G TKI group [16.7% (95% CI, 2.3-42.6%)] than in the IM group [62.5% (95% CI, 15.5-88.8%)] (Fig. 2c, Table 5), although the difference was not statistically significant (p=0.06).

Figure 2.

Kaplan-Meier analyses for patients in the progressing phase category. (a) The overall survival, (b) relapse-free survival, and (c) cumulative incidence of relapse according to the types of pretransplant TKI; (d) the overall survival according to the disease status at allo-SCT, (e) the overall survival according to achieving CHR vs. others, and (f) the post-relapse survival according to the types of pretransplant TKI.

Risk factors associated with transplant outcomes in the progressing phase category

To further investigate the factors associated with transplantation outcomes in the progressing phase category, we performed a univariate analysis for the progressing phase category (Table 6). The types of TKIs used before transplantation (2/3G TKI vs. IM) and the number of TKIs used before transplantation (1 vs. ≥2) were factors associated with the OS (Fig. 2a). The types of TKIs used before transplantation (2/3G TKI vs. IM), number of TKIs used before transplantation, and use of TKI after transplantation were associated with the RFS. Achieving CCyR at transplant and the use of TKI after transplantation were associated with the CIR. The types of TKIs used before transplantation (IM vs. 2/3G TKI) and the number of TKIs used before transplantation (1 vs. ≥2) were significantly associated with the 5-year CIR, although the difference was not significant (p=0.06) (Fig. 2c).

Table 6.

Risk Factor Analyses for OS, RFS, NRM, and CIR after allo-SCT in Progressin Phase Category.

Variables		n	5-y OS(95%CI)	p value	5-y RFS (95%CI)	p value
Age, years	<40	8	75.0 (31.5-93.1)	0.79	62.5 (22.9-86.1)	0.54
	≥40	13	69.2 (37.3-87.2)		46.2 (19.2-69.6)
Sex	Male	11	63.6 (29.7-84.5)	0.35	54.5 (22.9-78.0)	0.86
	Female	10	80.0 (40.9-94.6)		50.0 (18.4-75.3)
Donor recipient sex combination	Female donor, male recipient	1	-	-	-	-
	Others	20	70.0 (45.1-85.3)		50.0 (27.1-69.2)
Time from diagnosis to transplant, month	<12	9	88.9 (43.3-98.4)	0.14	77.8 (36.5-93.9)	0.06
	≥12	12	58.3 (27.0-80.1)		33.3 (10.3-58.8)
Disease status at transplant (AP/BP vs. CP2)	AP/BP	7	42.9 (9.8-73.4)	0.07	28.6 (4.1-61.2)	0.36
	CP2	14	85.7 (53.9-96.2)		64.3 (34.3-83.3)
Disease status at transplant	MR	2	100.0 (-)	0.30	100.0 (-)	0.53
	CCyR	8	75.0 (31.5-93.1)		62.5 (22.9-86.1)
	PCyR	3	100.0 (-)		66.7 (5.4-94.5)
	CHR	1	-		-
	Non-CHR	0	-		-
	AP/BP	7	42.9 (9.8-73.4)		28.6 (4.1-61.2)
CHR at transplant	Yes	14	85.7 (53.9-96.2)	0.07	64.3 (34.3-83.3)	0.36
	No	7	42.9 (9.8-73.4)		28.6 (4.1-61.2)
CCyR at transplant	Yes	10	80.0 (40.9-94.6)	0.53	70.0 (32.9-89.2)	0.24
	No	11	63.6 (29.7-84.5)		36.4 (11.2-62.7)
MR at transplant	Yes	2	100.0 (-)	0.39	100.0 (-)	0.23
	No	19	68.4 (42.8-84.4)		47.4 (24.4-67.3)
Types of TKIs used before transplant	None	1	-	0.01	-	0.01
	IM	8	37.5 (8.7-67.4)		12.5 (0.7-42.3)
	2/3G TKI	12	91.7 (53.9-98.8)		75.0 (40.8-91.2)
Numbers of TKIs used before transplant	0	1	-	0.01	-	0.01
	1	8	37.5 (8.7-67.4)		12.5 (0.7-42.3)
	≥2	12	91.7 (53.9-98.8)		75.0 (40.8-91.2)
Type of donor	HLA-matched related	4	50.0 (5.8-84.5)	0.36	25.0 (0.9-66.5)	0.43
	Others	17	76.5 (48.8-90.4)		58.8 (32.5-77.8)
Use of TKI after tranplant	No	16	75.0 (46.3-89.8)	0.60	68.8 (40.5-85.6)	<0.01
	Yes	5	60.0 (12.6-88.2)		0.0 (-)
EBMT score	0-4	7	71.4 (25.8-92.0)	0.93	57.1 (17.2-83.7)	0.57
	≥5	14	71.4 (40.6-88.2)		50.0 (22.9-72.2)

Table 6.

Risk Factor Analyses for OS, RFS, NRM, and CIR after allo-SCT in Progressin Phase Category.

Variables		n	5-y NRM (95% CI)	p value	5-y CIR (95% CI)	p value
Age, years	<40	8	12.5 (0.5-44.5)	0.86	25.0 (2.9-58.3)	0.54
	≥40	13	15.4 (2.1-40.2)		38.5 (12.9-64.0)
Sex	Male	11	27.3 (5.7-55.4)	0.08	18.2 (2.4-45.9)	0.13
	Female	10	0.0 (-)		50.0 (16.3-76.8)
Donor recipient sex combination	Female donor, male recipient	1	-	-	-	-
	Others	20	15.0 (3.5-34.2)		35.0 (15.0-55.9)
Time from diagnosis to transplant, month	<12	9	0.0 (-)	0.12	22.2 (2.8-53.1)	0.40
	≥12	12	25.0 (5.2-52.1)		41.7 (13.6-68.1)
Disease status at transplant (AP/BP vs. CP2)	AP/BP	7	14.3 (0.3-51.3)	0.95	57.1 (12.2-86.2)	0.20
	CP2	14	14.3 (2.1-37.6)		21.4 (4.7-45.9)
Disease status at transplant	MR	2	0.0 (-)	0.68	0.0 (-)	0.32
	CCyR	8	25.0 (3.7-55.8)		12.5 (0.7-42.3)
	PCyR	3	0.0 (-)		33.3 (0.9-77.4)
	CHR	1	-		-
	Non-CHR	0	-		-
	AP/BP	7	14.3 (0.3-51.3)		57.1 (12.2-86.2)
CHR at transplant	Yes	14	14.3 (2.1-37.6)	0.95	21.4 (4.7-45.9)	0.20
	No	7	14.3 (0.3-51.3)		57.1 (12.2-86.2)
CCyR at transplant	Yes	10	20.0 (2.6-49.0)	0.45	10.0 (0.4-37.9)	0.047
	No	11	9.1 (0.4-36.0)		54.5 (20.4-79.3)
MR at transplant	Yes	2	0.0 (-)	0.56	0.0 (-)	0.32
	No	19	15.8 (3.9-34.9)		36.8 (15.8-58.2)
Types of TKIs used before transplant	None	1	-	0.33	-	0.06
	IM	8	25.0 (2.3-60.0)		62.5 (15.5-88.8)
	2/3G TKI	12	8.3 (0.4-32.3)		16.7 (2.3-42.6)
Numbers of TKIs used before transplant	0	1	-	0.33	-	0.06
	1	8	25.0 (2.3-60.0)		62.5 (15.5-88.8)
	≥2	12	8.3 (0.4-32.3)		16.7 (2.3-42.6)
Type of donor	HLA-matched related	4	25.0 (0.9-66.5)	0.56	21.4 (4.7-45.9)	0.20
	Others	17	11.8 (1.8-32.0)		57.1 (12.2-86.2)
Use of TKI after tranplant	No	16	18.8 (4.3-41.1)	0.42	12.5 (1.9-33.7)	<0.01
	Yes	5	0.0 (-)		100.0 (-)
EBMT score	0-4	7	14.3 (0.3-51.3)	0.95	28.6 (3.0-63.9)	0.71
	≥5	14	14.3 (2.1-37.6)		35.7 (12.0-60.6)

AP: accelerated phase, BP: blastic phase, CP2: second chronic phase, CHR: complete hematologic response, PCyR: partial cytogenetic response, CCyR: complete cytogenetic response, MR: molecular response, HLA: human leukocyte antigen, OS: overall survival, RFS: relapse-free survival, NRM: non-relapse mortality, CIR: cumulative incidence of relapse, allo-SCT: allogeneic stem cell transplantation, CI: confidence interval, TKI: tyrosine kinase inhibitor, IM: imatinib

We then investigated whether or not achieving a better response at allo-SCT contributed to the improved prognosis of patients receiving 2/3G TKI in the progressing phase category (Table 6). The frequency of patients achieving CHR was considerably lower in the IM group (50%, 4 out of 8 patients) than in the 2/3G TKI group (83.3%, 10 out of 12 patients), although the difference was not significant (Fisher's exact test, p=0.16) (Table 2). In addition, significantly fewer patients achieved CCyR at transplant in the IM group than in the 2/3G TKI group. The 5-year OS was considerably higher with improved disease status at transplantation, especially when CHR or CP2 was achieved (p=0.07) (Table 6, Fig. 2d, e), whereas achieving CCyR or MR was not correlated with a better 5-year OS (Table 6). Having only one patient in a group was not sufficient for performing Kaplan-Meier analyses, so we excluded the patients who achieved CHR in the IM group. In terms of relapse, 5 out of 8 patients relapsed in the IM group, while 2 out of 12 patients relapsed after allo-SCT in the 2/3G TKI group (Table 3). Among those who relapsed, 3 out of 5 patients died in the IM group, but both patients in the 2/3G TKI group survived, resulting in a considerably lower 5-year PRS in the IM group [40.0% (95% CI, 5.2-75.3%)] than in the 2/3G TKI group (100.0%), although the difference was not significant (p=0.21) (Fig. 2f).

Impact of pretransplant use of TKIs on the development of GVHD

Next, we investigated the effect of pretransplant TKI treatment on GVHD development. We found that significantly fewer patients in the IM group developed acute GVHD than patients in the Non-TKI group (Fig. 3a), and the use of 2/3G TKI before allo-SCT did not have a significant effect on the development of acute GVHD. The cumulative incidence of acute (Grade >I) GVHD on day 100 was 77.8% (95% CI, 28.4-95.1%), 33.3% (95% CI, 6.7-64.0%), and 57.1% (95% CI, 26.7-78.9%) in Non-TKI, IM, and 2/3G TKI groups, respectively (Supplementary material). We further found that the incidence of chronic GVHD did not differ significantly among the TKI subgroups (Fig. 3b). The cumulative incidence of chronic GVHD at 1 year was 75.0% (95% CI, 23.3-94.5%), 66.7% (95% CI, 22.5-89.6%), and 43.9% (95% CI, 16.7-68.5%) in Non-TKI, IM, and 2/3G TKI groups, respectively (Supplementary material).

Figure 3.

Kaplan-Meier analyses of GVHD according to TKI subgroups. (a) Acute GVHD and (b) chronic GVHD.

Discussion

With the advent of TKIs, allo-SCT is no longer considered the first-line treatment for CML. However, according to the NCCN guidelines and European LeukemiaNet recommendations, allo-SCT is still considered for patients presenting with or progressing to BP, patients progressing to AP while on TKI therapy, and CP patients who are resistant or intolerant to multiple TKIs (10,11).

To elucidate how the pretransplant use of TKIs, including 2/3G TKIs, affects the outcome of CML patients receiving allo-SCT, we retrospectively analyzed data from our institute. Previous studies have shown that pretransplant use of the first TKI, IM, did not adversely affect transplantation outcomes (13-15). In terms of 2/3G TKIs, there has only been one large-scale study with long-term follow-up on allo-SCT in CML that evaluated the effect of pretransplant use of TKIs on transplantation outcomes (30). That study retrospectively evaluated 118 CML patients receiving allo-SCT between 2002 and 2017 in Sweden, where there were 47 and 67 patients receiving IM and 2/3G TKIs before allo-SCT, respectively (30). The study showed that the 5-year OS ranged from 36.9% to 96.2%, depending on the disease status at allo-SCT, and that the only risk factor associated with death was CP >1 at the time of allo-SCT. Nevertheless, the Swedish registry study did not evaluate the effect of pretransplant TKIs on the outcome of allo-SCT or the incidence of GVHD. Therefore, the exact effect of prior TKI use on the outcome of allo-SCT needs to be further evaluated.

Our results indicated that the 5-year OS was dependent on the disease state at transplant (achieving CHR vs. other), types of TKIs used before transplant, and number of TKIs used before transplant, while the 5-year RFS was dependent on the disease state at transplant (achieving CHR vs. other) (Table 4), which was consistent with the findings of the Swedish CML registry studies and the conventional EBMT scoring system (22,30). In each disease phage category, Non-TKI, IM, and 2/3G TKI subgroup patients were matched with patients who received 0, 1, and ≥2 TKIs before transplant, respectively. Therefore, these two variables represent the same clinical conditions. In a subgroup analysis according to disease stage, we further showed that the pretransplant use of 2/3G TKIs improved the prognosis of CML patients in the progressing phase category, i.e. patients who presented with AP/BP at the diagnosis or presented with CP but progressed to AP/BP. In the progressing phase category, achieving CCyR at transplant was associated with a lower 5-year CIR (Table 6), and significantly fewer patients achieved CCyR at transplant in the IM group than in the 2/3G TKI group (Table 2), which might be one reason why the 5-year CIR in the IM group was considerably higher than that in the 2/3G TKI group (Table 4, Fig. c).

It is difficult to elucidate the exact mechanisms underlying the better survival achieved by 2/3G TKIs in the progressing phase category than IM because of the retrospective nature of this study. One possible explanation is that there was a tendency for a better response at allo-SCT in the 2/3G TKI group than in the IM group (Table 2). The frequency of patients achieving CHR was considerably higher in the 2/3G TKI group than in the IM group (83.3% vs. 50%), although not to a significant degree (Fisher's exact test, p=0.16). Alternatively, relapse was considerably less common in the 2/3G TKI group than in the IM group (Table 6, Fig. c). Finally, the PRS was also considerably better in the 2/3G TKI group than in the IM group (Table 3, Fig. f). All of these findings might have led to the better survival in the 2/3G TKI group than in the IM group.

Our results further showed that pretransplant treatment with TKIs, including 2/3G TKIs, did not increase the development of acute and chronic GVHD (Fig. 3a, b, Supplementary material). Although some studies have shown that the pretransplant use of IM increases the incidence of acute GVHD (31) and decreases the incidence of chronic GVHD (14,32), little is known about the effect of 2/3G TKIs on the development of GVHD. Our results indicate that pretransplant use of 2/3G TKIs is safe and does not increase the incidence of GVHD, supporting the utility of 2/3G TKIs for patients who are undergoing allo-SCT.

Although we could not show the results of a multivariate analysis because of the limited number of patients, the population size was not small, considering the number of allo-SCT performed in patients with CML in Japan annually. Based on data from the JDCHCT, the number of allo-SCTs for CML was 54 in 2007 and has remained roughly the same since then (12), indicating that about 1 CML patient per prefecture undergoes allo-SCT annually in Japan. However, our study included 32 CML patients who received their first allo-SCT at our institute between 2001 and 2020, a 20-year period, showing that more than 1 patient underwent allo-SCT at our institute each year.

Several limitations associated with the present study warrant mention. First, it was retrospective in nature. Second, patients in the IM and Non-TKI groups received allo-SCT earlier than those in the 2/3G TKI group. Finally, indications for CML transplantation differed among eras. Over time, the clinical practice of allo-SCT may have changed, which might have led to heterogeneity of patient/donor and disease demographics, transplant procedures, and supportive care, thus affecting the treatment outcome of allo-SCT.

Conclusion

We demonstrated that pretransplant use of 2/3G TKIs before allo-SCT was safe, not increasing the frequency of GVHD, and that these newer TKIs improved the outcome of CML patients who presented with or progressed to AP/BP, partly because 2/3G TKIs make it possible for those patients to achieve a better response at transplant and receive allo-SCT in better settings and partly because these newer TKIs reduce the relapse rate and improve the survival after relapse. Thus, our results support the utility of 2/3G TKIs for CML patients in pre-transplant settings. The accumulation of more cases is necessary to confirm the effect of the pretransplant use of TKIs for CML.

Informed consent was obtained in the form of opt-out on the website. Patients who were rejected were excluded.

The authors state that they have no Conflict of Interest (COI).

Supplementary Material

Supplementary material

The effect of TKI on the development of GVHD