REVIEW ARTICLE
The role of hematopoietic stem cell transplantation in chronic
myeloid leukemia
Alois Gratwohl & Helen Baldomero & Jakob Passweg
Received: 15 July 2014 / Accepted: 7 December 2014
#
Springer-Verlag Berlin Heidelberg 2015
Abstract Allogeneic hematopoietic stem cell transplantation
(HSCT) is currently recommended as 2nd or 3rd line therapy
for patients with chronic myeloid leukemia (CML) in first
chronic phase or as salvage for patients with very advanced
disease. As a consequen ce, numbers of HSCT in chronic
phase have dropped significantly since the introduction of
tyrosine kinase inhibitors (TKI), numbers of transplants in
advanced disease to a lesser extent. These current recommen-
dations consider primarily disease risk, defined as failure of
TKI therapy; they might need to be adapted. We propose a
more balanced appraisal of HSCT for individual patients
which should include disease risk, transplant risk, and macro-
economic aspects. HSCT should be integrated into the treat-
ment algorithms from diagnosis and be considered very early
at first TKI failure for patients with high disease but low trans-
plant risk. For patients with very advanced disease and high
transplant risk in contrast, HSCT might only be recommended
in a restricted research setting.
Keywords Chronic myeloid leukemia
.
Hematopoietic stem
cell transplantation
.
Allogeneic
.
Autologous
.
Risk
assessment
Introduction
Chronic myeloid leukemia (CML) has seen unprecedented
changes over the last decade. The introduction of tyrosine
kinase inhibitors (TKI) has changed the outlook for patients
with this previously uniformly fatal disease. The ease of ap-
plication, the rapid response, and the mostly excellent tolera-
bility by the patients has focused interest on targeted drug
therapy [1–7]. Hematopoietic stem cell transplantation
(HSCT) has lost its former importance as the “only curative
therapy” [8–12]. This is reflected by the numbers of publica-
tions in medical journals or by the numbers of presentation on
the topic at scientific or promotional meetings. If any, it is
considered by many as tool of last resort when everything else
has failed. This phenomenon is not restricted to CML. Ease of
application and improved response to modern drug therapy
has almost halted HSCT for multiple myeloma and limited
HSCT to selected patients [13, 14]. Still, HSCT is the most
powerful intervention; it holds the potential for “cure” and
outcome has dramatically improved over the last years [15,
16]. It might be good to look at the past and to reconsider the
current status and the potential role of HSCT in the treatment
algorithm of CML today.
Evolution of HSCT for CML
Historical perspective: the role model of CML for HSCT
The first report of a successful HSCT from a syngeneic donor
to a patient with CML dramatically changed the concept on
how to look at CML. For the first time, it became possible to
achieve a Ph state, to eradicate the malignant BCR/ABL
clone, and to reverse the previously inexorable course of the
disease [17]. The concept was rapidly taken up and extended
to HSCT from an HLA-identical sibling donor. It did coincide
with the introduction of cyclosporine A as novel and most
powerful tool for the prevention of graft-versus-host disease
(GvHD) and the concept of HSCT in first complete remission
of acute leukemia; hence, it was introduced in patients with
CML early in their disease, in first chronic phase [8–10, 18].
A. Gratwohl (*)
:
H. Baldomero
:
J. Passweg
Hematology, University Hospital, Dittingerstrasse 4,
4053 Basel, Switzerland
e-mail: alois.gratwohl@unibas.ch
Ann Hematol (2015) 94(Suppl 2):S177–S186
DOI 10.1007/s00277-015-2313-3
The first allogeneic HSCT for CML was reported to the Eu-
ropean Group for Blood and Marrow Transplantation
(EBMT) database in 1975 from France, soon to be followed
in 1978 by a patient from Switzerland and by 10 patients in
1979 from France, Italy, and the UK (personal communica-
tion; EBMT database, Leiden NL). The concept proved to be
right and CML became soon the most frequent indication for
an allogeneic HSCT in Europe and worldwide (Fig. 1)[11,
19]. Of note, as of June 2014, 3 of these 12 patients were
reported to be alive at plus 35 years, one as lost to follow-up.
CML played a role model for HSCT in general in
many aspects. CML did provide the first example for
risk assessment with the EBMT risk score (see below)
[ 20 –22]. It became clear that disease stage was more
important than bulk of the disease. Splenectomy, consid-
ered initially as essential showed no advantage, nor did
splenic irradiation [23]. CM L was the first disease to
demonstrate a consistent graft-versus-leukemia effect.
Relapse risk was highest after T cell depletion in CML
compared to other diseases, in contrast, donor lympho-
cyte infusion (DLI) proved to be the most powerful tool
in CML. CML paved the way for reduced intensity con-
ditioning, specifically with the additional role of preemp-
tive DLI [24–26]. Last but not least, in no other disease
became the impact of macroeconomic factors on use of
HSCT as clear as in CML. Rates of HSCT for CML
dropped already in the year 2000, 2 years before the
release of imatinib in high income countries, illustrating
how expectations drive m edical decision making. They
remained at a stable level in middle and low income
countries whe re costs of drug therapy became higher
than costs for a transplant [11, 19, 27–30].
CML showed a s well a r ole model for autologous
HSCT. It was introduced in Europe early on, almost si-
multaneously with allogeneic HSCT. The first patient
was reported to the EBMT database in 1979 from
France, to be followed by 4 patients in 1980, from
France as well. None of them stayed alive. The concept
was clear, restore chronic phase in patients with ad-
vanced disease throu gh stem cel ls obtai ned in early
phase. Pilot studies proved to be promising and led to
the design of several multicenter prospective randomized
trials in Europe [31, 32]. None was completed; the intro-
duction of the TKI ended these trials prematurely and the
answer about the potential role of autologous HSCT re-
mains open. At least, a retrospective meta-analysis of six
multicentre trials in Europe and the US showed no ad-
vantage of such a procedure compared to concurrent
drug treatment [33]. Numbers of autologous HSCT al-
most vanished away since 2006 [11](Fig.1).
Current status in 2014
Data from the EBMT activity survey report a total of 377
allogeneic HSCT for CML in 2012, 167 in early phase of
the disease, 210 in advanced phase from 35 countries, and 8
autologous HSCT, 5 in early disease, 3 in advanced phase.
Their distribution over disease stage, donor type, and stem cell
source is illustrated in Table 1. Compared to previous years,
total numbers remained stable.
Allogeneic HSCT were performed in 35 countries. There
were significant differences in transplant rates (numbers of
HSCT per 10 million inhabitants) between reporting countries
(Fig. 2). They ranged from none to more than 10 in Belgium,
Estonia, Finland, Sweden, and Switzerland.
0
200
400
600
800
1000
1200
90 92 94 96 98 00 02 04 06 08 10 12
H
S
C
T
Year
CML early allo
CML advanced allo
0
50
100
150
200
250
H
S
C
T
Year
CML early auto
CML advanced auto
a
b
90 92 94 96 98 00 02 04 06 08 10 12
Fig. 1 Evolution of HSCT in Europe from 1990 to 2012. The graph
illustrates i ncrease and decrease of absolute numbers of allogeneic
(Fig. 1a) and autologous (Fig. 1b) HSCT in Europe over time. In blue
early disease (first chronic phase), in green advanced disease stage at time
of HSCT (accelerated phase or blast crisis). a Evolution of allogeneic
HSCT. b Evolution of autologous HSCT
S178 Ann Hematol (2015) 94(Suppl 2):S177–S186
Outcome of HSCT for CML
Factors associated with outcome
HSCT has been and still is associated with significant early
and late transplant-related mortality. In the early days, mortal-
ity appeared erratic, with some young patients dying, others
surviving. In the mid-eighties of last century, it became appar-
ent that outcome was related to specific pretransplant criteria,
independent of transplant technology. This was especially im-
portant for patients with CML who faced the difficult decision
to make, an early transplant with the significant risk for
immediate mortality versus t he risk of blastic transforma-
tion with minimal chances for rescue with HSCT. The
EBMT risk score, based on five pretransplant factors
didpermitarapidassessmentonascalefrom0–VII at
the physician’s desk and gained rapid acceptance
(Table 2). The risk score was validated in several inde-
pendent cohorts and proved to be valid, with some minor
modifications, for all acquired hematological disorders
and for autologous HSCT as well [ 20–22].
The difficulty in risk assessment lies in the fact that
some factors such as disease stage have congruent im-
pact on the two key endpoints, transplant-related mor-
tality and relapse, hence affect overall survival uniform-
ly in the same direction; others have discordant effects.
Thenetresultmightthendependonthesumofall
other risk factors. T cell depletion reduces the risk of
graft-versus- host disease bu t i ncreases the risk of re-
lapse. The net benefit on overall survival will differ
between patients transplanted in early disease compared
to those transplanted in advanced disease stage. Re-
duced intensity might be of benefit in an older patient
with comorbidities but early disease; it might be of no
benefit in the same patient with no comorbidities but
advanced disease (Table 3)[20–22].
As a general concept, risk factors act additively but not in a
symmetrical way. A negative CMV serostatus might further
improve outlook for a low-risk patient but will have no addi-
tional beneficial effect in a high-risk patient; in contrast, a
reduced Karnofsky score might be of minimal impact in a
low-risk patient but deleterious in a high-risk patient. Hence,
the general statement that the probability of survival after an
allogeneic HSCT for CML at 5 years is 60 % is of limited
value; it might range from more than 90 % to less than 5 %. As
we will see below, integration of all elements, including mac-
roeconomic factors of patient’s location, should impact on
choice of transplant technique and the final decision to pro-
ceed with HSCT or to abstain from it [34–39].
Impact of pretransplant treatment
Most patients will have pretreatment for their CML be-
fore HSCT. Earlier studies indicated a higher transplant-
related mortality in patients pretreated with busulfan
compared to hydroxyurea and in patients given interfer-
on alpha up to the day of the transplant. Today, all
patients will have had TKI before their transplant. There
are clear indications that no type of TKI given before or
after the transplant has a deleterious effect on outcome
after HSCT; in one study, results appeared even better
for patients with TKI prior to HSCT. In contrast, type
of response to TKI therapy will impact on post trans-
plant outcome with a good outcome for patients intol-
erant to TKI but with a higher likelihood of worse out-
come for those who failed T KI therapy before HSCT
[34, 40–45].
Impact of HSCT methodology
Despite now 30 years of experience, the best condition-
ing regime n an d the be st gr a ft - ver s us - hos t di s eas e pr e-
vention method remains to be defined. No other condi-
tioning has been documented to arrive at better long term
overall survival than cyclophosphamide and total body
irradiation or the combination of busulfan and cyclo-
phosphamide, no other graft-versus-host disease preven-
tionmethodthanthecombination of cyclosporine and
short methotrexate [ 8]. Reduced intensity conditioning
has extended application of HSCT to elderly patients or
to those with comorbidities [26]. In a large observational
retrospective study by the CIBMTR, it showed a better
overall survival in elderly patients compared to non-
myeloablative cond itioning; no compa rison was made
Table 1 HSCT for CML in Europe 2013 (preliminary data)
Allogeneic HSCT Autologous HSCT Total
Donor type Stem cell source Total Total
Syngeneic Family Unrelated BM PB CB
cP 0 60 80 34 102 4 140 0 140
Not cP 0 78 117 32 157 6 195 3 198
Total 0 138 197 66 259 10 335 3 338
Ann Hematol (2015) 94(Suppl 2):S177–S186 S179
with standard conditioning [42]. Bone marrow as stem
cell source appears to be of advantage in early low-risk
disease, peripheral blood in advanced disease [46, 47].
Of general importance to note, there are no indications
that impact of transplant technology in CML differs from
that in any other disease treated with HSCT [8, 15, 48].
Fig. 2 Transplant rates in Europe 1999 and 2012. The figure depicts
number of HSCT per 10 million inhabitants and illustrates the decrease
in transplant rates for CML over time, in contrast to the transplant rates for
all indications in general. It depicts as well the vast heterogeneity between
countries. a Transplant rates for CML in 1999. b Transplant rates for
CML in 2012. a Transplant rates for all allogeneic indications in 1999.
a Transplant rates for all allogeneic indications in 2012
S180 Ann Hematol (2015) 94(Suppl 2):S177–S186
Graft-versus-host–graft-versus-leukemia effects
There is no doubt about the astonishing graft-versus-
leukemia effects observed in CML as documented by
the powerful effects of DLI [25, 43]. Complete and
lasting molecular remissions can be obtained with even
one infusion of cells, some without any signs of graft-
versus-host disease. Many attempts have been made to
exploit this effect better; so far, no study has proven to
separate graft-versus-host from graft-versus-leukemia
Fig. 2 (continued)
Ann Hematol (2015) 94(Suppl 2):S177–S186 S181
