Professional Documents
Culture Documents
Estey 2020
Estey 2020
Correspondence Author:
Elihu H Estey, MD
Division of Hematology, University of Washington and Clinical Research
Division, Fred Hutchinson Cancer Research Center
825 Eastlake Ave E, #CE3-300, Seattle WA, USA
E eestey@uw.edu
This article has been accepted for publication and undergone full peer review but has not been
through the copyediting, typesetting, pagination and proofreading process which may lead to
differences between this version and the Version of Record. Please cite this article as doi:
10.1002/ajh.25975
When first seeing a newly- diagnosed patient physicians must decide whether to offer
conventional treatment or investigational therapy, the latter preferably in the context of a clinical
trial. As noted below, such trials have led to changes in what today is considered “conventional”
therapy compared to even 1-2 years ago. In older patients decision making has often included
inquiring whether specific anti-AML therapy should be offered at all, rather than focusing on a
purely palliative approach emphasizing transfusion and antibiotic support, with involvement of a
palliative care specialist.
In principle these decisions should be based on the comparative ratios of benefit to risk with
conventional, investigational, or purely palliative approaches. However, as implied by the word
“trial”, benefit/risk ratios of a given trial are generally unknown. Accordingly, a principal reason to
enroll on a trial is dissatisfaction with the benefit/ risk ratios associated with conventional and
purely palliative approaches.
Physicians might have varying opinions as what constitutes benefit and what constitutes risk. The
same is true of patients. Longer survival would likely qualify as a benefit but so might achievement
of a prolonged remission since it is plausibly associated with longer survival, reduced frequency
of transfusions, and better “quality of life” (QOL). The potential benefits, relative to conventional
CR, of newer less-stringently defined categories of remission, such as CRp, CRi, CRh, and of the
more-stringently defined CR without measurable residual disease (MRD) need to be accounted
for. Treatment-related mortality (TRM) resulting in death from treatment before expected death
from disease is an obvious risk but so are various toxicities of treatment and a decreased QOL.
Indeed, a major benefit of many of the recently- FDA approved therapies for AML is their putative
|Page
Physicians may be able to identify factors that although difficult to quantify (“I know it when I see
it”) motivated them to recommend treatment rather than purely palliation and were themselves
responsible for the longer survival in treated older patients. Randomization is a means to address
this possibility. Dombret et al.[6] asked physicians of 488 patients ≥ age 65 with newly-diagnosed
AML, >30% blasts, and performance status 0-2 whether they would typically assign these
patients to best supportive care (BSC) only, low-dose cytarabine(LDAC), or 7+3, which could
|Page
Although these randomized trials can be criticized for insufficient patients in the purely palliative
(BSC) arm or exclusion of patients with poor performance status, the preponderance of evidence
suggests little reason to offer a purely palliative approach to as many patients age ≥ 65 with newly-
diagnosed AML as has been the practice in the past. This is even more so today given improved
supportive care, especially better antifungal therapy[8], whose importance is magnified following
administration of, particularly, intensive chemotherapy. These developments have led to declines
in TRM rates following such therapy[9]. They may play a role in improvements in survival observed
over the 1997-2016 period (8 years median follow-up) in Sweden despite the introduction of no
new therapies, although for unclear reasons these improvements were limited to men aged 50-
75 years[10] (figure 2). A Danish population based study (3825 patients) accounting for the
increasing age of AML patients over the 2000-2015 period found improved 2- year survival rates
|Page
Venetoclax – Despite remaining issues, this BCL-2 inhibitor, when combined with azacytidine (as
approved by FDA in 2018), should be considered, rather than purely palliative care, or single-
agent azacytidine or decitabine, conventional management in many newly-diagnosed older
patients and, accordingly, the standard to which future therapies should be compared in these
patients.
FDA approval of venetoclax together with either azacytidine or decitabine (“hypomethylating
agents”, HMA) or with low-dose cytarabine (LDAC) was granted for newly-diagnosed patients age
≥ 75 or with co-morbidities “precluding use of intensive induction therapy”. The trials leading to
approval[13, 14]
were single-arm and, given the infrequency with which results seen in such
“promising” trials are confirmed in randomized trials[15], led to randomized trials comparing
azacytidine +/- venetoclax and LDAC +/- venetoclax. As reported at the 2020 European
Hematology Association meeting[16], the former had enrolled 431 patients (median age 76, range
49-91), with similar eligibility criteria as in the approval study[13], and with 2:1 double-blind
randomization between azacytidine + venetoclax or azacytidine + placebo. Azacytidine was
given as 75mg/m2 daily X 7 every 28 days and venetoclax as 400 mg daily on days 1-28 after a
3-day “ramp-up” in cycle 1, leading to an incidence of tumor-lysis syndrome of only 1%. With a
|Page
A parallel trial randomized 211 patients (median age 76, range 36-93) considered unfit for
intensive therapy, using similar criteria as in the azacytidine +/- venetoclax study, in 2:1 double
blind fashion between LDAC + venetoclax or + placebo[17]. The venetoclax dose was 600 mg daily
in 28 -day cycles after a ramp-up period. At the time of pre-planned analysis, median follow-up in
living patients was 12 months, and the risk of death in the venetoclax arm was three-quarters that
in the placebo arm (HR 0.75, 95% CI 0.52-1.07, p=0.11, with median survivals of 7.2 vs 4.1
months ). Two subsequent (post-hoc) analyses were performed : the first after accounting for
covariates influencing survival (performance status < 2 vs ≥ 2, age < vs > 75, intermediate vs.
adverse cytogenetics, secondary vs de novo AML) yielded hazard rate 0.67 (95% CI 0.47-0.96,
p = 0.03), and the second after an additional 6 months of follow-up yielded HR (0.70, 95%CI 0.50-
0.99, p = 0.04 and median survivals of 8.4 vs. 4.1 months). Various metrics of QOL were also
improved following treatment with venetoclax + LDAC.
The longer survival seen when venetoclax was combined with azacytidine rather than LDAC might
have reflected azacytidine’s superiority to LDAC, as noted by Dombret et al[6]. It also may have
reflected the eligibility of patients for the LDAC +/- venetoclax trial if they had not responded to
azacytidine (or decitabine) when given for MDS; such patients had worse outcomes following
venetoclax + LDAC than those who had never received azacytidine or decitabine. A comparison
|Page
CPX 351
In 2017 FDA approved this liposomal formulation of daunorubicin and cytosine arabinoside for
patients with therapy-related AML (t-AML) or with AML and “myelodysplasia related changes
(AML-MRC). Approval followed a trial randomizing (1:1) 309 patients age 60-75 between CPX
and 7+3, with daunorubicin at the now accepted minimally effective dose of 60 mg/m2 daily X
3[18]. Patients had t-AML, AML arising after myelodysplasia or CMML, or de-novo AML but with
MRC as defined by MDS-associated cytogenetic changes. PS 0-2 was required, 88% had PS 0-
1. CR rates were 37% with CPX vs. 26% for 7+3 (p=0.04) and CR+ CRi rates were 48% vs
33%(p=0.02). Event-free survival was longer with CPX (p=0.02, medians 2.5 vs 1.3 months.
Improved survival was cited as the grounds for approval. As updated at the 2020 ASCO meeting,
with a median follow-up of 61 months administration of CPX reduced the risk of death by 30%
(HR 0.7, 95% CI 0.6-0.9, and thus p < 0.05, medians 9.3 months CPX, 6.0 months 7+3)[19]. CPX
appears superior to 7+3 in AML-MRC regardless of whether defined by a history of MDS ,
including patients who have and have not received HMAs, or only by MDS-cytogenetic changes[20]
although it is unclear whether the same is true in patients whom MRC is defined only by
morphologic dysplasia. The latter appear to have a better prognosis than other AML-MRC
patients[21], possibly suggesting AML with MRC defined only by dysplasia is not truly secondary
AML. Because the annual rate of relapse or death in CR does not decline to <5% until 3 years
have elapsed from remission[22] the CPX data might be viewed as more stable than the venetoclax
data (median follow-up 20 months), although it is unknown whether the failure rates will decline
|Page
These results establish CPX, and not 7+3, as the standard “intensive” therapy for older adults
with secondary AML. The updated results presented at ASCO 2020[19] also noted the risk of death
in patients in whom survival was dated from allogeneic hematopoietic cell transplant (allo HCT)
date was less in recipients of CPX than of 7+3 (HR 0.51, 95% CI 0.3-0.9). This HR was also less
than that (0.7) favoring CPX considering all treated patients. This suggests that CPX may be
particularly useful in patients who subsequently receive allo HCT, perhaps by being more effective
at reducing MRD than 7+3. Of course, this assumes such reduction is particularly important in
recipients of allo HCT, which is unclear. Besides its potential disproportionate value in patients
who are allo HCT candidates, other questions about CPX are its effectiveness in patients age <
60 with secondary AML and in patients with secondary AML who met eligibility criteria for the trial
leading to approval but were not included. The possibility of such “selection bias” is an important
consideration in all trials.
Midostaurin
The randomized “RATIFY” trial of Stone et al.[23] established 7+3 + the FLT 3 inhibitor midostaurin
as conventional therapy for adults age < 60 with either a FLT3 tyrosine kinase domain (TKD)
mutation or a FLT3 internal tandem duplication (ITD) regardless of the ratio of affected to
unaffected alleles(allelic ratio, AR, with a ratio > 0.5 denoting a worse prognosis). The superiority
of midostaurin was independent of allo HCT with best results seen in patients receiving allo HCT
in CR1 after receiving midostaurin[24]. Questions remaining in the 2019 update[25] were the need
for midostaurin maintenance[26] and whether the benefit of midostaurin would also be seen in
adults age > 60. The German AMLSG addressed these questions by administering
7+3+midostaurin to 86 patients aged 61-70 and 198 patients aged 18-60 with a FLT3 ITD[27]. They
intended all patients entering CR or CRi to receive allo HCT followed by 12 months of midostaurin
maintenance. The CR/CRi rate was 76% and was similar in older and younger patients and
unaffected by FLT 3 AR < vs > 0.5. Allo HCT was performed in 58% of those entering CR with
the remaining patients receiving “high-dose” cytarabine(HDAC) post-remission therapy followed
|Page
|Page
Ivosidenib
FDA approved this IDH1 inhibitor as a single agent for IDH1 mutated AML occurring in a very
similar population for which it granted approval for venetoclax + HMA or LDAC. As recently
reported[38], at a dose of 500 mg daily given to 34 patients ,median age 77, who remained on
study for a median of 4.5 months (up to 40 months), CR rate was 30% (95% CI 16-49%) and CR
+ CRh (see below) rate was 42% (95% CI 26-61%) . 95% CIs overlapped but rates of CR (7/18
vs. 3/15) and of CR + CRh (10/18 vs. 4/15) were higher in patients who had not received HMA,
e.g. for MDS. Median time to response was 2.8 months for CR and for CRh (up to 4.6 months for
CR and 12.9 months for CRh). Median durations of response were not reached, with 62% of CR+
CRh and 78% of CR patients remaining in remission at 1 year. With a median follow-up time of 2
years, median survival was 12.6 months (95% CI 4.5-25.7). IDH1 mutations disappeared in 5/10
patients entering CR, 4/4 entering CRi, and 0/16 achieving neither. Durations of such molecular
remissions were not noted. 9 of 21 patients who were transfusion dependent pre-treatment
became transfusion independent for.at least 56 days, while 7 of 12 who were transfusion
independent at study entry remains so for at least 56 days. Consistent with the drug’s mechanism
of action a “differentiation syndrome”(DS) analogous to that which occurs following treatment of
|Page
Although a randomized trial in “unfit” patients similar to those in the venetoclax and ivosidenib
studies showed longer survival patients randomized to the hedgehog pathway inhibitor glasdegib
+ LDAC vs LDAC alone[39], leading to FDA approval of glasdegib, the drug has yet to garner the
same traction, as venetoclax, and I will it omit from further discussion, noting the issues pertaining
to venetoclax discussed below are pertinent to glasdegib.
Table 1 summarizes scenarios in which the recently approved FDA drugs should be considered
“standard of care”. These conform to FDA’s bases for approval, supplemented by data published
since approval in the cases of midostaurin[27]and gemtuzumab[32]. The table also notes some
remaining issues. Primary among these is the definition of “unfitness” for intensive chemotherapy.
|Page
|Page
Although it is often assumed patients given less intense induction would have superior QOL, El-
Jawhari et al. analyzing QOL, anxiety, and depression in patients ≥ 60 during the first 24 weeks
following initiation of remission induction therapy found that QOL and anxiety generally improved
(depression stayed the same) regardless of whether intensive (anthracycline+ cytarabine) or less
intense (HMA) induction was given[55]. Of course, it could be argued the results with neither
standard intensive therapy nor venetoclax + azacytidine as a new non-intensive standard are
such that trials combining more and less intensive induction would be out of place in appropriate
patients, a topic I address below.
Conventional therapy vs. investigational therapies
Assuming the advent of therapies such as venetoclax + azacytidine or ivosidenib means fewer
patients should be offered purely palliative care, the question remains whether a given patient
should be offered conventional therapy, including those newly identified as such, or investigational
therapy on a clinical trial. The latter often requires referral to a large academic center and the time
spent there may eventually represent a significant proportion of a patient’s remaining life
expectancy. Furthermore, the trial’s outcome cannot be known a priori. It follows the main reason
to refer a patient for a trial is dissatisfaction with the results of conventional therapy. For example,
median survival in patients given azacytidine + venetoclax arm in the recently reported
randomized trial was 14.7 months vs. 9.6 months for azacytidine + placebo; at 2 years probability
of survival was 40% vs. 30%[16]. CPX afforded a 9.6 months median survival vs. 6 months for the
prior standard, 7+3[18]. The median does not give a complete account, but represents a convenient
picture of probable outcome, particularly given our limited ability to know who will do much better
|Page
For some patients the 5- 6- month improvement in median survival afforded by addition of
venetoclax to azacytidine or the 3-4 month improvement in median survival following CPX351
rather than 7+3 is very meaningful, allowing them for example to see a grandchild graduate
college. For other patients the improvement will be less meaningful particularly if informed that
US Social Security Administration tables[58] indicate the average 75 year old man can expect to
live another 11.4 years and the average 75 year old woman another 13.1 years in the absence of
AML, admitting that patients who develop AML may have been susceptible to other life-
threatening conditions if they had not developed AML. The point is the choice between
conventional and investigational therapy is largely subjective, particularly because patients may
vary in what they value as benefit (e.g. survival vs. QOL) and the amount of risk they are willing
to accept to achieve a benefit[59]. As noted by Schiffer[60], the “alternative treatment” section of
typical written informed consent documents for clinical trials, while listing alternatives to the trial
such as conventional treatment or purely palliation, usually give very little specific information
about what might be expected with these. Provision of such information is thus incumbent on the
treating physician.
ELN 2017 prognostic system to assess expected prognosis with conventional therapy:
general remarks about “prognostic systems”
ELN 2017 provides a useful guide to aid in the choice between conventional and investigational
therapy[30]. Patients are divided into “favorable”, “intermediate” and “adverse” groups. (table 2).
The German AMLSG recently validated the prognostic significance of ELN2017 for remission,
survival, and relapse-free survival (figure 8)[61]. They also noted the particularly poor prognosis of
patients with TP53 mutations (TP53 is located on the short (p) arm of chromosome 17) or complex
|Page
Multivariate analyses indicate that even after accounting for its association with adverse
cytogenetics and thus disproportionate representation in the ELN adverse group, older age is
itself associated with poorer relapse-free, and overall, survival[61]. SWOG and UK MRC/NCRI data
indicate patients age > 65 with the presumed “favorable” constellation of an NPM1 mutation, no
FLT3 ITD, and normal cytogenetics have inferior survival and RFS survival than patients age 55-
65 and < 55 with the same findings[63].
Neither ELN2017[30] nor its modifications[61-64] incorporate data from patients treated with the new
therapies listed in table 1. However above I noted several examples (cytogenetics, TP53
mutations, prior HMA) where patients expected to do relatively poorly with previous conventional
therapies (e.g. 7+3) also did relatively poorly with newer conventional therapies (e.g. venetoclax
+ azacytidine, CPX351, ivosidenib). This suggests systems like ELN2017 are likely to be useful
in forecasting outcome even in patients given these newer therapies.
An important question is the accuracy of ELN 2017 and other prognostic systems. One way to
quantify accuracy is via areas under receiver operating characteristic curves (AUC). An AUC
value of 1.0 indicates every time a model predicts a given patient will fare better than another
patient the forecast is correct, although the amount of improvement is not specified. In contrast
an AUC value of 0.5 indicates no predictive ability (a coin flip). Although AUC values are
uncommonly reported, when reported the predictive value of most models has been found to be
approximately intermediate between certainty and a coin flip (AUC 0.70-0.75), which is a much
poorer performance than suggested by p-values often << 0.05[56].
|Page
|Page
A still unsettled question is whether 7+3 or a regimen including higher doses of cytarabine, such
as FLAG-ida, is the preferable intense induction regimen. SWOG trial 1203 randomized 754
adults age < 60 among 7+3 (with a daunorubicin dose of 90 mg/m2), idarubicin (12 mg/m2 daily
X 3) + cytarabine 1.5 g/m2 daily by continuous infusion X 4 days (IA) and IA + vorinostat.CR rates
were 75-79% although 24% of 7+3 patients required 2 courses to enter CR vs 9-11% for the IA
treatments. Allogeneic hematopoietic cell transplant rates in CR were similar in each arm (about
48%) and there were no differences in EFS, RFS, or survival among the 3 arms, including in
patients with NPM1, FLT3, or CEBPA mutations or intermediate or adverse cytogenetics, with
patients with “favorable” cytogenetics having inferior EFS,RFS, or survival if in the IA arms[74]. The
study has been criticized because post CR doses of cytarabine were higher in the 7+3 arms
(18g/m2 vs. 2.25g/m2 on each cycle for 1-4 cycles depending on HCT availability). In contrast the
NCRI/MRC group in the United Kingdom found reduced relapse rates with FLAG-ida than with 10
days of standard cytarabine + daunorubicin.[75] However, there were no differences in survival
possibly because of higher treatment-related mortality rate in CR in the FLAG- ida arm.
Further complicating the issue is the probability a given result may not be applicable to different
subsets, recognizing a principal characteristic of AML is its heterogeneity. For example, although
a 1205 patient randomized NCRI trial found no difference, other than more TRM in the higher
dose arm, between 7 + 3 using daunorubicin doses of 90 mg/m2 vs the usual 60 mg/m2 each daily
X 3, patients with FLT3 ITD (n = 200) had less relapse leading to longer survival if given the 90
[76]
mg/m2 dose . The same may apply when deciding on intensity of induction. For example,
|Page
With the increasing use of targeted therapy an important question is the practicality of “precision
medicine” in AML, i.e. assigning individuals to therapy based on their mutational pattern. Above
we noted the lack of obvious influence of time from diagnosis to treatment on outcome[70, 71].
Building on these findings the ongoing “BEAT AML” trial enrolled, at 14 sites, 497 newly-
diagnosed patients age > 60, of whom 395 were eligible, with ineligibility generally reflecting an
incorrect diagnosis[83]. “Next generation sequencing” (NGS) was done with the goal of treatment
assignment (TA) to a specific sub-study, for example anti TP53 or to a “marker-negative” therapy
within 7 days of enrollment (figure 10). This goal was accomplished in 57% of the 395 patients
with minimal inter-center variation. The remaining patients either received standard therapy (7+3
or azacytidine, decitabine), a non-specific trial, or a purely palliative approach. Survival analyses
adjusted for standard prognostic covariates indicated survival was similar following receipt of a
BEAT AML or a non-specific trial (figure 10), although it was not clear how results would have
been affected had only patients on a specific BEAT AML sub-study been included.
A major source of drugs for treatment of newly-diagnosed AML are drugs initially tested in
relapsed/refractory (R/R) AML. I discuss several of the latter including bi-specific antibodies
(BiTEs) and newer FLT3 inhibitors in the section on R/R AML
|Page
Recent years have seen the adoption of responses less than CR, for example CRp or CRi in
which absolute neutrophil count (ANC) or platelet count respectively meet the criteria for CR, but
both ANC and platelet count do not. The category CRh in which platelet count is > 50,000 and
ANC > 500 but not meeting requirements for CR (ANC > 1,000, platelet count > 100,000) has
appeared most recently. FDA approved ivosidenib and enasidenib based on rates of CR + CRh[40,
86],
. Particularly with new drugs, CRh, and CRi/CRp are often combined into “overall response rate
(ORR)”. However, patients undoubtedly care more about the effect of a given effect on their
survival (or quality of life) than on the nomenclature affixed to the response. After accounting for
the longer time needed to achieve CRp than CR and differences in standard prognostic
covariates, Walter et al. reported CR was associated with longer relapse-free survival, and to a
less extent survival, than CRp[87]. However, CRp was associated with longer survival than
resistance (neither CR nor CRp but living at least 30, 60, or 90 days after initiation of therapy),
thus validating CRp as a clinically relevant response[87]. The same has yet to be done for CRh.
Although blood count criteria for CR are typically considered to include only ANC > 1000 and
platelet count > 100 000, Danish investigators have noted the association between continued receipt of
red cell transfusions at time of CR and increased risk of relapse[88].
It has long been known disease recurs in most patients who enter CR. This suggested such
patients still had disease present in CR. But no means were available to detect such disease.
These days various methods, discussed below, are available to discover “measurable residual
disease” (MRD). In general, CR without MRD is associated with longer remissions, RFS, and
survival than is CR with MRD. This seems to be the case in patients given more[89] or less[90]
|Page
Post-Remission Therapy
It is broadly accepted post remission therapy is needed to prevent rapid occurrence of relapse.
Such therapy has usually involved hematopoietic cell transplant (HCT, allogeneic or autologous)
or continued chemotherapy. HCT, particularly allo HCT, is more likely to produce non-relapse
mortality and morbidity due, for example, to acute and chronic graft vs host disease (GVHD) and
development of secondary malignancies reflecting use of immunosuppressive to treat GVHD.
However, HCT, particularly allo HCT, typically reduces risk of relapse, with a graft vs. leukemia
(GVL) effect playing a major role in the allo case and leading to efforts to manipulate the
transplanted immune cells to minimize GVHD and maximize GVL.
Quantifying risks of relapse and non-relapse mortality (NRM) to inform the allo HCT
decision
The decision to recommend allo HCT is based on the presumption the reduced risk of relapse
will be greater than the increased risk of NRM and morbidities such as GVHD. Cornelissen et al.
proposed a projected > 10% improvement in RFS following allo HCT to justify the procedure[94],
although such a criterion likely varies according to individual preferences. The potential morbidity
from GVHD has been recognized by the introduction of a new endpoint (GVHD-free, relapse-free
survival [GRFS] [95].
|Page
|Page
Using Mantel-Byar methodology and landmark analyses to account for the “guarantee time” HCT
patients, but not chemotherapy only patients, must live before HCT,[110] 0stgård et al. used the
Danish Acute Leukemia Registry assuring complete data collection on virtually all 1031 adults with
intermediate—or adverse-risk cytogenetics age <70 who entered CR between 2000 and 2014, of
whom 19% received HCT in CR1[111]. HCT patients were younger with fewer co-morbidities but
more often had adverse cytogenetics; two-thirds of transplants used reduced intensity, rather
than myeloablative, conditioning. Median follow-up was 2.2 years. cumulative incidence of
relapse was lower, and survival (figure 12) and relapse-free survival longer, in HCT recipients with
either intermediate or adverse cytogenetics or age <60 or 60-69 (Figure 12). Relapses occurred
in 49% and deaths in 68% of chemotherapy patients vs 24% and 42% for HCT patients. Median
RFS was 293 days (interquartile range 142-910) for chemotherapy only vs 1067 (IQ range 397-
2696) for HCT with median survival times of 476 days (IQ 167-1589) and 1173 (IQ 474-2859).
These findings support the general recommendation to offer allo HCT to all ELN2017
intermediate and adverse risk patients. In contrast, the risk of relapse without allo HCT in the
ELN 2017 favorable subgroup is generally considered too low to justify the risk of allo HCT.
MRD assessments can refine these recommendations. Thus table 6 indicates the risk of relapse
without allo HCT in patients in the ELN 2017 favorable groups NPM1 mutated/FLT3 negative
and CBF varies substantially depending on whether there was a 3-log or greater reduction in the
|Page
A GIMEMA group study in younger adults with newly- diagnosed AML showed the potential
value of using MRD data in deciding whether to recommend allo HCT[113]. After an induction and
a post-remission course of standard therapy patients with intermediate-risk cytogenetics were
divided into MRD (by MFC) positive and negative groups. Allo HCT was recommended in the
former, autologous HCT in the latter. The MRD positive and negative patients had similar
survival and RFS, suggesting the decision to offer the MRD positive patients allo HCT played a
role in these unexpected findings. Analogous results have been reported in favorable risk
cytogenetic patients[114].
The relative contribution of NRM to RFS is much less with chemotherapy than allo HCT[49].
Although there are alternatives, the hematopoietic cell transplantation comorbidity index (HCT-CI)[115]
effectively estimates NRM post- HCT (table 6).
|Page
With these developments the use of allo HCT may increase. In addition to potentially benefiting
patients, this might decrease the potential effect of selection bias in allo HCT studies, a problem
that of course is not unique to allo HCT. Although in the study of Ostgard et al[111], it was
recommended that all patients with intermediate or adverse cytogenetics in CR1 receive allo
HCT, only 19% did so, although this proportion increased from11% to 27% in the later years of
the study. HCT rates of 60% in patients with adverse cytogenetics in CR1 have been reported
in a cooperative group setting[122]. As in any study it remains unknown how representative these
60% were of the entire population, a problem that becomes worse when studies do not report
the number of patients receiving various types of HCT as a proportion of those who might have
done so.
|Page
Risk of relapse might also be reduced by post HCT interventions. Madiarz et al. randomized 60
adults age < 70 with FLT ITDs in remission 1-2 months after allo HCT to receive usual care +/-
midostaurin 50 mg twice daily; median exposure was 10 months. Relapse rates were 24% in the
control and 11% in the midostaurin arm, with 18-month RFS probabilities 76% (95% CI 54-88%)
in the control and 89% (69-96%)[128]. Analogous trials of newer FLT3 inhibitors (see below) are
ongoing in this setting. Despite the overlapping 95% CI in the +/- midostaurin trial, the proportion
of patients who will be willing to accept a 50% chance of receiving placebo rather than a 100%
chance of receiving midostaurin (off study) is uncertain, especially given the results of Schlenk et
al. noted above[27]. Because of azacitidine’s ability to upregulate potential AML- associated
antigens thus inducing a CD8+ T cell response augmenting the GVL effect, Craddock
administered the drug to 51 patients undergoing allo HCT, finding a decreased relapse risk (HR
0.30, 95% CI 0.10-0.85) with similar improvement in RFS in patients demonstrating a CD8+ T-
cell response to tumor antigens[129]. Although azacytidine has yet to be accepted as prophylaxis
against post HCT relapse, it is well-tolerated and thus is often given to patients at high risk of this
complication. The availability of the potentially more effective oral form of azacytidine (CC-486,
see below) may further increase the use of azacytidine as prophylaxis of post HCT relapse[130].
More sophisticated attempts have also been made to augment the GVL effect, for example by
generating T cells expressing receptors (TCR) that target selected AML antigens, for example
WT-1. Chapuis et al. identified a high-affinity, HLA-A2-restricted WT1-specific TCR (TCRC4) from
|Page
MRD testing might prompt a re-evaluation of auto HCT. Specifically, patients in the ELN2017
intermediate (and possibly some in the ELN favorable group) who are MRD negative after initial
therapy might be considered for auto HCT rather than allo HCT.
|Page
Chemotherapy for Group A emphasizes cytarabine in individual doses of 1-3 g/m2 daily or BID
for 3 days per cycle for 3 cycles, often given as 3 g/m2 every 12 hours on days 1, 3, and 5.
However, a German- Austrian AML Study Group found administration of this dose on days 1, 2,
and 3 to 392 patients was associated with a median 4-day reduction in time to ANC >500 from
start of chemotherapy and with fewer platelet transfusions than in 176 historical patients given
the 1, 3, and 5 schedule[137]. Cytarabine doses of 1 g/m2 twice daily for 3-4 days for three courses
after CR is achieved are very probably as effective as higher doses and/or more courses
irrespective of cytogenetic group.[138] Furthermore, after randomizing 221 adults age < 60
between cytarabine in the usual 3 g/m2 schedule and cytarabine 1 g/m2 daily X5 plus clofarabine
30 mg/m2 daily X5 (CLARA) both preceded by G-CSF, Thomas et al. reported a lower incidence
of relapse (34% vs 46% at 2 years P = .025) accompanied by longer RFS (but not survival) with
CLARA regardless of risk group, raising the question whether addition of a purine analogs such
as fludarabine, cladribine, or clofarabine should be added to cytarabine for standard post-
remission therapy in such patients.[139] Perhaps most importantly decisions about doses and
schedules for subsequent courses of post-remission therapy cannot be stereotyped and must be
based on tolerance of the previous course, bearing in mind that relapse rather than death in CR
is the predominant problem.
“Maintenance” has generally referred to the prolonged administration of relatively less intense
therapy to maintain CR. It had until very recently fallen out of favor[140]. However, azacitidine
appears to offer a new maintenance option for patients who are candidates for neither HCT nor
cytarabine in the doses described in the preceding paragraph (table 6). Huls et al. randomized
between observation and azacytidine 116 patients age 60 or greater (median 69) with
|Page
The above study required 7.5 years to accrue 118 patients, a not uncommon occurrence in studies
of maintenance therapy in older patients[141]. A possible explanation is the need to receive
parenteral therapy for a long time. An oral preparation may not only be useful per se but the
[142]
prolonged exposure it affords may be therapeutically beneficial. Wei et al. randomized 472
patients age 55-86 (median 68) with performance status < 3), intermediate (86%) or adverse risk
(14%) cytogenetics between oral azacitidine (CC-486, 300 mg daily on days 1-14 of 28-day
cycles) and placebo. 81% were in CR and 19% in CRi. 45% had received one and 35% two
courses of “consolidation” therapy other than azacytidine. Azacytidine/placebo could continue
indefinitely barring excess toxicity or appearance of > 15% blasts. With a median follow-up of 3.5
years CC 486 reduced the risk of death (HR 0.69, 95% CI 0.55-0.86) with median survivals of
24.8 vs. 14.2 months and the risk of relapse and/or death (HR 0.65, 95% CI 0.52-0.81), with
median RFS of 10.2 vs. 4.8 months (figure 14). Nausea, vomiting, and diarrhea, generally mild-
moderate, were more common with CC486 occurring in 64%,59% and 49% of patients and were
the most common reasons for discontinuation. However, health- related QOL was similar in both
arms.
|Page
|Page
|Page
|Page
E-selectin ligand inhibition- E-selectin ligand is expressed on AML blasts and mediates adherence
to E-selectin on bone marrow endothelium resulting in resistance to chemotherapy. Uproleselan
disrupts adherence, putatively re-sensitizing blasts to chemotherapy. A single-arm trial of
uproleselan added to MEC reported[156] a 45% CR/CRi rate and 12.8 month median survival in
the 22 patients with >10% E-selectin ligand expression vs. 29% CR/CRi rate with 5.4 month
median survival in the 14 patients with less expression; 79% of the former and 72% of the latter
patients were in the Breems et al. poor prognostic group or receiving at least 2nd salvage
suggesting results in the higher expression group were better than those otherwise expected[150,
152]
. A randomized trial comparing FAI or MEC +/- uproleselan is ongoing in relapsed/ refractory
and newly- diagnosed patients.
Immunotherapeutic approaches
a) Bispecfic T-cell Engagers antibodies(BiTEs)- these attempt to redirect T cells to hematologic
malignancies. Blinatumomab and inotuzumab, which engage CD3 present on almost all T cells
|Page
b) Chimeric antigen receptor (CAR) T-cells – as with BiTEs, to date the success of CAR T cell
therapy in AML has been considerably less than in ALL. One likely reason is the absence of a
specific cell surface antigen, such as CD 19, that can be targeted with relatively little toxicity, such
as myelosuppression. Current approaches to address this problem include: 1) using transient or
removable CAR T-cells as a component of an allo HCT conditioning regimen, 2) gene-editing
(e.g. deleting CD33 from the donor’s cells) to allow safe and protracted anti-AML CAR T-cell
function[159].
c) immune checkpoint blockade -the success of this approach in solid tumors has motivated trials
in AML. Resistance to azacytidine or decitabine (“HMAs”) appears at least partially mediated by
these drugs’ ability to increase expression of PD-1 thus decreasing the immune response to AML
and suggesting the use of anti PD-1 antibodies, which by themselves are only minimally effective
in AML[160] This prompted Daver et al to administer azacytidine (75 mg/m2 daily X 7) and
nivolumab(3 mg/kg days 1 and 14) to 70 patients with relapsed/refractory AML[160].The ORR was
33%(6% CR) with median survival 6 months. Results appeared better in patients who had not
received HMAs than seen in analogous historical controls (ORR 52% vs. 22%), as did survival in
patients receiving azacytidine and nivolumab as “first salvage” (medians 10.6 vs. 5.3 months)
with similar observations for EFS, although no multivariate analysis was provided and follow-up
was naturally longer in the controls (medians 51 vs 13 months). Higher marrow CD3 and CD8
expression seemed to distinguish responders from non- responders and might be useful in
|Page
Allo HCT- like all therapies for AML allo HCT is much less effective when morphologic disease is
present. Improved conditioning regimens are one means to address this problem[127]. In a recent
example Gyurkocza et al. randomized 75 patients age > 55 not in remission, or in relapse and
with > 5% blasts, between investigator’s choice of chemotherapy including venetoclax, ivosidenib,
or enasidenib (“conventional care”) and [130]I apamistamab (IOMAB B)
followed 12-14 days later by reduced intensity conditioning (fludarabine (FLU)/TBI) and allo
HCT[162]. Patients in the conventional care arm could proceed to allo HCT in the event of CR or
crossover to the IOMAB-HCT arm if not in CR. Reflecting that 85% of patients had previously
failed at least 2 induction therapies, 31/38 conventional arm patients again failed chemotherapy,
including 11 receiving targeted therapy. However, patients were able to cross-over within a
median of 64 days (range 51-161) from randomization such that 50/74 patients (31 initially
randomized to HCT plus 19 cross overs) were able to receive HCT. The 100-day post HCT NRM
rate was 0% in the former and 7% in the latter, with a median of 2-3 weeks to neutrophil and
platelet engraftment in each arm. Although EFS, RFS, CR, and survival were not reported the trial
suggests the feasibility of allo HCT in patients typically not considered HCT candidates based on
older age, active AML, or number of prior salvage attempts.
Treatment of MRD
|Page
OTHER TOPICS
Need for hospitalization
Conventionally, AML patients have been hospitalized for 5-7 days to receive initial induction,
remaining in hospital for another 3-4 weeks until neutrophil and platelet count recovery. The
increasing use of less intense induction will almost certainly reduce the need for hospitalization.
However, it has been shown that patients can be safely discharged from hospital after completion
of intense induction therapy (“early discharge”), assuming availability of caregivers and an
adequate outpatient infrastructure, with patients living relatively near (e.g. 30-50 miles) of the
center. Early discharge patients were shown to have similar TRM as controls who met the same
medical eligibility criteria (doing well without fever etc. when chemotherapy completed) but who
lacked caregivers or who lived at too great a distance from the outpatient center[164]. Selected
|Page
Mutations present in hematopoietic cells at birth are germline mutations (i.e. present at birth in all
cells)[166, 167]
. Germline mutations often involve RUNX1, GATA2, and DDX41 and characterize
hereditary myeloid malignancy syndrome (HHMS). These frequently culminate in MDS/AML, which,
although usually presenting in childhood, can be seen in adults usually age <40 years but occasionally
older. Screening families for germline mutations is advisable if a family member is a potential donor for
allo- HCT, if a patient has a potentially germline mutation (e.g., RUNX1, GATA2, and DDX41), HHMS
organ manifestations, or a suggestive family history[166, 167]
.Skin biopsies or nail clippings are
commonly used to test for germline mutations.
Fertility preservation
Given the potential teratogenicity of AML therapy storage of sperm or eggs is an option that
should be offered every young patient, in consultation with appropriate specialists. Such storage
is eminently feasible given data suggesting the need to begin AML therapy as an emergency is
probably less than suspected (see above and references 70 and 71).
|Page
In any event the 20% blast cut-off used to distinguish “AML” from “MDS” is arbitrary. Based on the
possibility mutations and cytogenetics are more prognostically and therapeutically
relevant than the 20% cut-off, some centers treat patients with >10% blasts as AML, even if they
are considered as excess blasts 2 by the WHO.
Date-sharing: Data sharing is not applicable to this article as no new data were created or analyzed
in this study.
Funding: None
References
1. Meyers J, Yu Y, Kaye JA, et al. Medicare fee-for-service enrollees with primary acute myeloid
leukemia: an analysis of treatment patterns, survival, and healthcare resource utilization and costs. Appl
Health Econ Health Policy 2013;11:275-286.
2. Medeiros BC, Satram-Hoang S, Hurst D, et al. Big data analysis of treatment patterns and
outcomes among elderly acute myeloid leukemia patients in the United States. Ann Hematol
2015;94:1127-1138.
3. Oran B, Weisdorf DJ. Survival for older patients with acute myeloid leukemia: a population-
based study. Haematologica 2012;97:1916-1924.
4. Juliusson G, Lazarevic V, Hörstedt AS, et al. Acute myeloid leukemia in the real world: why
population-based registries are needed. Blood 2012;119:3890-3899.
5. Juliusson G, Antunovic P, Derolf A, et al. Age and acute myeloid leukemia: real world data on
decision to treat and outcomes from the Swedish Acute Leukemia Registry. Blood 2009;113:4179-4187.
6. Dombret H, Seymour JF, Butrym A, et al. International phase 3 study of azacitidine vs
conventional care regimens in older patients with newly diagnosed AML with >30% blasts. Blood
2015;126:291-299.
|Page
|Page
|Page
|Page
|Page
|Page
|Page
|Page
|Page
|Page
|Page
|Page
Fo
rP
ee
rR
ev
iew
Fo
rP
ee
rR
ev
iew
Fo
rP
ee
rR
ev
iew
Fo
rP
ee
rR
ev
iew
Frequencies, response rates, and outcome measures should be reported by risk category, and,
if sufficient numbers are available, by specific genetic lesions indicated.
*Prognostic impact of a marker is treatment-dependent and may change with new therapies.
†Low, low allelic ratio (,0.5); high, high allelic ratio ($0.5); semiquantitative assessment of FLT3-
ITD allelic ratio (using DNA fragment analysis) is determined as ratio of the area under the curve
“FLT3-ITD” divided by area under the curve “FLT3- wild type”; recent studies indicate that AML
with NPM1 mutation and FLT3-ITD low allelic ratio may also have a more favorable prognosis
and patients should not routinely be assigned to allogeneic HCT.57-59,77
‡The presence of t(9;11)(p21.3;q23.3) takes precedence over rare, concurrent adverse-risk gene
mutations.
§Three or more unrelated chromosome abnormalities in the absence of 1 of the WHO-designated
recurring translocations or inversions, that is, t(8;21), inv(16) or t(16;16), t(9;11), t(v;11)(v;q23.3),
t(6;9), inv(3) or t(3;3); AML with BCR-ABL1.
||Defined by the presence of 1 single monosomy (excluding loss of X or Y) in association with at
least 1 additional monosomy or structural chromosome abnormality (excluding core-binding factor
AML).116
{These markers should not be used as an adverse prognostic marker if they cooccur with
favorable-risk AML subtypes.
#TP53 mutations are significantly associated with AML with complex and monosomal
karyotype.37,66-69
Adverse Clinical trial, for example 5+2 Clinical trial, for example
or FLAG-ida each + “BEAT AML”
venetoclaxe
consider adding GO ;
azacitidine
Adverse MFC Negative 70- 80% ≤40% Clinical trial (see section on
(66% of pts) (≤3) relapsed/ref AML )
CR azacitidine