Treatment strategies for therapy-related acute myeloid leukemia

Abstract

Prospective evidence for management of therapy-related acute myeloid leukemia (t-AML) is limited, with evidence extrapolated from major AML trials. Optimal treatment is challenging and needs consideration of patient-specific, disease-specific, and therapy-specific factors. Clinical trials are recommended, especially for unfit patients or those with unfavorable cytogenetics or mutations. CPX-351 as an upfront intensive chemotherapy is preferred for fit patients; venetoclax with decitabine or azacitidine is an option for patients unfit for intensive chemotherapy. Hematopoietic cell transplant, the only curative option, should be offered to eligible patients with intermediate or unfavorable t-AML or patients with good risk AML with minimal residual disease. Ongoing clinical trials focusing on treatment of t-AML, including targeted agents and immunotherapy, bode well for the future.

Introduction

Optimal treatment for therapy-related acute myeloid leukemia (t-AML) is a matter of debate. The 2016 world health organization (WHO) classification lists t-AML as a distinct entity resulting from prior cytotoxic chemotherapy or radiation [1]. There is no specific definition for secondary AML (s-AML) but includes AML transformed from antecedent myelodysplastic syndrome or myeloproliferative neoplasm as well as t-AML [2]. Management is of t-AML challenging because of high risk factors associated with t-AML including older age, poor performance status, greater burden of co-morbidities, and high risk cytogenetics and mutations [3-7]. Prospective data specifically on t-AML or s-AML are limited, and the evidence is usually extrapolated from major AML trials. Some researchers have argued that t-AML may not always need a different treatment strategy than de novo AML, and upfront intensive chemotherapy in suitable patients, particularly fit patients with good risk t-AML, may lead to better outcomes [3]. The approval of 8 new AML drugs in the last 2 years has expanded our therapeutic armamentarium, thus making therapy selection more difficult. Here we discuss strategies to personalize treatment of adults with t-AML.

Biology and prognosis of therapy-related AML

Patients with t-AML are older and frequently have unfavorable cytogenetics compared to those with de novo AML[4, 8-10] (Table 1). Multiple studies have reported older age to be independently related to worse overall survival (OS) in t-AML patients [9, 11]. Although t-AML carries relatively worse prognosis than de novo AML, median OS is strongly influenced by cytogenetics [12]. The frequency of cytogenetic risk categories varies in t-AML based on different studies: 1-26% of favorable cytogenetics, 26-52% intermediate cytogenetics, and 31-67% of unfavorable cytogenetics [8, 9, 13-15]. Subsequently, OS for t-AML may range from > 2 years for favorable cytogenetics to <6 months for unfavorable cytogenetics group [9, 12, 16, 17]. The presence of core binding factor (CBF) rearrangements have shown to confer favorable prognosis in some but not other studies with t-AML and s-AML [13, 18]. CBF abnormalities conferred worse OS and EFS in s-AML than de novo AML; however, in multivariate analysis, CBF abnormalities only had marginal significance in outcome for s-AML [19].

Table 1.

Comparison of cytogenetics and overall survival in t-AML and de novo AML

AML risk category	No. of patients, t-AML vs de novo AML	OS, t-AML vs de novo AML
Favorable
Armand et al. [13]	2 vs 45	5-yr OS: 100% vs 70%
Boddu et al. [87]*	3 vs 226	-
Ostgard et al. [3]	5 vs 81	1-yr OS: 80 vs 91%
Kayser et al. [8] †	136 vs 1207	-
Schoch et al. [9]	24 vs 242	Median OS: 18 months vs NR
Intermediate
Armand et al.	11 vs 186	5-yr OS 54 vs 33%
Boddu et al.	118 vs 1159	-
Ostgard et al.	48 vs 821	1yr OS- 56 vs 69%
Kayser et al.	46 vs 1174	-
Scoch et al.	26 vs 626	Median OS: 11 vs 14 months
Unfavorable
Armand et al.	7 vs 42	5-yr OS: 13% vs 8%
Boddu et al.	109 vs 849	-
Ostgard et al.	35 vs 205	1-yr OS; 20 vs 41%
Kayser et al.	18 vs 272	-
Scoch et al.	43 vs 223	Median OS: 6 vs 6 months

^*ts-AML was defined as s-AML which develops after antecedent hematological disorder such as myelodysplasia, myeloproliferative neoplasm, or aplastic anemia, and received at least 1 therapy for that diagnosis. Median OS for ts-AML vs de novo AML for >60 year-old patient: 5 vs 24 months and for ≥60 year-old: 5 vs 9 months OS for ts-AML vs t-AML in <60-year-old patients: 5 vs 11.2 months; ≥60-year-old:4.7 vs 5.4 months, OS for Non-adverse vs adverse cytogenetics in <60 yr old patients- 7 vs 3 months; ≥60 yr old- 10 vs 3 months

^†4-yr OS for total t-AML patients vs de novo AML patients: 25% vs 38%

AML- acute myeloid leukemia, OS- overall survival, t-AML- therapy related acute myeloid leukemia, s-AML- secondary acute myeloid leukemia, ts-AML- treated secondary acute myeloid leukemia

A few studies have looked at the incidence of myeloid mutations in t-AML [20-22]. TP53 has been reported in 20-50% of patients with t-AML, which is significantly higher than patients with de novo AML [15, 20, 22-25]. TP53 mutation is associated with unfavorable cytogenetics, increased number of chromosomal abnormalities, intrinsic therapy resistance, and worse prognosis [15, 20]. In a study by Ok et al., t-AML, in comparison to de novo AML had higher incidence of PTPN11, lower incidence of FLT3 and NPM1, and similar incidence of IDH1/IDH2, DNMT3A, RAS, KIT, GNAS mutations [22]. Lindsley et al. compared specific genetic mutations associated with t-AML and s-AML versus novo AML [20]. The study concluded that the presence of mutations in SRSF2, SF3B1, U2AF1, ZRSR2, ASXL1, EZH2, BCOR, and STAG2 was >95% specific for s-AML. One-third of t-AML patients harbored these mutations; t-AML patients with these mutations were more likely to be older, have more recurrent driver mutations, and require multiple induction cycles, compared to t-AML patients with NPM1 mutations, CBF rearrangements, and KMT2A rearrangements [20]. NPM1 mutation, thus, may be considered a favorable factor in t-AML, specifically in association with wild type FLT3-ITD, and treated with intensive induction followed by consolidation chemotherapy [18]. However, the prognostic impact of each mutations in t-AML has not been extensively studied and needs further research.

Treatment options

The choice of therapy for t-AML differs based on age, comorbidities, performance status, and cytogenetic and molecular features, which influence response to treatment and OS. Clinical trials should be offered to patients with t-AML, if possible. Table 2 summarizes pivotal trials related to available treatment options for AML in general. Therapy for a patient with t-AML has to be tailored based on cytogenetic risk and fitness of an individual.

Table 2:

Selected key trials for approved agents for previously untreated acute myeloid leukemia

Author, year	Study arms	Study type, no. of patients	Patient characteristics	CR/CRi	OS*	Conclusion/Comments
Lancet 2018 [36]	CPX-351 vs 7+3	Phase III, 309	Age 60-75 yrs, s-AML or t-AML, MRC	38% vs 33% (p-0.1)	9.5 vs 5.9 month s (p-0.003)	OS better with CPX-351 in older, high-risk AML
DiNardo 2019 [42]	Venetoclax+Aza/Decitabine	Phase Ib, 145	Age ≥65 yrs, Unfit (25% s-AML)	67%	17.5 months	Venetoclax with aza or decitabine effective in older, unfit adults
Wei 2019 [43]	Venetoclax + LDAC	Phase I/II	Age ≥65 yrs, Unfit (49% s-AML)	54% s-AML-35% De novo AML-71%	10.1 months	Venetoclax with LDAC effective in older, unfit adults
Cortes 2019 [44]	LDAC with or without glasdegib	Phase II, 132	Age ≥ 55, Unfit, AML or high risk MDS	17% vs 2% (p<0.05)	8.3 vs 4.9 (p-0.003)	Glasdegib better in unfit or adults ≥75 years
Dombret 2015 [58]	Aza vs CCR (BSC, LDAC, or IC)	Phase III, 488	Age ≥65, Unfit intermediate-unfavorable risk cytogenetics, (10% s-AML)	28% vs 25%, p-0.5)	10.4 vs 6.5 months (p-0.1)	Aza better OS than BSC, similar to LDAC or IC†
Amadori 2016 [39]	GO vs BSC	Phase III, 237	Age ≥60, Unfit (31% s-AML)	27% with GO	4.9 vs 3.6 month s (p-0.005)	GO improves OS in favorable and intermediate risk cytogenetics

^*denotes median OS unless specified

^†Median OS in subgroup analysis: Aza vs BSC- 5.8 vs 3.7 p-0.02 ; Aza vs LDAC- 11.2 vs 6.4 p-0.4; Aza vs IC 13.3 vs 12.2 p-0.5

Aza- azacitidine, AML- acute myeloid leukemia, BSC- best supportive care, GO- gemtuzumab ozogamicin, IC- intensive chemotherapy, LDAC- low dose cytarabine, OS- overall survival, s-AML- secondary acute myeloid leukemia, t-AML- therapy related acute myeloid leukemia. 7+3-intensive induction regimen of cytarabine for 7 days with anthracycline for 3 days.

In young fit patients, options for intensive chemotherapy include CPX-351 or 7+3 regimen (cytarabine for 7 days and anthracycline for 3 days) citation [4, 8, 26-33]. CPX-351 is a liposomal encapsulation of daunorubicin and cytarabine in the 1:5 molar ratio [34, 35]. The molar ratio of chemotherapy agents delivered to cancer cells is important in enhancing apoptosis and killing of cancer cells. The molar ratio of 1:5 for daunorubicin to cytarabine was shown to be most effective at treating AML and the liposomal encapsulation maintains the ratio in plasma for 24 hours after injection [34]. In a phase III trial, CPX-351 was compared to standard 7+3 regimen in newly diagnosed secondary AML patients 60-75 years of age [36]. The study included patients with t- AML, AML with a history of myelodysplastic syndrome (MDS) with and without prior hypomethylating agents (HMA), AML with a history of chronic myelomonocytic leukemia (CMML), de novo AML with MDS-related cytogenetic abnormalities, as defined by the WHO criteria. However, the WHO criteria for AML with MRC does not include AML with multilineage dysplasia when a mutation of NPM1, biallelic mutation of CEBPA, or del(9q) cytogenetic abnormality is present [1]. In the study, OS was significantly increased with CPX-351 compared to 7+3 (9.7 vs 5.9 months, p=0.003). In addition to t-AML, OS in AML with antecedent MDS or CMML was better with CPX-351 than 7+3 (7 vs 6 months, HR 0.7 [0.5-0.99]). Within patients with prior MDS, only patients without prior HMA exposure had significantly better OS with CPX-351 than 7+3 (16 vs 5 months, HR 0.46 [0.21-0.97]); OS was similar in both arms among patients with prior HMA exposure (6 vs. 7 months, HR 0.9 [0.64-1.51]). There are no head to head comparisons of 7+3 regimen with CPX-351 in patients younger than 60 years, however, the FDA approval and the NCCN guidelines include CPX 351 as an option for younger patients.

Patients with CD33+ t-AML may benefit from addition of gemtuzumab ozogamicin (GO) to the induction therapy[37, 38]. GO is associated with sinusoidal obstruction syndrome (venoocclusive disease), thus, an interval of 2-3 months between last GO dose and HCT has been suggested [63, 65, 66]. This may limit GO only to induction therapy [5]. However, the addition of GO to CPX 351 in contrast to 7+3, has not been studied. GO as a single agent can be used in patients unable to tolerate intensive chemotherapy [39]. In patients with FLT3 mutation, which is present in 19% of s-AML, midostaurin may be used in combination with 7+3 induction and cytarabine consolidation [27, 40, 41].

Unfit patients or those with multimorbidity may not be able to tolerate intensive chemotherapy. Functional status may be measured with performance status or a geriatric assessment that allows more comprehensive evaluation and can help guide supportive care interventions such as physical therapy, nutrition, or treatment of underlying depression [7]. While the available data to guide treatment of unfit patients or those with multimorbidity are limited, low intensity chemotherapy can limit the risk of toxicities while providing an opportunity to achieve remission. In a single center retrospective study of s-AML patients, OS (6.9 months vs 5.4 months, p=0.04) was better with low intensity regimens (HMA and low dose cytarabine [LDAC]) than intensive chemotherapy (7+3, and CPX-351) among older adults deemed unfit for HCT [14]. There was no significant difference between lower intensity treatment groups, but there was a trend towards improved OS with HMA over intensive therapy (6.7 vs 5.2, p=0.05) or investigational agents (7.7 vs 4.6, p=0. 6). However, no difference in OS was detected with lower intensity therapy compared to CPX-351 (p=0.7). A subgroup analysis of OS for t-AML was similar with OS of 6 vs 5 months respectively for lower intensity vs intensive chemotherapy (p=0.92).

Venetoclax with HMA or LDAC, based on subgroup analysis of two early phase trials, has shown to be effective in s-AML [42, 43] In s-AML patients, OS for venetoclax and HMA was not reached (NR) at the time of analysis (14.6 months-NR), and OS for venetoclax and LDAC was 4 months (3-6.5). In a phase II trial, Glasdegib, a hedgehog pathway inhibitor, in combination with LDAC, resulted in CR rate of 18%, and median OS of 8 months [44]. The trial included 116 newly diagnosed AML patients: 78 patients got glasdegib with LDAC, out of which 51% had s-AML. Based on the results of this trial, glasdegib plus LDAC is approved by FDA for newly diagnosed adults ≥75 years or with comorbidities who cannot undergo intensive induction therapy[45]. In a phase Ib trial, glasdegib plus LDAC, decitabine, or 7+3 resulted in CR/ CR with incomplete hematologic recovery (CRi) of 9%, 29% and 54%, and median OS 4, 11 and 35 months, respectively; 38% out of total 52 patients in the study had s-AML [46].

Therapy-related AML with favorable risk features

Patients with t-AML with favorable cytogenetics, albeit uncommon, may be treated in an approach similar to that used for de novo AML. Intensive induction followed by consolidation chemotherapy without routine HCT in the presence of CR and especially minimal residual disease negativity may be appropriate [47]. The recommendations are based on the fact that response rates and outcomes have been comparable in t-AML and de novo AML with favorable cytogenetics [9, 12, 13, 48].

Intensive therapy

In younger or fit older adults with favorable -risk t-AML, the goal of treatment is to achieve complete remission with intensive chemotherapy followed by consolidation chemotherapy In some studies, median OS of more than 2 years has been seen after intensive chemotherapy with 7+3 regimen; most of survivors had favorable cytogenetics [12, 48]. In one study, 7+3 induction therapy led to 5-year OS of 47-54% in t-AML patients with favorable cytogenetics [13].

CPX-351 is considered the new preferred first line intensive induction therapy for t-AML or MDS-related changes (MRC) with category 1 recommendations from the National Comprehensive Cancer Network (NCCN) [36, 49]. In the CPX-351 study, 5% of patients in CPX arm had favorable cytogenetics compared to 4% in 7+3 arm. For favorable or intermediate risk group, CR/CRi was comparable for both CPX-351 and 7+3 (55% vs 45%, odds ratio (OR) of 1.71 [0.29-10.30]); median OS, however, was 15 months for CPX-351 vs 8 months for 7+3 (HR 0.64, 0.41-0.99). GO has been shown to improve event free survival (EFS) particularly in patients with favorable risk cytogenetics [38, 39, 50]. Induction chemotherapy in FLT3 mutated t-AML patients should include midostaurin. FLAG (fludarabine, high dose cytarabine, and granulocyte colony stimulating factor [G-CSF]) has also been suggested as induction chemotherapy, but the data on favorable cytogenetics is limited [51].

For favorable cytogenetics, consolidation chemotherapy follows intensive induction chemotherapy. If used as an induction chemotherapy, CPX-351 is continued for consolidation after the remission [36]. Intermediate to high dose cytarabine is a common post-remission option following the use of 7+3 induction [52].

Low intensity therapy

Venetoclax with HMA or LDAC, shown to be effective in s-AML in two early phase trials, can used for patients unfit for intensive chemotherapy [42, 43]. An ongoing phase II trial is evaluating a combination of 10-day decitabine and venetoclax (NCT03404193) in newly diagnosed AML, s-AML and R/R AML. Interim analysis showed CR/CRi of 71% and MRD negativity in 40% of 7 s-AML patients; 21% had favorable cytogenetics [53]. Venetoclax in combination with azacitidine, decitabine or LDAC was approved by the US food and drug administration (FDA) in 2018 for the treatment of newly-diagnosed AML in adults who are ≥75 years, or who have comorbidities that preclude use of intensive induction chemotherapy [54].

A single agent GO can be an option in patients unable to undergo intensive induction therapy [39]. A phase III trial compared GO with best supportive care (BSC) in 237 patients >60 years of age and ineligible for intensive chemotherapy. The median OS for GO and BSC were 4.9 and 3.6 months respectively (p=0.005) [39]. Thirty-percent of total patients had s-AML. Fifty percent of patients had favorable/intermediate cytogenetics, who derived significant benefit with GO compared with BSC (HR, 0.52; 95% CI, 0.34 to 0.77). Glasdegib may be considered in patients with favorable cytogenetics in combination with LDAC [44, 46]. Glasdegib has been used in LDAC in phase Ib and phase II trials with remission in as many as 19% of patients with favorable or intermediate cytogenetics [44, 46].

Low intensity chemotherapy is usually continued until progression in patients who achieve remission. For patients unfit for intensive chemotherapy, venetoclax with azacitidine or decitabine would be the preferred choice; however, the data is based on phase Ib trial and the results of a phase III trial are awaited.

Intermediate risk AML

For t-AML patients with intermediate-risk cytogenetics, intensive or low-intensity chemotherapy followed by HCT is recommended. HCT may offer longer OS in these patients than consolidation chemotherapy [55].

Intensive therapy

In the phase III trial comparing CPX-351 versus 7+3, intermediate risk AML comprised of 45% and 40% of patients, respectively [36]. Results showed better CR/CRi and OS for intermediate and favorable risk AML with CPX-351 compared to 7+3. GO with intensive chemotherapy can be an option in CD33+ patients, although benefits are less in patients with intermediate risk AML, and only de novo AML were included in the pivotal ALFA 0701 trial [37, 38]. Midostaurin is recommended in patients with FLT3 mutation.

FLAG regimen has been suggested in s-AML with intermediate risk cytogenetics, for possibly less toxicities than other intensive chemotherapy options. A retrospective study analyzed the benefits of using FLAG over standard 7+3 regimen in s-AML [51]. Intermediate risk AML comprised of 35% of patients. Thirty percent of total 106 patients had t-AML; t-AML was present in 48% of 40 patients in the FLAG group. FLAG had better overall response rate (70% vs 48%, p=0.04), but HCT (33% vs 15%, p=0.052), and 5-year OS (22% vs 6%, p=0.054) were similar. Treatment-related toxicity was less with FLAG with lower 30-day mortality than 7+3 (3% vs 8%, p=0.7). Other studies have reported CR rates of 58-75% and HCT in 20-29% of total patients with FLAG [56, 57]. In these studies, about 47-67% of patients had intermediate risk AML. Sixteen percent of patients had significant treatment related toxicities [56, 57]. However, FLAG should be avoided in severe renal disease due to accumulation of neurotoxic metabolite.

For post-remission consolidation, HCT is recommended. In patients who are unfit for HCT or do not have a suitable donor, consolidation chemotherapy should follow the induction chemotherapy.

Low intensity therapy

Dombret et al. conducted a phase III trial, in which azacitidine was compared with physician-selected conventional care regimens (standard induction therapy, LDAC, or supportive care) in patients ineligible for intensive chemotherapy [58]. Patients with intermediate or unfavorable cytogenetics were included; 11% had s-AML and median OS was better with azacitidine compared to other regimens (10 vs 6 months, p=0.1). A phase II study used 10-day decitabine and reported CR of 47% and overall response rate of 74% among patients with s-AML or t-AML [59]. CR was 52% for patients with intermediate risk cytogenetics. Median OS of all subjects was 55 weeks. Decitabine for 5 days has been used in 55 AML patients >60 years of age: 42% had s-AML, and 53% had intermediate cytogenetics. CR was achieved in 26% and 21% of patients with s-AML and intermediate cytogenetics, respectively [60]. In a phase Ib trial with venetoclax and HMA, CR/CRi rates for patients with intermediate cytogenetics was 74%, and median OS was NR; 51% of total patients had s-AML [42]. Fourteen percent of patients underwent HCT after remission. In a phase I/IIb trial with 82 patients, out of which 35% had s-AML, venetoclax with LDAC induced CR/CRi in 31% of patients with intermediate cytogenetics, compared to 11% in poor risk cytogenetics; median OS was 16 months vs 5 months [43]. Glasdegib has been used with decitabine or LDAC in few s-AML patients with intermediate-risk cytogenetics [44, 46]. In a phase Ib study with glasdegib, 10 patients with s-AML had intermediate risk cytogenetics- 1 patient had partial response with incomplete blood count recovery, 2 patients had stable disease, 3 patients had treatment failure, 4 patients were removed from protocol due to adverse effects [46].

We recommend hypomethylating agents with or without venetoclax in patients with intermediate cytogenetics, followed by HCT. Patients achieving remission after low intensity therapy, if ineligible for HCT, should continue the low intensity regimen until progression.

Unfavorable-risk AML

Clinical trials are suggested as the best option for t-AML with unfavorable-risk cytogenetics given their poor prognosis [9]. Outside of a clinical trial, induction chemotherapy followed by HCT may be used in fit transplant-eligible patient.

Intensive therapy

Induction chemotherapy with CPX-351 is preferred given its superiority over 7+3. However, remission rates and OS decline considerably with unfavorable cytogenetics and older age, despite intensive chemotherapy [6, 9, 10, 36, 56]. In the CPX-351 study, more than 50% of patients had unfavorable risk AML [36]. In a subset analysis of unfavorable risk AML, CR/CRi was 43% with CPX-351 and 22% with 7+3 (Odds ratio, OR 2.79 [1.34-5.82]), median OS was 7 vs. 5 months. Two retrospective studies reported use of FLAG as an induction therapy in s-AML patients; >50% of total patients had unfavorable risk cytogenetics, and median OS was 9 months [51, 56].

Low intensity therapy

Low intensity chemotherapy or best supportive care may be offered in patients unfit for aggressive therapy. A phase II trial with 10-day decitabine reported CR of 50% for unfavorable cytogenetics [59]. In patients with AML and MDS with mutations in TP53 or unfavorable-risk cytogenetics, 10-day decitabine demonstrated a response rate of 67% [61]. CR rate of 24% for unfavorable-risk cytogenetics, similar to CR rate of 26% for patients with s-AML, was reported in a phase II study using decitabine for 5 days [60]. In a study of venetoclax plus HMA, 25% of total patients had s-AML, and 49% of patients had unfavorable cytogenetics. For patients with unfavorable cytogenetics, CR/CRi was 60% and OS of 9.6 months [42]. Similarly, CR/CRi of 42% was reported for unfavorable risk AML treated with venetoclax and LDAC [43]. In a phase II study, glasdegib, in combination with LDAC induced CR rate in 14% of patients with unfavorable cytogenetics [44]. In a phase Ib study with glasdegib., 6 patients had s-AML with unfavorable cytogenetics: 2 patients had treatment failure, 1 patient had indeterminate response, and, 3 patients were removed from the study due to adverse effects [46].

Hematopoietic cell transplant

HCT is regarded as the only curative method for t-AML, and OS is better with HCT compared to consolidation chemotherapy [55]. In a study by Chang et al, HCT had similar OS in patients with secondary MDS or t-AML compared to those with de novo AML [62, 63]. OS after HCT for t-AML has been variable, ranging from 2-year OS of 27% to 10-year OS of 24% (Table 3).

Table 3:

Outcomes after hematopoietic cell transplant in t-AML patients

	Total patients	Age (median, range)	Overall survival (%)	Non-relapse or treatment- related mortality (%)	Relapse (%)
Sengsayadeth [64]	4997	58 (IQR, 50-64)	2-yr survival: 45	2-yr: 27	2-yr: 34%
Alam [16]	65	53 (19-69)	2-yr: 34*	2-yr: 31	2-yr: patients <60 years: 30, >60 years: 36
Litzow [26]†	868	40 (4-72)	1-yr: 32, 5-yr 21	1-yr: 41	1-yr: 27, 5-yr: 31
Yakoub-Agha [65]	70		2-yr: 30	2-yr: 49	2-yr: 42
Kroger [66]†	461	40 (3-69)	3-yr: 35	3-yr: 37	3-yr: 31
Finke [88]	79	58 (20-76)	5-yr: 38, 10-yr 24%	5-yr: 23, 10-yr 32	5-yr: 42, 10-yr: 44

^*Favorable, intermediate, and unfavorable cytogenetics had 2-year overall survival of 53%, 44%, and 21% respectively.

^†includes both t-AML and t-MDS

CR- complete remission, IQR- Inter-quartile range, t-AML- therapy related acute myeloid leukemia, t-MDS- therapy related myelodysplastic syndrome

In the CPX-351 study, 30% of patients underwent HCT after induction and median OS was after HCT was not reached in 30 months [36]. Survival after HCT was better with CPX-351 than with 7+3 (Hazard ratio [HR], 0.46 [0.24-0.89]). Among those patients who underwent HCT, 21% had t-AML in CPX-351 arm versus 23% with 7+3 arm.

In a study by Litzow et al, age >35 years, unfavorable risk cytogenetics, t-AML not in remission or advanced t-MDS, and donor other than an HLA-identical sibling or a partially or well-matched donor were associated with worse prognosis [26]. OS for patients with none of the aforementioned risk factors was 50% compared to 4% for those with all 4 risk factors. Other factors such as the presence of TP53 mutation, functional status, and cytomegalovirus positivity in recipient also affect the outcomes with HCT [26, 28, 64-66]. TP53 mutation can be an important tool to predict outcomes after HCT. Middeke et al. reported 3-year OS of 10% after HCT in patients with TP53 mutation and unfavorable cytogenetics [67].

Therapy-related AML with TP53 mutation

TP53 mutations represent a specific cohort with adverse prognosis in AML with lower frequency of remission, higher chances of relapse, and worse OS [68-71]. AML with TP53 mutations are commonly associated with complex karyotype and other high risk features and may be managed in a way similar to those with unfavorable-risk t-AML outside of clinical trials [72]. Clinical trials are preferred in all cases, if available. Options outside of clinical trials include intensive chemotherapy with CPX-351 and subsequent HCT in younger and fit patients.

In patients unable to tolerate intensive chemotherapy, options include HMAs or HMA or LDAC in combination with venetoclax. In one study, 10-day decitabine induced better response in AML patients with TP53 mutations than those with wild-type TP53; OS in TP53 mutant patients was similar to those with intermediate-risk cytogenetics [61]. Currently, decitabine as a single agent or in combination with other drugs including arsenic trioxide and cytarabine is being studied specifically in AML patients with TP53 mutations (NCT03063203, NCT03855371, and NCT03381781). Several novel agents to target mutant TP53 have been developed and are being tested in multiple pre-clinical and clinical studies. APR-246 is currently on phase Ib/II and phase II clinical trials with azacitidine for TP53 mutated AML patients, either newly diagnosed AML patients (NCT03072043) or those after HCT (NCT03931291). Other agents in development such as PK7088, PK11007, ZMC1, and different peptides may be important therapeutic options in future for t-AML with TP53 mutation.

HCT remains the only option with chances of long-term survival in TP53 mutated AML, although the remission rates are low with higher probability of early relapse after HCT [73, 74]. Patients with TP53 mutated AML with low HCT comorbidity index, better performance status, AML in first or second CR before HCT, and absence of complex karyotype have better outcomes with HCT [74, 75]. Close monitoring for sign of relapse and prompt treatment with donor lymphocyte infiltration or tapering of immunosuppressants has been suggested in TP53 mutated AML after HCT [73]. Patients should also be considered for clinical trials of post-transplant maintenance.

Therapy-related acute promyelocytic leukemia

Literature on therapy-related acute promyelocytic leukemia (t-APL) is limited, but t-APL has been reported to have comparable outcomes compared to de novo APL [76-78]. A large database analysis comparing secondary APL to de novo APL reported similar rates of 1-mortality (29% vs 23%, p=0.2) and 5-year OS (42% vs 50% p=0.2) [78]. Low risk patients (white blood cell count ≤10,000/mcL) are treated with all-trans-retinoic acid (ATRA) in combination with arsenic trioxide (ATO) [79]. Options for high risk patients (white blood cell count >10,000/mcL) include ATRA and ATO in combination with idarubicin or GO. GO instead of idarubicin is an option in t-APL with cardiac issues [79].

Novel agents and ongoing studies

Multiple novel agents are being studied in various clinical trials in combination with other drugs (Table 4). Pracinostat was studied in a phase 2 trial with azacitidine; 50 patients aged ≤65 years with untreated de novo or s-AML, intermediate or unfavorable cytogenetics and not eligible for intensive therapy were included [80]. Results showed composite complete response rate (CR+ CRi+ MLFS [morphologic leukemia-free response]) of 52%. Median OS at 21 months was not reached for s-AML. A phase Ib study evaluated pevonedistat with azacitidine in newly diagnosed AML patients ≥60 years and ineligible for intensive induction chemotherapy [81]. In a total of 64 patients, 44% had s-AML, 50% had intermediate cytogenetics, and 28% had unfavorable cytogenetics. CR/CRi/partial response was 50% without any difference in de novo vs s-AML. Median OS was 7 months. Combination of venetoclax, azacitidine and pevonedistat is currently being studied in a phase I/II trial (NCT03862157); the combination was shown to be more effective than single agent or combination of two agents in preclinical models of AML [82]. In a phase Ib trial, Hu5F9-G4, an antibody targeting CD47 which is a key macrophage checkpoint, has shown to be well tolerated and active in AML/MDS patients; the trial is currently ongoing expansion cohorts (NCT03248479) [83]. Another HMA, guadecitabine at a dose of 60-90 mg/m2 was used in a phase I/II trial with untreated, older AML patients; 37% had s-AML [84]. Intermediate and unfavorable cytogenetics were present in 53% and 42% patients respectively. Results showed CR/CRi of 53% with median OS of 10 months. There was no difference in CR/CRi and median OS among s-AML or high-risk AML. Nivolumab, in combination with idarubicin and cytarabine, has shown promising results in patients with s-AML and intermediate or unfavorable risk cytogenetics. A phase II trial with nivolumab in newly diagnosed AML or high risk MDS patients reported CR/CRi of 77%, minimal residual disease (MRD) in 53%, use of HCT in 41%, and median OS of 18 months [85]. In a phase I/II study, uproleselan, an E-selectin antagonist, was added to 7+3 for older untreated AML patients [86]. Fifty-two percent of patients had s-AML and 48% had unfavorable cytogenetics. CR/CRi rate for s-AML was 69% with median OS of 10 months

Table 4:

Ongoing trials for newly diagnosed AML patients, including therapy-related AML

Drug	Mechanism of action	Phase	Study arms	Patient population*	Clinical trial number
Pracinostat	HDAC inhibitor	III	Azacitidine with or without pracinostat	Untreated AML	NCT03151408
Etinostat	HDAC inhibitor	II	Etinostat + Azacitidine	Untreated, older patients with AML	NCT01305499
Pevonedistat	NEDD8-activating enzyme inhibitor	III	Pevonedistat with azacitidine or azacitidine only	Untreated, low-blast AML, CMML, high-risk MDS	NCT03268954
		I/II	Venetoclax with Pevonedistat and azacitidine	Newly diagnosed AML	NCT03862157
Nivolumab	PD-1 inhibitor	II/III	Azacitidine vs Azacitidine + Nivolumab vs Azacitidine + Midostaurin vs Decitabine + Cytarabine	Untreated older patients with AML	NCT03092674
Vosaroxin	Topoisomeras e II inhibitor	II	Vosaroxin + Cytarabine	Untreated AML	NCT02658487
Ruxolitinib	JAK2 inhibitor	I/II	Ruxolitinib + decitabine	Untreated, accelerated MPN or post-MPN s-AML	NCT02076191
Volasertib	Polo-like kinase inhibitor	III	LDAC with or without volasertib	Untreated, older, unfit patients with AML	NCT01721876
Tosedostat	Aminopeptidase inhibitor	II	Tosedostat with HMA or cytarabine	Untreated AML	NCT01567059
Veliparib	PARP inhibitor	II	Topotecan and carboplatin with or without veliparib or placebo	Untreated AML, including post-MPN AML	NCT03289910
Uproleselan	E-selectin antagonist	III	7+3 with or without uproleselan	Untreated, older patients with AML	NCT03701308
APR-246	Mutant TP53 activator	Ib/II	APR-246 with azacitidine	TP53 mutated myeloid neoplasms	NCT03072043
		II	APR-246 with azacitidine	TP53 mutated AML or MDS after HCT	NCT03931291
Hu5F9-G4	Anti-CD47 monoclonal antibody	I	Hu5F9 vs Hu5F9 with azacitidine	Hematological malignancies	NCT03248479

^*Patient population indicates AML unless otherwise stated, unfit denotes patient ineligible for intensive induction chemotherapy; untreated denotes patient not previous treated for AML

^†Interim analysis of the trial with 48 patients reported overall CR/CRi rate of 71%. CR/CRi rate for newly diagnosed, s-AML and R/R AML were 92%, 71% and 44%, respectively> MRD negativity was achieved in 49% of patients – 52% in newly diagnosed, 40% in s-AML and 50% in R/R AML. CR/CRi rate in TP53 mutated pts was 67% [53].

CCR included BSC only, azacitidine + BSC, LDAC + BSC, or IDAC + BSC.

AML- acute myeloid leukemia; BCL-2: B-cell leukemia/lymphoma-2; BSC- best supportive care; CCR- complete clinical response; FLAG-Ida: Fludarabine, Cytarabine, granulocyte colony stimulating factor-Idarubicin; FLAI- Fludarabine, Cytarabine, Idarubicin; FLAM- Flavopiridol, cytarabine, mitoxantrone; HCT- hematopoietic cell transplant; HDAC- histone deacetylase; HMA- hypomethylating agent; IDAC- intermediate dose cytarabine; LDAC- low dose cytarabine; MDS- myelodysplastic syndrome; MPN- myeloproliferative neoplasm; ORR- overall response rate; OS- overall survival; R/R- relapsed/refractory; s-AML- secondary acute myeloid leukemia.

Conclusion and recommendations:

Management of t-AML is challenging with limited prospective evidence. A personalized treatment plan should be based on various patient-specific, disease-specific and therapy-specific factors.

We recommend clinical trials for patients with t-AML, particularly unfit patients and those with unfavorable cytogenetics or mutations. For fit patients, CPX-351 represents the new standard. HCT is recommended for intermediate or unfavorable t-AML or patients with good risk AML with minimal residual disease. Midostaurin and GO are added for FLT3 mutation and CD33 expression, respectively for induction and consolidation as appropriate. Venetoclax with either decitabine or azacitidine may be the preferred low intensity therapy in patients who are unfit for intensive chemotherapy. Venetoclax with low dose cytarabine may be an alternative. If venetoclax combination cannot be used for any reason, 10-day decitabine may be utilized, especially in patients with TP53 mutation. Ongoing clinical trials focusing on treatment of t-AML, including targeted agents and immunotherapy, look promising for future.