Overall, the clinical management of R/R AML patients requires a therapeutic approach capable of providing rapid achievement of a high-quality response, together with a low burden of toxicity, as a bridge to HSCT. That clearly pertains to patients eligible for curative intent. Needless to say, the prognosis of elderly or unfit R/R patients is even more disappointing, so the approach to this patient subset is generally focused on different aims, such as disease control, transfusion dependency reduction, and improvement of quality of life.
2. From Preclinical Evidence to Early Clinical Experiences with Venetoclax in AML
The B-cell lymphoma 2 (BCL-2) family of proteins regulates mitochondrial outer-membrane permeabilization (MOMP), the crucial event in the regulation of the apoptotic process
[10]. BCL-2 antagonizes the activation of apoptosis by sequestering proapoptotic family members required for MOMP. A relevant role of BCL-2 has been implicated in the survival of AML cells. This has supported the development of oral BCL-2 inhibitors in diverse clinical contexts.
Venetoclax is a small molecule that acts as a potent, selective inhibitor of BCL-2. It is a first-in-class BH3 mimetic that acts by mimicking the BH3 domain of all proapoptotic BCL-2 family proteins. The drug displaces BCL-2 from binding to a set of proteins, specifically the pore-forming proteins BAX and BAK, thereby making them capable of permeabilizing the mitochondrial membrane. The dependence of the cell on the effect of VEN relies on the amount of BCL-2 that sequesters proapoptotic proteins, a sort of functional state of BCL-2, otherwise called “primed”, the entity of which is predictive of the ability of VEN to induce MOMP
[11].
In 2006, Konopleva et al.
[12] provided compelling preclinical evidence that BCL-2 inhibition may yield meaningful therapeutic potential in AML. By exposing AML cell lines and primary blasts from R/R AML patients to the BCL-2/BCL-X
L/BCL-W inhibitor ABT-737, they were able to demonstrate the disruption of BCL-2/BAX association, resulting in the freeing up of BAX and BIM, which, in turn, leads to cytochrome c release, caspase activation, and finalization of the apoptotic cascade. The antileukemic potential of ABT-737 was subsequently confirmed in murine models of AML
[13].
In 2014, Pan et al.
[14] reported on the efficacy of VEN (ABT-199) on multiple AML cell lines and patient-derived xenografts at nanomolar concentrations. Interestingly, the cytotoxicity of VEN was largely independent of mutational status and was maintained on chemorefractory primary AML blasts.
A phase II study assessed the clinical activity of VEN monotherapy (800 mg daily) in 32 R/R AML patients
[8]. The study population consisted of heavily pretreated patients with a median age of 71 years (range, 19–84) and was significantly enriched with high-risk cytogenetic/molecular features. While safety signals were encouraging, the single-agent activity of VEN was found to be modest, with 6/32 patients meeting the criteria for complete remission (CR) or CR with incomplete count recovery (CRi). Additionally, a disease burden reduction not qualifying for CR was observed in 19% of patients. Responses were mostly short-lived, with a median progression-free survival (PFS) of 2.5 months. The study also provided preliminary insights on possible biomarkers for VEN sensitivity. Specifically, the presence of
IDH1/
IDH2 mutations and the lack of blast dependence on BCL-X
L and MCL-1 were associated with response to VEN, whereas the presence of mutations in RAS pathway genes, particularly
FLT3-ITD and
PTPN11, correlated with treatment failure.
Venetoclax has been shown to synergize with a number of agents already in use in AML, including hypomethylating agents (HMAs) and cytotoxic chemotherapy
[15][16][17][18][19][15,16,17,18,19], providing the rationale for a combinational strategy.
Driven by the promising efficacy data on heavily pretreated patients and the relatively favorable toxicity profile, the clinical development of VEN in AML rapidly shifted towards the frontline setting for elderly/unfit patients, where therapeutic standards
[20][21][22][20,21,22] (HMAs and low-dose cytarabine (LDAC)) had been providing unsatisfactory outcomes in clinical practice
[23][24][25][26][27][23,24,25,26,27]. Moreover, the possibility of increasing treatment efficacy without excessive toxicity with an orally available drug was valuable. Clinical trials with VEN in combination with HMAs or LDAC in elderly/unfit patients in the frontline setting demonstrated striking efficacy with unprecedented overall response rates (ORRs) and prolonged overall survival (OS) and exhibited an overall manageable toxicity profile
[9][28][29][9,28,29], marking a paradigm shift in the therapeutic approach to this patient population.
3. Venetoclax in R/R AML
Although VEN-based regimens have been gaining popularity in the context of salvage therapy, their positioning in the setting of R/R AML is far less clear. The clinical development of VEN in this setting has suffered critically from the lack of interventional, multicenter clinical trials, not to mention the absence of randomized studies comparing VEN-based regimens to conventional strategies. Nonetheless, as the number of patients treated with VEN in the frontline setting in combination with HMAs or cytotoxic chemotherapy steadily increases, its role in the context of salvage regimens is bound to become more problematic.
3.1. Venetoclax + HMAs or LDAC for R/R AML
Preclinical evidence has highlighted the synergistic effect of combining BCL-2 inhibitors with HMAs and LDAC
[15][16][17][18][15,16,17,18]. The rationale for exploring the combination of VEN with such lower-intensity treatments in R/R AML lies in the possibility of achieving clinical efficacy with reasonable toxicity. This concept is of utmost relevance in the context of R/R disease, where patients have already been exposed to the toxicity of previous chemotherapy and might develop significant end-organ damage if exposed to further intensive treatment. In the case of younger/adult patients with R/R AML, the possibility of achieving a meaningful clinical response with limited toxicity is expected to translate into a higher rate of successful HSCT transition and, possibly, lower rates of transplant-related mortality (TRM). For elderly/unfit patients, who are not usually candidates for HSCT with curative intent, VEN in combination with lower-intensity regimens could provide an accessible platform aimed at achieving a survival benefit while preserving the quality of life. Finally, the possibility of disease control with limited toxicity could be beneficial for patients relapsing after HSCT, who are often not candidates for intensive reinduction approaches due to the persistence of transplant-related toxicities. In this setting, VEN-based lower-intensity approaches could interact positively with donor lymphocyte infusions (DLIs) or provide a bridging platform for a second HSCT.
Initial experiences with VEN + HMAs/LDAC were mainly derived from retrospective case series including heavily pretreated patients with relatively advanced median age enriched in high-risk cytogenetic and molecular features and treated outside of clinical trials. In 2017, DiNardo et al.
[30] provided a retrospective analysis of 43 patients treated off-protocol with VEN in combination with HMAs or LDAC for R/R AML (n = 39), myelodysplastic syndrome (MDS) (n = 2), or blastic plasmacytoid dendritic cell neoplasm (BPDCN) (n = 2). The median age was 68 years (range, 25–83). In the AML cohort, more than 30% of patients had received treatment for an antecedent hematological neoplasm. Most patients (84%) were treated in a second or further salvage setting, and 77% of patients had received previous HMA therapy. Almost 50% of patients had adverse cytogenetic abnormalities, and high-risk molecular features (mutations in
TP53,
RUNX1, and
ASXL1) were highly represented. The overall response rate was 21% (CR + CRi, 12%; morphologic leukemia-free state (MLFS), 9%). The median OS was 3 months for the whole cohort and 4.8 months for responding patients. As previously reported
[8],
IDH1/2 and
RUNX1 mutations appeared to confer a higher likelihood of response (ORR, 27% and 50%, respectively). The main adverse events (AEs) were cytopenias (mainly grade 3–4 neutropenia; 100%) and infections (mainly pneumoniae; 40%). It must be noted that 47% of patients were already receiving intravenous antimicrobials for active infections at the time of treatment initiation, not surprisingly for a heavily treatment-experienced patient cohort. These results have fostered subsequent explorations of VEN + HMAs/LDAC in R/R AML worldwide.
Aldoss et al.
[31] reported on the outcomes of 33 consecutive adult patients (median age, 62 years; range, 19–81) with R/R AML treated with VEN + HMAs at City of Hope Medical Center between 2016 and 2017. Previous HMA treatment was reported in 60.6% of patients, while 39.4% had previously received HSCT. High-risk genetic features were reported in 54.5% of patients. Most patients (31/33) received VEN in combination with decitabine (DAC), 51.6% of them with a longer 10-day schedule. The median number of cycles was 2 (range, 1–10), with a relatively short median follow-up of 6.5 months. An ORR of 64% was observed (CR, 30%; CRi, 21%; MLFS, 9%). Consistent with the previous report by DiNardo et al.
[30], the best response was observed after a median of two cycles. Interestingly, 53% of patients with available MRD evaluation were MRD-negative, suggesting that VEN + HMA combination may yield deep responses in a subset of patients with R/R AML. A ten-day DAC schedule was not associated with improved outcomes. While previous HSCT and HMA exposure did not negatively affect the response rate, de novo AML, the absence of high-risk cytogenetics, and the presence of
IDH1/2 mutations were associated with improved ORR. Interestingly, 67% of
TP53-mutated and 44% of
FLT3-mutated (TKD or ITD) patients achieved an objective response. One-year OS was 53% (73% for de novo AML patients), and median disease-free survival (DFS) was 8.9 months. Safety signals were in line with previous experiences, and most serious AEs were deemed unrelated to therapy.
These data were confirmed in a subsequent extension
[32] of the original case series (n = 90; median age, 59 years; range, 18–81), with additional insights into the correlation between molecular features and response to therapy. In this larger cohort (previous HMA exposure was reported in 51% of patients), a 46% ORR was reported. Despite the limitations of the small sample size for individual genetic groups, the authors reported objective responses across all genetic subtypes. In multivariate analysis, European Leukemia Net (ELN) genetic risk was associated with differential ORR, while the presence of either
ASXL1 or
TET2 mutations was associated with better CR/CRi. In univariate analysis of mutations and genetic functional pathways, the presence of
TP53 mutations (
p = 0.049) and alterations in chromatin-modifying genes (
p = 0.002) adversely influenced the OS. However, in multivariate analysis, neither had an independent impact on OS.
A 10-day DAC schedule was also explored by DiNardo et al.
[33] in a phase II trial conducted at the MD Anderson Cancer Center (MDACC) including 55 AML patients R/R after a median of 2 previous lines (range, 1–3); the median age was 62 (range, 43–73). A very promising ORR of 62% was reported, with 54% of MRD-negative responses. However, 18% of patients received concomitant genotype-targeted agents (FLT3 inhibitors, enasidenib, and ponatinib), which may have influenced ORR favorably.
The Mayo Clinic experience with VEN + HMAs in AML was reported by Morsia et al.
[34] in 2020. This retrospective analysis included 42 patients with R/R AML (35.7% were HMA-experienced). Sixty-six percent of patients were classified into the ELN high-risk group. A 33% ORR was observed with frequent CRi (14.3%). The median number of cycles to best response was one, and the median duration of response was 2.0 months. In keeping with the City of Hope Medical Center experience
[31], responses were observed across all genetic groups, but the sample size for individual genetic subsets was too small to draw conclusions. Eight patients (19%) were able to proceed to HSCT after VEN-based treatment. The median OS was 5 months (15 months for patients achieving CR/CRi; 16 months for patients proceeding to HSCT).
Owing to the lack of randomized comparisons of AZA, DAC, and LDAC as partner drugs in VEN-based regimens, the choice is often driven by clinicians’ preference and convenience, the only exception being the use of LDAC, for which available data from the frontline setting suggest
[28][29][28,29] inferior activity compared to HMAs. Notably, evidence of synergism with VEN through MCL-1 downregulation is available for all three drugs
[17][18][19][35][17,18,19,36].
Stahl et al.
[36][37] reported the clinical outcomes and biological correlates of 86 R/R AML patients treated with VEN in combination with AZA (n = 35), DAC (n = 20) or LDAC (n = 27) at Memorial Sloan Kettering Cancer Center (MSKCC) between August 2016 and February 2021. The median age was 67 years (range, 29–86). Previous exposure to HMA was reported in 57% of patients, who were predictably older and predominantly allocated to VEN + LDAC. Seventeen percent of patients relapsed after HSCT. An ORR of 31% was reported for the whole cohort. Although the use of AZA was associated with significantly higher ORR and survival outcomes compared to DAC and LDAC (ORR, 49% vs. 25% vs. 15%, respectively,
p = 0.02; median OS, 25 vs. 5.4 vs. 3.9 months, respectively,
p = 0.003), it must be stressed that the non-randomized, non-interventional nature of the study might have introduced a fair amount of selection bias regarding the choice of partner drug. In fact, patients receiving DAC had a higher incidence of high-risk genetic features, while patients receiving LDAC were significantly older (median age, 74) and mostly R/R after frontline HMAs (88%), most likely reflecting different patient populations. Interestingly, the number of previous salvage lines (0, 1, or 2) did not seem to exert a significant detrimental effect on ORR for the whole cohort, suggesting a substantial biological difference between the antileukemic activity of VEN + HMAs/LDAC compared to conventional chemotherapy and the ability to overcome, at least in part, traditionally defined chemoresistance. However, failure of ≥3 previous lines of therapy was associated with a lower ORR (
p = 0.04), even in the setting of VEN-based treatment. Median OS and DFS were 6.1 and 7.8 months, respectively. Fifteen patients were transitioned to HSCT after VEN-based treatment, achieving a significant OS benefit.
The effect of previous HMA exposure on the response to VEN + HMAs was assessed by Feld et al.
[37][38] in a retrospective analysis including 44 R/R AML patients (median age, 61 years); 59.1% had previous HMA exposure. A 38.5% CR + CRi rate was reported for the whole population. Interestingly and in contrast to the MSKCC experience
[36][37], previous HMA exposure translated to a lower ORR (exposed, 14.3%; unexposed, 66.7%).
Recently, the PETHEMA group presented the results of a multicentric, retrospective analysis of 51 R/R AML patients (median age, 68 years; range, 25–82) treated with VEN in association with AZA (n = 30), DAC (n = 15) or LDAC (n = 6) in Spain
[38][39]. Previous HMA treatment was reported in 51% of patients, and 61% had received at least two previous lines of therapy. A rather unsatisfactory 12.4% CR + CRi rate was observed. The choice of AZA as partner drug seemed to yield a higher CR rate. The probability of achieving a CR was affected by the presence of
NPM1 mutations and mono- or biallelic
CEBPA mutations.
In conclusion, a definitive assessment of the efficacy of VEN-based lower-intensity regimens in the context of R/R AML is extremely challenging, the main limitations being the retrospective nature of most studies, selection biases affecting the comparability of different study populations, and the frequent inclusion of patients with previous HMA exposure. Although limited clinical evidence suggests the superiority of AZA over DAC and LDAC
[36][38][37,39], the lack of randomized clinical trials must be taken into consideration. The safety profile of VEN + HMAs/LDAC in R/R AML is generally regarded as manageable and overall favorable overall
[28][29][30][31][32][33][34][36][38][39][28,29,30,31,32,33,34,35,37,39] compared to conventional salvage chemotherapy. The risk of tumor lysis syndrome (TLS) in AML during treatment with VEN is low, with most reported cases falling into the “laboratory TLS” definition rather than “clinical TLS”
[40][41][42][40,41,42]. While treatment is often initiated in inpatient facilities, it has been shown that it can be safely conducted in an outpatient setting
[39][42][35,42], where the availability of properly trained home care services provides invaluable support to the patients and their families
[42]. In the case of R/R AML patients, hospitalization is often mandated by unresolved toxicities and profound cytopenias resulting from previous treatments.
Aside from cytopenias, an expected consequence of any form of treatment for R/R AML, infectious complications represent the main category of AE across reports. Patients with R/R AML are generally frail and carry the burden of immunosuppression resulting from previous treatments. As most of the retrospective studies examined so far included mostly heavily pretreated patients with a median age >60, it is not surprising that infectious complications were frequently reported. A definitive evaluation of the risk of infection in R/R AML is challenging, mostly due to the risk of underreporting resulting from the retrospective nature of most studies and different conduits regarding antibacterial prophylaxis, especially in the outpatient setting. In a recent update of
our
esearchers' experience
[43], only 24 febrile neutropenia episodes were observed on a total of 276 recorded VEN + HMA/LDAC cycles.
Invasive fungal infections (IFIs) deserve special consideration. Prolonged neutropenia in the context of induction/salvage chemotherapy has been considered an indication of mold-active antifungal prophylaxis in AML
[44]. Unfortunately, azole antifungals interact with VEN via inhibition of CYP3A4
[45], increasing systemic exposure to VEN. Despite the availability of pharmacokinetic evidence and recommendations
[46][47][48][46,47,48] for VEN dose reduction with concomitant azole administration, clinical conduits regarding antifungal prophylaxis vary greatly among different institutions. Although mainly derived from experiences with VEN + HMA as first-line treatment for AML, currently available data suggest that the rate of IFIs in patients receiving VEN + HMA is generally low (5%)
[49]. However, in a report by Aldoss et al.
[50], a 19% incidence of IFIs was reported among patients receiving VEN + HMA for R/R AML, with both R/R status and refractoriness to VEN + HMA independently associated with an increased risk of IFI.
3.2. Venetoclax + Intensive Chemotherapy for R/R AML
The demonstration of improved clinical outcomes when VEN is added to lower-intensity regimens
[9][28][29][9,28,29] has led to the question of whether its addition to higher-intensity chemotherapy could yield higher ORRs translating into a tangible survival benefit. Clinical trials assessing the efficacy and safety of VEN in combination with intensive chemotherapy (IC) for previously untreated AML patients have reported remarkably high CR rates
[51][52][53][54][51,52,53,54]. However, combining VEN with IC comes with challenges. Specifically, the effect of VEN on the grade and duration of cytopenias appears to be particularly relevant when combined with IC, translating into a higher risk of neutropenia-related infectious complications. For example, in the MDACC trial combining VEN with FLAG-Ida
[52], a protocol amendment reduced VEN exposure duration from 21 to 14 days and cytarabine dose from 2000 to 1500 mg/m
2 following the observation of pronounced grade 3 and 4 neutropenia-related infections (including a case of typhlitis) in the original phase Ib study. Subsequent experiences with similar regimens have led to the adoption of an even shorter 7-day VEN schedule
[51].
In the R/R AML cohorts included in phases Ib (n = 16; median age, 51 years; range, 20–73) and IIB (n = 23; median age, 47 years; range, 22–66) of the aforementioned MDACC trial
[52], the median time to peripheral count recovery after cycle 1 was 37 days and was prolonged across all cohorts following cycle 2 despite the use of G-CSF. Extension of the cycle length over 40 days and dose reductions were eventually required for a large portion of patients proceeding after cycle 1, particularly for secondary and R/R AML patients. The CR + CRi rate for R/R patients was 67% (69% were MRD-negative), and 46% of R/R patients proceeded to consolidative HSCT with survival benefit. Febrile neutropenia and pneumonia were frequently observed (50% and 28%, respectively) and equally affected newly diagnosed and R/R patients, although bacteremia was significantly more frequent in R/R patients (46% vs. 21%,
p = 0.04). Nevertheless, the overall safety profile in R/R AML patients was not significantly different from what is usually expected with conventional salvage regimens
[55][56][57][58][59][60][55,56,57,58,59,60].
Wolach et al.
[61] retrospectively analyzed the outcomes of 24 R/R AML patients (median age, 53.4 years; range, 30.1–72) treated with VEN + FLAG-Ida in Israel in a real-world setting. The median number of previous lines of therapy was one (range, 0–3), and 44% of patients had previously received HSCT. The observed CR + CRi rate was 72% (91% in the post-HSCT group); DFS and OS at 12 months were 67% and 50%, respectively. Thirty-day mortality was 12%, and 48% of patients developed bacteremia. Notably, the authors reported a relatively high (32%) incidence of IFI. Count recovery occurred at a median of 31 days (95% CI, 17.6–38.3) for platelets and 23 days (95% CI, 20–28) for neutrophils.
A retrospective single-center analysis by Shahswar et al.
[62] compared the outcomes of 37 patients receiving FLA-Ida + VEN (100 mg daily for 7 days with concomitant posaconazole) with those of a cohort of 81 patients treated with FLA-Ida without VEN between 2000 and 2018. The two populations were balanced regarding median age (54 vs. 52 years, respectively), genetic features, R/R status, and previous HSCT. Patients in the FLAVIDA group had a significantly higher ORR (78% vs. 47%,
p = 0.001), although without a significant impact on OS. Interestingly, times to count recovery did not differ significantly between the two groups.
A combination of VEN + high-dose cytarabine and mitoxantrone (HAM) was explored in the phase I/II Alliance Leukemia (SAL) Relax trial
[63]. Twelve patients (median age, 57 years; range, 40–70) with relapsed AML were enrolled in the dose escalation part. The combination of VEN 400 mg daily (after a 3-day ramp-up) on days 3 to 14 + HAM was shown to be safe and effective, with 11 out of 12 patients achieving a CR/CRi.
Overall, these data provide an encouraging outlook on the efficacy of VEN in combination with IC in R/R AML. While ORRs are generally impressive, toxicity, particularly myelosuppression, is significant and requires careful consideration. Shorter VEN schedules are warranted in combination with IC, and such salvage regimens should be reserved for younger AML patients under the age of 60.
3.3. Venetoclax as Bridge-to-Transplant and Salvage Approach for Post-HSCT Relapse
Responses to VEN-based regimens in R/R AML are mostly short-lived in the absence of consolidative HSCT. Nonetheless, evaluating the clinical usefulness of VEN-based regimens as bridge-to-transplant platforms is problematic. First, most retrospective case series mainly include very advanced AML patients with previous exposure to multiple treatment lines, including HSCT, with extremely limited clinical options. Secondly, especially in the setting of R/R AML patients treated with VEN-based lower-intensity regimens, previous HMA exposure is a very frequent finding. Although not clearly stated by the authors, it is reasonable to assume that those were predominantly transplant-ineligible, elderly patients and/or high-cytogenetic-risk patients treated with frontline single-agent HMAs as standard clinical practice. All this considered, it is unsurprising that very few patients were reported to be able to proceed to HSCT.
Additionally, assessing the efficacy of a VEN-based bridging strategy would require an explicit statement regarding the number of patients treated with an intention to transplant (ITT), which is very rarely found in the context of retrospective studies. In
reseaour
chers' experience
[43] involving 67 R/R AML patients with a relatively low median age (58 years; range, 33–74), 39 patients (58%) were treated with VEN + HMAs/LDAC with an ITT. Ultimately, the HSCT actualization rate was 66%, and the only reason for failure to proceed with HSCT was refractory disease. This rather high rate of HSCT bridging compares favorably with other experiences, but it must be noted that
our
esearchers' case series included mostly young/adult patients with ELN favorable/intermediate-risk R/R AML and very few post-HMA patients.
Zappasodi et al.
[64] reported the outcomes of 10 R/R AML patients treated with VEN + AZA specifically as a bridge to transplant. The ORR was 60%, and all responding patients were able to transition to HSCT. Infectious complications were observed in 4/10 patients (including a case of sinonasal aspergillosis), all of which were successfully managed. In the MDACC VEN + FLAG-Ida trial
[52], the transplantation rate in the R/R cohort was 46%, with a clear OS benefit for patients receiving consolidative HSCT. However, patients enrolled in the trial were predominantly young/adults and in first salvage. Although no randomized study comparing VEN-based strategies with conventional salvage chemotherapy is currently available, a propensity score-matching analysis by Maiti et al.
[65] examined the outcomes of 65 patients with R/R AML treated with VEN + 10-day DAC compared to 130 IC recipients. Although this study was not powered to specifically address the question of which salvage platform might provide better HSCT bridging and the HSCT rate did not differ significantly between the two groups, overall efficacy outcomes were in favor of VEN + DAC.
Another relevant question regarding VEN-based bridging strategies is whether the addition of IC to VEN
[52][63][66][67][52,63,66,67] can provide additional benefits in terms of efficacy compared to VEN + HMAs. Although reported ORRs tend to be higher for patients treated with VEN combined with IC, no randomized trials are currently available to conclusively address this question, and the patient populations involved are profoundly different in terms of demographic, clinical, and molecular features. While VEN + HMAs is an intuitively favored choice in the case of elderly R/R AML patients, it is unclear to which extent younger patients benefit from further escalation of intensity. This consideration might be particularly relevant from a safety perspective, since the use of IC is expected to be associated with a higher risk of end-organ damage, possibly jeopardizing HSCT eligibility.
The overall favorable safety profile of VEN-based regimens represents an appealing option for patients relapsing after HSCT. However, data from current scientific literature are conflicting. Several retrospective experiences with VEN-based lower-intensity regimens
[36][39][43][35,37,43] have reported lower ORRs in the post-HSCT setting, although this observation might have been biased by the enrichment in high-risk cytogenetic/molecular features in relapsed HSCT recipient cohorts. Conversely, other experiences have reported encouraging results. Aldoss et al.
[31] showed no detrimental effect of previous HSCT exposure on ORR following VEN-based salvage (previous HSCT, 46.2%; no previous HSCT, 55%;
p = 0.73). Byrne et al.
[68] reported the outcomes of 21 AML patients who relapsed after HSCT and were treated with VEN (mainly combined with HMAs). The ORR was 42.1%, and the median OS was 7.8 months, with significantly longer OS in patients achieving CR/CRi (
p = 0.005). Infectious events were relatively frequent, as were grade 4 cytopenias. VEN dose reductions were ultimately applied to all responding patients. Similarly, Joshi et al.
[69] reported a 38% ORR among 29 patients relapsing with AML following HSCT. In responders, the median OS was 403 days, and the median DFS was 259 days. The safety profile was in keeping with previous reports, and shortening of VEN exposure to 21 or 14 days was applied to mitigate the duration of cytopenias.
Zhigarev et al. investigated the immunologic landscape of T cells after exposure to VEN + HMA
[70]. Treatment with HMA/VEN resulted in a greater fraction of T cells with effector memory phenotype, inhibited IFN-γ secretion by CD8+ T cells, upregulated perforin expression in NK cells, downregulated PD-1 and 2B4 expression on CD4+ T cells, and stimulated T-regulatory cell proliferation and CTLA-4 expression. Based on these findings, one could speculate that VEN might exert a beneficial effect in the context of adoptive cellular therapies for patients relapsing after HSCT. The combination of AZA with DLIs has been a commonly adopted salvage/preemptive strategy for post-HSCT relapse in AML
[71], and the role of the addition of VEN in this setting was explored prospectively by Zhao et al.
[72] in 26 AML patients who relapsed after HSCT. An encouraging 61.5% ORR (CR, 26.9%; PR, 34.6%) was observed. Graft-versus-host disease (GVHD) developed in six patients (23.1%), with a median time to GVHD onset of 77 days. Amit et al.
[73] investigated the efficacy of DLI combined with VEN monotherapy for patients with early AML relapse after HSCT (n = 22). Treatment was generally well-tolerated, and the ORR was 50%, with a median duration of response of 135 days. The incidence of acute and chronic GVHD was 18% and 27%, respectively.
Zucenka et al.
[74] retrospectively compared the outcomes of 20 patients receiving VEN in combination with LDAC and D-actinomycin (followed by VEN + DLI maintenance), with 29 patients receiving FLAG-Ida for R/R AML after HSCT. Patients receiving VEN had superior ORR (70% vs. 34%,
p = 0.02) and OS (13.1 vs. 5.1 months,
p = 0.032). Notably, treatment-related mortality was 0% in the VEN group and 34% in the FLAG-Ida group (
p = 0.003), supporting the idea that VEN-based lower-intensity regimens might represent an advantageous clinical option in this patient population.
Taken together, these data support a larger-scale exploration of VEN-based salvage for post-HSCT relapse.
As VEN is becoming increasingly available for prescription across different countries worldwide and the therapeutic offering is broadening both in previously untreated and R/R AML with the introduction of novel molecularly targeted agents, the identification of reliable predictive biomarkers associated with sensitivity is key to rationally driving the allocation of patients to VEN-based therapies and avoiding pointless (and potentially harmful) therapeutic attempts.
Molecular lesions affecting recurrently mutated genes in AML have been reported to modulate VEN sensitivity. However, the role of single-gene lesions is often modulated by concomitant additional mutations and may vary in a context-dependent manner. Unfortunately, with very few exceptions, specific comutational patterns tend to be detected in a relatively low number of patients in most VEN clinical trials and retrospective studies in AML, which are generally not powered to investigate the predictive value of mutations/comutational patterns. Consequently, data regarding the role of single mutations are often inconclusive or even conflicting, especially in the R/R setting. Although responses to VEN-based therapies have been observed across all genetic subgroups in most studies
[9][29][30][32][34][9,29,30,32,34], a restricted number of mutations has been consistently shown to exert a particularly significant favorable impact.
NPM1 mutations occur in 30% of AML patients
[75] and are considered predictive of improved response to induction chemotherapy
[75]. Mutations in
NPM1 have been shown to correlate with higher ORR and survival outcomes in the context of VEN-based therapies in different settings. In an early phase I trial assessing the safety and efficacy of VEN + HMAs, the CR + CRi rate in 23 previously untreated
NPM1-mutated patients was 91.5%
[29]. The reported CR + CRi rate for 27
NPM1-mutated patients enrolled in the VIALE-A trial was 66.7%
[9]. Similarly, promising efficacy data in this molecular setting emerged from the use of VEN + LDAC, with a reported CR + CRi rate of 78%
[28]. Very encouraging CR + CRi rates were reported in clinical trials combining VEN with FLAG-Ida
[53] (100%) or the “2 + 5” regimen (80%)
[54]. Data from real-world experiences convey a similar picture
[76].
Although to a lesser extent compared to the frontline setting, the presence of
NPM1 mutations has been associated with higher ORR and improved survival outcomes in R/R AML patients
[36][39][43][35,37,43]. Tiong et al.
[77] reported the outcomes of 12 patients with
NPM1-mutant AML with MRD persistence/relapse treated with VEN + LDAC. All five patients with molecular persistence achieved complete MRD clearance and durable OS without subsequent HSCT. Additionally, six out of seven patients with molecular relapse achieved MRD eradication, and most of them transitioned to HSCT. A clinical trial conducted by the Gruppo Italiano Malattie Ematologiche dell’Adulto (GIMEMA) is currently investigating the use of VEN + AZA specifically as bridge to transplant in
NPM1-mutated AML with MRD persistence or molecular disease recurrence (NCT04867928).
Mutations in genes encoding the isocitrate dehydrogenase isoforms 1 and 2 (
IDH1,
IDH2) are found in approximately 15% of AML patients
[78] and have been shown to confer exquisite susceptibility to BCL-2 inhibition via cytochrome C oxidase inhibition mediated by the oncometabolite 2-hydroxyglutarate, ultimately resulting in the lowering of the mitochondrial threshold for the triggering of apoptosis
[79]. The favorable impact of
IDH1/2 mutations on response to VEN, even as a single agent
[8], has been reported in several clinical trials and real-world experiences both in the frontline
[9][28][29][76][9,28,29,76] and R/R settings
[30][36][39][43][30,35,37,43]. Overall,
IDH1/2 mutations are associated with higher ORR and longer DFS and OS, especially in previously untreated patients. Interestingly, in R/R AML, some authors have reported less of an impact for
IDH1/2 mutations, with no significant differences in outcomes compared to unmutated patients.
[32][34][38][32,34,39] As the use of IDH1 and IDH2 inhibitors available for clinical use in R/R AML is associated with similar ORRs
[80][81][82][83][80,81,82,83], the current positioning of VEN-based regimens in this population warrants further investigations.
Data regarding the beneficial impact of other mutations are difficult to interpret due to the low number of patients harboring the same genetic lesions in most study populations. Consequently, it is often not possible to establish whether specific mutations have an actual impact on clinical outcomes per se or rather in a comutational, context-dependent manner. For example, mutations affecting the epigenetic modifier genes (i.e.,
DNMT3A and
TET2)
[32][34][32,34] and the spliceosome machinery (i.e.,
SRSF2)
[84] were shown to correlate with higher ORR. However, such lesions are often found in comutational contexts including mutations in genes with a known favorable impact on VEN sensitivity, such as
NPM1,
IDH1, and
IDH2 [84][85][84,85]. Moreover, some mutations (i.e.,
TET2) may be associated with improved susceptibility to HMA rather than VEN
[86].
Overall, it is quite evident that mutations alone are not sufficiently accurate in predicting VEN sensitivity or resistance. Newer predictive tools not based on traditional cytogenetic/molecular biomarkers, such as BH3-profiling
[8], might provide clinicians with useful information that might eventually drive therapeutic decisions if prospectively validated in a clinical setting.
Encyclopedia
