Outcomes of patients with higher-risk myelodysplastic syndromes/neoplasms treated with hypomethylating agents + venetoclax—an analysis from the International Consortium for MDS (icMDS) VALIDATE database

Abstract

Monotherapy with hypomethylating agents (HMA) remains the standard of care for patients with higher-risk myelodysplastic neoplasms (HR-MDS). Recently, the randomized phase III VERONA study evaluating azacitidine plus venetoclax (VEN) versus azacitidine plus placebo in newly diagnosed HR-MDS showed no difference in overall survival (OS) between the two arms. However, whether the addition of VEN to HMA improves outcomes among subsets of patients with HR-MDS remains debated. We analyzed 1907 patients with HR-MDS from 31 centers in 9 countries who were treated with HMA monotherapy or HMA/VEN in the frontline setting (HMA monotherapy: n = 1773; HMA/VEN: n = 134). Responses were assessed centrally by two investigators using the IWG 2023 response criteria. Addition of VEN improved composite complete remission (cCR) rates (48.8% vs. 27.7%; p < 0.001) but not CR rates (17.1% vs. 11.7%; p = 0.16). In multivariable logistic regression analysis, cCR remained favorable for HMA/VEN vs. HMA monotherapy (Odds Ratio [OR]: 2.49; 95% CI: 1.56-3.96; p < 0.001). However, we did not observe a statistically significant difference in OS for HMA/VEN vs. HMA monotherapy (Hazard Ratio [HR]: 0.83; 95% CI: 0.64-1.07; p = 0.15). In subgroup analyses, patients with TP53 wild-type disease (HR: 0.47; 95% CI: 0.29-0.74; p = 0.002) had a significant improvement in OS and those with ≥10% bone marrow blasts (HR: 0.73; 95% CI: 0.53-1.01; p = 0.06) had a trend towards OS benefit with HMA/VEN.

Introduction

Monotherapy with a hypomethylating agent (HMA) including azacitidine (AZA), decitabine (DEC), or oral decitabine-cedazuridine (DEC-C) has remained the standard of care for patients with higher-risk myelodysplastic syndromes/neoplasms (HR-MDS) for almost two decades [1–3]. AZA monotherapy showed an overall survival (OS) benefit compared to conventional care regimens in the AZA-001 trial, and DEC improved leukemia-free survival [1, 2]. Nevertheless, rates of complete remission (CR) with either agent are typically 15-30%, and median OS with HMA monotherapy is <2 years in both real-world studies and contemporary clinical trials [1, 2, 4–6]. Moreover, a multitude of clinical trials investigating combinations of HMA with novel agents have failed to improve OS in patients with HR-MDS compared with HMA alone [7–10].

The addition of the BCL2 inhibitor venetoclax (VEN) to AZA improved OS in AML compared to AZA plus placebo, and showed encouraging efficacy in a single-arm phase I trial in HR-MDS patients [11, 12]. While the recently completed phase III, placebo-controlled VERONA trial of AZA+/- VEN in HR-MDS showed improved overall response rates with AZA/VEN over AZA monotherapy, there was no difference in OS between the two treatment arms [13]. Although the overall results of the VERONA trial did not show a survival benefit from AZA/VEN, it remains unclear whether specific HR-MDS subgroups may have improved outcomes with the addition of VEN.

Here we used the VALIDATE database, a collaborative effort across 31 centers in 9 countries under the International Consortium for MDS (icMDS), to compare the outcomes of HR-MDS patients treated with HMA or HMA/VEN as frontline therapy and to evaluate the real-world efficacy of HMA/VEN (including response assessments based on the novel IWG 2023 response criteria) [14]. We also examined subgroups of patients who might benefit from the addition of VEN based on blast percentage, TP53, ASXL1, and RUNX1 mutation status, and treatment intent (palliative with medical therapy alone vs. curative with allo-HCT).

Methods

Patient selection

The icMDS VALIDATE database includes over 4000 adult patients with MDS who received HMA-based therapy (Supplementary Table 1). For this analysis, we included patients with HR-MDS defined by an International Prognostic Scoring System [IPSS] score ≥1.5 or an IPSS-revised [IPSS-R] score ≥4, at diagnosis or HMA initiation, who were treated with HMA monotherapy or HMA/VEN as frontline therapy for HR-MDS. IPSS, IPSS-R, and IPSS-molecular (IPSS-M) scores were centrally computed by two independent investigators [15–17]. For patients with missing molecular profiling at time of diagnosis, the IPSS-M score was calculated according to the original published methodology using the publicly available code, which estimates risk under best-case, average-case, and worst-case scenarios for untested variables, thereby permitting assignment of a risk score and corresponding risk category despite absent mutational data [15].

Patients in the HMA monotherapy group could have received either subcutaneous or intravenous AZA, intravenous DEC, or oral DEC-C. Administration schedules and dose adjustments for subsequent treatment cycles and concurrent medications were at the discretion of the treating physician. Patients who received HMA in combination with agents other than VEN were excluded, as were patients who received HMA after any other form of MDS treatment except for erythropoiesis-stimulating agents. Furthermore, patients were excluded if survival status at the time of last follow up was unknown.

Response assessment

Treatment responses were assessed centrally by two independent investigators. We assessed the best response to treatment at any time from treatment initiation based on the IWG 2023 criteria [14]. Composite complete remission (cCR) was defined as a composite of CR, CR with bilineage hematologic recovery (CRbi), CR with unilineage hematologic recovery (CRuni), CR with partial hematologic recovery (CRh), and CR equivalent (CRequi).

Statistical analysis

The primary endpoint of this analysis was OS stratified by treatment type (HMA monotherapy vs. HMA/VEN). We calculated OS from time of HMA initiation to time of death from any cause. Patients who had not experienced the event (death) were right censored at the time of last follow-up. We used the Kaplan-Meier method to estimate OS and compared OS between subgroups using the log-rank test. We fitted Cox proportional hazards models for time-to-event outcomes and logistic regression models for response endpoints, adjusting for baseline variables. As survival curves crossed in a subset of analyses, suggesting possible violation of the proportional hazards assumption, restricted mean survival time (RMST) over 1- and 2-year time horizons was included as a sensitivity analysis [18].

We compared rates of cCR between patients treated with HMA monotherapy vs. HMA/VEN via multiple logistic regression, adjusting for baseline patient and disease characteristics, such as age category, sex and IPSS-M risk category. We performed subgroup analyses stratified by baseline bone marrow blast percentage (<10% vs. 10-19%), achievement of cCR, and the presence of TP53, ASXL1 or RUNX1 mutations. TP53 allelic state was defined as biallelic vs. monoallelic based on 5th edition of the World Health Organization Classification of Hematolymphoid Tumors: Myeloid and Histiocytic/Dendritic Neoplasms [19]. Baseline characteristics were compared with the Wilcoxon rank sum test or two-sample t-test for continuous variables and Fisher’s exact test for categorical variables.

To account for the impact of allo-HCT and avoid immortal time bias [20], we modeled allo-HCT as a time-dependent intermediate event within a multi-state framework [21]. This model comprised three states: treatment initiation (baseline), allo-HCT, and death. We defined time to transplant as the interval from treatment initiation to the date of allo-HCT and OS as the time from therapy initiation to death or last follow-up. Patients who did not undergo allo-HCT remained in the baseline therapy state until death or censoring. This approach ensures that all patients contribute time to the ‘pre-transplant’ state until the time of allo-HCT, after which they move to the ‘post-transplant’ state. Unlike landmark analysis, this method preserves the full longitudinal dataset and avoids selection bias induced by conditioning on survival to a fixed time point [22].

The VALIDATE protocol was exempted by Institutional Review Board (IRB) of the coordinating Center (Yale University) due to minimal risk, and the need for informed consent was waived. The protocol was exempted, acknowledged, or approved by the other centers according to their local policies. All participating centers signed a data use agreement. VALIDATE is supported by an independent grant from AbbVie. AbbVie had no role in the study design, data analyses, interpretation, or writing of the manuscript.

Results

Patient population

Out of 4732 patients included in the VALIDATE database, a total of 1907 patients met eligibility criteria and were included in this analysis (Supplementary Fig. 1). Most patients were treated with AZA monotherapy (n = 1552 patients; 81.4%) or DEC monotherapy (n = 221; 11.6%), with a smaller percentage treated with HMA/VEN combination (n = 134; 7.0%). Subsequent allo-HCT was performed in 479 patients (25.1%; 430 after HMA monotherapy [24.3%] and 49 after HMA/VEN [37.1%]). Our cohort had high rates of patients with adverse disease characteristics, including complex karyotype (33.6%) and TP53 mutations (28.7%). Compared to patients treated with HMA/VEN, patients treated with HMA monotherapy were more likely to be male (p = 0.012), to have lower baseline platelet counts (p = 0.037), to have lower peripheral blood blasts (p = 0.004), and to have lower bone marrow blast percentage (p < 0.001). Furthermore, patients treated with HMA monotherapy were less likely to have IPSS-M very high-risk disease (p < 0.001), less likely to have TP53-mutated disease (p < 0.001), and less likely to have a complex karyotype (p = 0.020). Baseline patient and disease characteristics are presented in Table 1. The mutational and cytogenetic profile of our patient population is shown in Supplementary Fig. 2.

Table 1.

Baseline patient, disease, and treatment characteristics.

	HMA monotherapy (N = 1773 patients)	HMA/VEN (N = 134 patients)	P value
Age (mean; SD)	69.03 (11.26)	68.08 (10.30)	0.342
Male sex (n; %)	1163 (65.6)	73 (54.5)	0.012
Baseline CBC parameters at diagnosis (mean; SD)
WBC (x 109/L)	4.34 (6.16)	4.55 (5.42)	0.697
ANC (x 109/L)	2.53 (13.06)	2.65 (7.66)	0.916
Hemoglobin (g/dL)	9.11 (1.77)	9.17 (1.83)	0.715
Platelets (x 109/L)	103.05 (108.77)	123.64 (129.52)	0.037
Peripheral blood blasts (%)	1.46 (3.03)	2.32 (4.22)	0.004
Bone marrow blast percentage at diagnosis (mean; SD)	9.25 (5.05)	11.25 (4.79)	<0.001
IPSS-M risk category (n; %)			<0.001
Very Low, Low, or Moderate Low	170 (9.6)	11 (8.2)
Moderate High	269 (15.2)	15 (11.2)
High	673 (38.0)	32 (23.9)
Very High	660 (37.2)	75 (56.0)
TP53 mutant (n; %)	277 (26.8)	55 (45.5)	<0.001
Complex karyotype (n; %)	578 (32.9)	57 (43.2)	0.020
Number of treatment cycles (median; IQR)	6 (3–11)	3 (2–6)	<0.001
Allo-HCT recipients (n; %)	430 (24.3)	49 (36.6)	0.292
Time from diagnosis to allo-HCT (median; IQR)	8.5 months (4.9–17.5)	5.8 months (3.5–12.6)	<0.001

Allo-HCT allogeneic hematopoietic cell transplant, ANC absolute neutrophil count, CBC complete blood count, HMA hypomethylating agent, IQR interquartile range, IPSS-M International prognostic scoring system—molecular, IQR interquartile range, SD standard deviation, VEN venetoclax, WBC white blood cell count.

Effect of the addition of VEN to HMA on composite CR rates

We first evaluated the rates of cCR by treatment type. In univariable logistic regression, the rate of cCR was significantly higher in HMA/VEN treated patients compared to HMA monotherapy treated patients (48.8% vs. 27.7%; OR = 2.49; 95% Confidence Interval [CI]: 1.56 – 3.96; p < 0.001). However, CR rates were not significantly different between groups (17.1% vs. 11.7% for HMA/VEN and HMA monotherapy, respectively; OR: 1.56; 95% CI: 0.81 – 2.81; p = 0.157; Fig. 1A). After adjusting for age, sex, and IPSS-M risk category in a multivariable model, patients treated with HMA/VEN remained more likely to achieve cCR compared to HMA monotherapy (odds ratio [OR]: 2.49; 95% CI: 1.56-3.96; p < 0.001; Fig. 1B).

Fig. 1. Composite CR rates by treatment type.

Composite CR rates by treatment type. A shows rates of composite complete remission (cCR) by treatment type. CCR was defined as a composite of CR and non-CR responses, which comprised CR with limited hematologic recovery (CR_L), CR with partial hematologic recovery (CR_h), and CR equivalent (CR_equi). B shows the results of a multivariable logistic regression model for achieving a cCR.

Effect of addition of VEN to HMA on OS among patients with HR-MDS

At a median follow-up of 42 months, median OS from treatment initiation was comparable at 17.4 months (95% CI: 16.7 - 18.6 months) for HMA monotherapy and 19.6 months (95% CI: 14.9 - 37 months) for HMA/VEN, respectively (Fig. 2A). In unadjusted analysis, OS did not differ significantly by treatment type (HR: 0.89; 95% CI: 0.69 - 1.13; p = 0.33). For patients treated with HMA monotherapy, the OS rates at year 1, 2, and 3 were 0.65 (95% CI: 0.63 - 0.67), 0.38 (95% CI: 0.36 - 0.40), and 0.27 (95% CI: 0.25 - 0.29), respectively. For patients treated with HMA/VEN, the OS rates at year 1, 2, and 3 were 0.64 (95% CI: 0.56 - 0.73), 0.45 (95% CI: 0.36 - 0.56), and 0.37 (95% CI: 0.27 - 0.51), respectively.

Fig. 2. Overall survival by treatment type.

A shows overall survival (OS) from time of HMA initiation for patients treated with HMA monotherapy and HMA/VEN. B shows the results of a multivariable Cox regression model of OS adjusted for IPSS-M disease risk category, age, sex, treatment type and allogeneic hematopoietic cell transplant as a time-dependent variable.

In a multivariable Cox model adjusted for age, sex, peripheral blood blast percentage, IPSS-M risk category, and allo-HCT as a time-dependent variable, treatment with HMA/VEN vs. HMA monotherapy did not significantly influence OS (HR for HMA/VEN vs HMA monotherapy: 0.83; 95% CI: 0.64 - 1.07; p = 0.15; Fig. 2B).

When restricting the analysis to a fixed time period of 1 year, the RMST was 9.8 months (95% CI: 9.2 - 10.4 months) for patients treated with HMA/VEN and 10.1 months (95% CI: 9.9 - 10.2 months) for those who received HMA monotherapy. At 2 years, the RMST in the HMA/VEN cohort was 16.1 months (95% CI: 14.6 - 17.6 months) and in the HMA monotherapy cohort was 16.0 months (95% CI: 15.6 - 16.4 months). There was no significant difference between HMA/VEN and HMA monotherapy at either 12 months (RMST difference: −0.30 months; 95% CI: −0.91 to 0.32; p = 0.34) or 24 months (RMST difference: 0.06 months; 95% CI: −1.48 to 1.60; p = 0.94).

Overall survival by cCR status

Next, we analyzed OS by cCR status as defined by IWG 2023 criteria. Among patients treated with HMA monotherapy, we found that patients who achieved a cCR had favorable outcomes compared to patients who did not achieve a cCR (median OS: 23.9 months [95% CI: 19.8 - 31.3 months] vs. 15.6 months [95% CI: 13.8 - 17.2 months]; HR: 0.69 [95% CI: 0.57 - 0.84]; p < 0.001; Supplementary Fig. 3A). Similarly, patients who received HMA/VEN and achieved a cCR had improved OS compared to patients who did not achieve a cCR (median OS: not reached vs. 12.7 months [95% CI: 7.4 - 22.4 months]; HR: 0.33; 95% CI: 0.18 - 0.64; p < 0.001; Supplementary Fig. 3B). To account for immortal time bias, we conducted 4-month and 6-month landmark analyses, which also showed improved OS for patients who achieved a cCR compared to patients who did not, irrespective of treatment type (Supplementary Fig. 3C–F).

Overall survival by transplant status

Median time to allo-HCT from treatment initiation was 172 days and 135 days for patients treated with HMA monotherapy and HMA/VEN, respectively. In a multistate model adjusted for age category, sex, and IPSS-M, patients treated with HMA/VEN were more likely to proceed to allo-HCT compared with patients treated with HMA monotherapy (HR: 1.68; 95% CI: 1.24 - 2.28; p < 0.001). Among patients who underwent allo-HCT, OS from time of transplant was better among patients treated with HMA/VEN compared with HMA monotherapy (HR: 0.49; 95% CI: 0.29 - 0.85; p = 0.01). Conversely, among patients who did not undergo allo-HCT, there was no difference in OS from time of HMA initiation for patients treated with HMA/VEN vs. HMA monotherapy (HR: 1.10; 95% CI: 0.83 - 1.46; p = 0.50; Fig. 3).

Fig. 3. Multistate transition model.

This figure presents transition probabilities between different states for patients treated with HMA/VEN vs. HMA monotherapy. Hazard ratios with associated 95% confidence intervals are reported.

Outcomes by TP53 mutation status

As TP53 mutations in HR-MDS are associated with adverse prognosis and lower benefit from the addition of VEN in AML [16, 17, 23], we analyzed response rates separately by TP53 mutation status. In multivariable logistic regressions adjusted for age, sex, and IPSS-M risk category, we found that the OR for cCR for HMA/VEN vs HMA monotherapy was 4.17 (95% CI: 2.14 - 8.31; p < 0.001) and 1.27 (95% CI: 0.54 - 2.86; p = 0.57) for TP53 wild-type and TP53-mutant patients, respectively (Fig. 4A, B). In addition, a significant treatment-by-TP53 mutation interaction was observed (interaction OR = 3.39; 95% CI: 1.21 - 9.72; p = 0.022), suggesting that the cCR benefit associated with HMA/VEN was greater in TP53 wild-type patients than in TP53-mutant patients. When stratified by TP53 mutation status, we found that addition of VEN improved OS only for patients with TP53 wild-type disease in a multivariable Cox model adjusted for age, sex, PB blast, IPSS-M and allo-HCT as a time-varying covariate (HR of HMA/VEN vs. HMA monotherapy: 0.47; 95% CI: 0.29 - 0.74; p < 0.001) but not for patients with TP53-mutant disease (HR = 0.96; 95% CI: 0.66 - 1.41; p = 0.57; Supplementary Fig. 4). Adjusting for other variables in the model, HR for patients who underwent allo-HCT vs. those who did not were 0.65 (95% CI: 0.49 - 0.85) and 0.36 (95%CI: 0.24 - 0.53) in patients with TP53 wild-type and mutated disease, respectively.

Fig. 4. Response by TP53 mutation status and treatment type.

A, B show results of a multivariable logistic regression model of cCR for patients with TP53 mutation and TP53 wild-type disease, respectively.

As previous studies have suggested that TP53 allelic status is associated with differences in prognosis among MDS patients, with biallelic TP53 mutations associated with inferior outcomes compared with monoallelic mutations, we analyzed outcomes by treatment type separately by TP53 allelic state [24, 25]. In our cohort there were 94 patients (n = 13 patients treated with HMA/VEN and n = 81 patients with HMA monotherapy) and 238 patients (n = 42 patients treated with HMA/VEN and n = 196 patients with HMA monotherapy) with monoallelic and biallelic TP53 mutations, respectively. When analyzing OS only among patients with monoallelic TP53 mutations, there was no difference in OS by treatment type (HR: 1.04; 95% CI: 0.50 - 2.19; p = 0.91; Supplementary Fig. 5A). Similarly, there was no difference in OS among patients with biallelic TP53 mutations by treatment type (HR: 0.94; 95% CI: 0.64 - 1.39; p = 0.76; Supplementary Fig. 5B).

Outcomes by ASXL1 and RUNX1 mutation status

As prior studies including the VERONA trial suggested that mutations in ASXL1 and RUNX1 are associated with improved outcomes with the addition of VEN to HMA [26, 27], we performed dedicated subgroup analyses defined by ASXL1 and RUNX1 mutation status. In a multivariable logistic regression adjusted for age, sex, and IPSS-M risk category, we found that the OR for cCR for HMA/VEN vs HMA monotherapy was 5.29 (95% CI: 2.02 - 14.50; p = 0.001) and 2.15 (95% CI: 1.18 – 3.91; p = 0.012) for ASXL1 mutant and ASXL1 wild type patients, respectively (Supplementary Fig. 6A, B). In additional analyses, the treatment-by-ASXL1 mutation interaction did not reach statistical significance but suggested a potentially stronger association between HMA/VEN and cCR among ASXL1-mutant patients compared with ASXL1 wild-type patients (OR = 2.66; 95% CI: 0.87– 8.12; p = 0.09). When stratified by ASXL1 mutation status, we found that in a multivariable model adjusted for age, sex, peripheral blood blast percentage, IPSS-M risk category, and allo-HCT as a time-varying covariate that the addition of VEN did not improve OS for patients with neither ASXL1-mutant disease (HR = 0.586; 95%CI: 0.33 - 1.04; p = 0.069) nor ASXL1 wild-type disease (HR of HMA/VEN vs. HMA monotherapy: 0.85; 95% CI: 0.61 - 1.71; p = 0.32; Supplementary Fig. 6C, D).

In a multivariable model that included age, sex, and IPSS-M risk category, we found that the OR for cCR for HMA/VEN vs HMA monotherapy was 0.69 (95% CI: 0.09 - 3.52; p = 0.68) and 3.06 (95% CI: 1.80 – 5.24, p < 0.001) for RUNX1-mutant and RUNX1-wild type patients, respectively (Supplementary Fig. 7A, B). In addition, we did not observe a statistically significant treatment-by-RUNX1 mutation interaction (OR = 0.23; 95% CI: 0.04–1.35; p = 0.15). When stratified by RUNX1 mutation status, we found in a multivariable model adjusted for age, sex, peripheral blood blast percentage, IPSS-M risk category, and allo-HCT as a time-varying covariate that addition of VEN did not improve OS neither among patients with RUNX1-mutant disease (HR = 0.42; 95% CI: 0.15 - 1.22; p = 0.11) nor among patients with RUNX1 wild-type disease (HR of HMA/VEN vs. HMA monotherapy: 0.77; 95% CI: 0.58 – 1.03; p = 0.081; Supplementary Fig. 7C, D).

Outcomes by baseline bone marrow blast percentage

Lastly, we investigated whether patients with MDS/AML classification by the ICC system (i.e., patients with 10-19% bone marrow blasts) would benefit from the addition of VEN to HMA [28]. When comparing patients with 10-19% bone marrow blasts to patients with <10%, we found a trend towards an OS benefit for treatment with HMA/VEN vs. HMA monotherapy (HR: 0.73; 95% CI: 0.53 - 1.01; p = 0.056; Fig. 5A) in the subgroup of patients with 10-19% bone marrow blasts but not among patients with <10% bone marrow blasts (HR: 1.26; 95% CI: 0.87 - 1.82; p = 0.23; Fig. 5B).

Fig. 5. Overall survival by baseline bone marrow blast percentage.

A, B show overall survival (OS) from time of treatment initiation for patients with ≥10% bone marrow blasts and patients with <10% bone marrow blasts at baseline, respectively.

Discussion

In this large, international retrospective study of 1907 patients with HR-MDS who received HMA monotherapy or HMA in combination with VEN, we found that the addition of VEN did not lead to an OS benefit in the overall group of patients. However, we found that patients with (i) TP53 wild-type disease and (ii) patients with blasts of 10-19% may benefit from the addition of VEN to HMA. Furthermore, those who received HMA/VEN were more likely to undergo subsequent allo-HCT compared with those receiving HMA monotherapy.

Our findings are generally in line with the results from the randomized phase III VERONA trial of AZA + /- VEN, which also showed higher modified response rates (i.e., CR + mCR per IWG 2006 criteria) but no improvement in CR rates or OS [13]. While randomized, placebo-controlled trials are the gold standard to assess the efficacy of a new therapy, our study provides important real-world evidence supporting the conclusions of the VERONA trial in a large, international cohort study. Moreover, our analysis extends the conclusions of the VERONA trial to key subgroups of patients that were excluded from the VERONA trial, such as those with therapy-related MDS [26].

Patients in the VALIDATE cohort had different molecular characteristics compared with patients enrolled in the VERONA trial, including higher prevalence of TP53 mutations (28.7% vs. 21.6%), and lower rates of ASXL1 (21.3% vs. 31.8%) and RUNX1 mutations (13.1% vs. 19.3%) [26]. These differences might explain the longer median OS in both arms of the VERONA trial compared to our study. In fact, clinical trials of HMA-based treatments have consistently demonstrated outcomes that are poorly reproducible in real-world studies. This is likely explained by a combination of differences in patient characteristics, more vigorous patient monitoring, and tighter adherence to treatment schedules [1, 4, 6, 13, 29].

While there was no OS benefit with the addition of VEN in HR-MDS patients, it is unclear if there are any subsets of patients that may benefit from HMA/VEN combination therapy [10]. From our analysis, it appears that patients with TP53-mutant disease, patients with less than 10% blasts in the bone marrow, and patients who do not undergo a subsequent allo-HCT might not achieve a survival benefit from the addition of VEN. However, these patients represent a significant proportion of HR-MDS both in the VALIDATE cohort and the VERONA trial. In VERONA and VALIDATE, 21.6% and 28.7% of patients were TP53 mutated, 53.4% and 52.9% had less than 10% blasts, and only 14.9% and 25.1% of patients proceeded to allo-HCT, respectively. In a multistate model, we found that patients treated with HMA/VEN had a higher likelihood of proceeding to allo-HCT and better OS after allo-HCT, compared with patients treated with HMA monotherapy. Although this suggests a potential benefit for HMA/VEN in patients intended for and proceeding to allo-HCT, this subgroup analysis needs to be interpreted cautiously. Selection bias may occur with patients receiving more aggressive therapy and then proceeding to allo-HCT, as time to transplant and transplant characteristics were not standardized, which could have affected outcomes [30]. Furthermore, it is difficult to ascertain whether the higher likelihood of proceeding to allo-HCT after HMA/VEN compared with HMA monotherapy in our study was due to the higher response rates observed with HMA/VEN or related to differences in baseline characteristics (e.g., lower bone marrow blast count and disease risk with HMA monotherapy) or patient and/or physician preferences. However, our findings are in line with the VERONA trial that also reported higher rates of allo-HCT in the AZA/VEN arm compared with the AZA monotherapy arm [26].

Our results are in contrast to another retrospective, multicenter cohort study from the US, which suggested an improvement in OS for HMA/VEN over HMA monotherapy in the overall MDS population studied [31]. Differences in the overall study population between these two retrospective studies might account for the divergent findings regarding OS benefit. Still, both studies suggest a potential benefit with the addition of VEN in patients with TP53 wild-type disease, allo-HCT recipients and blasts 10-19%. These results highlight the need for individualized patient selection when deciding between HMA monotherapy vs. HMA/VEN for patients with HR-MDS. Furthermore, it remains to be seen whether certain subsets of patients with TP53-mutant disease (e.g., single-hit mutation status) might still benefit from the addition of VEN to HMA.

A major strength of our analysis is the large, international sample of patients with HR-MDS with detailed molecular annotations. This global representation reflects real-life practice patterns and increases the generalizability of our results. Another notable strength of our study is that all responses and disease risk scores (IPSS, IPSS-R, and IPSS-M) were centrally verified by two independent investigators. However, our study is also subject to certain limitations. First, data were largely derived from academic centers, which might have introduced selection bias and potentially explains the higher frequency of TP53 mutations and allo-HCT, especially among HMA/VEN-treated patients. However, we accounted for this by conducting analyses adjusted for baseline disease risk, as well as subgroup analyses stratified by TP53 mutation status and undergoing subsequent allo-HCT. Second, as this was a real-world retrospective study, there was a certain degree of missing data. For example, molecular variables were missing in 38.6% of patients, which required imputation of IPSS-M scores and might have led to uncertainty regarding risk stratification for some patients. Similarly, we did not have access to serial hematologic parameters, which limited or precluded response assessments, especially with regards to hematologic improvement in some patients. Furthermore, treatment schedules, dose adjustments, and timing of response assessments were not standardized in our study. We also did not have information on post-progression treatments, which might have influenced outcomes. Lastly, the results from the subgroup analysis of patients who underwent a subsequent allo-HCT need to be interpreted cautiously as there is potential confounding by differences in the time to transplant between arms.

In conclusion, this large, international, retrospective analysis from the icMDS VALIDATE database showed that treatment with HMA/VEN was associated with higher cCR rates — but not CR rates — compared to HMA monotherapy among patients with HR-MDS. However, we did not observe a statistically significant difference in OS with the addition of VEN in the overall study cohort. However, the addition of VEN might lead to a benefit in some subgroups of patients such as those with TP53 wild-type disease, patients with ≥10% bone marrow blasts at baseline, and those who undergo subsequent allo-HCT. Our results highlight the importance of individualized treatment decision-making but are limited by the retrospective nature of the dataset and the relatively small number of HMA/VEN-treated patients. Additional randomized studies are needed to explore whether certain subgroups of patients with HR-MDS would benefit from the addition of VEN, or other novel BCL-2 inhibitors, to HMA.

Author contributions

JPB, TK, MS, and AMZ conceptualized the project, provided patient data, analyzed the data, and wrote the original manuscript. WW and TRG provided statistical and scientific input. All authors provided patient data, analyzed data, revised the manuscript, and provided scientific input. All authors approved the current version of the manuscript, including figures and tables.

Funding

VALIDATE is supported by an independent grant from AbbVie. AbbVie had no role in the study design, data analyses, interpretation, or writing of the manuscript.

Data availability

Anonymized or aggregated study data will be made available upon reasonable request from qualified researchers, to the extent permitted by applicable laws and data use agreements signed with study sites. Approval of such requests is at the discretion of the VALIDATE study team and is dependent on the nature of the request, the merit of the research proposed, the availability of the data, and the intended use of the data.

Competing interests

The authors declare the following conflict of interests: AED: Sobi: Consultancy; Novartis: Membership on an entity’s Board of Directors or advisory committees; Geron: Membership on an entity’s Board of Directors or advisory committees; Caribou: Membership on an entity’s Board of Directors or advisory committees; Bristol Myers Squibb: Consultancy; Appellis: Consultancy, Membership on an entity’s Board of Directors or advisory committees. Celgene/BMS, Agios, and Novartis: Advisory boards and/or consultancy; Geron, Servier, Keros, Shuattuck labs, and Celgene/BMS: Clinical trial committees or DSMB. MAS: Geron: Membership on an entity’s Board of Directors or advisory committees; Kurome: Consultancy, Current holder of stock options in a privately-held company; Novartis: Consultancy, Membership on an entity’s Board of Directors or advisory committees; BMS: Consultancy, Membership on an entity’s Board of Directors or advisory committees. GLU: Incyte: consultancy; Geron: consultancy; Syndax: consultancy; Autolus: consultancy; Vironexis: consultancy; HEC: Agios: Consultancy, Speakers Bureau; BMS: Consultancy, Research Funding, Speakers Bureau; Jazz Pharmaceuticals: Consultancy, Other: Travel, Accommodations, Expenses, Speakers Bureau; Takeda: Other; Genentech: Consultancy; Daiichi: Consultancy; Celgene: Research Funding; Syndax: Other: DSMB; Astex Pharmaceuticals: Other; AbbVie: Other; Stemline Therapeutics: Consultancy, Speakers Bureau; Novartis: Consultancy, Other: Travel, Accommodations, Expenses, Speakers Bureau. PD: Abbvie: Consultancy, Other: Advisory role; BMS: Consultancy, Other: Advisory role, Research Funding; Janssen Research & Development: Research Funding; Servier: Consultancy, Other: Advisory role; Janssen Pharmaceuticals: Current Employment. Syndax: Consultancy; Kura Oncology: Consultancy. EAG: Alexion Pharmaceuticals/AstraZeneca Rare Disease: Consultancy, Research Funding; Apellis Pharmaceuticals: Consultancy, Research Funding; Blueprint Medicines, Inc: Research Funding; Bristol-Myers Squibb: Consultancy/Research Funding; Celldex Therapeutics: Research Funding; Cogent Biosciences: Consultancy; Curio Science: Honoraria; Geron Inc.: Consultancy; MediCom Worldwide, Inc.: Honoraria; MedscapeLIVE: Honoraria; NextCure, Inc: Research Funding; Novartis: Consultancy, Research Funding; Pfizer Inc: Consultancy; Refined Oncology: Consultancy; Servier Pharmaceuticals: Consultancy; Taiho Oncology/Astex Pharmaceuticals: Consultancy/Research Funding; Vera and Joseph Dresner Foundation: Membership on an entity’s Board of Directors or advisory committees. VJHemeOnc: Honoraria. EMS: Calithera: Consultancy; Daiichi: Consultancy; Foghorn: Consultancy; Servier: Consultancy; CTI Biopharma: Consultancy; Menarini: Consultancy; Genentech: Consultancy; Genesis: Consultancy; Syndax: Consultancy; OnCusp: Consultancy; Abbvie: Consultancy; Neoleukin: Consultancy; Jazz: Consultancy; Bristol Myers Squib: Consultancy, Research Funding; Agios: Consultancy; Novartis: Consultancy; Gilead: Consultancy; Eisai: Research Funding; Janssen: Consultancy; PinotBio: Consultancy; Aptose: Consultancy; Syros: Consultancy; Astellas: Consultancy; Ono Pharma: Consultancy; Blueprint: Consultancy. AMB: Agios: Consultancy, Research Funding; AstraZeneca: Research Funding; Celgene/BMS: Consultancy, Research Funding; Gilead: Consultancy; Taiho: Consultancy; Novartis: Consultancy, Research Funding; Janssen: Research Funding; Keros Therapeutics: Consultancy; Takeda: Consultancy; Acceleron: Consultancy; GSK: Research Funding. CM: Syndax: Advisory board; Kura Oncology: Safety monitoring committee. RMS: Bristol Myers Squibb: Consultancy; Cardinal Health: Consultancy, Honoraria; Curio Science: Consultancy; Geron: Consultancy, Honoraria; Gilead Sciences: Consultancy, Honoraria; Kura Oncology: Consultancy; Rigel: Consultancy; Servier: Consultancy; Syndax Pharmaceuticals: Consultancy; TScan Therapeutics: Consultancy. JFZ: AbbVie: Consultancy, Research Funding; Akesobio: Research funding; Aptevo: Research funding; Arog: Research Funding; Ascentage: Research funding; Astellas: Consultancy; AstraZeneca: Consultancy, Research Funding; Auron: Research Funding; Crossbow Therapeutics: Consultancy; Daiichi Sankyo: Consultancy, Research Funding; Faron: Research Funding; Foghorn: Consultancy; Genentech: Consultancy; Genmab: Consultancy; Geron: Consultancy; Ipsen: Consultancy; Jazz: Consultancy, Research Funding; Johnson & Johnson: Consultancy; Kura Oncology: Consultancy, Research Funding; Lin Biosciences: Research Funding; Loxo: Research Funding; Merck: Research Funding; Novartis: Consultancy, Research Funding; Relmada Therapeutics: Consultancy; Rigel: Consultancy; Sellas: Consultancy, Research Funding; Servier: Consultancy; Shattuck Labs: Consultancy, Research Funding; Stemline: Research Funding; Sumitomo Pharma: Consultancy, Research Funding; Syndax: Consultancy; Zentalis: Research Funding. MRS: Karyopharm Therapeutics Inc.: Consultancy, Current equity holder in publicly-traded company; Empath Biosciences: Current equity holder in publicly-traded company; Incyte Corporation: Research Funding; Sierra Oncology, Inc.: Consultancy; Taiho: Research funding and travel costs; ALX Oncology: Research Funding; AbbVie Inc.: Consultancy; Boehringer Ingelheim: Patents & Royalties; TG Therapeutics, Inc.: Research Funding; Takeda Pharmaceutical Company: Consultancy, Research Funding; Ryvu Therapeutics: Consultancy, Current equity holder in publicly-traded company; Novartis: Consultancy; Geron Corporation: Consultancy; Forma Therapeutics Inc.: Consultancy, Consultancy; Bristol Myers Squibb: Membership on an entity’s Board of Directors or advisory committees; CTI BioPharma Corp.: Consultancy. CNL: Azurity: Advisory board. GJR: Agios: Consultancy; Actinium: Consultancy; Amgen: Consultancy; Astellas: Consultancy; AstraZeneca: Consultancy; Bristol Myers Squibb: Consultancy; Celgene: Consultancy; Caribou Biosciences: Consultancy; Daiichi Sankyo: Consultancy; Ellipses Pharma: Consultancy; Geron: Consultancy; Glycomimetics: Consultancy; Jazz Pharmaceuticals: Consultancy; MEI: Consultancy; Mesoblast: Consultancy; Molecular Partners: Consultancy; Morphosys: Consultancy; Oncoprecision: Consultancy; Oncoverity: Consultancy; AbbVie: Consultancy; Novartis: Consultancy; Pfizer: Consultancy; Rigel: Consultancy; Roche: Consultancy; Syndax: Consultancy; Blueprint: Consultancy; Jasper Pharmaceuticals: Consultancy; Bluebird bio: Consultancy; GlaxoSmithKline: Consultancy; Janssen: Consultancy and research support. Funding; Takeda: Consultancy. ESW: Gilead: Consultancy; GlaxoSmithKline: Consultancy; Jazz: Consultancy; Kite: Consultancy, Speakers Bureau; Novartis: Consultancy, Speakers Bureau; Pfizer: Consultancy, Speakers Bureau; PharmaEssentia: Consultancy; Takeda: Consultancy; Dava oncology: Speakers Bureau; Kura Oncology: Speakers Bureau; BMS: Consultancy; Astellas: Consultancy, Speakers Bureau; Abbvie: Consultancy. ZX: Novartis: Speakers Bureau; Moffitt Cancer Center: Current Employment. JPM: Regeneron: Consultancy, Honoraria; Alexion: Membership on an entity’s Board of Directors or advisory committees; Novartis: Honoraria, Speakers Bureau; Omeros: Consultancy. DS: Aprea, Jazz: Research Funding; AbbVie, Affimed Gmbh, Gilead, Incyte, Intellisphere, LLC, Molecular Partners AG, PGEN Therapeutics, Inc., Takeda, Zentalis; Advisory board for AvenCell, BlueBird Bio, BMS, Intellia, Jasper Therapeutics, Kite, Magenta Therapeutics, NKARTA, Novartis, Orbita: Consultancy. AA-K: Novartis: Research funding to institution, consultancy; Astex: Research funding to institution; ALX Oncology: Research funding to institution; BMS: Research funding to institution; Aprea: Research funding to institution; H3B/Hemavant: Research funding to institution, Celgene: Research funding to institution. AE: Otsuka: Consultancy, honoraria, research funding to institution; Servier: Consultancy; Jazz: Consultancy; RaceOncology: Consultancy, honoraria; Amgen: Consultancy; AbbVie: Consultancy. YFM: Honoraria/consulting fees from BMS, Kura Oncology, BluePrint Medicines, Geron, OncLive and MD Education, VJHemOnc and Medscape Live. Advisory boards and received honoraria from Stemline Therapeutics, Blueprint Medicines, Taiho Oncology, SOBI, Rigel Pharmaceuticals, Geron, Cogent Biosciences. Received travel reimbursement from MD Education. YA: Consultancy, Honoraria: Astellas, Bristol Myers Squibb, Daiichi Sankyo, Geron, KiTE, Pfizer, Rigel, and Servier. Research funding to institution: AbbVie, ALX Oncology, Biomea, Biosight, Curis, and Novartis. TB: Honoraria from Amgen, Ascentage, Syndax, Geron, Takeda. Research funding from Takeda, Amgen. TH: MLL Munich Leukemia Laboratory: equity ownership. IK: Abbvie: Consultancy, Honoraria; Janssen: Consultancy, Honoraria, Research Funding; BMS: Consultancy, Honoraria; Novartis: Consultancy, Research Funding. VP: Research Support /PI : Celgene, Janssen, Abvie; Honoraria/Advisory Board: Genesis Pharma, Janssen, Roche, Servier, Abvie, Novartis, Gilead, BMS, Astra Zeneca, GSK, Amgen, Sobi, Pfizer. CH: MLL Munich Leukemia Laboratory: equity ownership. MS: Kymera: Membership on an entity’s Board of Directors or advisory committees; Rigel: Membership on an entity’s Board of Directors or advisory committees; Curis Oncology: Other: GME activity; Clinical care options: Other: GME activity; GSK: Membership on an entity’s Board of Directors or advisory committees; Novartis: Membership on an entity’s Board of Directors or advisory committees, Other: GME activity; Haymarket Media: Other: GME activity; Boston Consulting: Consultancy; Sierra Oncology: Membership on an entity’s Board of Directors or advisory committees; Dedham group: Consultancy. MGDP: Bristol Myers Squibb: Honoraria, Membership on an entity’s Board of Directors or advisory committees. RSK: Geron: Consultancy; BMS: Honoraria, Membership on an entity’s Board of Directors or advisory committees, Research Funding; AbbVie, CTI biopharma, Jazz, Pharma Essentia, Servio: Honoraria, Membership on an entity’s Board of Directors or advisory committees, Speakers Bureau; Rigel, Taiho, DSI: Honoraria, Membership on an entity’s Board of Directors or advisory committees; Novartis: Membership on an entity’s Board of Directors or advisory committees. AMZ: Incyte: Consultancy, Honoraria; Agios: Consultancy, Honoraria; Boehringer-Ingelheim: Consultancy, Honoraria; Schrödinger: Consultancy, Honoraria; Celgene/BMS: Consultancy, Honoraria; Astellas: Consultancy, Honoraria; Syros: Consultancy, Honoraria; Jazz: Consultancy, Honoraria; Regeneron: Consultancy, Honoraria; Novartis: Consultancy, Honoraria; Servier: Consultancy, Honoraria; Zentalis: Consultancy, Honoraria; Tyme: Consultancy, Honoraria; Astex: Research Funding; Shattuck Labs: Research Funding; AbbVie: Consultancy, Honoraria; Lox Oncology: Consultancy, Honoraria; Pfizer: Consultancy, Honoraria; Mendus: Consultancy, Honoraria; Foran: Consultancy, Research Funding; Otsuka: Consultancy, Honoraria; BioCryst: Consultancy, Honoraria; ALX Oncology: Consultancy, Honoraria; BeyondSpring: Consultancy, Honoraria; Chiesi: Consultancy, Honoraria; Takeda: Consultancy, Honoraria; Kura: Consultancy, Honoraria; Ionis: Consultancy, Honoraria; Gilead: Consultancy, Honoraria; Syndax: Consultancy, Honoraria; Amgen: Consultancy, Honoraria; Epizyme: Consultancy, Honoraria; Janssen: Consultancy, Honoraria; Genentech: Consultancy, Honoraria; Seattle Genetics: Consultancy, Honoraria; Notable: Consultancy, Honoraria; Taiho: Consultancy, Honoraria; Geron: Consultancy, Honoraria; Orum: Consultancy, Honoraria; Daiichi Sankyo: Consultancy, Honoraria.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41408-026-01543-6.