Supplementary Figures S1, S2, S3, S4, S5 with legends

PURPOSE:Mutant IDH1 (mIDH1) defines a therapeutically targetable subtype of intrahepatic cholangiocarcinoma (ICC), with the mIDH1 inhibitor ivosidenib approved for advanced disease. A subset of patients experiences prolonged disease stabilization; however, the molecular basis for eventual progression remains poorly defined. EXPERIMENTAL DESIGN:We performed molecular profiling of matched baseline and postprogression circulating tumor DNA (ctDNA) samples from patients with mIDH1 ICC enrolled in the ClarIDHy phase III trial. Functional studies were conducted to characterize candidate resistance mechanisms. RESULTS:Longitudinal ctDNA analysis of 18 patients treated with ivosidenib for >6 months revealed emergent genomic alterations in multiple cases. Acquired mutations in mitogen-activated protein kinase (MAPK) pathway genes (KRAS, NRAS, MAP2K1, NF1) were identified in five cases, with instances of concurrent alterations and/or high variant allele fractions (VAF). Additional candidate resistance events included a secondary IDH1 mutation and a hotspot IDH2 mutation, detected at low VAF in the same patient. Functional studies showed that these IDH mutations conferred sustained 2-hydroxyglutarate production and ivosidenib resistance, whereas MAPK activation blunted gene expression induced by ivosidenib plus IFNγ, a key therapeutic output of mIDH1 inhibition. In parallel, baseline ctDNA profiling of 81 patients revealed that ARID1A mutations and elevated mIDH1 VAF were associated with reduced clinical benefit. CONCLUSIONS:MAPK pathway alterations represent a recurrent mechanism of resistance to mIDH1 inhibition in ICC, whereas emergent IDH1/IDH2 mutations appear infrequent. Functional data suggest that MAPK-mediated resistance may involve impaired IFN signaling. These results support MAPK-directed combination strategies and highlight the utility of ctDNA profiling to identify predictive and resistance biomarkers in mIDH1-driven ICC.

Introduction Ivosidenib (IVO) is approved as monotherapy and in combination with azacitidine for frontline treatment of patients (pts) with mIDH1 acute myeloid leukemia (AML) unfit for intensive chemotherapy (chemo). In this ph 1 study IVO was combined with intensive induction and consolidation chemo in fit pts with ND mIDH1 AML (NCT02632708). The regimen was as well tolerated as induction/consolidation chemo alone. There were no new safety signals compared with IVO monotherapy (Stein, Blood 2021). The rate of complete remission (CR)+CR with partial hematologic recovery (CRh) was 77%. We present long-term follow-up response/safety data. Methods This multicenter, open label ph 1 study was previously described (Stein, Blood2021). Pts with ND mIDH1 AML received induction therapy: cytarabine 200 mg/m2/d × 7 d and either daunorubicin 60 mg/m2/d or idarubicin 12 mg/m2/d × 3 d (up to 2 cycles of induction were permitted) and IVO 500 mg once daily starting on d 1 of induction therapy. Pts with at least partial remission at end of induction could receive ≤4 cycles of consolidation chemo while continuing IVO 500 mg daily. Those completing, or ineligible for, consolidation could receive maintenance with IVO 500 mg daily until relapse, unacceptable toxicity or allogeneic hematopoietic stem cell transplantation (HSCT). Pts could proceed to HSCT at any point. Results From Jan 2016–Jul 2018, 60 pts received IVO. Median age was 62.5 yr (range 24–76); 30 (50%) were female. Of 60 pts who received induction therapy, 9 (15%) received 2 induction cycles, 35 (58%) proceeded to consolidation, and 19 (32%) received IVO maintenance. At the data cutoff date of 16 Jan 2025, 3/19 pts who received IVO maintenance were still in maintenance and 16 had discontinued treatment due to: HSCT (n=2), adverse event (n=1), progressive disease (n=4), pt/physician decision (n=6), and other reasons (n=3). A total of 29 (48%) underwent HSCT in CR1: 13 (22%) after induction, 14 (40%) after consolidation, and 2 (11%) during maintenance. Ten pts are still in survival follow-up. Overall, CR was 70% (42/60; 95% CI: 57, 81); CR+CRh was 77% (46/60; 95% CI: 64, 87). Median time to CR was 34 d (range 23–246); 34 d to CR+CRh (range 22–246). Median duration of CR was 24.9 mo (95% CI: 16.0, 41.6) and of CR+CRh was 25.1 mo (95% CI: 8.3, 41.6). After the 1st induction cycle, absolute neutrophil count (ANC) recovery rate (ANC >500/µL) was 95.1% (39/41 pts eligible for analysis) and platelet count recovery rate (>50,000/µL) was 97.6% (40/41); median duration to recovery was 28 days for both. After the first consolidation cycle, ANC recovery rate was 90% (9/10 eligible pts for analysis) and platelet count recovery rate was 80% (8/10); median duration to recovery was 32 days and 26 days, respectively. Durable responses were noted across all comutations, including TP53. Median overall survival (OS) was not reached (NR; 95% CI: 39.4 mo, NR); OS rate at 3/5 yr was 67/61%. For 4 TP53 mutation positive pts, OS ranged from 18 to 66 mo; 1 pt had OS of 0.8 mo. Treatment-emergent adverse events (TEAEs) during induction and consolidation were consistent with previously reported TEAEs (Stein, Blood 2021). Incidence of G≥3 thrombocytopenia, leukopenia, and neutropenia after IVO-based induction was 62% (37/60), 33% (20/60), and 18% (11/60), respectively; after IVO-based consolidation, incidence was 43% (15/35), 26% (9/35), and 17% (6/35). G≥3 QT prolongation was reported in 7 (12%) pts during induction, 1 (33%) during consolidation, and 0 pts during maintenance. Median duration of IVO treatment in maintenance was 589 d (range 12–2871). During maintenance, at least 1 TEAE was reported in 17/19 (90%) pts, and G≥3 TEAEs were reported in 6 pts (32%), including acute renal failure and neutropenia in 1 pt each (5%). TEAEs leading to treatment interruption during maintenance were reported in 6 pts (32%), and discontinuation in 1 pt (5%); no TEAE-related deaths were reported. Conclusions Addition of IVO to intensive induction and consolidation chemo followed by single-agent IVO maintenance produces long-term responses with an acceptable safety profile and can serve as bridge to HSCT if needed. IVO maintenance has an acceptable safety profile, is associated with stable normalization of blood counts, and results in durable responses and long-term survival across comutational profiles. The benefit of this frontline regimen is being assessed in a phase 3 randomized, blinded trial (NCT03839771).

Abstract: In the phase 3 AGILE study, after a 12.4-month median follow-up, ivosidenib, a mutant isocitrate dehydrogenase 1 (IDH1) inhibitor, combined with azacitidine significantly improved event-free survival, overall survival (OS), and complete remission rates compared with placebo-azacitidine in patients with newly diagnosed IDH1-mutated acute myeloid leukemia (AML), who were unfit for intensive chemotherapy. This post hoc analysis reports long-term follow-up results from AGILE after a median follow-up of 28.6 months. Overall, 148 patients were randomized to receive ivosidenib-azacitidine (n = 73) or placebo-azacitidine (n = 75). Median OS was significantly longer with ivosidenib (29.3 months; 95% confidence interval [CI], 13.2 to not reached) than with placebo (7.9 months; 95% CI, 4.1-11.3; hazard ratio, 0.42 [95% CI, 0.27-0.65]; P < .0001). Hematologic recovery was faster, more durable, and conversion to transfusion independence (53.8% vs 17.1%; P = .0004) was more common with ivosidenib than with placebo. Of 33 ivosidenib-treated patients evaluable for molecular measurable residual disease (MRD), 10 converted to MRD negativity. Although OS did not differ significantly between MRD-negative and MRD-positive responders at the 0.1% variant allele frequency (VAF) threshold, MRD-negative patients had numerically longer survival. MRD status appeared more predictive of long-term OS when an exploratory 1% VAF threshold was applied. MRD response was not associated with IDH1 variant, VAF, inferred clonality, or number of baseline comutations. The previously reported safety profile was maintained. These long-term efficacy and safety results confirm the benefit of ivosidenib-azacitidine in this challenging-to-treat population and support its use as a standard of care with the longest reported survival benefit for intensive chemotherapy–ineligible patients with IDH1-mutated AML. This trial was registered at www.ClinicalTrials.gov as #NCT03173248.

Vorasidenib and ivosidenib inhibit mutant forms of isocitrate dehydrogenase (mIDH) and have shown preliminary clinical activity against m IDH glioma. We evaluated both agents in a perioperative phase 1 trial to explore the mechanism of action in recurrent low-grade glioma (IGG) and select a molecule for phase 3 testing. Primary end-point was concentration of d -2-hydroxyglutarate (2-HG), the metabolic product of mIDH enzymes, measured in tumor tissue from 49 patients with m IDH1 -R132H nonenhancing gliomas following randomized treatment with vorasidenib (50 mg or 10 mg once daily, q.d.), ivosidenib (500 mg q.d. or 250 mg twice daily) or no treatment before surgery. Tumor 2-HG concentrations were reduced by 92.6% (95% credible interval (CrI), 76.1–97.6) and 91.1% (95% CrI, 72.0–97.0) in patients treated with vorasidenib 50 mg q.d. and ivosidenib 500 mg q.d., respectively. Both agents were well tolerated and follow-up is ongoing. In exploratory analyses, 2-HG reduction was associated with increased DNA 5-hydroxymethylcytosine, reversal of ‘proneural’ and ‘stemness’ gene expression signatures, decreased tumor cell proliferation and immune cell activation. Vorasidenib, which showed brain penetrance and more consistent 2-HG suppression than ivosidenib, was advanced to phase 3 testing in patients with m IDH LGGs. Funded by Agios Pharmaceuticals, Inc. and Servier Pharmaceuticals LLC; ClinicalTrials.gov number NCT03343197.

70 genes selected for validation.

Supplementary Figure 1. Synthesis and pharmacokinetic characterization of AG-221. Supplementary Figure 2. Biochemical attributes of AG-221. Supplementary Figure 3. Biochemistry of AG-221 with respect to substrate and cofactor. Supplementary Figure 4. TF-1 IDH2R140Q cells treated with AG-221. Supplementary Figure 5. AG-221 can reverse the block in EPO-induced differentiation caused by the expression of IDH2R140Q in the TF-1 erythroleukemia cell line. Supplementary Figure 6. Pharmacokinetics/pharmacodynamics of AG-221 in IDH2R140Qmutant U87MG xenograft tumor-bearing mice. Supplementary Figure 7. AG-221 does not affect intrinsic hematological parameters or body weight. Supplementary Figure 8. AG-221 strongly reduces the number of human IDH2R140Q blasts in the liver and spleen in AML-1, AML-2, and AML-3. Supplementary Figure 9. Flow cytometry analyses of bone marrow-derived hCD45+ cells in primary human AML xenograft models. Supplementary Figure 10. Affinity of the IDH2R140Q homodimer for NADPH. Supplementary Figure 11. Electron density map diagrams for bound ligands for IDH2R140Q co-complex structures determined by X-ray crystallography.

X-ray data collection and refinement statistics for DHODH•FMN•ORO•AG-636

Background: Ivosidenib (IVO) + azacitidine (AZA) improved complete remission (CR) rates and overall survival (OS) relative to placebo (PBO) + AZA in patients (pts) with newly diagnosed mutant isocitrate dehydrogenase 1 (m IDH1) acute myeloid leukemia (AML) in the pivotal AGILE study (Montesinos et al. NEJM 2022). Prior analyses showed clinical responses to IVO+AZA were associated with deep clearance of m IDH1 (limit of detection [LOD]: 0.02-0.04% variant allele frequency [VAF]), as well as clearance of baseline co-mutations below the threshold of conventional next-generation sequencing (NGS) (LOD: 2.0% VAF) (Döhner et al. Blood 2022; Daigle et al. Blood 2021). Measurable residual disease (MRD), a negative prognostic marker, following chemotherapy is often detected via multiparametric flow cytometry or quantitative polymerase chain reaction although both techniques have their limitations. (Heuser et al. Blood 2021). NGS can detect a wide range of AML mutations and studies show that NGS-assessed MRD is highly predictive of both relapse and survival (Thol et al. Blood 2018; Tsai et al. Blood Adv 2021). However, a limitation of NGS is difficulty in distinguishing between somatic leukemia-associated mutations and clonal hematopoiesis of indeterminate potential (CHIP) (Bacher et al. Blood Cancer J 2018). We aimed to perform an NGS analysis of MRD responses in the AGILE study. Methods: All AGILE pts who had a best overall response (BOR) of CR, CR with incomplete count recovery, or CR with incomplete platelet recovery, and had ≥1 on-treatment bone marrow mononuclear cell (BMMC) sample available were included. Suggested timepoints for MRD testing included day 1 of treatment cycles 3, 5, 7, 9, 11, 14, 20, 26, and 32. Baseline and on-treatment BMMC DNA samples were analyzed using a diagnostic 51-gene myeloid NGS panel (LOD: 3% VAF) and a 26-gene AML MRD panel, respectively. For variants detected at baseline, the LOD was 0.1% VAF; for variants not detected at baseline (either emerging mutations or if baseline samples were unavailable), the LOD was 0.5% VAF. All variants above the applicable LOD with known or potential clinical significance were considered evidence of positive MRD, aside from DNMT3A, TET2 and ASXL1 (“DTA”) mutations to reduce the risk of false-positive MRD due to CHIP. Results: The analysis set comprised 173 samples from 33 IVO+AZA-treated pts and 10 PBO+AZA-treated pts (Fig. 1). Clinical response data were based on a March 18, 2021 primary analysis data cut; OS data were based on an updated June 30, 2022 data cut. The median number of baseline mutations was 4 (range: 1-10) per patient, median number of MRD assessments was 3 (range: 1-9) and median follow-up period was 189 days (range: 49-875). Ten (30.3%) IVO+AZA-treated pts became MRD-negative (MRD neg), compared with 2 (20.0%) PBO+AZA-treated pts. All pts attaining an MRD neg response had a BOR of CR (10 [34.5%] of 29 IVO+AZA-treated pts in CR; 2 [22.2%] of 9 PBO+AZA-treated pts in CR). Seven (70.0%) of the 10 MRD neg IVO+AZA-treated responding pts converted to an MRD neg response by day 1 of cycle 7 (C7D1). Fewer pts with an MRD neg response had an Adverse risk at baseline according to European LeukemiaNet 2022 genetic risk classification versus those pts with an MRD-positive (MRD pos) response (26.3% vs 73.7%) (Table 1). In the IVO+AZA arm, rates of event-free survival at 12 months and OS at 24 months were numerically higher in MRD neg vs MRD pos responders (80.0% vs 60.6% and 88.9% vs 72.3%, respectively) but did not reach statistical significance. All 5 (15.2%) IVO+AZA-treated pts with confirmed relapses had detectable MRD at the last completed assessment prior to relapse. One pt converted from an MRD neg to an MRD pos response (“MRD relapse”) 169 days prior to overt clinical relapse. Among the 23 IVO+AZA-treated pts without an MRD neg response, 16 were alive as of the updated data cut; OS duration ranged from 20.3-48.9 months. Conclusions: IVO+AZA induced molecular MRD negativity in approximately one third of responding pts with newly diagnosed m IDH1 AML; MRD neg responses were most often observed on or before C7D1. Interestingly, several pts had durable clinical responses and OS despite ongoing MRD positivity, potentially due to persistent CHIP rather than true MRD, illustrating the complexity of interpreting NGS MRD data. Additional analyses, including associations between specific molecular alterations, MRD response, and OS are ongoing and will be presented.

Genome-wide CRISPR depletion screen data

7036 Background: Ivosidenib (IVO) twas approved based on a CR/CRh rate of 31.8% in IDH1 mutant (IDH1m) R/R AML; median duration of CR/CRh (DOR) was 8.2 months (mo). The purpose of this study is to elucidate the clinical and molecular characteristics of exceptional responders to single agent IVO. Methods: We analyzed all patients (pts) with IDH1m R/R AML who received IVO 500mg QD on the phase 1 dose escalation/expansion study and had a DOR of >12 mo. Results: Of 179 pts who received IVO 500 mg QD, 57 (31.8%) achieved a CR/CRh. Of these, 20 (35.1%) had a DOR >12 mo. After excluding 7 pts who went to HSCT, 13 (22.8% of responders, 7.3% of cohort) had DOR >12 mo (exceptional response) and 8 (14% of responders, 4.5% of cohort) had DOR >24 mo. The 13 exceptional responders all achieved CR as best response, and median DOR was 42.6 mo. Median OS was not reached; estimated OS at 48 mo was 67.1%. Median EFS was 44.4 mo. Reasons for discontinuing treatment included relapse (n=4), adverse event unrelated to IVO (n=3), and patient (pt) decision (n=3). Three pts remain on treatment. All pts who relapsed had DOR between 1-2 years; none of the 8 pts with DOR >2 years relapsed. Baseline characteristics are in the table. The most common co-mutations were DNMT3A, ASXL1, SRSF2, and JAK2 (3 each, 23%). Five pts (39%) had a splicing factor mutation. No pts had FLT3 or RTK pathway mutations other than JAK2. Only 1 pt with NPM1 had an exceptional response, and 1 pt had TP53 (VAF 3.6%). Six pts (46%) had an abnormal karyotype, 6 normal (46%), and 1 missing (8%). Most pts had intermediate-risk cytogenetics (10 pts, 77%), 2 poor-risk, and 1 missing. The median IDH1 VAF was 25% (range 10-50%). The median number of co-mutations was 1 (range 0-6). Having 0-1 co-mutations was associated with a longer response. Eleven pts (85%) had an IDH1 R132C mutation; 1 each (7.7%) had R132H and R132L mutations. Mutation clearance of IDH1 was noted in 8 pts (61.5%). Conclusions: A subset of pts with IDH1m R/R AML have prolonged CR on single agent IVO without HSCT (22.8% of CR/CRh responders) and no pts in CR for >2 years (14% of CR/CRh responders) relapsed. A low mutational burden, lack of RTK pathway mutations and canonical AML drivers, and co-occurrence of mutations associated with clonal hematopoiesis appear to be associated with exceptional response. Clinical trial information: NCT02074839 . [Table: see text]

Supplementary Table 4 from Comparison of Human and Rat Uterine Leiomyomata: Identification of a Dysregulated Mammalian Target of Rapamycin Pathway

Supplementary methods, references and figures

Abstract INTRODUCTION Vorasidenib is an oral, brain-penetrant, inhibitor of mutant isocitrate dehydrogenase (mIDH) 1/2 enzymes. The INDIGO study (NCT04164901) showed significantly improved radiographic progression-free survival (PFS) by blinded independent review committee (BIRC) with vorasidenib, compared with placebo, in patients with mIDH1/2 adulttype diffuse glioma (hazard ratio [HR] 0.39, 95% CI 0.27–0.56; one-sided P = 0.000000067 [Mellinghoff N Engl J Med 2023]). The key secondary endpoint of time-to-next-intervention was also met. METHODS In this double-blind Phase 3 study, patients aged ≥ 12 years with residual/recurrent grade 2 mIDH1/2 oligodendroglioma or astrocytoma, measurable non-enhancing disease, and no prior treatment for glioma were randomized 1:1 to receive vorasidenib 40 mg or placebo daily in 28-day cycles. An investigational clinical trial assay centrally confirmed IDH1 R132H/C/G/S/L or IDH2 R172K/M/W/S/G mutation variants. Pretreatment (archival) tumor tissue was analyzed by next-generation sequencing for CDKN2A/B homozygous deletion and other co-mutations (ACE Extended Cancer Panel). RESULTS As of Sep 6, 2022 (preplanned second interim analysis), 168 patients were randomized to vorasidenib and 163 to placebo (median age, 40.0 years; Karnofsky performance scale = 100, 53.5%; oligodendroglioma, 172; astrocytoma, 159; mIDH1, 315; mIDH2, 16). Two patient subgroups were defined by baseline IDH1/2 variant allele frequency (VAF): lower/higher than the median value (0.377). The median PFS results favored vorasidenib over placebo in both subgroups (low VAF, one-sided P = 0.0116; high VAF, one-sided P < 0.0001). Vorasidenib and placebo groups had similar and a low number of co-mutations in known/likely oncogenic genes at baseline (median number of co-mutations across both groups was 4). CDKN2A homozygous deletion was detected in only two participants (both in the placebo group). Additional data will be presented. CONCLUSION In the first randomized Phase 3 study of a targeted therapy in grade 2 mIDH1/2 glioma, vorasidenib prolonged median PFS by BIRC, relative to placebo, irrespective of IDH1/2 VAF.

Supplemental Methods. DNA constructs, Gel electrophoresis and western blotting, Metabolite extraction and analysis

Supplementary Table 1. Drug metabolism and pharmacokinetic attributes of AG-221. Supplementary Table 2. Selectivity of AG-221 confirmed by testing against a panel of kinases. Supplementary Table 3. Clinical characteristics of patients with IDH2R140Q-mutated AML. Supplementary Table 4. Treated NSG mice (AML-1, AML-2, AML-3) engrafted with human IDH2R140Q mononuclear cells display stable levels of AG-221 in serum. Supplementary Table 5. AG-221 inhibits 2HG production in models AML-1, AML-2, and AML-3. Supplementary Table 6. Summary of pharmacokinetics/pharmacodynamics in primary human acute myeloid leukemia xenograft model (AML-4). Supplementary Table 7. Summary of data collection and refinement statistics. Supplementary Table 8. Percentage of human chimerism in peripheral blood in models AML-1 and AML-2.

Mutations in isocitrate dehydrogenase 1 and 2 (IDH1/2) occur in ~23% of patients with acute myeloid leukemia (AML) and ~7% with myelodysplastic syndrome (MDS).1, 2 These mutations lead to the accumulation of the oncometabolite D-2-hydroxyglutarate (2-HG), which competitively inhibits α-ketoglutarate-dependent enzymes, causing epigenetic dysregulation and impaired hematopoietic differentiation, ultimately promoting oncogenesis.2, 3 Selective mutant IDH (mIDH) 1 and 2 inhibitors (ivosidenib, olutasidenib and enasidenib) are approved for patients with hematologic cancers, but isoform switching between mIDH1 and mIDH2 has emerged as a mechanism of resistance.4 Vorasidenib (AG-881) is an investigational, dual mIDH1/2 inhibitor that may prevent isoform switching.5 We conducted an open-label, first-in-human, phase 1, dose-escalation study of vorasidenib in patients with mIDH1/2 advanced hematologic malignancies (NCT02492737). Eligibility criteria included: aged ≥18 years, having AML and failed prior treatment with an approved IDH inhibitor, or having other relapsed/refractory advanced hematologic malignancies and having failed previous standard therapy, and having a documented IDH1 and/or IDH2 mutation. Ineligibility criteria were: hematopoietic stem cell transplant within 60 days or systemic anticancer therapy or investigational agent within 14 days of vorasidenib first dose. All patients provided written informed consent. The study adhered to the Declaration of Helsinki and the International Council for Harmonisation Guidelines for Good Clinical Practice. The protocol was approved by each study site's institutional review board/ethics committee. All patients received vorasidenib on day −3, followed by safety and pharmacokinetic/pharmacodynamic (PK/PD) assessments over 72 h. Vorasidenib was then administered orally, once daily (QD), in 28-day cycles. Seven vorasidenib dose levels were tested (from 25 to 1100 mg). Criteria for dose escalation and dose-limiting toxicities are detailed in the Supplementary Material. Maximum tolerated dose (MTD) was assessed in all patients who experienced a dose-limiting toxicity (DLT) during cycle 1 or who received ≥75% of vorasidenib doses during cycle 1 and had sufficient safety data to conclude that a DLT had not occurred. The full analysis set included patients who received ≥1 vorasidenib dose. Primary objectives were to determine the MTD, or recommended phase 2 dose, and to assess the safety and tolerability of vorasidenib in this population. Patients were enrolled from July 23, 2015, to March 21, 2018, at sites in the United States and France (Table S1). Baseline demographics and clinical characteristics are summarized in Table S2. Overall, 46 patients received vorasidenib: 34 with AML, 11 with MDS, and 1 with angioimmunoblastic T-cell lymphoma (AITL). In total, 31 (67.4%) patients had IDH2 mutations only (mean variant allele frequency [VAF] 34.92% [standard deviation (SD), 8.59%; range, 13.32–45.03%; n = 18]); 9 (19.6%) patients had IDH1 mutations only (mean VAF 37.49% [SD, 2.58; range, 34.59–41.13%; n = 5]); and 4 (8.7%) patients had co-occurring IDH1 and IDH2 mutations (of these, 2 had received prior enasidenib therapy, 1 had received prior ivosidenib therapy, and 1 patient with MDS had not received prior IDH-inhibitor therapy). Overall, 35/46 (76.1%; 34 with AML, 1 with MDS) patients received an IDH inhibitor before enrollment; 16/46 (34.8%) patients received an IDH inhibitor as their last prior regimen (all had AML). The median number of vorasidenib treatment cycles was 3 (range, 0–24); median duration of treatment was 2.2 (range, 0.0–22.6) months. Overall, 8/46 (17.4%) patients had exposures to vorasidenib lasting ≥6 months (Figure 1A), of whom 5 were prior IDH-inhibitor naive. Of the patients with AML, 21/34 (61.8%) discontinued vorasidenib due to progressive disease (PD), 10/34 (29.4%) due to adverse events (AEs), and 3/34 (8.8%) due to withdrawal of consent. The patient with AITL discontinued vorasidenib due to PD. Of patients with MDS, 4/11 (36.4%) discontinued due to PD, 2/11 (18.2%) due to AEs, 2/11 (18.2%) to proceed to stem cell transplant, 1/11 (9.1%) withdrew consent, and 2/11 (18.2%) were transferred to single-patient FDA Investigational New Drug Applications and continued to receive vorasidenib for >5 years. All patients experienced ≥1 AE, and 4/46 (8.7%) patients experienced serious AEs assessed as treatment-related (Table S3). The most common AEs were fatigue, increased ALT, increased blood alkaline phosphatase, and diarrhea (Table S4). Although no DLTs occurred, 1100 mg QD was deemed intolerable due to AEs of increased alanine aminotransferase (ALT) in 5/6 patients (83.3%, Table S4), with 4/6 (66.6%) experiencing grade 2 elevations. Vorasidenib 600 mg QD was the last tolerable dose evaluated. No on-treatment deaths (n = 9) or discontinuations due to AEs (n = 10) were considered treatment-related (Tables S5 and S6). The best overall response (BOR) for patients with AML was morphologic leukemia-free state (MLFS) in 2/34 (5.9%) patients; 22/34 (64.7%) patients experienced stable disease; 6/34 (17.6%) patients had PD; response was not evaluable in 4/34 (11.8%) patients (Figure 1A). In patients with MDS, BOR was complete remission (CR) in 3/11 patients (27.3%) and marrow CR in 1/11 (9.1%) patients; 5/11 (45.5%) patients experienced stable disease; treatment failure and PD were experienced by 1/11 (9.1%) patients each. Objective response rates were 5.9% (95% confidence interval [CI], 0.7–19.7) for patients with AML, 36.4% (95% CI, 10.9–69.2) for patients with MDS, and 27.3% for prior IDH inhibitor−naive patients. Overall, 6/31 (19.4%) patients who were transfusion−dependent at baseline became transfusion independent, per the definition of ≥28 days between consecutive transfusions. PK/PD were evaluated in all patients who received ≥1 vorasidenib dose and had ≥1 evaluable blood sample (Table S7). At cycle 1 day 15, vorasidenib had a median Tmax of 0.97–2.0 h across dose levels, and geometric means for Cmax and AUC0–24 were correlated with increasing vorasidenib dose levels up to 600 mg QD; Cmax and AUC0–24 were lower for the 1100 mg QD dose than 600 mg QD (Table S7, Figure S1A). Across the study, plasma 2-HG levels generally decreased with increasing plasma vorasidenib concentrations (Table S7, Figure 1B, Figure S1B). In patients who received vorasidenib 400 mg QD, mean reductions in plasma 2-HG BRmin of −87.0% and −89.3% were observed at cycle 1 day 15 and cycle 2 day 1, respectively. Decreased plasma 2-HG levels were more pronounced in patients with MDS without prior IDH-inhibitor treatment than in those who had previously received IDH inhibitors, regardless of vorasidenib dose received (Figure 1C), though the numbers in this analysis were small. Of the patients with IDH1 and IDH2 comutations, 3/4 experienced a BOR of stable disease; plasma 2-HG levels reduced from 195, 1320, and 1960 ng/mL at screening to on-treatment minimums of 75.3, 65.3, and 109 ng/mL, respectively. Figure S2 shows baseline comutation data for patients grouped by BOR. Figure S3 illustrates VAF–PK/PD relationships in patients with MDS or AITL and AML, and lists their BOR. There was no change in IDH mutation burden (VAF) in most patients who received vorasidenib. Of the 3 patients who achieved a CR, all displayed a decrease in 2-HG, 1 displayed a decrease in VAF and no VAF data were available for the other 2, 2 did not report myeloblasts at baseline, and 1 reported 14% myeloblasts at baseline. Although no DLTs were reported with vorasidenib 1100 mg QD, the clinical study team determined this dose was not tolerable. Plasma pharmacokinetic parameters for vorasidenib are generally negatively correlated with plasma 2-HG pharmacodynamic parameters, with decreases in plasma 2-HG greatest in patients who received vorasidenib 400 mg QD. This may have been a result of CYP3A4 auto-induction occurring at the higher dose levels and/or the interrupted vorasidenib dosing in the 1100 mg QD cohort owing to ALT AEs. Based on the former, it is plausible that further plasma 2-HG decreases would not be achieved with doses >600 mg QD. A dose-exposure/2-HG relationship could not be established in patients who received >600 mg QD, suggesting doses from 400 to 600 mg QD could be feasible. For most patients with AML, the BOR on vorasidenib treatment was stable disease, while 3 patients with MDS achieved CR. The difference in objective response rate for patients with AML and MDS is likely related to the enrollment criteria, which required patients with AML to have failed prior IDH-inhibitor therapy. Thus, all patients with AML had received prior enasidenib or ivosidenib, versus only 1/11 patients with MDS. It is likely that the patients with AML in this study were more resistant to IDH-inhibitor monotherapy, resulting in less favorable outcomes with vorasidenib treatment. Given its ability to inhibit both mIDH1 and mIDH2, vorasidenib is predicted to prevent isoform switching from evolving as a resistance mechanism if administered as initial IDH-inhibitor therapy. For instance, in 1 patient with a dual IDH mutation at baseline who previously achieved CRi on enasidenib treatment and then relapsed with a new IDH1 second-site mutation, a reduction in IDH1 VAF with vorasidenib was observed. In conclusion, in this patient population, vorasidenib treatment was tolerable up to doses of 600 mg QD. 2-HG suppression and evidence of clinical activity were observed in previously mIDH inhibitor−naive patients, but the suboptimal dose-efficacy profile in AML patients previously treated with mIDH inhibitors limited the evaluation of vorasidenib's dual mutant-enzyme inhibitor properties to prevent isoform switching in this population; therefore, development of vorasidenib in AML was not pursued. Based on vorasidenib's brain-penetrant activity,5 its development is currently focused on patients with mIDH glioma.6 C.D.D., S.d.B., D.A.P., R.M.S., J.K.A., A.T.F., and E.M.S. participated in the recruitment and treatment of patients in the trial and the collection of data. T.L., M.L., S.C., M.H., A.E.T., Q.M., S.M.K., and S.S.P. performed the data analysis. S.C., M.H., A.E.T., Q.M., and S.M.K. oversaw drafting of the manuscript. All authors participated in clinical data interpretation and manuscript development, and approved the submitted version. Medical writing assistance was provided by David Pertab, PhD, and Christine Ingleby, DPhil, CMPP, Excel Medical Affairs, Glasgow, UK, and supported by Servier Pharmaceuticals LLC. The study was supported by Agios Pharmaceuticals, Inc. Servier Pharmaceuticals LLC, which funded medical writing assistance, has completed acquisition of Agios' oncology business. C.D.D. has consulted for AbbVie and Servier; received research funding from AbbVie, Astex, BMS, Cleave, Foghorn, Forma, Immune-Onc, Loxo, and Servier; received honoraria from Astellas, Bluebird Bio, BMS, Foghorn, Gilead, Immune-Onc, Jazz, Novartis, Kura, Servier, and Takeda; has a membership of the board of directors or advisory roles with Genmab, GSK, Kura, and Notable Labs; and has stock options with Notable Labs. S.D.B. has received honoraria from AbbVie, Astellas Pharma, Bristol Myers Squibb, Jazz Pharmaceuticals, and Servier; has acted as a consultant or advisor to Bristol Myers Squibb, GlaxoSmithKline, Servier, and Syndax; has participated in speakers' bureau for AbbVie, Astellas Pharma, Bristol Myers Squibb, Jazz Pharmaceuticals, and Servier; has received research funding to his institution from Auron Therapeutics and Forma Therapeutics; and has received travel, accommodations, or expenses from AbbVie and Servier. D.A.P. has acted as a consultant or advisor to AbbVie, Celgene, Bristol Myers Squibb, Takeda, Foghorn, Aprea, Genentech, Syros, Novartis, Gilead, Astellas, Karyopharm, Syndax, Jazz, Bergen Bio, Arcellx, AstraZeneca, Kura, Ryvu, Magenta, Qihan, Zentalis, Medivir, HiberCell, and has received research funding from AbbVie, Bristol Myers Squibb, Teva, and Karyopharm. R.M.S. reports grants and personal fees from AbbVie, Agios, and Novartis; personal fees from Actinium, Astellas, BioLineRx, Celgene, Daiichi-Sankyo, Elevate, GEMoaB, Janssen, Jazz, MacroGenics, Onconova, Syndax, Syntrix, Syros, Takeda, Trovagene, BerGenBio, Foghorn Therapeutics, GlaxoSmithKline, Aprea, Innate, Amgen, BMS, Boston Pharmaceuticals, Aptevo, Epizyme, Kura Oncology, and grants from AROG, outside the submitted work. J.K.A. has acted as a consultant or advisor for AbbVie, Astellas Pharma, BioSight, Bluebird Bio, Curio, Gilead, Kura Oncology, Kymera, Stemline Therapeutics, and Syros; received research funding from Servier; and received travel, accommodations, or travel-related expenses from BioSight. A.T.F. has consulted for AbbVie, Agios, Amgen, Astellas, Bristol Myers Squibb, Celgene, EnClear, Forma, Genentech, Immunogen, Ipsen, Kite, Mablytics, Novartis, Orum, PureTech, Servier, and Takeda; and has received funding for clinical trials from AbbVie, Agios/Servier, and Celgene/Bristol Myers Squibb. E.M.S. has acted as a consultant or advisor to Novartis, Janssen, BMS/Celgene, Agios, Jazz Pharmaceuticals, Menarini, Genentech, Genesis Pharma, AbbVie, Neoleukin Therapeutics, Gilead Sciences, Syndax, OnCusp Therapeutics, Immunogen, CTI BioPharma Corp, Foghorn Therapeutics, Servier, Calithera Biosciences, Daiichi Sankyo, Aptose Biosciences, Ono Pharmaceutical, Blueprint Medicines, GEMoaB, Jnana Therapeutics, and Debiopharm Group; has stock or ownership in Auron Therapeutics; and has received research funding to his institution from Eisai, BMS/Celgene, Bayer, Agios, BioTheryX, Syros Pharmaceuticals, Servier, Foghorn Therapeutics, Syndax, Gilead Sciences, Cleave Biosciences, Prelude Therapeutics, and Loxo/Lilly. T.L., S.C., M.H., A.E.T., Q.M., S.M.K., and S.S.P. are employees of Servier. M.L., S.C., M.H., A.E.T., S.M.K., and S.S.P. were employees of Agios at the time of conducting these studies. All patients provided written informed consent before participation. Study-level clinical data from this study (including the protocol) will be made available upon reasonable request from a qualified medical or scientific professional for the specific purpose laid out in that request and may include deidentified individual participant data. The data for this request will be available after a data access agreement has been signed. Please send your data sharing request to https://clinicaltrials.servier.com/data-request-portal/. Data S1. Supporting Information. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.