MOSAIC: An Artificial Intelligence-Based Framework for Multimodal Analysis, Classification, and Personalized Prognostic Assessment in Rare Cancers.

PURPOSE: Rare cancers constitute over 20% of human neoplasms, often affecting patients with unmet medical needs. The development of effective classification and prognostication systems is crucial to improve the decision-making process and drive innovative treatment strategies. We have created and implemented MOSAIC, an artificial intelligence (AI)-based framework designed for multimodal analysis, classification, and personalized prognostic assessment in rare cancers. Clinical validation was performed on myelodysplastic syndrome (MDS), a rare hematologic cancer with clinical and genomic heterogeneities. METHODS: We analyzed 4,427 patients with MDS divided into training and validation cohorts. Deep learning methods were applied to integrate and impute clinical/genomic features. Clustering was performed by combining Uniform Manifold Approximation and Projection for Dimension Reduction + Hierarchical Density-Based Spatial Clustering of Applications with Noise (UMAP + HDBSCAN) methods, compared with the conventional Hierarchical Dirichlet Process (HDP). Linear and AI-based nonlinear approaches were compared for survival prediction. Explainable AI (Shapley Additive Explanations approach [SHAP]) and federated learning were used to improve the interpretation and the performance of the clinical models, integrating them into distributed infrastructure. RESULTS: UMAP + HDBSCAN clustering obtained a more granular patient stratification, achieving a higher average silhouette coefficient (0.16) with respect to HDP (0.01) and higher balanced accuracy in cluster classification by Random Forest (92.7% ± 1.3% and 85.8% ± 0.8%). AI methods for survival prediction outperform conventional statistical techniques and the reference prognostic tool for MDS. Nonlinear Gradient Boosting Survival stands in the internal (Concordance-Index [C-Index], 0.77; SD, 0.01) and external validation (C-Index, 0.74; SD, 0.02). SHAP analysis revealed that similar features drove patients' subgroups and outcomes in both training and validation cohorts. Federated implementation improved the accuracy of developed models. CONCLUSION: MOSAIC provides an explainable and robust framework to optimize classification and prognostic assessment of rare cancers. AI-based approaches demonstrated superior accuracy in capturing genomic similarities and providing individual prognostic information compared with conventional statistical methods. Its federated implementation ensures broad clinical application, guaranteeing high performance and data protection.

Personalized Timing for Allogeneic Stem-Cell Transplantation in Hematologic Neoplasms: A Target Trial Emulation Approach Using Multistate Modeling and Microsimulation.

PURPOSE: Decision about the optimal timing of a treatment procedure in patients with hematologic neoplasms is critical, especially for cellular therapies (most including allogeneic hematopoietic stem-cell transplantation [HSCT]). In the absence of evidence from randomized trials, real-world observational data become beneficial to study the effect of the treatment timing. In this study, a framework to estimate the expected outcome after an intervention in a time-to-event scenario is developed, with the aim of optimizing the timing in a personalized manner. METHODS: Retrospective real-world data are leveraged to emulate a target trial for treatment timing using multistate modeling and microsimulation. This case study focuses on myelodysplastic syndromes, serving as a prototype for rare cancers characterized by a heterogeneous clinical course and complex genomic background. A cohort of 7,118 patients treated according to conventional available treatments/evidence across Europe and United States is analyzed. The primary clinical objective is to determine the ideal timing for HSCT, the only curative option for these patients. RESULTS: This analysis enabled us to identify the most appropriate time frames for HSCT on the basis of each patient's unique profile, defined by a combination relevant patients' characteristics. CONCLUSION: The developed methodology offers a structured framework to address a relevant clinical issue in the field of hematology. It makes several valuable contributions: (1) novel insights into how to develop decision models to identify the most favorable HSCT timing, (2) evidence to inform clinical decisions in a real-world context, and (3) the incorporation of complex information into decision making. This framework can be applied to provide medical insights for clinical issues that cannot be adequately addressed through randomized clinical trials.

Clinical and Genomic-Based Decision Support System to Define the Optimal Timing of Allogeneic Hematopoietic Stem-Cell Transplantation in Patients With Myelodysplastic Syndromes.

PURPOSE: Allogeneic hematopoietic stem-cell transplantation (HSCT) is the only potentially curative treatment for patients with myelodysplastic syndromes (MDS). Several issues must be considered when evaluating the benefits and risks of HSCT for patients with MDS, with the timing of transplantation being a crucial question. Here, we aimed to develop and validate a decision support system to define the optimal timing of HSCT for patients with MDS on the basis of clinical and genomic information as provided by the Molecular International Prognostic Scoring System (IPSS-M). PATIENTS AND METHODS: We studied a retrospective population of 7,118 patients, stratified into training and validation cohorts. A decision strategy was built to estimate the average survival over an 8-year time horizon (restricted mean survival time [RMST]) for each combination of clinical and genomic covariates and to determine the optimal transplantation policy by comparing different strategies. RESULTS: Under an IPSS-M based policy, patients with either low and moderate-low risk benefited from a delayed transplantation policy, whereas in those belonging to moderately high-, high- and very high-risk categories, immediate transplantation was associated with a prolonged life expectancy (RMST). Modeling decision analysis on IPSS-M versus conventional Revised IPSS (IPSS-R) changed the transplantation policy in a significant proportion of patients (15% of patient candidate to be immediately transplanted under an IPSS-R-based policy would benefit from a delayed strategy by IPSS-M, whereas 19% of candidates to delayed transplantation by IPSS-R would benefit from immediate HSCT by IPSS-M), resulting in a significant gain-in-life expectancy under an IPSS-M-based policy (P = .001). CONCLUSION: These results provide evidence for the clinical relevance of including genomic features into the transplantation decision making process, allowing personalizing the hazards and effectiveness of HSCT in patients with MDS.

Toward a more patient-centered drug development process in clinical trials for patients with myelodysplastic syndromes/neoplasms (MDS): Practical considerations from the International Consortium for MDS (icMDS).

Notable treatment advances have been made in recent years for patients with myelodysplastic syndromes/neoplasms (MDS), and several new drugs are under development. For example, the emerging availability of oral MDS therapies holds the promise of improving patients' health-related quality of life (HRQoL). Within this rapidly evolving landscape, the inclusion of HRQoL and other patient-reported outcomes (PROs) is critical to inform the benefit/risk assessment of new therapies or to assess whether patients live longer and better, for what will likely remain a largely incurable disease. We provide practical considerations to support investigators in generating high-quality PRO data in future MDS trials. We first describe several challenges that are to be thoughtfully considered when designing an MDS-focused clinical trial with a PRO endpoint. We then discuss aspects related to the design of the study, including PRO assessment strategies. We also discuss statistical approaches illustrating the potential value of time-to-event analyses and their implications within the estimand framework. Finally, based on a literature review of MDS randomized controlled trials with a PRO endpoint, we note the PRO items that deserve special attention when reporting future MDS trial results. We hope these practical considerations will facilitate the generation of rigorous PRO data that can robustly inform MDS patient care and support treatment decision-making for this patient population.

Accelerated DNA replication fork speed due to loss of R-loops in myelodysplastic syndromes with SF3B1 mutation.

Myelodysplastic syndromes (MDS) with mutated SF3B1 gene present features including a favourable outcome distinct from MDS with mutations in other splicing factor genes SRSF2 or U2AF1. Molecular bases of these divergences are poorly understood. Here we find that SF3B1-mutated MDS show reduced R-loop formation predominating in gene bodies associated with intron retention reduction, not found in U2AF1- or SRSF2-mutated MDS. Compared to erythroblasts from SRSF2- or U2AF1-mutated patients, SF3B1-mutated erythroblasts exhibit augmented DNA synthesis, accelerated replication forks, and single-stranded DNA exposure upon differentiation. Importantly, histone deacetylase inhibition using vorinostat restores R-loop formation, slows down DNA replication forks and improves SF3B1-mutated erythroblast differentiation. In conclusion, loss of R-loops with associated DNA replication stress represents a hallmark of SF3B1-mutated MDS ineffective erythropoiesis, which could be used as a therapeutic target.

Molecular and clinical presentation of UBA1-mutated myelodysplastic syndromes.

Mutations in UBA1, which are disease-defining for VEXAS syndrome, have been reported in patients diagnosed with myelodysplastic syndromes (MDS). Here, we define the prevalence and clinical associations of UBA1 mutations in a representative cohort of patients with MDS. Digital droplet PCR profiling of a selected cohort of 375 male patients lacking MDS disease-defining mutations or established WHO disease classification identified 28 patients (7%) with UBA1 p.M41T/V/L mutations. Using targeted sequencing of UBA1 in a representative MDS cohort (n=2,027), we identified an additional 27 variants in 26 patients (1%), which we classified as likely/pathogenic (n=12) and unknown significance (n=15). Among the total 40 patients with likely/pathogenic variants (2%), all were male and 63% were classified by WHO2016 as MDS-MLD/SLD. Patients had a median of one additional myeloid gene mutation, often in TET2 (n=12), DNMT3A (n=10), ASXL1 (n=3), or SF3B1 (n=3). Retrospective clinical review where possible showed that 83% (28/34) UBA1-mutant cases had VEXAS-associated diagnoses or inflammatory clinical presentation. The prevalence of UBA1-mutations in MDS patients argues for systematic screening for UBA1 in the management of MDS.

Olutasidenib in post-venetoclax patients with mutant isocitrate dehydrogenase 1 (mIDH1) acute myeloid leukemia (AML).

Olutasidenib, a potent, selective, oral, mutant isocitrate dehydrogenase 1 (mIDH1) inhibitor, is FDA-approved for relapsed/refractory (R/R) acute myeloid leukemia (AML). Here we report efficacy and safety of olutasidenib in 18 patients with mIDH1 AML who were relapsed (10), refractory (6) or had complete remission with incomplete hematologic recovery (CRi; 2) to a venetoclax combination. Of the 16 patients who were R/R, 4 (25%) achieved complete remission (CR), one (6.3%) achieved CR with partial hematologic recovery (CRh), and 7 (43.8%) achieved a composite complete remission (CRc). Median time to CRc was 1.9 months (range 1-2.8). As of data cutoff (18 June 2021), median duration of CRc was not reached (range, 1.2-NR, ongoing at 30.4+ months). Both patients with CRi at study entry achieved a CR. Safety was consistent with the overall profile of olutasidenib. Olutasidenib offers a valuable treatment option for patients with mIDH1 AML previously treated with venetoclax.

Venetoclax and Cobimetinib in Relapsed/Refractory AML: A Phase 1b Trial.

BACKGROUND: Therapies for relapsed/refractory acute myeloid leukemia remain limited and outcomes poor, especially amongst patients who are ineligible for cytotoxic chemotherapy or targeted therapies. PATIENTS AND METHODS: This phase 1b trial evaluated venetoclax, a B-cell lymphoma-2 (BCL-2) inhibitor, plus cobimetinib, a MEK1/2 inhibitor, in patients with relapsed/refractory acute myeloid leukemia, ineligible for cytotoxic chemotherapy. Two-dimensional dose-escalation was performed for venetoclax dosed daily, and for cobimetinib dosed on days 1-21 of each 28-day cycle. RESULTS: Thirty patients (median [range] age: 71.5 years [60-84]) received venetoclax-cobimetinib. The most common adverse events (AEs; in ≥40.0% of patients) were diarrhea (80.0%), nausea (60.0%), vomiting (40.0%), febrile neutropenia (40.0%), and fatigue (40.0%). Overall, 66.7% and 23.3% of patients experienced AEs leading to dose modification/interruption or treatment withdrawal, respectively. The composite complete remission (CRc) rate (complete remission [CR] + CR with incomplete blood count recovery + CR with incomplete platelet recovery) was 15.6%; antileukemic response rate (CRc + morphologic leukemia-free state/partial remission) was 18.8%. For the recommended phase 2 dose (venetoclax: 600 mg; cobimetinib: 40 mg), CRc and antileukemic response rates were both 12.5%. Failure to achieve an antileukemic response was associated with elevated baseline phosphorylated ERK and MCL-1 levels, but not BCL-xL. Baseline mutations in ≥1 signaling gene or TP53 were noted in nonresponders and emerged on treatment. Pharmacodynamic biomarkers revealed inconsistent, transient inhibition of the mitogen-activated protein kinase (MAPK) pathway. CONCLUSION: Venetoclax-cobimetinib showed limited preliminary efficacy similar to single-agent venetoclax, but with added toxicity. Our findings will inform future trials of BCL-2/MAPK pathway inhibitor combinations.

Guadecitabine vs TC in relapsed/refractory AML after intensive chemotherapy: a randomized phase 3 ASTRAL-2 trial.

ABSTRACT: Guadecitabine is a novel hypomethylating agent (HMA) resistant to deamination by cytidine deaminase. Patients with relapsed/refractory acute myeloid leukemia (AML) were randomly assigned to guadecitabine or a preselected treatment choice (TC) of high-intensity chemotherapy, low-intensity treatment with HMAs or low-dose cytarabine, or best supportive care (BSC). The primary end point was overall survival (OS). A total of 302 patients were randomly assigned to guadecitabine (n = 148) or TC (n = 154). Preselected TCs were low-intensity treatment (n = 233 [77%; mainly HMAs]), high-intensity chemotherapy (n = 63 [21%]), and BSC (n = 6 [2%]). The median OS were 6.4 and 5.4 months for guadecitabine and TC, respectively (hazard ratio 0.88 [95% confidence interval, 0.67-1.14]; log-rank P = .33). Survival benefit for guadecitabine was suggested in several prospective subgroups, including age <65 years, Eastern Cooperative Oncology Group performance status 0 to 1, refractory AML, and lower peripheral blood blasts ≤30%. Complete response (CR) + CR with partial hematologic recovery rates were 17% for guadecitabine vs 8% for TC (P < .01); CR+CR with incomplete count recovery rates were 27% for guadecitabine vs 14% for TC (P < .01). Safety was comparable for the 2 arms, but guadecitabine had a higher rate of grade ≥3 neutropenia (32% vs 17%; P < .01). This study did not demonstrate an OS benefit for guadecitabine. Clinical response rates were higher for guadecitabine, with comparable safety to TC. There was an OS benefit for guadecitabine in several prespecified subgroups. This study was registered at www.clinicaltrials.gov as #NCT02920008.

Sabatolimab plus hypomethylating agents in previously untreated patients with higher-risk myelodysplastic syndromes (STIMULUS-MDS1): a randomised, double-blind, placebo-controlled, phase 2 trial.

BACKGROUND: Sabatolimab is an immunotherapy targeting T-cell immunoglobulin domain and mucin domain-3 (TIM-3), an immuno-myeloid regulator expressed on immune cells and leukaemic stem cells. In this trial, we compared the efficacy and safety of sabatolimab plus hypomethylating agent with placebo plus hypomethylating agents in previously untreated patients with higher-risk myelodysplastic syndromes. METHODS: STIMULUS-MDS1 was a multicentre, randomised, double-blind, placebo-controlled, phase 2 study done at 54 investigational sites in 17 countries. Adult patients (aged ≥18 years) with intermediate-risk, high-risk, and very high-risk myelodysplastic syndromes (according to Revised International Prognostic Scoring System criteria) who had not received previous treatment were included. Patients were randomly assigned (1:1) to intravenous sabatolimab (400 mg on day 8 and 22) or placebo plus a hypomethylating agent (intravenous decitabine 20 mg/m2 on day 1-5 or intravenous or subcutaneous azacitidine 75 mg/m2 on day 1-7 or day 1-5 and day 8 and 9) every 28 days until treatment discontinuation. The two primary endpoints were complete response rate and progression-free survival, assessed in the full analysis set, which included all randomly assigned patients. Complete response was analysed, as prespecified, 7 months after the last patient was randomly assigned. All other analyses presented, including progression-free survival, were done at the final data cutoff prespecified via a protocol amendment on Sept 2, 2021. Safety was assessed in in all patients who received at least one dose of study treatment. This study is registered with ClinicalTrials.gov, NCT03946670, and is ongoing. FINDINGS: Between July 29, 2019, and Aug 10, 2020, 127 patients were randomly assigned to sabatolimab plus a hypomethylating agent group (sabatolimab group; n=65) or placebo plus a hypomethylating agent (placebo group; n=62). The median age of participants was 73 years (IQR 69-77), of whom 86 (68%) of 127 patients were male and 77 (61%) were White. The primary endpoints were not met. Complete response (cutoff date of March 10, 2021) was achieved in 14 (22%; 95% CI 12·3-33·5) of 65 patients in the sabatolimab group vs 11 (18%; 9·2-29·5) of 62 patients in the placebo group (p=0·77). At the cutoff date of the final analysis (March 1, 2022), median follow-up for progression-free survival was 17·8 months (IQR 16·6-19·4) in the sabatolimab group and 19·2 months (17·7-22·3) in the placebo group, and the median progression-free survival was 11·1 months (95% CI 7·6-17·6) in the sabatolimab group vs 8·5 months (6·9-11·3) in the placebo group (hazard ratio 0·75 [95% CI 0·48-1·17]; p=0·1022). The most common adverse events of any grade were neutropenia (35 [56%] of 62 patients in the sabatolimab group vs 43 [68%] of 63 patients in the placebo group), thrombocytopenia (30 [48%] vs 32 [51%]), constipation (29 [47%] vs 24 [38%]), diarrhoea (27 [44%] vs 14 [22%]), anaemia (22 [35%] vs 34 [54%]), febrile neutropenia (22 [35%] vs 15 [24%]), and leukopenia (15 [24%] vs 20 [32%]). One patient developed a serious potential treatment-related immune-mediated adverse event in the sabatolimab group. There was one treatment-related death in the sabatolimab group due to pneumonitis. INTERPRETATION: The addition of sabatolimab to hypomethylating agents in this study did not result in a significant improvement in complete response rates or progression-free survival. Sabatolimab had a manageable safety in most patients with higher-risk myelodysplastic syndromes. A randomised phase 3 trial is ongoing to assess the potential benefit of sabatolimab plus azacitidine on overall survival in this setting. FUNDING: Novartis Pharmaceuticals.

MDS/CMML from resource-limited region: Characteristics and comparison to tertiary reference European center.

INTRODUCTION: Myelodysplastic Syndromes (MDS) and Chronic Myelomonocytic Leukemia (CMML) are clonal myeloid malignancies, characterized by bone marrow failure leading to cytopenias (and possible myeloproliferation for CMML) and a high propensity to evolve to Acute Myeloid Leukemia (AML). OBJECTIVE AND METHODS: The aim of our retrospective study was to evaluate the clinical and hematological features; the prevalence of MDS subtypes, R-IPSS, and the outcome of 106 Armenian MDS/CMML patients diagnosed over the 2008-2020 period in a single Armenian Hematology center and compare them to French MDS patients included in the GFM registry. RESULTS: Median age in the Armenian cohort was 64 years (range 19-84) and 55% were males. The main MDS subtypes were MDS-MLD (29.2%) and MDS-SLD (27.3%), the least frequent was del 5q (0.9%). By comparison, a higher prevalence of MDS-MLD, MDS-EB2, and MDS-RS was found in the French cohort. Armenian patients' cohort generally had poor access to standard MDS treatment and 42.3% of the patients were transfusion dependent. Overall survival, however, did not significantly differ between Armenian and French cohorts. CONCLUSION: Our study stresses issues regarding epidemiology, access to diagnosis, difficulties of risk stratification, and access to treatment.

Imetelstat in patients with lower-risk myelodysplastic syndromes who have relapsed or are refractory to erythropoiesis-stimulating agents (IMerge): a multinational, randomised, double-blind, placebo-controlled, phase 3 trial.

BACKGROUND: Unmet medical needs remain in patients with red blood cell transfusion-dependent (RBC-TD) lower-risk myelodysplastic syndromes (LR-MDS) who are not responding to or are ineligible for erythropoiesis-stimulating agents (ESAs). Imetelstat, a competitive telomerase inhibitor, showed promising results in a phase 2 trial. We aimed to compare the RBC transfusion independence (RBC-TI) rate with imetelstat versus placebo in patients with RBC-TD LR-MDS. METHODS: In phase 3 of IMerge, a double-blind, placebo-controlled trial conducted in 118 sites including university hospitals, cancer centres, and outpatient clinics in 17 countries, patients (aged ≥18 years) with ESA-relapsed, ESA-refractory, or ESA-ineligible LR-MDS (low or intermediate-1 risk disease as per International Prognostic Scoring System [IPSS] criteria) were randomly assigned via a computer-generated schedule (2:1) to receive imetelstat 7·5 mg/kg or placebo, administered as a 2-h intravenous infusion, every 4 weeks until disease progression, unacceptable toxic effects, or withdrawal of consent. Randomisation was stratified by previous RBC transfusion burden and IPSS risk group. Patients, investigators, and those analysing the data were masked to group assignment. The primary endpoint was 8-week RBC-TI, defined as the proportion of patients without RBC transfusions for at least 8 consecutive weeks starting on the day of randomisation until subsequent anti-cancer therapy, if any. Primary efficacy analyses were performed in the intention-to-treat population, and safety analyses were conducted in patients who received at least one dose of trial medication or placebo. This trial is registered with ClinicalTrials.gov (NCT02598661; substudy active and recruiting). FINDINGS: Between Sept 11, 2019, and Oct 13, 2021, 178 patients were enrolled and randomly assigned (118 to imetelstat and 60 to placebo). 111 (62%) were male and 67 (38%) were female. 91 (77%) of 118 patients had discontinued treatment by data cutoff in the imetelstat group versus 45 (75%) in the placebo group; a further one patient in the placebo group did not receive treatment. Median follow-up was 19·5 months (IQR 12·0-23·4) in the imetelstat group and 17·5 months (12·1-22·7) in the placebo group. In the imetelstat group, 47 (40% [95% CI 30·9-49·3]) patients had an RBC-TI of at least 8 weeks versus nine (15% [7·1-26·6]) in the placebo group (rate difference 25% [9·9 to 36·9]; p=0·0008). Overall, 107 (91%) of 118 patients receiving imetelstat and 28 (47%) of 59 patients receiving placebo had grade 3-4 treatment-emergent adverse events. The most common treatment-emergent grade 3-4 adverse events in patients taking imetelstat were neutropenia (80 [68%] patients who received imetelstat vs two [3%] who received placebo) and thrombocytopenia (73 [62%] vs five [8%]). No treatment-related deaths were reported. INTERPRETATION: Imetelstat offers a novel mechanism of action with durable transfusion independence (approximately 1 year) and disease-modifying activity for heavily transfused patients with LR-MDS who are not responding to or are ineligible for ESAs. FUNDING: Janssen Research & Development before April 18, 2019, and Geron Corporation thereafter.

Chronic Myelomonocytic Leukemia Patients With Lysozyme Nephropathy and Renal Infiltration Display Markers of Severe Disease.

Introduction: Chronic myelomonocytic leukemia (CMML) is a hematologic disorder that is an overlap syndrome between myelodysplastic syndromes and myeloproliferative neoplasms, and can be associated with autoimmune and inflammatory diseases. This study aimed to describe kidney involvement in patients with CMML, their treatments, and outcomes. Methods: We conducted a French and American multicenter retrospective study in 15 centers, identifying patients with CMML with acute kidney injury (AKI), chronic kidney disease (CKD), and urine abnormalities. Results: Sixteen patients (males, n = 14; median age 76.5 years [71.9-83]) developed a kidney disease 6 months [1.6-25.6] after the diagnosis of CMML. At the time of kidney disease diagnosis, median urinary protein-to-creatinine ratio was 2 g/g [1.25-3.4], and median serum creatinine was 2.26 mg/dl [1.46-2.68]. Fourteen patients (87.5%) underwent a kidney biopsy, and the 2 main pathological findings were lysozyme nephropathy (56%) and renal infiltration by the CMML (37.5%). Ten patients received a new treatment following the CMML-associated kidney injury. Among patients with monitored kidney function, and after a median follow-up of 15 months [9.9-34.9], 4 patients had CKD stage 3, 4 had CKD stage 4, 1 had an end-stage kidney disease. In our patient series, 2 patients evolved to an acute myeloid leukemia (AML), and 5 died. Compared with 116 CMML controls, patients who had a kidney involvement had a higher monocyte count (P < 0.001), had more CMML-1 (P = 0.005), were more susceptible to develop an AML (P = 0.02), and were more eligible to receive a specific hematologic treatment, with hydroxyurea, or hypomethylating agents (P < 0.001), but no survival difference was seen between the 2 groups (P = 0.6978). Conclusion: In this cohort of patients with CMML with a kidney injury, the 2 most frequent renal complications were lysozyme-induced nephropathy and renal infiltration by the CMML. Kidney involvement should be closely monitored in patients with CMML.