Background: Structural Blind Spots in the Watch-and-Wait Paradigm and the Clonal Evolution Data Bottleneck in Smoldering Multiple Myeloma (SMM) R&D

Smoldering Multiple Myeloma (SMM), which has a high-risk group with an annual conversion rate of approximately 10% to active Multiple Myeloma (MM), has been governed by global hematology guidelines (IMWG 2014 revision) that adhere to the static observation doctrine of "deferring therapeutic intervention until the onset of SLiM-CRAB end-organ damage." The critical flaw in this linear governance is evident: at the time the 20/2/20 risk model (bone marrow plasma cells >20%, serum M-protein >2 g/dL, involved/uninvolved light chain ratio >20) is met, subclonal branching and epigenetic drift are already in progress within the bone marrow clonal plasma cells. This cannot be captured by a single-timepoint bone marrow aspiration-based morphological assessment. Previous attempts at early intervention based on lenalidomide monotherapy maintenance (E3A06, QuiRedex, etc.) have demonstrated prolonged progression-free survival (PFS) but have structurally failed to achieve overall survival (OS) benefits and minimal residual disease (MRD) negativity, leading to a data bottleneck that necessitates the introduction of novel therapeutic modalities to achieve a level of immune eradication.

Discovery: Ciltacabtagene Autoleucel (Cilta-cel) BCMA-CAR T Single-Dose Infusion and Bone Marrow MRD Negativity Tensor Synchronization Demonstration

The CAR-PRISM (PRecision Intervention Smoldering Myeloma) Phase 2 clinical trial, led by the Omar Nadeem research team at Dana-Farber Cancer Institute, has produced the first human proof-of-concept data that disruptively surpasses this bottleneck. Twenty high-risk SMM patients (bone marrow plasma cells >10-40%, meeting 20/2/20 criteria) received a single infusion of ciltacabtagene autoleucel (Carvykti, Janssen/Legend Biotech), a BCMA-targeting dual-epitope binding CAR T cell construct, after lymphodepletion conditioning without any induction chemotherapy or bridging therapy. As a result, all 20 patients (100%) achieved MRD negativity within 2 months after infusion. At a median follow-up of 15.3 months, all patients remained in MRD-negative status, and complete remission (CR) and MRD negativity were sustained in all six patients with a follow-up of 18 months or longer. This is the first time that universal MRD negativity has been reported in a plasma cell disease clinical trial, and it represents a tensor surface that absolutely surpasses the BCMA-CAR T response rate baseline established in the CARTITUDE-1 (97.9% ORR, 82.5% sCR) and KarMMa (73% ORR) trials in active MM.

In terms of safety profile, all patients experienced grade 1-2 cytokine release syndrome (CRS), but no patients experienced grade 3 or higher CRS, and four cases of facial nerve palsy (all with complete recovery) and three cases of residual mild motor symptoms occurred, indicating dose-dependent neurotoxicity. As a result, a preemptive dexamethasone administration protocol and CAR T cell dose reduction were implemented in patients with high lymphocyte expansion and elevated CD4:CD8 ratios, which subsequently suppressed the occurrence of neurological events.

Establishment of a Model for Fine-Tuning Clonal Burden Gradients and Reversible Bone Marrow Immune Homeostasis Precision Stratification

The core architectural contribution of CAR-PRISM is the establishment of a functional precision stratification matrix based on the molecular phenotype of patients in the pre-disease stage. The design, which excludes patients with bone marrow plasma cell infiltration exceeding 40% in the eligibility criteria, is a rate-limiting step that proactively blocks the risk of CRS surge and tumor lysis syndrome (TLS) in silico by clamping the upper limit of tumor burden, given the protocol's characteristic of directly administering CAR T without induction/bridging therapy. At the same time, correlation analysis of each of the 20/2/20 model's sub-variables (bone marrow plasma cell ratio, absolute M-protein value, free light chain ratio) with the depth of response, and multidimensional tensor mapping of lymphocyte expansion dynamics (CD4:CD8 ratio, CAR T cell amplification peak time), have yielded a prototype of a companion biomarker panel that predicts which patients will achieve sustained clonal eradication with a single infusion and which patients will require dose modification or preemptive steroid intervention. The median follow-up data of 15.3 months, which has already surpassed the expected progression-free survival under the existing observation strategy, provides the backbone evidence that demonstrates that reversible bone marrow immune homeostasis re-ignition based on the depth of immune response can fundamentally reset the natural history of high-risk SMM, rather than static observation.

Prospects: Establishment of a Programmable Pre-Disease Immuno-Oncology Standard and Launch of Next-Generation IND Digital Governance

The CAR-PRISM data represents a declarative turning point in the governance of hematological pre-disease management, shifting from a static, post-hoc observational system to a fully programmable early intervention infrastructure based on AI-powered multidimensional tensor analysis. Despite the statistical limitations of a single-arm design and a small cohort of 20 patients, the triple absolute indicators of 100% MRD negativity, zero progression, and zero death provide sufficient signal strength to justify a multi-center, randomized Phase 3 expansion. Janssen/Legend Biotech's cilta-cel is already on track for FDA approval (2022) and expansion to first-line treatment (CARTITUDE-4, 2024 ASCO LBA) for active MM, and the expansion of the SMM indication is a pipeline asset that will disruptively expand the addressable population of the global CAR T market (estimated at ~8 billion USD in 2026) to the pre-disease spectrum. In the future, the establishment of an MRD monitoring panel linked to high-throughput single-cell RNA sequencing (scRNA-seq)-based clonal lineage tracing will allow real-time clamping of batch-to-batch CAR T cell manufacturing deviations (transduction efficiency, vector copy number variations) as a correction factor, ensuring lot-to-lot consistency in cGMP commercial-scale production. Furthermore, the combination of BCMA with alternative targets such as GPRC5D and FcRH5 in dual/triple CAR constructs, or the convergence with in vivo CAR T generation technologies (Umoja Biopharma, Capstan Therapeutics, etc., IND entry in 2025-2026), will function as a master asset that will resolve the rate-limiting bottleneck of current ex vivo CAR T supply chains (autologous cell collection, manufacturing, and transportation) and disruptively shorten the IND approval timeline under the digital healthcare companion diagnostics (CDx) regulatory framework.