2026 ASH®
Annual Meeting & Exposition December 12-15, 2026 ▪ New Orleans, Louisiana

Paper Details

Abstracts, authors, affiliations, and disclosures.

Monday, December 14, 2026
06:00 p.m. - 08:00 p.m.
MCCNO - Hall F
Access, Representation, Population Health, Technology and Procedures, Serologic Tests, Bioinformatics, Clinical Procedures, Emerging Technologies, Treatment Considerations, Supportive Care, Biological Therapies, Transplantation, Monoclonal Antibody Therapy, Biological Processes, Genomics, Clinical Practice (Health Services and Quality), Diseases, Bleeding and Clotting, Lymphoid Malignancies, Plasma Cell Disorders, Measurable Residual Disease, Research, Translational Research, Clinical Research, Clinical Trials, Epidemiology, Health Outcomes Research, Study Population, Human, Adult
PAPER

5712 A new validated dynamic model to assess the risk of smoldering multiple myeloma: A multinational study involving 2453 patients

Annika Werly, Cecilie Velsoe Maeng2, Floris Chabrun, MD3, Ben Patrick4, Thomas Hielscher5, Tobias Dittrich6, Sophia K. Schmidt7, Carsten Utoft Niemann, MD, PhD8, Hira Mian, MD9, Mareike Hampel10, Shaji Kumar, MD11, Omar Nadeem, MD12, Elizabeth O'Donnell, MD13, Catherine Marinac, PhD14, Mark Bustoros, MD15, Niels Weinhold, PhD16, Martin Dugas17, Kosima Zuern18, Anna Jauch19, Hartmut Goldschmidt, MD20, Carsten Mueller-Tidow, MD21, Marc-Steffen Steffen Raab, MD22, Alissa Visram, MD, MPH23, Sigrun Thorsteinsdottir, MD,PhD24, Irene M Ghobrial, MD25, Elias K. Mai, MD26
1 Heidelberg Myeloma Center, Internal Medicine V, Hematology, Oncology and Rheumatology, Heidelberg University Hospital and Medical Faculty Heidelberg, Heidelberg, Germany
2 Department of Hematology, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark
3 Dana Farber Cancer Institute, Boston, Massachusetts, United States
4 The Ottawa Hospital, Ottawa, Ontario, Canada
5 Division of Biostatistics, German Cancer Research Center (DKFZ), Heidelberg, Germany
6 Institute of Medical Informatics, Heidelberg University Hospital, Heidelberg, Germany
7 Heidelberg Myeloma Center, Internal Medicine V, Hematology, Oncology and Rheumatology, Heidelberg University Hospital and Medical Faculty Heidelberg, Heidelberg, Germany
8 Department of Hematology, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark
9 Juravinski Cancer Center at McMaster University, Hamilton, Ontario, Canada
10 Heidelberg Myeloma Center, Internal Medicine V, Hematology, Oncology and Rheumatology, Heidelberg University Hospital and Medical Faculty Heidelberg, Heidelberg, Germany
11 Department of Hematology, Mayo Clinic, Rochester, Minnesota, United States
12 Dana Farber Cancer Institute, Boston, Massachusetts, United States
13 Dana Farber Cancer Institute, Boston, Massachusetts, United States
14 Dana Farber Cancer Institute, Boston, Massachusetts, United States
15 Weill Cornell Medical College, New York City, New York, United States
16 Heidelberg Myeloma Center, Internal Medicine V, Hematology, Oncology and Rheumatology, Heidelberg University Hospital and Medical Faculty Heidelberg, Heidelberg, Germany
17 Institute of Medical Informatics, Heidelberg University Hospital, Heidelberg, Germany
18 Heidelberg Myeloma Center, Internal Medicine V, Hematology, Oncology and Rheumatology, Heidelberg University Hospital and Medical Faculty Heidelberg, Heidelberg, Germany
19 Institute of Human Genetics, Heidelberg University Hospital, Heidelberg, Germany
20 Heidelberg Myeloma Center, Internal Medicine V, Hematology, Oncology and Rheumatology, Heidelberg University Hospital and Medical Faculty Heidelberg, Heidelberg, Germany
21 Heidelberg Myeloma Center, Internal Medicine V, Hematology, Oncology and Rheumatology, Heidelberg University Hospital and Medical Faculty Heidelberg, Heidelberg, Germany
22 Heidelberg Myeloma Center, Internal Medicine V, Hematology, Oncology and Rheumatology, Heidelberg University Hospital and Medical Faculty Heidelberg, Heidelberg, Germany
23 Juravinski Cancer Center at McMaster University, Hamilton, Ontario, Canada
24 Department of Hematology, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark
25 Dana Farber Cancer Institute, Boston, Massachusetts, United States
26 Heidelberg Myeloma Center, Internal Medicine V, Hematology, Oncology and Rheumatology, Heidelberg University Hospital and Medical Faculty Heidelberg, Heidelberg, Germany

INTRODUCTION:

To date, the Mayo2018 (Lakshman A et al., BCJ, 2018) and IMWG2020 (Mateos MV et al., BCJ, 2020) represent commonly applied risk stratification models for patients (pts) with smoldering multiple myeloma (SMM). Risk stratification is increasingly shifting towards capturing individual disease dynamics over time to more accurately identify high-risk pts eligible for early individualized therapeutic intervention (Chabrun F et al., Nat Med., 2026). In 2025, we optimized the static Mayo2018/ IMWG2020 risk models by incorporating the dynamic and time-independent risk factor “evolving M-protein >20% and >0.4g/dl and/or evolving involved serum FLC >40% and >40mg/l within 12 months (eMP/eFLC)”. In a large-scale external validation, we aimed to confirm that our dynamic models outperform the static standard models for risk stratification at any time during pt follow-up (FU).

METHODS:

The training cohort of this retrospective study included 447 pts, diagnosed with SMM between 2009 and 2023 at the Heidelberg Myeloma Center, meeting the 2014 IMWG diagnostic criteria (Rajkumar SV et al., Lancet Oncol., 2014). The model was validated by three independent cohorts comprising 2,006 pts (cohort 1: 826 pts from DFCI, Boston, USA; cohort 2: 1090 pts from DALY-CARE (population-based), Denmark; cohort 3: 90 pts from Ottawa Hospital, Ottawa, Canada).

The criteria of the static Mayo2018 and IMWG2020 models, and additionally the point-based scoring system for the gradual version of the IMWG2020 (IMWG2020grad) model, were the base for the respective dynamic models. Pts fulfilling the novel dynamic eMP/eFLC-criterion were reclassified into the next higher risk-category. To improve risk stratification regarding c-index, the 3 risk groups of the static Mayo2018 model were expanded by a new risk group “low-intermediate” for the dynamic Mayo2018 model, while the dynamic IMWG2020 models retained their 4 risk groups.

Dynamic models were compared to the standard models regarding c-indices and absolute 2-year-risk of progression in the high-risk groups, serially 1, 2 and 3 years post-diagnosis.

 

RESULTS:

Across all cohorts, the dynamic Mayo2018, IMWG2020 and IMWG2020grad models consistently outperformed the static models regarding c-statistics. In the training cohort, c-indices increased significantly from 0.778 to 0.817 (static vs. dynamic; p=0.005) for the Mayo2018, from 0.739 to 0.781 (p=0.033) for the IMWG2020 and from 0.760 to 0.811 (p=0.005) for the IMWG2020grad model. Similarly, in validation cohorts 1, 2 and 3, c-indices increased from 0.792 to 0.805 (static vs. dynamic), from 0.743 to 0.765 and from 0.671 to 0.823 for Mayo2018. In validation cohort 1, c-indices increased from 0.757 to 0.818 for IMWG2020 and from 0.777 to 0.832 for IMWG2020grad.

When applying the dynamic Mayo2018 model, the 2-year risk of progression in the high-risk group was consistently and substantially higher compared to the static model across training and all validation cohorts. For example, at the 1-year landmark, the 2-year risk of progression among high-risk pts was 57% [95%CI, 36;77] vs. 67% [95%CI, 40;93] (static vs. dynamic) for the training cohort, 12% [95%CI, 0;24] vs. 43% [95%CI, 6;80] for validation cohort 1, and 39% [95%CI, 16;62] vs. 58% [95%CI, 27;88] for validation cohort 3.

In contrast, the dynamic IMWG2020/ IMWG2020grad models showed only partial increase in 2-year progression risk within the high-risk groups compared to the static model.

 

CONCLUSION:

Our study demonstrates that not only can the dynamic Mayo2018 model be applied at any time during pt FU, but it also substantially outperforms the standard model in identifying pts at high risk for progression. These findings were confirmed through large-scale external validation across three independent cohorts comprising a total of 2,006 pts.

To quickly and easily assess an individual pt risk profile based on dynamic eMP/eFLC trajectories within the previous 12 months, we developed an easy-to-use web calculator for clinicians (rshiny.med.uni-heidelberg.de/smm_score/).

Limitations of the IMWG2020/ IMWG2020grad models across training and validation cohorts highlight the need to revise the current high-risk cytogenetic and genomic components within the IMWG2020 model, as malignant transformation by genomic markers plays an essential role in predicting aggressive disease dynamics in pts with SMM (Maura F et al., J Clin Oncol., 2026; Aktas Samur A et al., J Clin Oncol., 2026; Alberge JB et al., Nat Genet., 2025).

Annika Werly: No relevant conflicts of interest to declare.