Construction and Validation of a Progression-Free Survival Prediction Model for Newly Diagnosed Multiple Myeloma Patients With 1q21 Gain/Amplification.
The HLBP model demonstrated superior predictive performance compared with existing staging systems and previously reported prognostic models for predicting progression-free survival in newly diagnosed multiple myeloma patients with 1q21 gain/amplification.
Key Findings
Results
Higher hemoglobin was independently protective while elevated lactate dehydrogenase, increased β2-microglobulin, and TP53 deletion were independent adverse prognostic factors for PFS in NDMM patients with 1q21+.
Independent prognostic factors were identified using multivariate Cox regression analysis
The four factors (hemoglobin, lactate dehydrogenase, β2-microglobulin, and TP53 deletion) were combined to construct the HLBP nomogram model
Study population consisted of 186 newly diagnosed multiple myeloma patients with 1q21 gain/amplification treated between 2018 and 2024
Patients were retrospectively analyzed from a single institution
Results
The HLBP model achieved AUCs of 0.805, 0.885, and 0.847 at 6, 12, and 24 months respectively in the training cohort.
Training cohort consisted of 131 patients (70% of total cohort)
Time-dependent receiver operating characteristic (ROC) curves were used to assess predictive accuracy
Model performance was also evaluated through bootstrap resampling and calibration plots
The 12-month AUC of 0.885 represented the highest predictive accuracy across the three time points in the training cohort
Results
The HLBP model achieved AUCs of 0.747, 0.827, and 0.712 at 6, 12, and 24 months respectively in the internal split-sample validation cohort.
Internal validation cohort consisted of 55 patients (30% of total cohort)
Patients were randomly assigned at a 7:3 ratio to training and validation cohorts
The 12-month AUC of 0.827 was the highest across time points in the validation cohort
Bootstrap resampling was used in addition to split-sample validation to evaluate model performance
Results
Patients stratified into high- and low-risk groups by the HLBP model showed significantly different PFS outcomes.
Risk stratification was performed using Kaplan-Meier survival analysis
The model provided binary risk group classification (high-risk vs. low-risk)
The HLBP model demonstrated superior predictive performance compared with existing staging systems and previously reported prognostic models
Conventional staging systems evaluated for comparison likely included ISS and R-ISS staging
Results
Risk-stratified analyses suggested heterogeneous prognostic patterns associated with induction treatment regimens and autologous stem cell transplantation across different risk groups.
Exploratory analyses assessed treatment-associated outcomes across HLBP-defined risk groups
Both induction treatment regimens and autologous stem cell transplantation (ASCT) showed differential associations with outcomes depending on risk group
These analyses were described as 'exploratory' and 'risk-stratified'
The study period spanned 2018 to 2024
Methods
The study population of 186 NDMM patients with 1q21+ was retrospectively assembled and split into training and validation cohorts for model development.
All patients were newly diagnosed multiple myeloma patients with 1q21 gain/amplification
Patients were treated between 2018 and 2024
The 7:3 random split yielded 131 patients in the training cohort and 55 in the internal split-sample validation cohort
The study design was retrospective
What This Means
This research suggests that a new prediction tool called the HLBP model can help forecast how long newly diagnosed multiple myeloma patients with a specific chromosomal abnormality (1q21 gain/amplification) will remain free of disease progression. The model uses four factors measurable at diagnosis — hemoglobin levels, lactate dehydrogenase, β2-microglobulin, and the presence of a TP53 gene deletion — to generate individualized risk estimates. The study analyzed 186 patients treated between 2018 and 2024, using about 70% of them to build the model and 30% to test it.
The HLBP model performed well at predicting progression-free survival at 6, 12, and 24 months, with accuracy measures (AUC values) ranging from approximately 0.71 to 0.89 across both the development and validation groups. These scores indicate good to strong discriminative ability, and the model outperformed existing staging systems like ISS and R-ISS when applied to this specific patient population. Patients classified as high-risk by the model had meaningfully worse survival outcomes than those classified as low-risk.
This research also suggests that patients in different risk groups may respond differently to specific treatments, including various induction drug regimens and stem cell transplantation, though these treatment-related findings were described as exploratory. The HLBP model may offer a more tailored approach to assessing prognosis in this high-risk subgroup of myeloma patients who carry the 1q21 chromosomal abnormality, potentially helping to guide future treatment planning and clinical trial design. External validation in larger, independent cohorts would be needed before broader clinical application.
Rao Y, Li S, Yu A, Yu W, Yao Z, Zhang H, et al.. (2026). Construction and Validation of a Progression-Free Survival Prediction Model for Newly Diagnosed Multiple Myeloma Patients With 1q21 Gain/Amplification.. Cancer medicine. https://doi.org/10.1002/cam4.72077