Comparison of IMWG and PANGEA risk models predicting MGUS to multiple myeloma progression in a safety net hospital.

Abstract

e19534 Background: The Personalised progression prediction model (PANGEA) and International Myeloma Working Group (IMWG) models offer different approaches to assessing monoclonal gammopathy of Undetermined Significance (MGUS) progression risk to multiple myeloma (MM). The IMWG model categorizes patients into discrete risk groups based on clinical and laboratory parameters, while PANGEA offers personalized risk probabilities using time-varying biomarkers. While PANGEA offers better individualization, the IMWG model's simplicity remains valuable in routine practice. Further research is crucial to validate both models in diverse populations and assess long-term clinical outcomes. Methods: This retrospective chart study, included a total of 116 African American (AA) patients with diagnosed MGUS. These patients' records were observed in a 5 year timeline to observe for any progression to MM, and their predicted risk of progression was evaluated with the IMWG and PANGEA scales. The following variables were collected from each chart to be used to the risk calculators: Kappa/ Lambda light chain serum, M protein concentration, Bone marrow plasma cell percentage, hemoglobin, & creatinine. In patients who progressed to MM, their time to progression was recorded in years and the duration of this progression was compared to the prediction of the IMWG and PANGEA scales. Results: A total of 12 patients progressed to MM during the 5 year period. A spearman correlation (ρ) was done among those that had progressed. The ρ for IMWG was .095 ( P value of .770), and for the Pangea model: 1-year prediction: ρ = -0.445 (P = 0.147) - 2-year prediction: ρ = -0.466 (P = 0.127) - 5-year prediction: ρ = -0.435 (P = 0.157) - 10-year prediction: ρ = -0.347 (P = 0.269). The Mann-Whitney U test was used to compare the risk scores between progressed and non-progressed groups, and both scales in our data set showed statistically significant differences (p < 0.001), confirming their ability to differentiate between the two groups. Next a logistic regression model was done to assess the relation between risk scores and progression. The odds ratios indicated that for each unit increase in risk score, the odds of progression increased by 25.9% for IMWG (r 2 = 0.6034 ) and 24.3% for PANGEA ( r 2 = 0.5459). Conclusions: Ultimately, both models failed to show significant spearman correlation, however, for the PANGEA model the negative correlations observed across different timeframes suggest a more nuanced capture of progression risk, likely due to sample size limitations. For the IMWG model, its (ρ = 0.466) Showed weak positive correlation with progression time, given the 2 year limited window. The complementary nature of both models suggests potential benefit in combined application; PANGEA's multiple time horizons offer more detailed progression risk assessment and IMWG's established framework provides valuable baseline risk stratification.