Abstract
8088
Background: The updated (2010) International Myeloma Working Group criteria emphasize individual follow-up and development of early treatment trials for high risk SMM patients. To facilitate these goals, it is important to have reliable models for risk of progression to MM derived from prospective clinical studies. Current clinical risk models by the Mayo Clinic and the Spanish PETHEMA group stratify SMM patients as low (LR), intermediate (IR), or high risk (HR) of progression using bone marrow plasma cell fraction and protein levels (Mayo) and abnormal plasma cell fraction on flow cytometry with or without immunoparesis (Spanish). We report the first comparison of these models in a prospective natural history study. Methods: We evaluated 70 patients with SMM and determined the risk classifications of each using the Mayo and Spanish models. P values of comparisons between risk groups were calculated using Fisher’s exact test. Results: Among SMM patients classified as HR by the Spanish model (n=31), 2 patients (6%) were HR by the Mayo model; the remaining 17 (55%) and 12 (39%) patients were classified as IR and LR by the Mayo model, respectively. Among 2 SMM patients classified as HR by the Mayo model, both patients (100%) were HR by the Spanish model. There was significant disagreement between the models in identifying HR patients (HR v. non-HR, p<0.0001). Similarly, of SMM patients classified as LR by the Spanish model (n=17), 11 patients (65%) were LR by the Mayo model; 6 (35%) were IR and 0 (0%) were HR by the Mayo model. Among SMM patients classified as LR by the Mayo model (n=36), 11 (31%) patients were classified as LR by the Spanish model; 13 (36%) were IR and 12 (33%) were HR by the Spanish model. There was significant disagreement between the models in identifying LR patients (LR v. non-LR, p=0.0016). Conclusions: The significant discordance between the Mayo Clinic and Spanish PETHEMA clinical risk models underscores the urgent need for biologically derived models that rely on molecular markers to predict disease progression. Such models are critical for counseling individual patients and for the development of early treatment studies that seek to prevent or delay progression to MM.