The Impact of S-Li-M Criteria in Myeloma in a Real-Life Population: Patient & Disease Characteristics, Treatment and Outcomes from the Australian and New Zealand Myeloma and Related Diseases Registry (MRDR)

BACKGROUND

For many years active myeloma (MM) was defined by 4 CRAB criteria: hyperCalcemia, Renal impairment, Anaemia, Bone lesions. In 2014, 3 additional biomarkers that were associated with ≥70% chance of activation to symptomatic myeloma (CRAB) within 2 years were adopted as myeloma defining events: clonal bone marrow (BM) plasma cell (PC) percentage ≥60%, involved: uninvolved serum free light chain ratio (SFLCR) ≥100, and >1 focal lesion on MRI - the SLiM criteria (SixtyLightchainMRI). We assessed the application of these criteria in a real-life population to determine their impact on clinical practice and treatment outcomes.

METHODS AND AIMS

We evaluated clinical characteristics, treatment & outcomes of pts registered in the Australian & New Zealand Myeloma Registry (MRDR) from 2013 to 2018, comparing pts diagnosed with active MM by CRAB (n=1686) vs SLiM (n=132) criteria.

RESULTS

Patient & Disease characteristics: In 132 SLiM pts, 58 had BMPC≥60%, 48 SFLCR≥100 and 26 were known to have both, nil had MRI lesions. CRAB pts included a higher proportion with advanced stage ISS-3 (35.6% vs 10.7%, p<0.001), R-ISS-3 (16.0% vs 2.5%, p=0.001) and poor performance status (ECOG 2 to 4: 24.7% v 8.3%, p<0.001). Difference in ISS is mainly due to higher b2microglobulin (CRAB vs SLiM med 4.0 vs 3.1, p<0.001) with no difference in albumin. Renal impairment was more prevalent in CRAB vs SLiM (eGFR 67 vs 79 ml/min, p<0.001) and mean Hb was lower (108 vs 117 g/L, p<0.001). Median BM PC (60% vs 50%, p = 0.004) and SFLCR (100 vs 39; p<0.001) were higher in SLiM than CRAB; this applied to both isotypes – kappa: 147 vs 54 (p = 0.002) & lambda: 100 vs 25 (p<0.001). LC levels (involved minus uninvolved) were higher in SLiM: 831 vs 452 mg/L (p=0.07), with a higher incidence of FISH anomalies in CRAB: 65.8% vs 55.3% (p=0.07).

Outcomes: There was no difference in therapy between SLiM and CRAB. Median OS of SLiM pts was longer than CRAB (76.3 vs 65.2 m, p=0.02, HR 1.75 Fig 1a), with no difference in PFS (30.8 vs 30.2 m, p=0.30, HR 1.17, Fig 1b). However PFS2 was superior for SLiM of 48.7 m vs 38.2 m for CRAB (HR 1.38, p=0.06, Fig 1c). No difference in best response rates (≥PR & ≥VGPR) was seen in SLiM vs CRAB. Within SLiM, comparing 2 groups who only satisfied the criteria of SFLCR≥100 (SFLC group, n=28) or PC≥60% (PC group, n=34) alone and with treatment responses, there was a difference in the distribution. CR was lower in SFLC group (7 v 20%), but VGPR +PR was higher (89% vs 61%). More pts in PC group had only minimal response (MR) or stable disease (SD) as best clinical response (17.6%) compared to SFLC group (3.6%). The rate of pts with <PR was 3.6% in SFLC group, 17.6% in PC group, and 20% for pts with both (n=20) (p=0.04), indicating that pts who satisfy PC criteria (alone or together with SFLCR) have a higher risk for suboptimal response compared to pts who satisfy SFLCR alone.

DISCUSSION

Apart from the expected differences distinguishing SLiM and CRAB pts eg. renal impairment, we found additional differences in pt & disease characteristics between the 2 groups – higher BM PC infiltration and SFLCR in SLiM, and a higher incidence of FISH anomalies in CRAB. Not surprisingly a higher proportion of CRAB pts had advanced stage and poor performance status. SLiM pts had a median OS 11.1 m longer than the CRAB group, but no difference in PFS1 nor response rates. The OS difference of ~1 yr could reflect the earlier initiation of therapy in SLiM (with OS calculated from time of first therapy, corresponding to the SLiM criteria derived from ≥70% chance of transformation to CRAB within 2 yrs if untreated). However the trend for a longer PFS2 for SLiM indicates an improved treatment outcome to therapy for first relapse, suggesting that the shorter PFS2 of CRAB pts may result from residual and re-emerging clones that are more drug resistant at relapse. Within the SLiM cohort, while PFS & OS did not differ between pts treated by SFLCR, PC criteria or both, supporting their equal impact on disease behavior, a higher proportion of pts with PC criteria (alone, or with SFLCR) had a suboptimal response (<PR) compared to pts who satisfied SFLCR alone.

CONCLUSIONS

Our findings support the benefit of earlier initiation of therapy by SLiM criteria in a real-life population, indicating an improved OS and a trend towards improved salvageability at first relapse (PFS2). Patients with PC criteria in the SLiM cohort are associated with an increased risk for suboptimal response compared to those who satisfy SFLC criteria alone.