Page 232 - 2021_03-Haematologica-web

878
Letters to the Editor
than in the UKF one. Age, KPS, bone marrow plasma cells and cytogenetics were comparable in the two groups (Table 1).
In all 284 MM patients, the R-MCI, IMWG-frailty score, CCI and other functional tests (Table 1) were expe- ditiously assessable. The R-MCI confirmed that patients in the multicenter cohort were fitter than those in the UKF cohort with a mean score of 3 versus 4, respectively. In contrast, according to the IMWG-frailty score, there was no difference, with a mean of 1 in both cohorts. It is noteworthy that the IMWG-frailty score and CCI were increased in this prospective analysis compared to the prior IMWG description,11 confirming our previous, prospective validation analysis of the R-MCI and IMWG- frailty score.12 Frailty and functional tests confirmed that patients were fitter in the multicenter cohort than in the UKF cohort with regards to severe frailty (6% vs. 19%), and mean activities of daily living (ADL) score (6 vs. 5), IADL score (8 vs. 7), physician-rated fitness (2 vs. 3) and TUGT results (10 s vs. 12 s; respectively) (Table 1).
We also assessed and compared the patients’ character- istics and functional differences in the three R-MCI sub- groups of fit (R-MCI 0-3), intermediate-fit (R-MCI 4-6) and frail (R-MCI 7-9) patients of the entire, UKF and mul- ticenter cohorts (Table 2). We considered patients’ char- acteristics (age, KPS), MM risk factors (renal function, bone marrow plasma cells), frailty scores and functional fitness tests separately for intermediate-fit and frail patients as compared to fit patients. This confirmed that single R-MCI components and the R-MCI itself were of relevance to defining risks in both the UKF and multicen- ter cohorts.6,12
Comparison of the fit, intermediate-fit and frail R-MCI subgroups in the multicenter vs. UKF cohorts confirmed that there more fitter (62% vs. 27%) and fewer frail patients (4% vs. 12%) in the former than the latter cohort. The differences between the two cohorts were more impressive for risk scores (frailty, CCI, IMWG) and functional tests (ADL/IADL, physician-rated fitness, TUGT) than for age, KPS, renal function or bone marrow plasma cell results (Table 2).
We also compared the frequencies of fit, intermediate- fit and frail patients identified directly via the R-MCI, CCI and IMWG-frailty score (Figure 1A-F): according to the R-MCI, 33% of the entire cohort were considered fit, 56% intermediate-fit and 11% frail (Figure 1A). Similarly, according to the CCI, 41% were fit, 51% intermediate-fit and 8% frail (Figure 1B), whereas according to the IMWG-frailty score, 27% were fit, 38% intermediate-fit and 35% frail (Figure 1C). The allocation into fit, inter- mediate-fit and frail patients was consequently compara- ble via the R-MCI and CCI, whereas the IMWG-frailty score identified fewer intermediate-fit and more frail patients (Figure 1A-C).3,11,13
When comparing the R-MCI, CCI and IMWG-frailty score allocations in the UKF versus multicenter cohorts (Figure 1D-F), the R-MCI identified more fitter patients (27% vs. 62%, respectively) and fewer intermediate-fit and frail patients in the multicenter cohort (Figure 1D).
According to the CCI assessment, the differences between the UKF and multicenter cohorts were least marked for fit and intermediate-fit patients with 43% ver- sus 33% and 52% versus 44%, respectively, but substan- tial for frail patients with the allocation being 4% versus 23%, respectively. Via the CCI, more multicenter than UKF patients were unsustainably defined as frail (Figure 1E), which did not reflect the results of the patients’ char- acteristics (Table 1A and B) or the R-MCI-defined group differences (Figure 1D).
When comparing subgroups of fit, intermediate and frail patients via the IMWG-frailty score (Figure 1F), we observed a similar proportion of fit patients in the UKF versus multicenter cohorts (26% vs. 31%), fewer interme- diate-fit patients (33% vs. 60%) and more frail patients (41% vs. 9%, respectively), with the largest proportion of UKF patients being misguidedly assigned to the frail sub- group. Since the allocations via the IMWG-frailty score in the UKF versus multicenter cohorts were very different from those via the R-MCI and CCI, a difference which had already been perceivable in the comparison of the entire cohort of MM patients (Figure 1A-C), we postulate that the IMWG-frailty score overestimates frail patients when prospectively assessed (as here). This was most apparent when the proportions of frail patients defined by the different scores were compared, since this propor- tion was four times greater with the IMWG-frailty score than with the R-MCI in the UKF group (41% vs. 12%, respectively) (Figure 1C and E) and, again in the UKF group, ten times greater with the IMWG-frailty score than with the CCI (41% vs. 4%, respectively) (Figure 1D and E).
Since IMWG-frailty scores in the UKF and in the mul- ticenter cohorts did not differ (mean: 1) and the CCI was higher in the latter (2 vs. 3, respectively) (Table 1), the R-MCI was of interest and verified cohort differences (4 vs. 3, respectively). Moreover, the R-MCI was in line with all functional/frailty tests (Table 1). Thus, the R-MCI and functional fitness tests confirmed greater fitness in the multicenter cohort than in the UKF cohort, and the com- parative analysis of results from five DSMM/EMN myeloma centers showed that a risk score and functional assessment may indeed help to better define patients’ dif- ferences. Fitter patients in the multicenter compared to the UKF cohort were best clarified via R-MCI and only this score was consistently in line with frailty, ADL/IADL, physician-rated fitness and TUGT results.
The strength of this study was the prospective assess- ment in the five EMN/DSMM centers. Furthermore, frailty and functional assessments were done by the same skilled person in all centers, which excluded differences in handling the patients’ assessment and data acquisition. The R-MCI was compared with the IMWG-frailty score, CCI and fitness tests, the former including a few comor- bid conditions that are readily obtainable from the collec- tion of the medical history and were obtained from the multivariate risk analysis of a large prospective sam- ple.6,12,13,18,19 Additional advantages of the R-MCI are that it: (i) allows more accurate assessment of physical condi- tions than via clinical judgment, age or KPS/ECOG alone; (ii) precisely divides patients into fit, intermediate-fit and frail patients with different progression-free and overall survival risks;12,13,18,19 (iii) allows the inclusion of biological risks, namely cytogenetics; and (iv) consistently divides risk groups of R-MCI 0-3 (=fit), 4-6 (=intermediate-fit) and R-MCI 7-9 (=frail) patients, irrespective of age and treatment (i.e. <65/≥65; <70/≥70; or <75/≥75 years, and with/without novel agents or with/without transplanta- tion; as reported elsewhere12,13,18,19).
A limitation of the present study was the application of different anti-myeloma therapies (typically bortezomib, cyclophosphamide, dexamethasone or bortezomib, lenalidomide, dexamethasone). It is, however, worth not- ing that in prior analyses, we demonstrated that the R- MCI distinguished highly significant risk groups despite different anti-myeloma treatments,12,13 a finding con- firmed by others.5,11,20 Another criticism that could be raised concerns the limited number of patients in the multicenter cohort, which was a consequence of our
haematologica | 2021; 106(3)