Page 152 - Haematologica3

C.H. Blimark et al.
Survival according to response
Overall, better response to first-line treatment was sig- nificantly associated with superior survival (P<0.05) (Online Supplementary Table S5). In younger patients, there was no significant difference in 5-year RS in patients in PR, VGPR and CR (Figure 3)
Survival according to year of diagnosis
Patients diagnosed in the period 2011-2015 had a trend to better 1-, 3- and 5-year RS compared to patients diag- nosed 2008-2010. In patients aged over 65 years, this trend was more evident than in younger patients (Online Supplementary Tables S5 and S6, and Figure S3). In a propor- tional hazard model of RS by year of diagnosis in all patients, later calendar year of diagnosis was significantly associated with improved RS, with an HR of 0.93 (95%CI: 0.92-0.95; P<0.05).
Survival according to treating hospital
The 1, -3 – and 5-year survival was significantly higher in university hospitals (Online Supplementary Table S7). In a proportional hazards model for the RS, the HR was 0.93 (95%CI: 0.87-0.99; P<0.005). Even when adjusting for age, sex, and ISS-stage, the HR was of borderline statistical sig- nificance (HR=0.91; 95%CI: 0.83-1.0; P=0.04). Similar results were obtained when analyzing centers that treated 10 or more MM patients per year (data not shown).
Discussion
In this study from the Swedish Myeloma Registry, we report incidence, baseline characteristics and survival of an unselected population comprising more than 97% of all myeloma patients diagnosed in Sweden in the period 2008-2015. We found an age-adjusted incidence of 6.8 per 100,000 inhabitants; this translates into 4.8 and 3.2 per 100,000 inhabitants in European and World standards, respectively. This is higher than figures previously report- ed by most population-based studies,22,23 but is in agree- ment with data from a previous large Swedish study.24 The high age-adjusted incidence might be explained by better case ascertainment in the elderly. Overall, the proportion of elderly (65 years and older) myeloma patients at diag- nosis was 72%, and this exceeds the number of reported elderly patients in most known registries today, but is sup- ported by population-based data from the Danish Myeloma Registry25 and a recent report on a large cohort of European patients.26 We observed a median age of 71 years at diagnosis, which is higher than other myeloma studies,8 and a steep increase in age-specific incidence extending to the oldest age cohorts. This indicates that our population, given the very high coverage provided by the Swedish Myeloma Registry, reflects the 'real-world' situa- tion in myeloma today.
Our study shows encouraging survival rates in the MM population. In our population-based study, the 5-year OS was 38%, similar to the data from the EUROCARE study.27 In a 2014 report from the Mayo clinic based on 1084 MM patients (median age 66 years), the median OS from diagnosis was 5.2 years and the 6-year OS estimate was 45%.8 We show that with the increased use of novel agents there was an improvement in response rates. We also show that, over the study period, the proportion of elderly patients receiving novel drugs increased. The dif- ference in survival between the different age cohorts was
less pronounced in RS compared to OS, which demon- strates the importance of including RS in survival analyses inMM.
In the European Registry data from 2008 (EURO- CARE),28 a 2% survival advantage was seen in women. However, in our more recent study covering the period from 2008 to 2015, after age standardization, there was no difference in survival between men and women.
As expected, and as shown before,29,30 achievement of response was predictive of prolonged survival. There was a significant difference in survival in patients aged over 65 years. Given this, we investigated the impact of response grade on survival in different age cohorts in patients with MM at diagnosis. The analysis revealed that responding patients in all age groups had a better outcome than non- responding patients, and that patients achieving CR had the longest survival. However, interestingly, in patients aged 65 years or under there was no significant difference in survival according to the degree of response (CR, VGPR or PR). This is contrary to results from many randomized studies,31-33 and may indicate that achievement of a high quality response to first-line treatment may not have the same importance for survival in a young, unselected myeloma population where the majority of patients will eventually receive multiple lines of treatment.
We found a survival benefit in patients reported from university hospitals and those hospitals treating a large number of MM patients. This is not surprising given the speed of progress in diagnostics and the new treatments of recent years, and has, in fact, also been reported in other studies.34,35 We could not detect a significant difference in referral patterns, but in spite of this, our results should be interpreted with caution, as residual confounding factors may have influenced outcome. However, this does under- line the importance of high volume centers with expert knowledge in MM treatment and the need for further studies to monitor access to care for myeloma patients.
The strength of this study is the large, population-based cohort and excellent coverage provided by the Swedish Cancer Registry. Another strength is the public Swedish health care system. In Sweden, all patients with a diagno- sis of cancer are treated in public hospitals, enabling pub- licly financed and equal treatment for all MM patients; this reduced the risk of information- and selection-bias in this study. The Swedish Myeloma Registry has provided valuable information on how new treatments have been introduced and have been established as standard of care in clinical practice, leading to improved response rates in all age groups. Importantly, we have been able to show that there is good adherence to guidelines in all regions of Sweden, both with regards to diagnostics and to manage- ment, and the registry has helped define areas where improvement is needed. The proportion of patients with prognostic classification according to ISS and for whom FISH was performed as part of diagnostic workup has increased; however, FISH has still not been established as standard clinical practice in all hospitals. One limitation is that treatment data on 8% of patients were incomplete, and some baseline characteristics, such as ISS-stage, were also missing. In addition, we do not have detailed data on cytogenetics and comorbidities. Finally, we did not have sufficient follow-up data to perform analyses on progres- sion-free survival after first-line treatment, which is a fur- ther limitation of this study.
Many large and important studies on characteristics and
512
haematologica | 2018; 103(3)