Validation of a New Clinical Prediction Model for Outcome in Newly Diagnosed Multiple Myeloma Patients Not Eligible for Autologous Stem-Cell Transplantation; A Population-Based Study from the Danish National Multiple Myeloma Registry

Background: The UK Myeloma Research Alliance recently introduced a new clinical prediction model for outcome in newly diagnosed multiple myeloma (MM) patients not eligible for autologous hematopoietic stem-cell transplantation (ASCT) (Lancet Haematology 2019; 6: e154-66). The score or Myeloma Risk Profile, MRP, includes WHO performance status (PS), the International Staging System (ISS), age, and C-reactive protein (CRP) as prognostic variables. First a score is calculated by the formula: Score = (PS - 2) * 0.199 + (age - 74.4) * 0.0165 + (ISS - 2) * 0.212 + (log(CRP + 1) - 2.08) * 0.0315, where PS and ISS are defined as numbers between 0-4 and 1-3, respectively, and CRP is in mg/L. Next, three risk groups are defined as 1) low risk: score < -0.256, 2) medium risk: -0.256 ≤ score ≤ -0.0283, or 3) high risk: score > -0.0283. The MRP score was generated based on two prospective clinical trial cohorts, the NRCI-Myeloma XI study (ISRCTN49407852) as training set or internal validation, and the NRCI-Myeloma IX study (Blood 2011; 118, 1231-38) as test set or external validation. Both trials investigated conventional oral alkylating agents, cyclophosphamide or melphalan, in combination with thalidomide, lenalidomide, and/or bortezomib; thus including drugs typically used in treatment of elderly MM patients. Establishment of the model included 1852 patients in the training set, and 520 patients in the test set. All patients were recruited as part of clinical trials and therefore fulfilled defined inclusion and exclusion criteria. To validate the MRP score in a population-based setting we performed a study of the entire cohort of transplant ineligible MM patients in the Danish National MM Registry.

Methods: The Danish MM registry started 01 January 2005. It includes registration of all diagnosed MM patients in Denmark and given first- and second-line treatment. A data validation study has been performed (J Clin Epidemiology, 2016; 8: 583-587). At 31 December 2014, 2,926 newly diagnosed treatment demanding MM patients were registered, hereof 1,803 patients were above 65 years and found ineligible for ASCT, and constituted the patient population for this study.

Results:Of 1,803 transplant in-eligible but treatment demanding newly diagnosed MM patients above 65 years 426 patients had one or more missing values for calculation of the MRP score, most often this was caused by missing ISS. Thus, 1,377 patients were evaluable with a median follow-up of 40.9 months. Patients were treated according to standard of care in Denmark during the 10-years registration period which included upfront conventional alkylating agent, mostly melphalan in 37.7%, thalidomide-based in 25.6%, bortezomib-based in 26.1%, lenalidomide based in 2.7%, and only palliative, mostly steroid-based in 7.9%. The distribution of the risk groups according to MRP was as follows: low risk 28.5%, medium-risk 25.1%, and high-risk 46.4%. Ccompared to the UK datasets we had a higher proportion of high-risk patients which undoubtedly reflects that our cohort is population based. Median survivals for the 3 risk groups are presented in Table 1 and overall survival curves illustrated in Figure 1. The model performed well in separating the patients into subgroups with different survival risks.

In conclusion, our real life population-based data confirm that the MRP score is a robust and valuable risk assessment tool for elderly newly diagnosed MM patients older than 65 and not eligible for ASCT. An important advantage of the MRP score is that it is calculated from simple parameters that should be part of everyday diagnostic work-up.

View largeDownload slide

View largeDownload slide

Close modal