Life expectancy and incidence of cancer have substantially increased, the latter being closely interlinked to our longevity. Today, 617 million people are ≥65 years; by 2050, this number will have reached 1.6 billion, nearly 20% of the world’s population, and the number of "very old" (>80 years) will have more than tripled.1 This aging of the population involves enormous changes to patient care. For the moment, the most profound changes are to be seen in Japan, Europe and North America. Major risk factors associated with aging include cancer (also multiple cancers in a single patient),2 and cardiovascular and neurodegenerative diseases, all requiring long-term care. Therefore, especially high-income countries are obliged to meet the challenges.1
Multiple myeloma (MM), as one example of cancer, and the 2nd most frequent hematologic malignancy, affects adults of all ages, but is primarily a disease of the elderly. The highest burden of MM-related deaths occurs among persons between 65 and 84 years of age.3–7 Similarly to the situation in several other types of cancers, management of older MM patients is more demanding due to their often impaired organ function, underlying comorbidities, and co-existing frailty, which may increase therapy-related toxicity, and lead to dose reduction and shorter treatment endurance.3,4,6–9 The high prevalence of geriatric impairments is increasingly being recognized, but is not always easily detectable without an objective assessment.3,6,7 Our goal today involves reducing the risk of under-treating fit patients and over-treating those who are frail.5,10–12 Although eligibility criteria for studies of anti-cancer/-MM agents have traditionally relied on age cut-offs and performance status, geriatric and MM-specific frailty assessments are just beginning to be incorporated into more accurate stratification plans of treatment algorithms.6,7,11,12 Similarly to MM patients, geriatric assessments (GA) have been defined for patients with chronic lymphocytic leukemia (CLL)8,13,14 and myelodysplastic syndrome (MDS),15,16 where determination of frailty versus fitness has moved into clinical practice. However, solutions as to how they might be more uniformly used and valued in their daily pratice have not been fully determined.
Recommendations of the geriatric oncology working groups (i.e. German Society of Geriatrics/German Society of Hematoloogy&Oncology) have suggested GA-tools to check comorbidity in patients aged ≥70 years via the Charlson Comorbidity Index (CCI), cognition via the Mini Mental test (MMST), activity/instrumental activity of daily living (ADL/IADL), mobility via the Timed Up and Go test, depression via the geriatric depression scale (GDS), and nutrition via body mass index (BMI) and Mini Nutritional Status.6,7,11,12,17 While these GA-tools have been established and validated, their execution is time-consuming, an additional workforce is needed, and the involvement of a geriatric team is advisable.6,7,9,11,12,17 Whether shorter frailty scores in cancer patients may substitute and/or add to GA-tools is being pursued in single- and multi-center trials (Table 1).
The aim of this commentary is to define strategies in MM patients, and explore how frailty assessment may be employed in clinical practice and clinical trials.
The epidemiologic and biologic considerations of elderly MM patients, with widely expanding treatment options, have motivated global efforts to validate simple instruments to assess vulnerability of patients, test them in their clinical significance to predict treatment outcome [overall survival (OS) and progression-free survival (PFS)], occurrence of severe adverse events, and to tailor treatment with more or less intensified regimens.11,12,18
Under-treatment of fit elderly patients has been demonstrated to occur more frequently than over-treatment.12 Under-treatment may prevent improvement of organ function, while over-treatment of frail patients can induce unnecessary morbidity and mortality. Both instances reduce patients’ health-related quality of life (HRQOL). In a study that assessed HRQOL across >16,000 cancer survivors, those with MM were among those with the lowest HRQOL scores, highlighting the urgent need for this to be improved and for frequent reassessment of HRQOL in cancer patients.19 The art of managing elderly MM patients involves balancing competing disease-related and patient-specific factors and to make adequate treatment decisions.
Numerous induction (and relapse) MM-treatment options are available today. These include bortezomib-cyclophosphamide-dexamethasone (VCD), bortezomib-lenalidomide-dexamethasone (VRD or VTD), bortezomib-melphalan-prednisone (VMP) or antibody-combinations, autologous stem cell transplantation (ASCT) and 2-drug combinations, such as lenalidomide-dexamethasone (Rd), bortezomib-dexamethasone (Vd), and others.3,20–22 These largely expanded therapeutic strategies, including immunotherapies,23 have significantly evolved in recent years, but the beneficial effect is not seen across the age spectrum, with intermediate-fit or frail patients not obtaining the maximal benefit from such new treatment. Part of this failure to achieve benefit relates to the host biology of older patients. Therefore, there is an unmet need to give the right therapy to the patient most suited to benefit from it; the starting point for this approach is an appropriate classification of who is fit and who is frail.
That age alone is a much less well-suited discriminator for treatment designation has been shown via various risk parameters and comorbidity scores, that are usually described as patient-related factors.11,12,24–26 These involve simple measures of daily activity, such as constitution via Eastern Cooperative Oncology Group (ECOG) or Karnofsky performance status (KPS), organ function, and comorbidities. However, because the KPS/ECOG do not reflect the entire functional status of cancer patients, advances in defining patient fitness more precisely are warranted. In an analysis of 466 consecutive MM patients, the median KPS was determined to be 90%, although a precise reassessment showed this was actually 60%, i.e. 30% lower than that estimated by physicians. This clearly demonstrates that both KPS/ECOG are often over-estimated, and a more precise frailty assessment is valuable.25 In a subsequent analysis, 13 comorbid conditions were assessed in 801 patients. These were graded and rated according to the Common Terminology Criteria for Adverse Events (CTCAE 4.03), which included: renal-, lung- and KPS-impairment, cardiac, liver or gastrointestinal disease, disability, frailty, infection, thromboembolic events, peripheral neuropathy, pain, and secondary malignancies. In addition, age, cytogenetics via fluorescence in situ hybridization, renal function and lung disease were determined. The multivariate Cox proportional hazard model based on backward selection revealed five highly significant risks as relevant for OS: renal and lung function, KPS, age, and frailty (Fried definition). Score weights for comorbidities were determined on the basis of regression coefficients of the prognostic factors.12 Although impairment of organ function such as lung disease had been defined as having nothing to do with MM, in line with other large MM study groups (such as both the German GMMG and DSMM study groups), patients with moderate and severe lung impairment and continued smoking habit were at substantial risk for treatment complications.11,12 We would, therefore, refrain from ASCT/allogeneic-SCT, triplet and quadruplet therapies in heavy smokers and/or those with impaired lung function.11,12,24–26 Moreover, disease-related factors add additional complexity in MM, such as cytogenetics, International Staging System (ISS)/revised (R)-ISS stage, bone marrow infiltration, and number of CRAB (C, hypercalcemia; R, renal impairment; A, anemia; B, bone lesions) symptoms. In addition, treatment-related factors, such as how quickly and for how long the disease responds to therapy, are critical.27–29
In prior organ function analyses,11,12,24–26,30 the extent of frailty in MM patients was substantial: 60% for entire (mild to severe) and 40% for severe frailty.11,12 This led to the development of the revised myeloma comorbidity score (R-MCI). This R-MCI uses weights generated via multivariate risk factor assessment with the essential risks being included therein, such as: renal and lung function, KPS, frailty and age, with the option to add cytogenet-ics.11,12 Apart from organ impairment, cytogenetic aberrations corroborate with impaired OS in MM patients. The analysis confirmed that cytogenetics provide independent additional information,31–35 and that patients with unfavorable cytogenetics had higher disease stages, adverse laboratory values, and reduced organ and physical function. Although cytogenetics proved to be a relevant risk factor, the analysis confirmed that others, such as physical and organ conditions, are equally important.6,7,11,12,18 Moreover, development of the R-MCI showed that the multivariate risks (renal, lung function, KPS, age, frailty) defined patients as fit, intermediate-fit, and frail, which could be improved with inclusion of cytogenetics, (but which could still be used even if this information was unavailable). Weighting of the R-MCI verified that this 9-point score defines three patient groups with clearly different survival,12 which remained true regardless of treatment or age subgroups.
Comparison of the R-MCI with numerous others [CCI, Kaplan Feinstein (KF), Hematopoietic Cell Transplantation Comorbidity Index (HCT-CI/Sorer),12 the Satariano Index24–26 or the International Myeloma Working Group (IMWG frailty score11) showed that they all divide patients into risk groups with substantially different OS. However, Brier scores determined the smallest prediction errors with the R-MCI. One reason for the comparability of the R-MCI with others was that most include risks that have some relevance in MM, namely renal and lung function, and physical condition. Compared to the initial non-weighted MCI,24–26 the R-MCI led to an improvement in group distinction, which highlights the relevance to further improve a risk score, as performed in subsequent analyses.11,12,24–26,30 Various risk scores that are used in different institutions and within clinical trials in MM patients are summarized in Table 1.18,36–38
The question is, therefore, whether one comorbidity score in MM should be put forward or whether more should be developed. Moreover, would harmonization and inclusion of biomarkers improve them?39 Another question is if MM experts will use these scores and whether treatment decisions are being improved.40,41 Whether risks determined by a score result in changes in treatment decision has not been fully addressed. Given that MM primarily affects the elderly, whose vulnerabilities may change over time, it is also reasonable to incorporate serial GA throughout treatment in order to potentially modify therapy over time and incorporate this into tumorboards and treatment guidelines. For older, fit patients, intensive treatment with ASCT may be appropriate, while in the very frail, with GA and high R-MCI-scores, end-of-life care discussions can be facilitated.10,42
Training and validation analyses of the R-MCI showed well-discriminated risk profiles in terms of both PFS and OS for fit, intermediate fit, and frail patients.12 This was true both for more intensively and less intensively treated MM patients.12 Moreover, if MM patients were risk-assessed via the R-MCI rather than the IMWG-frailty score, Kaplan-Meier analysis produced more clearly separated PFS and OS curves with the the R-MCI than with the IMWG-frailty score.11
Importantly, if patients are intermediate-fit or frail by R-MCI, precautions for dose reduction of systemic treatment can be made: i.e. if advanced frailty in MM patients is observed, dose reductions can be discussed, including whether the disease aggressiveness needs effective anti-MM treatment to be performed in spite of the patient’s frailty. Today, it seems important, given the widely different anti-MM treatment options, that the frailty scoring specifically warns MM experts that complications with intensive treatment may occur. As many precautions as possible can then be taken while treatment is being given, such as inpatient rather than outpatient treatment, observation on the ward until complications no longer occur, prophylactic medications, etc. In line with this, in their joint EMN-paper,7 the European Myeloma Network (EMN) consensus has stated that in fit MM patients, efficient antimyeloma therapy with the aim of deep remission is key, whereas in unfit or frail patients, the priority is to maintain a good balance between therapy efficacy and safety.
Although the IMWG-frailty score is a “reference” comorbidity index,18 others are more straightforward to use. The inclusion of “Lung function” in the R-MCI had been repeatedly requested by reviewers as a more objective measure than via the GOLD criteria, smoking status or dyspnea upon exertion, and is included in the diagnostic workup at our center (i.e. before intensive treatment, such as SCT).11,12,21,44–47 If unavailable, smoking status, its mandatory cessation before SCT/intensive treatment, no advanced GOLD criteria, and no dyspnea upon exertion have been used as substitutes in prior analyses.24–26
We have demonstrated the validity of the R-MCI as a valid prognostic instrument in large MM cohorts treated according to current standards. Based on existing recommendations, the R-MCI can be applied in routine clinical care, multicenter analyses and future clinical trials. It may be used in research to compare risk profiles of MM cohorts, to adjust for imbalanced risks, and to provide a basis to establish new clinical or biologic prognostic factors. Moreover, the R-MCI might be considered to be an integral part in the development of individualized risk-adapted treatment strategies to further improve outcome in MM. This includes correct use of resources, higher inclusion rates of older patients into clinical studies, and avoidance of under-supply for older but fit patients.
In the future, the R-MCI could also help to support treatment decisions, tolerability, and to avoid toxicity. Since any prospective comorbidity, frailty and disability evaluation can be time-consuming, we have implemented the R-MCI within a web-based technology application which allows a quick turnaround of results.48 Routine measurement of frailty in MM patients is, therefore feasible, and several analyses via R-MCI11,12 (or IMWG frailty scores18 with various adaptations6,47) are available.
All current developments in the field are enthusiastically welcomed, because more effective and individualized treatment offers an opportunity to further improve clinical outcomes, especially among older patients with hematologic malignancies.21,27,28,49 We have proven tools for a functional, more objective assessment to help guide every-day treatment, and these should be incorporated into tumor boards and may allow better trial comparability, as well as helping to guide trial design. It will be interesting to see whether, in the near future, these risk tools are readily implemented into clinical care and can improve patient management. We are entering an exciting era for research on aging, which holds unprecedented promise for increased patient lifespan, delaying pathologies of aging, allowing patients to grow old, and living a life full of purpose and well-being (Figure 1).1,50 Future clinical trials that target the aging process and that study biomarkers and intervention programs face considerable challenges, but the potential rewards will far outweigh their risks.1,50
The authors thank distinguished IMWG, EMN, DSMM and GMMG experts, especially Drs. Thierry Facon (Lille), Hervé Avet-Loiseau (Toulouse), Alessandra Larocca, Sara Bringhen, Francesca Gay (Turin), Gordon Cook (Leeds), Sonja Zweegman (Amsterdam), Heinz Ludwig (Vienna), Hermann Einsele (Würzburg) and Hartmut Goldschmidt (Heidelberg) for their advice and recommendations. We are also grateful of the insightful comments of all myeloma experts and colleagues at the University Clinic Freiburg (UKF), for their inspiring work and thrive for R-MCI topics. The paper is dedicated to Dr. Sandra Maria Woerner, Maximillian Holler, Sophia Scheubeck, Katja Schoeller, Irina Surlan, Mandy Möller, Stefanie Ajayi, Amelie Rösner, Drs. Martina Kleber, Laura Gengenbach, Giulia Graziani, Maximillian Schinke, Heike Reinhardt, Stefan Müller, Matthias Weiß, Johannes Jung, Christine Greil, Michael Rassner, Cornelius Miething, Milena Pantic and Barbara Deschler (UKF, University of Basel and University of Würzburg) for their exceptional support. We are indebted to the interdisciplinary MM tumorboard, the CCFF for incorporation of the R-MCI in its application modus and thank all MM patients who participated in our MM studies. We also thank the 3 insightful, knowledgeable reviewers for their critical comments, which have further improved this article.
This work was supported by the Deutsche Krebshilfe (grants 1095969 and 111424 [to ME and RW]) and Verein Senioren-Krebshilfe (R.S.).