Improving the identification of frailty in elderly newly diagnosed multiple myeloma patients
The level of frailty, as determined by the International Myeloma Working Group frailty index (IMWG-FI), clearly predicts the clinical outcome of elderly non-transplant eligible newly diagnosed multiple myeloma patients. However, with this index, age >80 years alone automatically classifies a patient as frail, which is at least questionable and possibly leads to under-treatment. Moreover, the IMWG-FI is not suitable yet to select the most vulnerable patients in whom treatment effect will be negligible or whom even will be done harm. Therefore, we investigated the impact of the different frailty factors on treatment discontinuation, PFS and OS with the aim to improve the prognostic value of the IMWG-FI.
The multicenter, non-randomized, phase 2, HOVON 123 trial (NTR4244) included patients aged 75 years or over with a newly diagnosed multiple myeloma. Patients were treated with 9 cycles dose-adjusted melphalan, prednisone and bortezomib. Geriatric assessments were performed at baseline. The primary endpoint was premature treatment discontinuation (within 9 cycles). Frailty factors (age, CCI, ADL, iADL) predicting treatment discontinuation, PFS and OS were identified by Akaike’s Information Criterion (AIC) using a backward selection procedure in a logistic regression model.
Two revised FIs (RFI) based on the IMWG-FI were developed, using a novel cut-off for frailty (≥3 versus ≥2; RFI-1) plus a modified assigned weight of comorbidities (CCI≥3 2 points versus 1 point; RFI-2), given their superior prognostic value for treatment discontinuation. These revised FIs identified a 45%-smaller frail population, without patients formerly defined frail based on an age >80 alone, who had more comorbidities and impairments in (i)ADL, and an inferior survival as compared to the original IMWG-FI: median PFS 12.7 and 14.4 months versus 16.6 months; OS 23.7 and 26.4 months versus 35.0 months, with the RFI-1, RFI-2 and IMWG-FI, respectively.
By reducing the importance of age >80 and increasing the weight of comorbidities, we identified a 45%-smaller but more vulnerable population of frail patients with an inferior clinical outcome, as compared to the original IMWG-FI. Our model can be used as a novel validation platform in differently treated patients in order to move from prognostic to predictive geriatric models.