Symptomatic Progression Free Survival (SPFS): A New Endpoint for Multiple Myeloma (MM) Combining Patient-Reported Outcomes (PROs) and Progression-Free Survival (PFS)

BACKGROUND

In other therapy areas, endpoints have been established which combine progression and its impact on patients (pts). An example of this is the ‘time to symptomatic progression’, which is considered pt-relevant in metastatic hormone-sensitive prostate cancer.

We sought to combine PROs and PFS to create ‘Symptomatic Progression-Free Survival’ (SPFS), a novel endpoint in MM that allows for a pt-centric analysis of progression.

METHODS

First, we identified the PROs that are both important to pts and could be expected to be linked to the risk of progression. To do this a systematic literature review (SLR) was performed to identify relevant symptoms, which were then tested by selected MM experts using Delphi methodology.

Delphi is a technique that uses multiple iterations designed to develop a consensus. Three rounds of electronic surveys were undertaken with 10 experts in MM. All responses were anonymous. The threshold for consensus was set at ≥70% agreement, or for categorical scales (ie, none, low, moderate, high, very high), consensus required all answers to fall within a range of ≤1.

A joint model using the MagnetisMM-3 study of elranatamab (ELRA; BCMA-directed therapy naive pts with triple-class exposed MM) was then developed to test whether these symptoms (as captured by repeatedly measured PRO scores) had a statistically significant impact on the risk of progression or death over time. Data on pts with a PRO assessment at baseline and ≥1 PRO assessment post baseline were used. Any symptoms without a statistically significant relationship with progression were dropped from the final generation of the SPFS endpoint.

Lastly, a composite endpoint for SPFS was created combining the identified symptoms and PFS (defined as clinical progression or death accompanied by a 10-point worsening in PRO scale [corresponding to the minimally important difference], within 28 days of the PFS event). Kaplan Meier (KM) curves and median times were generated for pts treated with ELRA.

RESULTS

The SLR yielded 3253 publications, resulting in 450 publications for full-text review after title and abstract screening, and 34 publications selected for inclusion. Symptoms identified as being important to MM pts and related to progression included: anxiety, appetite loss, constipation, depression, diarrhea, difficulties remembering, drowsiness, dyspnea/breathlessness, fatigue, insomnia, mouth problems, nausea/vomiting, pain, peripheral neuropathy, poor mobility, and tingling in the hands and feet.

The Delphi consensus was that ‘pain’ and ‘poor mobility’ have a high or very high impact on a pt’s quality of life and that ‘drowsiness’, ‘fatigue’, ‘pain’ and ‘poor mobility’ are linked to MM disease progression.

From the joint model, 3 of the 4 symptoms identified by the Delphi were found to have a significant association with PFS (effect size measures [credible intervals]): ‘pain’ (0.022 [0.008, 0.036]), ‘poor mobility’ (-0.027 [-0.042, -0.013]) and ‘fatigue’ (0.018 [0.002, 0.034]), with ‘drowsiness’ not reaching statistical significance (0.011 [-0.020, 0.041]). For each of these symptoms the impact of a 10-point change in the symptom on the risk of progression or death was calculated. A 10-point worsening in QLQ C-30 ‘poor mobility’ score had the biggest impact (increasing the risk of progression or death by 31%) with ‘pain’ and ‘fatigue’ increasing risk by 23% and 20%, respectively.

Once the final composite endpoint was developed, median SPFS (mSPFS) was not reached (NR) (95% CI, 12.2 – NR; KMs to be provided), and there was a strong link to PFS, with 84% of pts with clinical progression or death also experiencing an SPFS event (ie, PFS and a 10-point worsening in one of the symptoms).

CONCLUSIONS

A systematic approach was taken to develop a pt-centric PFS endpoint. Results from the joint model suggest there is a relationship between the identified symptoms and progression, except for ‘drowsiness’. Inclusion of the 3 statistically significant symptoms in the composite endpoint with PFS resulted in a new pt-relevant endpoint of SPFS. After a median follow up of 7.5 months (range, 0.46-34.99), mSPFS was NR. While this work is a unique standalone study, our SPFS endpoint can be leveraged in future studies to provide additional value in understanding the efficacy of treatment options in MM. Further research will involve replicating the Delphi methodology with pts to evaluate concurrence with the consensus of the MM experts.