Multidimensional Quantification of Exclusion Criteria to Reduce the Gap between Randomized Clinical Trial and Real-World Outcomes in Multiple Myeloma

Randomized clinical trials (RCTs) are considered the highest level of evidence to define the efficacy of newly developed treatments before their adoption into clinical practice. RCTs incorporate exclusion criteria that eliminate specific patient populations in order to reduce the incidence of serious adverse events and enhance the efficacy of a given anti-cancer strategy. However, exclusion criteria may lead to a significant gap between patients (pts) enrolled on RCTs and real world pt populations, which represent the ultimate stakeholders in cancer treatment. The analysis of real-world evidence to answer clinical questions has recently gained increased interest. Assessing different dimensions of this gap may help overcome barriers in trial recruitment and enhance the applicability of RCTs in daily practice.

There has been significant advancement in treating multiple myeloma (MM) over the past two decades bringing multiple new mechanisms of action to the bedside. We selected ten recent RCTs: ASPIRE, TOURMALNE-MM01, ELOQUENT-2, ENDEAVOR, POLLUX and CASTOR, OCEAN, ICARIA, APOLLO and ELOQUENT-3 studies, which are pharma-sponsored landmark trials that provided the basis for FDA approval of anti-myeloma agents. Our objective was to quantify the gap in eligibility criteria between the ten RCTs and real world populations by examining these trials using a single institution database.

Methods: Pts with relapsed MM that were initiated on a second (or later line) of therapy that were recognized, retrospectively. Eligibility criteria of the ten landmark RCTs was applied during the 21 day period before the index treatment date. Pts that received Len-containing regimens were tested as to be enrolled on trials with Len/Dex control arm, patients that received Bor-containing regimens were examined to be enrolled on Bor/Dex trials and subjects who had Pom-containing regimen were screened for Pom/Dex trials. Pts were then classified as "Trial eligible" or "Trial ineligible", accordingly and were monitored longitudinally from the index treatment date until death, loss to follow-up, or end of the follow up period. Ten commonly used eligibility criteria were examined (Fig. 1). Any cancer in the three years prior to the index treatment date was counted as "history of other malignancies", i.e., skin and prostate cancer were excluded. Concurrent infection was defined as use of any antibiotic other than acyclovir, ciprofloxacin or bactrim. To calculate area under the curve of the polygon graphs Shoelace algorithm was used.

Results: 516 pts were studied between 2010 and 2020 and 153 were excluded due to missing values. 224, 136 and 98 pts were treated with Len-, Bor- or Pom-containing regimens, respectively. Overall, the trial-eligible cohort was more likely to have autologous stem cell transplant and to have had longer treatment-free period before index treatment date (p-value: 0.009). There was a substantial variation in the ineligibility rate for these ten RCTs among the study population (Fig. 1). The most common items that excluded a patient from a RCT were: other malignancy, current infection and renal dysfunction. Differences between trial-eligible and trial-ineligible pts stratified by trial are listed in Tables 1, 2 and 3 for trials with Len, Bor and Pom as control arms, respectively. The median follow-up for the Len, Bor and Pom cohorts was 31, 30 and 22 months, respectively. Trial-ineligible pts displayed a significantly worse OS (2-year rate 58% vs. 78%, p-value: 0.001) and 49% higher chance of death (HR 1.69, 90%, CI: 1.17-2.62) compared with trial-eligible cohort.

Conclusion: Here, we assessed the multidimensional gap that exists between patient cohorts enrolled on RCTs and real world cohorts for ten landmark MM trials. We present a quantitative deviation score as a tool to calibrate the generalizability of these landmark trials against a single institution. Importantly, we show that trial-eligibility alone significantly correlates with superior OS across a variety of MM clinical trials across all ten MM RCTs. Furthermore, our results reveal that ineligibility rates were quite different among the ten trials which significantly limit cross-trial comparisons. We propose a uniform methodology to assess patient exclusion criteria and narrow the efficacy gap observed between RCTs and real world evidence.

Figure 1

View largeDownload slide

Figure 1

View largeDownload slide

Close modal