Thank you, Rich, and thanks to the ASCO organising committee for selecting this abstract for presentation on the impact of broadening trial eligibility in advanced non-small cell lung cancer, a real world analysis.
This work and this analysis came out of a project between ASCO and Friends of Cancer Research that created and crafted a set of recommendations for broadening eligibility criteria and that group included a number of stakeholders, including patient advocates, the FDA sponsors and investigators among others. We believe that broadening trial eligibility will allow more patients to participate in clinical research, that that data that’s generated will then be more generalizable to the population of individuals we see in the clinic and that it will accelerate accrual to these trials and optimally bring new therapies to patients more quickly.
These recommendations were published in a series of papers in The Journal of Clinical Oncology in 2017. They led a lot of different groups brought forward attributes of these recommendations. The ASCO TAPUR study is an example that already uses these broader criteria, for example age range down to 12. The FDA has stated that sponsors may be able to enrol these broader populations and analyse them separately. Other groups, including the NCI and cooperative groups, are starting to implement these recommendations across different areas with broader criteria and the FDA has issued draft guidance documents for commentary.
The primary objective of this analysis was to discuss or demonstrate the impact of broadened criteria versus traditional criteria on the eligibility of patients with advanced non-small cell lung cancer. The outcome measures were the number of patients that were excluded by traditional versus these broadened criteria and the characteristics of those populations and cohorts that would be eligible by traditional versus those broadened criteria.
It was a retrospective study on real world data of an eight year period between January 2011 and December 2018. The dataset was the ASCO CancerLinQ Discovery dataset of de-identified electronic health record data. It was conducted in advanced non-small cell lung cancer patients because these are patients with advanced disease and comorbidities may also be common. It’s certainly scientifically relevant because we have many trials available for this population of patients and, just as importantly, the data was curated within CancerLinQ Discovery allowing for an analysis. The inclusion criteria was advanced non-small cell lung cancer stage 3a or higher patients who had received therapy for systemic disease.
Three of the eligibility criteria were used for the analysis as part of the group’s recommendations from 2017. The three areas were prior and concurrent cancers and presence or absence of those cancers, presence or absence of brain metastases and then kidney function, specifically historically criteria mandated that patients with creatinine clearance of 60 or higher would be eligible for studies. The new recommendations suggested that the criteria should be lowered to 30ml/minute of creatinine clearance. This translated to the data that was analysed by including all cases of patients with a prior or concurrent cancer, allowing eligibility of patients regardless of presence or absence of brain metastases and then, finally, including patients that had renal function of a creatinine clearance of 30ml or higher.
This is the dataset, 10,500 cases were selected. By traditional criteria for brain metastases 21% of patients were excluded. Patients who had prior or concurrent cancers, 14% were excluded by traditional criteria. For renal function patients were excluded with creatinine clearance lower than 60ml/minute at a rate of 21.5%. Overall that excluded just under 48% of patients in the cohort. When broadened criteria were applied no patients were excluded due to brain metastases or prior or concurrent cancers and patients excluded for creatinine clearance dropped from 21.5% to 1.5%. Overall this increased the number of patients that would have been eligible for a trial from 5,495 to 10,346.
Age groups that were represented in these cohorts were altered slightly, particularly in the group of patients above 75 years. Those patients went from 16% of the traditional cohort to 22% of those in the broadened criteria. Patients in geriatric oncology and older patients are under-represented in trials overall so expanding these three criteria will increase that population by 6%.
Limitations to our data are here. The generalisability of these results were not assessed. These were 50 sites within the CancerLinQ practices that are geographically diverse and generally non-academic settings. It was difficult at times to match trial eligibility criteria to EHR structured data within the health record. We simplified the brain metastases definition of presence or absence, implemented a rule to consider the primary cancer code at the likely metastatic site, not as another primary cancer, which affected less than 1% of the overall population, and similarly performance status is not routinely reported in EHR structured data making that eligibility criteria more challenging to analyse. Also the analysis was not inclusive of all of the ASCO and Friends’ recommendations from the papers published in 2017, for example, HIV status, other organ function and age.
In conclusion, the use of expanded criteria would enable nearly twice as many non-small cell lung cancer patients to consider participation in clinical trials. These broadened criteria are likely to result in participants that are more reflective of the patients we see in clinics, improving generalisability of the data from these trials. Narrower criteria to exclude patients should only be used based on compelling scientific rationale. Additional ASCO and Friends of Cancer Research recommendations are currently in progress. At this point this group urges all clinical trial sponsors to adopt these criteria. Thank you.
