Measuring Expectations: Real-World Survival With Osimertinib

Over the past few years, osimertinib has received much fanfare among lung cancer experts. 

Results from the phase 3 LAURA trial were met with a standing ovation at the 2024 American Society of Clinical Oncology annual meeting, after the trial demonstrated a 33.5-month progression-free survival benefit in patients with EGFR-mutated unresectable stage 3 non–small cell lung cancer (NSCLC).

Several years earlier, the FLAURA trial established AstraZeneca’s EGFR-directed tyrosine kinase inhibitor as frontline therapy in patients with EGFR-mutated advanced NSCLC. The trial reported a 7.2-month median overall survival advantage in patients receiving osimertinib vs other EGFR inhibitors, gefitinib or erlotinib.

In 2018, osimertinib gained an FDA approval based on findings from the FLAURA trial and has garnered several other indications in advanced NSCLC as well as earlier-stage disease in the adjuvant setting. 

As of 2020, more than 90% of patients in the United States receive the drug as part of first-line treatment for locally advanced or metastatic EGFR-mutated NSCLC. 

However, a recent analysis in a real-world population of patients with advanced disease provided a bit of a reality check.

The phase 3 findings, published in Lung Cancerearlier this year, explored outcomes with first-line osimertinib among patients who would have been eligible for FLAURA as well as patients who would not have been eligible.

Ineligibility was based on five core criteria: Worse ECOG performance status (≥ 2); symptomatic brain metastases, or remained on steroids or spinal cord compression; hemoglobin levels under 90 g/L; platelet count under 100×109/L; or creatinine levels over 1.5-fold the upper limit of normal, with a clearance of < 50 mL/min.

In this real-world setting, 44% of patients (137 of 311) would have been ineligible for the FLAURA trial, predominantly due to low performance status (n = 120).Patients who would have been ineligible also had significantly higher rates of de novo stage IV disease and were more likely to have stage IVB disease as well as more than 3 sites of metastases.

At a median follow-up of 26.5 months, 70% (96 of 137) of FLAURA-ineligible patients had died vs 36% (63 of 174) of FLAURA-eligible patients. 

Patients who would have been eligible for FLAURA had similar median overall survival to patients in the FLAURA trial — 34.2 months (vs 39.1 months in FLAURA). The ineligible cohort, however, fared much worse, with a median overall survival of 15.8 months — more than 18 months shorter than those in the eligible cohort.

The median time to treatment discontinuation among FLAURA-eligible patients was 26.8 months vs 11.9 months among noneligible patients.

Because patients in the FLAURA-ineligible group have a worse prognosis overall, the difference between survival in this group and in the FLAURA-eligible patient group was not necessarily a surprise, said study author J. Connor Wells, MD, of the BC Cancer Agency in Vancouver, British Columbia, Canada.

“Patient performance status is often a good indicator of prognosis, so it is not entirely unexpected given that most of the patients in the ineligible group had a performance status of 2 or worse,” Wells told Medscape Medical News. Nevertheless, “it was a fairly stark difference compared to the eligible group.”

The results should help clinicians provide more measured expectations to patients.

“When patients want to discuss expected survival duration, we often fall back to what the large phase 3 clinical trial reported with the caveat that every patient is different and each cancer will respond to treatment in a different way,” Wells said. “However, our study provides more realistic numbers for patients who are more unwell to begin with.” 

Should FLAURA’s eligibility criteria have been broader to reflect the range of real-world patients?

Not necessarily, according to Stephen Liu, MD, who wasn’t involved in the analysis.

Liu explained that using more conservative inclusion criteria is important in clinical trials to provide a clear picture of treatment-related issues, such as toxicities. In many cases, for instance, complications occur in this harder-to-treat population because of comorbidities, not directly because of the cancer or the treatment, said Liu, director of Thoracic Oncology, head of Developmental Therapeutics of the Georgetown Lombardi Comprehensive Cancer Center, Washington, DC. 

And, even with the “striking” difference in survival, these real-world findings “do not change practice,” Liu noted.

Osimertinib remains the best option for many patients with NSCLC. What this real-world trial may ultimately indicate is, for patients who are sicker, even the best therapy likely won’t significantly change their trajectory.

But these findings should reframe expectations for oncologists and for patients.

The trial “serves as a sobering reminder that real-world outcomes often fall short of trial results,” Liu said.

Wells had no disclosures to report. Two study authors reported relationships with AstraZeneca. Liu reported consulting or research relationships with AstraZeneca as well as a range of other companies, including Abbvie, Boehringer Ingelheim, Bristol Myers Squibb, Daiichi Sankyo, Genentech/Roche, Gilead, Guardant Health, Janssen, Jazz Pharmaceuticals, Merck, and Pfizer.