This article is more than 5 years old. Information may no longer be current.
High ORR may be appropriate endpoint to signal breakthrough activity
High overall response rates appeared associated with regulatory approval of single-agent anticancer agents, according to the results of a systematic database search published in JAMA Oncology.
However, the connection between overall response rate (ORR) and regulatory approval of combination regimens appeared less strong, results showed.
Geoffrey R. Oxnard, MD
Since the FDA created the ‘breakthrough therapy’ designation in 2013 to facilitate the review of active therapies for serious conditions, nine anticancer therapies have been approved under the breakthrough therapy designation.
Single-arm trials — although limited by the lack of a comparator group and the potential for selection bias — have been discussed as having potential to adequately support regulatory approval via accelerated development. ORR on tumor imaging may be one endpoint used in single-arm trials.
“However, such single-arm trials studying ORR have not historically been used as definitive studies to support regulatory approval, and the optimal statistical design of such definitive single-arm studies has not been established,” Geoffrey R. Oxnard, MD, thoracic oncologist at Dana-Farber Cancer Institute, and colleagues wrote. “Toward the aim of identifying appropriate statistical endpoints for definitive single-arm trials, we characterized the relationship between ORR and regulatory approval of anticancer therapies for common solid tumors.”
Oxnard and colleagues performed a systematic review of 1,800 oncology clinical trials registered at clinicaltrials.gov between October 2007 and September 2010. Eligible trials enrolled at least 20 patients per trial arm and reported ORR based on Response Evaluation Criteria in Solid Tumors.
The researchers identified 874 eligible trial arms from 578 clinical trials evaluating anticancer therapies for colorectal cancer, melanoma, non–small cell lung cancer and renal cell carcinoma.
Of these, 542 trial arms had sufficient ORR data, representing 294 therapeutic regimens.
NSCLC comprised the largest proportion of eligible trial arms (46%). The majority of studies were phase 2 (60%) or phase 3 (22%), with a median trial arm enrollment of 52 patients (range, 20-5,394; interquartile range, 36-98).
The researchers then calculated the mean and maximum ORR of each single-agent or combination regimen. They evaluated the association between ORR and regulatory approval, defined as approval in any country for the treatment of that cancer.
Overall, increased ORR appeared associated with greater likelihood of regulatory approval (maximum ORR, Kendall tau = 0.27; P < .001; mean ORR, Kendall tau = 0.12; P = .01).
This relationship appeared stronger for single-agent therapies (maximum ORR, Kendall tau = 0.49; mean ORR, Kendall tau = 0.41) than for combination therapies (maximum ORR, Kendall tau = 0.28; mean ORR, Kendall tau = 0.17).
Richard Pazdur, MD
The researchers performed an evaluation of ORR thresholds to improve future single-arm trial designs. Results showed that an ORR statistically exceeding 30% for a single-agent therapy had a 98% specificity and 89% positive predictive value for identifying regimens achieving regulatory approval. Specificity and positive predictive value reached 100% with an ORR of 45%.
However, the researchers did not identify a predictive ORR threshold for combination regimens, with a maximum positive predictive value of 47% to 60%.
“Sensitivity was generally poor across all thresholds for both types of regimens, highlighting that high ORR is not a sensitive metric for identifying all regimens that can achieve regulatory approval,” the researchers wrote.
Study limitations included the lack of data from trials outside the specific search periods and the restriction of the analysis to trials studying drug regimens for solid tumors. The researchers also noted that newer trials focusing on immunotherapy agents were largely not included, and that these results may not be generalizable to other solid tumor types, such as breast cancer, prostate cancer or ovarian cancer.
“We present a large and unbiased analysis of the relationship between tumor response and regulatory approval,” Oxnard and colleagues wrote. “For single-agent regimens, maximum ORR statistically exceeding 30% to 45% was associated with regulatory approval, with positive predictive values in the range of 89% to 100%. This suggests that high ORR is an appropriate statistical endpoint for single-arm trials aiming to demonstrate breakthrough activity of a single agent.”
Although ORR is a valuable endpoint by which to measure clinical benefit, other aspects related to therapeutic response are being investigated for drug approval, Gideon M. Blumenthal, MD, clinical team leader of thoracic and head and neck oncology at the FDA and Richard Pazdur, MD, director of the Office of Oncology Drug Products at the FDA, wrote in an accompanying editorial.
“For immunotherapies and other cytostatic therapies, other factors may need to be considered when evaluating tumor response, including tumor growth kinetics, depth of response, durability of response and tumor volume,” Blumenthal and Pazdur wrote. “Change in tumor volume has been used as an approval endpoint for everolimus [Afinitor, AstraZeneca] in subependymal giant-cell astrocytomas. We at the FDA and other stakeholders are actively pursuing investigation into alternate metrics of response to better describe clinical benefit.” – by Cameron Kelsall
Disclosure: Oxnard reports consulting fees or honoraria from Ariad, AstraZeneca, Boehringer Ingelheim, Chugai and Clovis Oncology. Please see the full study for a list of all other researchers’ relevant financial disclosures. Blumenthal and Pazdur report no relevant financial disclosures.