In the Journals

FDA using surrogate endpoints more frequently to approve cancer drugs

The FDA increasingly is using surrogate endpoints to support approval of cancer drugs, according to results of a retrospective review in a research letter published in JAMA Internal Medicine.

Use of surrogate endpoints can expedite trial completion but often leaves “substantial uncertainty” about whether a therapy improved quality or quantity of life, Emerson Y. Chen, MD, associate professor of medicine and medical oncology at Oregon Health & Science University, and colleagues contend.

“Surrogate endpoints are inherently less reliable and less certain,” Chen told Healio. “For example, just because someone’s blood pressure is above normal doesn’t mean that person will develop heart attack, heart failure, stroke or need dialysis in his or her lifetime. Likewise, if a tumor shrinks on a CT scan or blood test, that patient may or may not live longer with their cancer. They could logically but, time and time again, we learn it is not always the case. What if the tumor shrinks but then grows in an accelerated way as it becomes resistant to the therapy later?”

The FDA approves drugs based on direct measures of patient benefit — including OS or quality of life — or on surrogate measures, such as change in biomarker level or tumor size on imaging studies. However, surrogate endpoints may not correlate with survival.

Chen and colleagues used the FDA website and a prior systematic review to retrospectively review cancer drugs that received approval between 1992 and July 2019 based on surrogate endpoints, including PFS and response rate. Researchers also reviewed data related to the drug, cancer type, approval basis, dates of approval and post-marketing follow-up.

Determining whether use of the surrogate endpoint occurred for the first time or subsequent times in treatment of the cancer type served as the study’s primary outcome.

Results showed the FDA approved 194 unique drug applications for 132 drugs based on surrogate endpoints during the period analyzed. These included 89 accelerated approvals and 105 regular approvals.

Additionally, the FDA used a surrogate endpoint for the first time for a specific cancer type in 64 (32.9%) of the approvals. The remaining 130 approvals were based on subsequent uses of the surrogate endpoints.

Researchers observed an increase during the period in the number of unique surrogate measure-cancer combinations accepted by the FDA for drug approval, from two approvals (1%) between 1992 and 1995 to 70 approvals (36.1%) between 2016 and 2019. Most of the combinations had unknown post-marketing OS data at the time of approval.

Of the 64 first-surrogate-based approvals, 39 (61%) had no documented correlation with OS, 10 (16%) had a poor correlation ( 0.7), one (2%) had a medium correlation (0.7 < < 0.85) and three (5%) had a high correlation ( 0.85).

Among the 49 approvals with low or unknown correlations with OS, 23 were accelerated approvals and 26 were regular approvals.

“I believe the FDA uses surrogate endpoints in other fields of medicine, but not all diseases are created equal,” Chen said. “Some disease conditions are less common and less lethal than cancer, and so the number of therapies in development is less. Consequently, the balance issue may not be as apparent in other fields.” – by John DeRosier

For more information:

Emerson Y. Chen, MD, can be reached at Knight Cancer Institute, Oregon Health & Science University, Building CHH2, 3181 SW Sam Jackson Park Road, Portland, OR 97239; email: cheem@ohsu.edu.

Disclosures: Chen reports honoraria from Horizon CME. Please see the study for all other authors’ relevant financial disclosures.

The FDA increasingly is using surrogate endpoints to support approval of cancer drugs, according to results of a retrospective review in a research letter published in JAMA Internal Medicine.

Use of surrogate endpoints can expedite trial completion but often leaves “substantial uncertainty” about whether a therapy improved quality or quantity of life, Emerson Y. Chen, MD, associate professor of medicine and medical oncology at Oregon Health & Science University, and colleagues contend.

“Surrogate endpoints are inherently less reliable and less certain,” Chen told Healio. “For example, just because someone’s blood pressure is above normal doesn’t mean that person will develop heart attack, heart failure, stroke or need dialysis in his or her lifetime. Likewise, if a tumor shrinks on a CT scan or blood test, that patient may or may not live longer with their cancer. They could logically but, time and time again, we learn it is not always the case. What if the tumor shrinks but then grows in an accelerated way as it becomes resistant to the therapy later?”

The FDA approves drugs based on direct measures of patient benefit — including OS or quality of life — or on surrogate measures, such as change in biomarker level or tumor size on imaging studies. However, surrogate endpoints may not correlate with survival.

Chen and colleagues used the FDA website and a prior systematic review to retrospectively review cancer drugs that received approval between 1992 and July 2019 based on surrogate endpoints, including PFS and response rate. Researchers also reviewed data related to the drug, cancer type, approval basis, dates of approval and post-marketing follow-up.

Determining whether use of the surrogate endpoint occurred for the first time or subsequent times in treatment of the cancer type served as the study’s primary outcome.

Results showed the FDA approved 194 unique drug applications for 132 drugs based on surrogate endpoints during the period analyzed. These included 89 accelerated approvals and 105 regular approvals.

Additionally, the FDA used a surrogate endpoint for the first time for a specific cancer type in 64 (32.9%) of the approvals. The remaining 130 approvals were based on subsequent uses of the surrogate endpoints.

Researchers observed an increase during the period in the number of unique surrogate measure-cancer combinations accepted by the FDA for drug approval, from two approvals (1%) between 1992 and 1995 to 70 approvals (36.1%) between 2016 and 2019. Most of the combinations had unknown post-marketing OS data at the time of approval.

PAGE BREAK

Of the 64 first-surrogate-based approvals, 39 (61%) had no documented correlation with OS, 10 (16%) had a poor correlation ( 0.7), one (2%) had a medium correlation (0.7 < < 0.85) and three (5%) had a high correlation ( 0.85).

Among the 49 approvals with low or unknown correlations with OS, 23 were accelerated approvals and 26 were regular approvals.

“I believe the FDA uses surrogate endpoints in other fields of medicine, but not all diseases are created equal,” Chen said. “Some disease conditions are less common and less lethal than cancer, and so the number of therapies in development is less. Consequently, the balance issue may not be as apparent in other fields.” – by John DeRosier

For more information:

Emerson Y. Chen, MD, can be reached at Knight Cancer Institute, Oregon Health & Science University, Building CHH2, 3181 SW Sam Jackson Park Road, Portland, OR 97239; email: cheem@ohsu.edu.

Disclosures: Chen reports honoraria from Horizon CME. Please see the study for all other authors’ relevant financial disclosures.

    See more from HemOnc Today's PharmAnalysis