Conference Coverage

Survival seen in mCRC trials has limited generalizability to real world




SAN FRANCISCO – Survival benefits seen with first-line therapy for metastatic colorectal cancer in clinical trials have limited generalizability to patients treated in real-world practice, suggests a study reported at the 2015 Gastrointestinal Cancers Symposium.

Investigators compared 16,614 patients from 31 randomized phase III trials of first-line therapy published between 2005 and 2010 with similar matched patients from the Surveillance, Epidemiology, and End Results (SEER) database.

Findings showed that the patients on trials as a whole had a median overall survival of 18.9 months and 1- and 2-year overall survival rates of 70% and 36%, respectively, Ziwei Wang reported in a poster session.

Ziwei Wang

Ziwei Wang

Relative to their real-world counterparts, trial patients had better median overall survival in 95% of trial arms (average difference, 5.4 months), better 1-year overall survival in 94% of trial arms (average difference, 16.7%), and better 2-year overall survival in 71% of trial arms (average difference, 7.2%) (P less than .0001 for each).

“When patients go to their physicians, they want to know what’s the prognosis, what benefit does the drug that you are giving me translate into, how does that help my overall survival,” Ms. Wang commented in an interview. “And most physicians get their information just from the clinical trial data that comes out.”

Taken together, the study’s findings are “a good message to physicians just to be mindful that these differences do exist and not to just draw from clinical trials to inform your patients,” she said. “You can sort of think about what you are telling them to let them know that the information you are giving them is coming from trials that might not be reflective of what we see in the general population.”

The survival gaps likely stem in part from the often-strict eligibility criteria used in clinical trials, according to Ms. Wang, who is a medical student at the University of California, San Diego.

“When you are designing clinical trials, you have a set of criteria that a patient has to meet before they are even allowed to enroll, so already, you might kind of be drawing from an unrepresentative sample when you start these trials,” she elaborated. “And so we think that a lot of that plays into why you see these differences between what the trial concludes and what SEER shows.”

A previous similar study compared survival among patients with a variety of cancers treated in versus out of clinical trials (J. Natl. Cancer Inst. 2014 [doi:10.1093/jnci/dju002]), Ms. Wang noted at the meeting cosponsored by the AGA Institute, the American Society of Clinical Oncology, ASTRO, and the Society of Surgical Oncology.

“They found that initially, the clinical trial patients did better, but after 5 years or so, it kind of evened out,” she said. “What we found was up until 5 years and even longer, you do see a difference.”

Next Article: