KEYNOTE-177: What the Colorectal Cancer Trial Found

A trial about a test result, not an organ

Most colorectal cancer does not respond to immunotherapy. A small slice does, and KEYNOTE-177 was built around finding that slice first.

The slice is defined by two related lab terms. Mismatch repair is the cell's spellchecker for DNA copying errors. When it is broken, the tumor is called mismatch repair deficient, shortened to dMMR. The result is microsatellite instability, where short repeating stretches of DNA come out the wrong length. A tumor with a lot of that is called microsatellite instability high, or MSI-H.

Those tumors carry an unusually large number of mutations, which makes them easier for the immune system to recognize.

Before this trial, checkpoint blockade was already known to help MSI-H tumors after chemotherapy had failed. Nobody had tested it as the first treatment.

The design

KEYNOTE-177 was a phase 3, open-label trial. It enrolled 307 people with metastatic MSI-H or dMMR colorectal cancer who had not been treated for it before.

They were randomly split in half. One group got pembrolizumab, a PD-1 checkpoint inhibitor, by infusion every three weeks. The other got chemotherapy every two weeks: a 5-fluorouracil-based regimen, with or without bevacizumab or cetuximab.

People on chemotherapy could cross over to pembrolizumab if their cancer grew. Progression-free survival and overall survival were both primary endpoints.

Open-label means everyone knew which treatment they were getting. That is normal when the two treatments look and feel nothing alike, but it can color how side effects get reported.

What it found

At the second interim analysis, after a median follow-up of 32.4 months, pembrolizumab beat chemotherapy on the first endpoint.

Median progression-free survival was 16.5 months against 8.2 months. The hazard ratio was 0.60, with a 95% confidence interval of 0.45 to 0.80 and a p-value of 0.0002.

Tumors shrank in 43.8% of the pembrolizumab group and 33.1% of the chemotherapy group. The difference in how long those responses lasted was starker: at 24 months, 83% of responses in the pembrolizumab group were still going, against 35% in the chemotherapy group.

The side effect gap ran the other way from what many people expect. Treatment-related side effects of grade 3 or higher occurred in 22% of the pembrolizumab group and 66% of the chemotherapy group, including one death.

At the data cutoff, 56 people in the pembrolizumab group and 69 in the chemotherapy group had died. Overall survival data were still accumulating, with 66% of the required events reached, and remained blinded.

Why the test comes first

None of this applies to a colorectal cancer that is not MSI-H or dMMR, and most are not.

That makes tumor testing the gatekeeper. NCI notes that tissue taken at biopsy may be checked for the gene changes behind Lynch syndrome, an inherited condition that causes mismatch repair problems. Our page on biomarker testing explains what these tests look for and why the result changes the plan.

When to get checked

Colorectal cancer is one of the few cancers where screening can prevent the disease, not just find it early, because polyps can be removed before they turn. Most expert groups advise starting at age 45 for average risk. Our page on colorectal cancer screening covers the options.

Between screenings, NCI says to check with a doctor about:

• Blood in the stool, either bright red or very dark.
• A change in bowel habits, including diarrhea or constipation.
• A feeling that the bowel does not empty completely.
• Stools that are narrower or a different shape than usual.
• Frequent gas pains, bloating, fullness, or cramps.
• Weight loss with no known cause.
• Fatigue or vomiting.

A family history of colorectal cancer, or a known Lynch syndrome in the family, means starting earlier. That is a conversation to have before symptoms, not after.

The wider picture

SEER, NCI's cancer surveillance program, republishes an American Cancer Society projection of about 158,850 new colorectal cancers and 55,230 deaths in the United States in 2026. Five-year relative survival across all stages was 65.4% for people diagnosed from 2016 through 2022.

By stage, SEER records 34% found while local, at 91.3%. Thirty-seven percent are found in nearby lymph nodes, at 75.2%. Twenty-three percent are found after distant spread, at 16.9%.

KEYNOTE-177 enrolled people from that last category, and only the MSI-H portion of it. These registry figures describe a whole population over past years, not any individual.

What this trial cannot tell you

The headline result is progression-free survival. When the paper was published, overall survival data were still blinded and incomplete, so the trial had not shown that pembrolizumab helps people live longer.

Crossover complicates that question permanently. People on chemotherapy could switch to pembrolizumab when their cancer grew, which is the right thing to do for them and makes a clean survival comparison harder to read.

It also cannot tell you anything about colorectal cancer that is not MSI-H or dMMR, which is the large majority. And a 43.8% response rate means the majority of people in the pembrolizumab arm did not see their tumor shrink.

Sources

• Andre T et al., Pembrolizumab in Microsatellite-Instability-High Advanced Colorectal Cancer, N Engl J Med 2020 (NCBI E-utilities record)
• ClinicalTrials.gov record for NCT02563002 (KEYNOTE-177)
• NCI PDQ: Colon Cancer Treatment (Patient Version)
• NCI: Screening Tests to Detect Colorectal Cancer and Polyps
• NCI SEER Cancer Stat Facts: Colorectal Cancer

How this article was prepared

An AI-assisted editorial system helped prepare this page. No named medical reviewer has reviewed it unless one is listed.

See an error, old source, or unclear wording? Tell us.