Early-stage tests of the drug BMS-936558, known as an anti-PD-1 treatment, showed it was relatively safe and shrank tumors in three of the five cancer types studied, the team reported at the American Society of Clinical Oncology (ASCO) meeting on Saturday and published online in the New England Journal of Medicine.

The team saw significant tumor shrinkage in 18 percent of the 76 lung cancer patients, 28 percent of the 94 melanoma patients, and 27 percent of the 33 patients with kidney cancer.

"These are significant regression rates, especially considering the kind of patient we are treating," said Dr. Suzanne Topalian, professor of surgery and oncology at Johns Hopkins Kimmel Cancer Center, who presented the findings at the ASCO cancer meeting, noting that many of the patients in the study had been treated with at least three other drugs.

Cancer specialists and Wall Street analysts have high hopes for the treatment, the second in a class of so-called checkpoint modulators that rev up the immune system to fight tumors.

Bristol-Myers' drug Yervoy, or ipilimumab, the first in this class, has already won approval for the treatment of advanced melanoma, and researchers are eager to add a new weapon to their growing arsenal of immunotherapy drugs.

These drugs work by removing a natural brake on immune system cells, but cancer experts said this drug, at least in early trial, appears to be less toxic and have higher response rates than Yervoy.

In an editorial published in the New England Journal of Medicine, Dr. Antoni Ribas of the University of California Los Angeles noted that patients in the study had long-lasting response rates, exceeding the 10 to 15 percent "ceiling" for lasting responses seen in other immunotherapy treatments.

Cancer experts are especially excited about the compound's activity in lung cancer, a leading cause of cancer death that so far has shown limited response to immune system therapies.

"This is really the first time we've seen this rate of response in lung cancer. That makes us think we don't know the full spectrum of activity of this drug. Part of the future will be to use it in different cancer types," Topalian told the meeting.

BMS-936558 is an antibody that blocks the activity of a receptor on immune cells called programmed death 1 or PD-1.

"Its normal function is to turn off immune responses at the proper time so you don't get collateral normal tissue damage as a result of the immune response," Topalian said in an interview.

"Cancers can co-opt the PD-1 pathway so they can, in essence, fly below the radar of the immune system," she said.

Topalian and colleagues tested this drug in 296 patients with various advanced cancers who had not responded to standard therapies. Of those receiving the anti-PD-1 therapy, 236 who started treatment by July 2011 were analyzed for tumor response.

In the study, there were 31 patients in the trial whose tumors had regressed at least 12 months prior to being analyzed, and among that group, two-thirds had a response that persisted for more than a year.

"It's the durability of these responses that is particularly impressive," Topalian said.

James Allison of Memorial Sloan-Kettering Cancer Center, who did pioneering work on ipilimumab, said lasting cancer response is really important.

"The key thing isn't just shrinking the tumor. Other drugs do that. The key is, does the tumor come back?"

About 14 percent of the patients reported serious side effects, and three patients died from a lung swelling condition called pneumonitis. Two of the patients who died had lung cancer and the third had colon cancer.

Topalian said this was about 1 percent of patients and overall, the side effects were largely manageable. Only 5 percent of patients in the trial dropped out because they could not tolerate the side effects, she added.

Fouad Namouni, head of Bristol's anti-PD-1 program, said in a telephone interview that the company has learned to identify patients at risk of lung reactions and manage the side effect with steroids.

"We hope that our learning from the activity and safety will hold in the registration trials," Namouni said.

He said the company wants to advance the drug as quickly as possible into late-stage clinical trials.

As part of the study, the team also screened 42 patients before they were treated with PD-1 to see if their tumors expressed a companion molecule called PD-L1. They noted that none of the patients whose tumors were negative for PD-L1 had tumor shrinkage, suggesting a possible strategy for predicting which patients would benefit from the drug.

Cancer researchers say the drug will likely be tested with various other treatments to elicit an even more robust response.

(Additional reporting by Bill Berkrot in New York; Editing by Lisa Shumaker)

• A Permanent Economic Emergency