Tom Telford’s stomach ached. The New York City teacher had been drinking cup after cup of coffee as he labored to finish year-end grading and coach his high-school baseball team through the playoffs. He worried he might have an ulcer.When school let out, though, Mr. Telford looked forward to relaxing on a 25th anniversary cruise with his wife. But once in the Caribbean, he struggled to swim and climbing from one deck to another exhausted him. Back at home, he collapsed while running a TV cable in his bedroom.His family doctor told him he had lost two pints of blood. Further tests revealed a tumor the size of a quarter on his small intestine. He had surgery at Memorial Sloan Kettering Cancer Center, followed by months of chemotherapy. But the disease spread to his liver and kidneys. The diagnosis: Stage 4 melanoma, a skin cancer typically fatal within a year.

“Death is not an option,” he told his doctor.

Nine years later, against all odds, Mr. Telford is still alive. What saved him was an experimental immunotherapy drug—a medication that unleashes the body’s own immune system to attack cancer.

When his tumors began melting away more than eight years ago, Mr. Telford’s good fortune was largely an anomaly amid a mostly dreary landscape for advanced cancer. But his remarkable survival caught the attention of researchers, who began to realize that the way immunotherapy drugs were affecting tumors was unlike almost anything seen with conventional treatments.

Today Mr. Telford is among a growing group of super-survivors who are transforming the world of oncology. In both total numbers and duration of survival, they are charting new territory. And they are reviving hopes that the long-maligned idea of enlisting the power of the immune system against cancer may help to turn the tide against some of the most lethal and resistant forms of the disease.

“It’s the most exciting thing I’ve ever seen,” says David Lane, scientific director of New York’s Ludwig Institute for Cancer Research. “It’s the long-term survival of people who have advanced disease. This is very unusual.”

Cancer immunotherapy comes in several forms. The drugs sparking the most interest are called checkpoint inhibitors. They work by releasing the natural brakes on the immune system, enabling its foot soldiers, called T cells, to attack tumors.

It is hard to know how many patients whose cancers have metastasized, or spread, have enjoyed sustained survival following immunotherapy treatment. An analysis of 4,846 advanced melanoma patients treated with one checkpoint inhibitor—Bristol-Myers Squibb Co.’s Yervoy—found that 21% were still alive three years later. That amounts to more than 1,000 people, most of whom experts say almost certainly would have died otherwise. Especially striking is how good the long-term prospects were for people who survived at least three years.

“The people that make it after three years don’t die of melanoma,” says James Allison, head of immunology at MD Anderson Cancer Center in Houston, whose seminal discovery about the immune system and cancer in the mid-1990s laid the groundwork for many of the current advances.

Newer drugs that work similarly to Yervoy, but on different immune-system brakes, are getting even better early results and are extending the benefits beyond melanoma to other cancers.

Immunotherapy is still in its early stages, and more rigorous studies are needed. Oncology is filled with tales of advances that held promise only to be thwarted by cancer’s uncanny ability to develop resistance to medicine’s attacks. There isn’t any assurance that the new immunotherapy will be different.

Researchers and drug makers are striving to overcome huge obstacles to a lasting cure. For one, most patients don’t respond the way the super-survivors have, and researchers are just beginning to understand why. Another mystery is why some patients relapse while remaining on therapy while others go into prolonged remission after undergoing just one course of treatment.

Most experts believe it will take combinations of immunotherapy drugs—or combinations of immunotherapy with other cancer treatments—to optimize their impact. But finding safe and effective combinations is a daunting undertaking.

While side effects of the new drugs are relatively mild for some patients, others have developed potentially devastating complications caused by an out-of-control immune system. Some patients have died as a result. Researchers are devising ways to minimize such problems.

“It’s extremely exciting that so many patients are responding” to checkpoint inhibitors, says Bert Vogelstein, director of the Ludwig Center at Johns Hopkins Kimmel Cancer Center in Baltimore. “But the reality is that most are not.”

The drugs, which are costly to develop, are certain to fuel the debate about the cost of innovative drugs. Yervoy costs more than $120,000 for a four-course treatment, while Merck & Co’s Keytruda, approved in September for advanced melanoma, costs $12,500 a month, or $150,000 for a year.

More than 25 companies ranging from the pharmaceutical industry’s biggest names to a group of startups are pursuing some form of immunotherapy.

Bristol-Myers’s drug Yervoy, which is based on Dr. Allison’s discoveries, blocks an immune system brake called CTLA-4. Merck’s Keytruda inhibits a brake called PD-1. Bristol, Roche Holding AG and AstraZeneca PLC are among several companies also testing agents against checkpoint targets.

Another approach involves genetically modifying certain T cells outside the body, creating what are called CAR T cells, and infusing them back to attack targets on the surface of cancer cells.

Novartis AG, closely held Juno Therapeutics Inc., Kite Pharma Inc., and a collaboration between Bluebird bio and Celgene are pursuing this strategy. Amgen Inc. is developing yet another T-cell approach while several other companies are reviving efforts to develop cancer vaccines.

“I divide pharmaceutical companies into two categories,” says Drew Pardoll, co-director of immunology at Johns Hopkins. “They’re in immunotherapy up to their eyeballs, or they want to be.”

The efforts are a departure from most current cancer treatments, including chemotherapy, radiation and the new crop of medicines that target genetic mutations underlying a tumor’s growth. These strategies take direct aim at the tumor—generally to only limited effect in patients with advanced cancer. Even the recent excitement about genetically targeted drugs has been tempered by the ability of tumors to mutate and grow resistant to tailored attacks.

With immunotherapy, “We’re treating the immune system, not the cancer,” says Dr. Allison.

Richard Logan, a 59-year-old veterinarian in Ozark, Ala., embraced the alternative approach in 2009, soon after melanoma that had started as a bleeding mole on his back progressed to his lung and liver. He already had endured about a year of treatment with interferon, an immune-system booster with harsh side effects that left him exhausted.

In addition, his father had recently died of melanoma, succumbing about six months after the cancer had spread from its original site. “I pretty much knew the old treatments weren’t going to be much help,” he says.

An Internet search led him to the melanoma center at Boston’s Dana Farber Cancer Institute. In July 2009 he enrolled in a trial there for a drug that later became Yervoy. Within two months, his tumors started shrinking and by that December, he says, “I felt pretty confident I was going to have extended relief from this disease.”

Now nearly five years later, Dr. Logan says his cancer has stabilized. He remains on Yervoy, flying regularly to Boston for infusions. He rates the side effects on Yervoy as a “two or three” compared with a 10 on interferon. He continues to run his veterinary practice, saw his son graduate from college and recently remarried.

He doesn’t use the word “cure” to describe his status, saying only that he is “confident” in how he is doing on the drug. “If this starts to pale, I’m keeping aware of whatever else is available out there,” he says. “I’m pretty even-keel and take it as it comes.”

Getting the immune system to “see” and attack tumors has stymied researchers for decades. Unlike, say, a flu virus—an invader that quickly catches the immune system’s attention—a tumor cell may reflect only “modest changes in cells that the body has been taught to leave alone and tolerate,” says Walter J. Urba, director of cancer research at Providence Cancer Center in Portland, Ore. Moreover, he says, “cancer cells are pretty smart and they change in ways that can avoid the immune system.”

For years, scientists thought the immune system didn’t recognize tumors at all. Then research on biopsy specimens revealed that T cells often succeed in infiltrating the environment around tumors, but either fail to mount an adequate response or hold the cancer at bay for years before finally being overmatched.

When the T cell does recognize the tumor, it amounts to an ignition switch for the immune response. But just “turning the key” isn’t enough. T cells generally need a second so-called co-stimulatory signal to become activated against the cancer. That is the gas pedal.

For two to three decades, immunotherapy efforts have focused on the gas pedal. Researchers and companies have tested vaccines that educate T cells to find cancer cells and developed treatments to boost activated T cells and direct them at vulnerable targets on tumors. Despite some success against relatively rare cancers, results have proved lackluster.

“No matter what you did, only 10% to 15% of patients seemed to get a benefit,” says F. Stephen Hodi, director of the melanoma center at Dana Farber.

The insight that changed the equation came from Dr. Allison. In the early 1990s, researchers racing to unravel how T cells function homed in on a molecule called CTLA-4. They had suspected it was a “gas pedal” that activated the immune system. Dr. Allison, then at University of California Berkeley, was among those who determined that it actually inhibited the immune response—it was a brake.

That led him to a question: Would blocking CTLA-4 with a drug—in effect, suppressing the suppressor—unleash the immune system to attack cancer?

He and his colleagues fashioned an antibody to CTLA-4. In a study published in Science in 1996, they showed that the strategy, which prevented the immune system from turning off, led to the eradication of tumors in mice.

Translating that discovery into an FDA-approved medicine took 15 years. After shopping the antibody to skeptical drug companies for two years, Dr. Allison joined forces with a small company called Medarex Inc., then in Princeton, N.J. The initial human studies were encouraging enough to draw Bristol-Myers into a collaboration with Medarex to mount a large-scale study.

Early results were less than impressive. Tumors shrank sufficiently in only about 10% of patients—no better than was seen in other forms of immunotherapy, raising doubts the drug could be approved.

But researchers also noticed something unusual: Some patients were living much longer than expected, including some who had gone off therapy. Others reported feeling better even though their tumors hadn’t quickly responded by conventional measures.

One of them was Mr. Telford, the high-school teacher with advanced melanoma. In June 2006, with the encouragement of Jedd Wolchok, chief of melanoma and immunotherapy at Sloan Kettering, he enrolled in a study of the Bristol-Medarex anti-CTLA-4 drug. The protocol called for four treatments, one every three weeks. The goal: tumor shrinkage at 12 weeks.

It didn’t work. After Mr. Telford’s last infusion, his scans showed tumors in his liver had gotten much larger. Dr. Wolchok prepared to deliver the message oncologists dread: The drug wasn’t working and there weren’t any other options.

But in the exam room, Mr. Telford told the doctor: “This is the best I’ve felt in months.”

When Dr. Wolchok insisted the scans showed no improvement, Mr. Telford recalls saying, “I don’t care what your scans say, I feel better.” His energy levels were up and his night sweats had stopped.

A mystified Dr. Wolchok told him to come back for another checkup in two months. This time, the tumors were getting smaller. By May 2007, his scans showed no evidence of cancer.

Mr. Telford remained on the drug until last December. He is essentially free of disease. Now 60 years old, he is still teaching and coaching baseball.

© Provided by The Wall Street Journal. So exceptional was Mr. Telford’s case that he became known among immunotherapy researchers as “the liver guy,” Dr. Wolchok says. His experience “was a pivotal moment” that helped prompt Medarex and Bristol-Myers to change the primary measurement of the trial, Dr. Wolchok says. Instead of looking at how long the therapy kept cancer from progressing, a measure that can lead to approval of drugs for metastatic cancer, they opted to test for overall survival.

It took some 700 patients and extended the trial to 2010. In the end, fewer than 10% met standard criteria for tumor shrinkage. But 23% survived at least three years, making the drug the first to ever show a survival benefit in patients with advanced melanoma. The drug, now owned by Bristol-Myers and known as Yervoy, was approved in 2011.

The challenge was to widen the benefit to more patients and more cancers.

For reasons of biology, melanoma is more vulnerable than most other tumor types to attacks by the immune system. For immunotherapy to become a mainstay of cancer treatment, it needs to prove effective against more tumors.

That is beginning to happen. Data presented at the American Society of Clinical Oncology meeting in June and at the European Society of Medical Oncology in September offered encouraging results for bladder, head and neck, kidney and other cancers. Checkpoint inhibitors are being tested in breast and pancreatic cancers, and Hodgkin lymphoma, multiple myeloma and other cancers.

The disease attracting intense interest is lung cancer, which causes more than 200,000 deaths a year in the U.S.

David Gobin, a retired Baltimore policeman, was diagnosed with lung cancer at 58. He lost much of his right lung to surgery and much of his spirit to aggressive chemotherapy, which failed.

In late 2010, he joined a trial at Johns Hopkins for an experimental Bristol-Myers drug that blocks the immune system brake called PD-1. The protocol called for a one-hour infusion every two weeks for two years.

Four months later, a scan showed his tumor had significantly improved. Remnants of the tumor still showed up on scans when his treatment ended after two years, but it hasn’t grown since.

“Every time we scanned him, things were decreasing,” says Julie Brahmer, his oncologist at Hopkins. “With the rate his disease was progressing back when I first met him, he shouldn’t be around.”

Mr. Gobin gets short of breath because of the lung surgery, and he needs to take frequent rests. He isn’t complaining.

Before being treated with the drug—now called nivolumab—he was twice told he had less than a year to live. He hasn’t taken the drug for 23 months.

“I still have a little cancer. It’s still sitting there,” he says. “It’s not doing anything.”

Most researchers believe that the key to expanding immunotherapy lies in combination treatments.

A 53-patient Bristol-Myers study combining Yervoy and nivolumab resulted in a two-year melanoma survival rate of 79%, but that came at a cost of serious side effects.

Researchers also are experimenting with pairing such drugs with vaccines or with treatments such as chemotherapy, radiation and genetically targeted drugs. The hope is that attacking the tumor with these approaches will make it more visible to T cells, a necessary step if releasing the brakes is to have a significant effect on the cancer.

“Almost every combination is appealing in some way,” says Roy Herbst, chief of medical oncology at Yale Cancer Center, but “how to sort through all that is incredibly difficult.”

Richard Murphy, a father of three from Marshfield, Mass., didn’t know what a biopsy was when he was diagnosed in 2008, at 43, with a rare form of melanoma. He underwent two surgeries and several bouts of radiation and chemotherapy, but his tumors spread to his lung and surrounded his spine. He enrolled in a study of Yervoy in February 2011.

After his fourth treatment at Dana Farber, scans showed his tumors had shrunk. But by August, he says, “everything kind of had grown.” Dr. Hodi soon concluded the drug wasn’t working and took him off the study.

Dr. Hodi had another idea. His colleagues at Dana Farber had played key roles in identifying PD-1 as an immune checkpoint, and a slot opened up in a study of an anti-PD-1 agent being developed by Merck that became known as Keytruda. In March 2012, Mr. Murphy had his first dose.

The day of his scheduled sixth treatment in June, a blood test showed “off-the-chart” levels of an enzyme that suggested possible muscle damage. He was hospitalized out of concern for possible kidney failure. Whether the PD-1 drug was responsible isn’t certain, but Dr. Hodi took him off the study.

The next day, an ultrasound to check his kidneys revealed something else: Mr. Murphy’s tumors were shrinking. By the end of August 2012, there wasn’t any evidence of disease on his scans, just shadows where tumors had been. His checkup this October showed the same.

Mr. Murphy, a real-estate agent, stages and participates in triathlons to raise money for cancer research.

“I was on two clinical trials and I was kicked off two clinical trials,” he says. “I don’t think you’d expect the outcome that we have. You wouldn’t expect that would be the pot of gold.”