Gains in cancer research are maddeningly hard to understand, and many new drug regimens are extremely toxic. That's what analysts from BMO Capital seized upon, when they issued a negative report on biopharma Bristol-Myers Squibb (BMY -8.51%). The report was one factor that sent the stock plunging 6% last Thursday morning.

Good, but not good enough?
BMO analysts downgraded the stock amid concerns about the high toxicity for Bristol's drug combo nivolumab+Yervoy. They also cited lowered expectations for the combo's efficacy against lung and kidney cancers, based on their interpretation of studies released at the world's largest meeting of cancer doctors, the American College of Clinical Oncology meeting in Chicago.

According to ISI's Mark Schoenebaum, Bristol's nivo+Yervoy combo in previously treated kidney cancer showed data that was "quite good, but perhaps not as mind blowing as the combo drugs were in melanoma at last year's ASCO." Schoenebaum was more supportive of nivolumab in first-line lung cancer, with one-year overall survival rates "stellar" at between 59% and 87%.

Competition heating up
While Bristol has outlined how nivo+Yervoy could be a potential megablockbuster that could earn billions of dollars a year, it has some competition to consider. While Bristol is the biggest spender on cancer immune therapies ($649 million in trial spending, with $483 million consumed by trials in Phase 3), Merck (MRK 2.93%) has a competing drug -- pembrolizumab -- in the clinic, and Roche Holding (RHHBY -1.70%) has its own PD-L1 drug in the race. All three drugs unleash a T-cell attack on cancer cells. In essence, they deactivate a cloaking mechanism that mutes the immune system's response.

Roche is the smallest spender on high-stakes patient trials of the three companies, at $189 million per year, while Merck spends $327 million per year. But Roche believes it may have a safer approach, and early data shows that investigators could use the maximum dose safely, making it a strong candidate.

Another drug from Merck, pembrolizumab, showed median progression-free survival of 5.5 months, with an overall response rate of 40% in some patient groups. So many patients in the trial are still alive that researchers can't yet put a number on how much of a boost they're getting in overall survival from the drug, according to Dr. Antoni Ribas, an oncologist at the University of California.

Merck also released early findings on a Phase 1b study evaluating the drug in patients with advanced head and neck cancer, which is very difficult to treat and, and the news was good enough to support "further study", according to Dr. Tanguy Seiwert, associate director of the Head and Neck Cancer Program at the University of Chicago.

Final thoughts
Cancer research is an investment opportunity, but it's important to remember that there are real people being affected -- people like Kim Sherman--a 48-year-old mother of three with few prospects of survival who joined a clinical trial of Bristol's nivo+Yervoy. According to a report, in Sherman's trial, 79% of the initial group of 53 patients are still alive after two years. In Sherman's case, within three weeks she was seeing pain reductions, and three months later her tumor disappeared, restoring her to her family.

With Merck, Bristol, AstraZeneca, and Roche all racing to test immune therapy drugs on as broad an array of tumors as possible, the long term picture for cancer patients looks more and more promising. That may be a distinction lost on those whose interest in cancer research is strictly in its investing potential, but it shouldn't be missed. Citigroup analysts predict that immune therapies could lead to a market of $35 billion a year. But cancer is still the second most common cause of death in the United States, accounting for nearly 1 out of every 4 deaths. The real excitement for most of us is that we may have finally made an important step toward a cure.

 Editor's Note: A previous version of this article referred to pembrolizumab by its former name, lambrolizumab. The Fool regrets the error.