Each year, the ASCO annual meeting serves as a forum for some of the most important clinical studies in oncology, giving physicians important insights into how best to treat patients with cancer.

But the results are not always clear cut, particularly in the rapidly advancing field of immunotherapy.

Below are three pressing questions researchers are still grappling to answer:

1. Are Keytruda and Opdivo really all that different?

In April, a top executive from Bristol-Myers Squibb made headlines by claiming the pharma’s immunotherapy Opdivo (nivolumab) and Merck & Co. ‘s rival Keytruda (pembrolizumab) were more similar than different — a pharmaceutical version of Coke and Pepsi, in his words.

Both drugs inhibit a checkpoint molecule known as PD-1. Both have also won a long list of approvals from the Food and Drug Administration to treat many of the same advanced cancers.

Yet the remarks seemed to ignore the divergent paths taken by the two drugs in lung cancer.

Opdivo, which still earns more than Keytruda, famously failed a large study testing the drug as a first-line treatment of non-small cell lung cancer. That left the door open for Merck, which has now won two approvals in the lucrative market.

At this year’s ASCO, Merck presented data from a study roughly analogous to Bristol-Myers’ failed attempt two years ago: pitting Keytruda against chemotherapy in a broad range of NSCLC patients whose tumors expressed PD-L1 levels greater than 1%.

Unlike Opdivo, however, Keytruda succeeded. Merck’s drug cut the risk of death in the overall study population by 19%. While high PD-L1 patients benefited the most, the result begs the question of why the two drugs performed so differently.

Leena Gandhi, director of Thoracic Medical Oncology at the New York University School of Medicine, believes the trials are actually more similar than different, chalking up the split outcomes to study design.

“Overall survival, although it was seen to be significant in [Merck’s] Keynote-042 study and not in [Bristol-Myers’] CheckMate-026 study, I think is in part explained by the differences in crossover,” Gandhi said, referring to protocols that let patients who fail chemo go on to receive immunotherapy while on study.

More than 60% of patients in the chemo arm of CheckMate-026 received subsequent treatment with Opdivo, compared to only 20% in the chemo group of Keynote-042 who later received immunotherapy.

“I think the drugs are very similar. I think it is study design and patient selection which has led to the differences,” said Jill O’Donnell Tormey, head of the Cancer Research Institute.

Whether that’s true or not is still an open question. Either way, however, Bristol-Myers faces the uphill battle of unseating Merck in a market in which Keytruda now enjoys a commanding position.

2. Which first-line immunotherapy regimen should lung cancer patients receive?

Entering ASCO, oncologists had evidence Keytruda helped certain patients with metastatic NSCLC live longer.

But questions remained around which patients were best served by Keytruda alone and which would see greater benefit from receiving Keytruda together with chemotherapy.

Results from two Phase 3 studies of Keytruda presented at this year’s conference help paint a clearer picture, highlighting the mainstay role Keytruda looks set to play in initial treatment of the disease.

Currently, Keytruda is the only immunotherapy approved for first-line treatment of NSCLC. The PD-1 inhibitor is indicated as a monotherapy for patients with greater than 50% expression of the PD-L1 protein and in combination with chemo for all patients with nonsquamous NSCLC.

New data showed pairing Keytruda together with a different chemo regimen also extended survival in patients with the harder-to-treat squamous form of NSCLC, regardless of PD-L1 expression. This suggested that, for all NSCLC patients without a driver mutation like EGFR or ALK, Keytruda plus chemotherapy is a superior option than chemo alone.

A second study, testing Keytruda against chemo in NSCLC patients with greater than 1% PD-L1 expression, also attracted much attention at the conference. Results from the trial found Keytruda monotherapy reduced the risk of dying versus chemo in that broadly selected patient population.

Much of the survival benefit seen, however, was due to responses among patients with high PD-L1 levels.

“This benefit is clearly being driven by the higher PD-L1 subgroup,” remarked NYU’s Gandhi. “The benefit is not at all as clear cut for those with 1% to 49%, where a much larger percentage of patients are not benefiting from immunotherapy.”

Keytruda alone, however, was much less toxic than chemotherapy, suggesting that it could play a role for patients who aren’t good candidates for chemo.

Gilberto Lopes, who led the second Keytruda study, said in an interview the medical consensus will likely coalesce around giving immunotherapy together with chemo for patients who express between 1% and 49% PD-L1, while opting for Keytruda by itself in PD-L1 high individuals.

That approach could be complicated, however, by ongoing studies conducted by Merck’s rivals Roche and Bristol-Myers of their own checkpoint inhibitors.

Bristol-Myers, in particular, is developing a combination of two immunotherapies, Opdivo and Yervoy, which could offer a chemo-free option for patients who express a different biomarker known as tumor mutation burden.

What is clear, however, is the role immunotherapy now plays in first-line lung cancer treatment.

“Chemotherapy alone is no longer a first-line standard of care in non-small cell lung cancer,” said Gandhi in remarks at a plenary session.

3. Have companies moved too fast to advance immunotherapy combinations into large trials?

Pairing immunotherapies with other drugs has been a central focus of the industry’s recent investment in cancer R&D. More than 1,100 ongoing combination studies involving the five approved checkpoint inhibitors are now underway.

So far, all of that research activity has yielded few clear-cut successes outside of immunotherapy and chemo pairings.

Earlier this year, for example, a closely watched Phase 3 study of Incyte’s IDO inhibitor epacadostat and Keytruda in melanoma failed — showing Incyte’s drug added no benefit whatsoever. It was a crushing blow to one of the industry’s most advanced combinations and sparked questions of whether the two drugmakers moved too quickly.

Investors now appear worried that another high-profile combination could follow the same pattern. On Monday, Nektar Therapeutics lost $6 billion from its market capitalization as investors digested a confusing update on the biotech’s combination of its experimental drug NKTR-214 with Opdivo.

Encouraged by promising data among just several dozen patients, the partners greenlighted three late-stages studies of the two drugs in melanoma, kidney and bladder cancers.

The decision raised doubts of whether the companies had enough data to merit advancing into Phase 3, particularly as a second group of patients didn’t appear to experience the same high level of responses.

It’s a question that faces the entire field, as companies are forced to weigh waiting for more (or any) randomized data against advancing quickly with pairings that present a strong biologic rationale for combining.

“Is there enough of the translational data [generated]? Or are we going down another pathway of what happened with IDO — jumping from early to now spending millions on many patients,” said CRI’s O’Donnell-Tormey.

At the same time, others argue that when there is enough evidence to suggest strong potential, it’s worth investing in.

“It is really easy to say every asset needs a randomized Phase 2 before moving into Phase 3,” explained Kim Blackwell, head of early-phase oncology development at Eli Lilly, in an interview.

“But there is also a time [element] here — if you have good science and you have a safe agent in the early phase, sometimes going directly to a large clinical trial might get you there faster. Our patients need better therapies.”