Favorable Cancer-Free Survival Rates With Nivolumab in Proliferative Leukoplakia

Glenn J. Hanna, MD

A phase 2 study (NCT03692325) evaluating nivolumab (Opdivo) in patients with proliferative leukoplakia (PL) yielded positive safety and efficacy results, according to findings published in the Annals of Oncology.

PL is an aggressive precancerous condition which can transform into oral squamous cell carcinoma in patients. Within the single-arm, phase 2 study, enrolled 33 patients with PL. Patients were administered nivolumab every 28 days for up to 4 cycles.

Findings revealed a decrease in composite score based on size and degree of dysplasia of at least 40% in 36% of patients. Additionally, improvement continued in patients several months after they had been given their last dose of nivolumab.

At 2 years, there was a 78% rate of cancer-free survival, which is encouraging given the high risk of PL transforming into cancer. Investigators on the study plan to follow patients to see 5-year cancer-free survival.

This study is the first to show the efficacy benefit of using preventative immunotherapy for patients with high-risk oral precancerous disease. Longer-term treatment may further improve outcomes.

“What this shows is that immunotherapy has a potential to be a preventative strategy, whether it's preventing a first cancer event as primary prevention, or even a second event in someone who's had a prior oral cancer who's at risk for another. This data shows us that there is safety and feasibility, and that there is probably a group of patients who would benefit from these therapies,” said Glenn J. Hanna, MD, in an interview with Targeted Oncology^TM.

In the interview, Hanna, director, Center for Salivary and Rare Head and Neck Cancers, physician at Dana-Farber Cancer Institute, and assistant professor of Medicine, at Harvard Medical School, discussed the phase 2 study evaluating treatment with nivolumab in patients with high-risk oral PL.

Targeted Oncology: Can you provide some background on the phase 2 study of nivolumab for high-risk oral leukoplakia?

Hanna: As we know that oral leukoplakia, which is a whitish lesion in the mouth of undetermined malignant potential affects somewhere between 2% and 4% of the global population. But there's this subgroup of patients, a small percentage of those who have a very high risk, oral precancerous leukoplakia called proliferative leukoplakia. And it looks just as if it sounds very verrucous or heaped up, sometimes erythematous, and it's usually multifocal throughout the mouth.

What we've learned from our research and observational studies is that these patients have a very high risk of malignant transformation, meaning a high risk that they'll develop a cancer event within a 3- to 5-year period. Some people have said up to 60% to 80% of the patients will develop oral cancer within about 5 years. We did some preclinical work to understand how the immune environment might be dysregulated amongst the highest risk leukoplakias like PL and we found that there was a strong infiltration of T cells and maybe a microenvironment and immune environment that would lend itself to potential benefit from immune checkpoint inhibitors like immunotherapy.

Can you talk about the design of the study? What kind of patients were included?

We tried to be specific about the proliferative leukoplakia requirement. Based on the oral medicine providers that we work with, we chose patients who were noted to have a clinical diagnosis of prolific proliferative leukoplakia. They could have been a former smoker, they just could not have had prior immunotherapy. Importantly, they could have had a primary small prior oral cancer like an early stage I or II oral cancer that was removed surgically. They could not have had prior radiation or chemotherapy for head and neck cancer, the reason being those patients sometimes develop cancers and then still have other areas of concern that need attention and are at risk for developing cancer separate from the original area

We biopsy up to 3 sites in any given patient who enrolls because we understand that patients can have multifocal disease. Between 1 and 3 sites are biopsied, usually the largest and most concerning lesions, and if they have any degree of dysplasia, mild, moderate, or severe in any of the 3 lesions, they would be eligible. Then once at each monthly visit, we get 4 doses of nivolumab before 480 mg flat dosing 4 times. Then, we re-biopsy the same lesions that we started with at the end. We measure each lesion and take photographs at each of the monthly visits, or the Oral Medicine providers do, whereas we as the medical oncologists are making sure that therapy is safe and checking in for any immune toxicity.

At the end of the study, the patients all have a sort of composite score change. We measure the initial lesions that they start with, and we grade them based on histology. Then we do the exact same thing at the end of the study, and we see if there's a difference. If there's a sizable difference in the measurements and the degree of dysplasia, that gets us to a response because we can't use traditional RESIST criteria, because we're not doing scans and there aren't tumors to measure as they regress. It's a little complex in how we decided to evaluate response. One of the most important pieces is the time to a cancer event. We were monitoring everyone for the first sign that they had developed an oral cancer anywhere. That was 1 of the most important secondary end points, cancer-free survival or time to index cancer events after starting the immunotherapy.

What are the highlights from the results you presented at the conference?

We saw a 36% regression rate or response rate. So, this means that upwards of 36% of patients had more than 40% regression in their lesion sizes and/or the degree of dysplasia in those lesions; remember [they had] up to 1 to 3 lesions [biopsied]. So that was exciting to see. Importantly, we would not expect PL lesions to regress spontaneously. Some people would say, “Could that just be the ebbs and flows of the disease?” No, we don't expect these lesions to get better. These are so high risk that they would likely evolve over time into cancer.

The other thing we noticed was that there were some patients who had clinically visible regression, and we showed clinical photographs in our slides at ESMO [2022 European Society of Medical Oncology Annual Congress], showing that some lesions regressed and that even carried on well beyond exposure to nivolumab. So even after the 4 doses, months and months later, which we can observe in advanced cancer patients with immunotherapy, we saw some improvements in clinical leukoplakia. That means this may take a while to work, and so the question of whether we should be giving more doses comes up.

I think the other thing was the 2-year cancer free survival. It's a short interval. But...many of these patients will develop cancer events. It was approaching 78%. That's strong compared [with] our historical data. We certainly will follow these patients out to the median 5-year mark, but at least for a 2-year time point, it's encouraging that almost 80% of patients have not suffered a cancer event.

The safety was acceptable. Keep in mind, this was a non-cancer population so to see a 20% or so immune-related adverse event rate of grade 3 or 4 was a little higher than we'd expect from advanced cancer patients treated with immunotherapy. I argue that may be because these patients are overall healthy. Sometimes when people are very sick with cancer and you put them on a trial and you grade their adverse events, you might be catching some of the cancer-related issues, not so much treatment-related issues. I think these people were generally healthy.

The last thing that I think was exciting, at least in the preliminary analysis, was in patients with immune characterization for PD-L1 status on the epithelial dysplasia in the immune environment before we treated the patients, we saw some evidence that higher scores, similar with what we see in advanced head and neck cancer, on these dysplastic tissues might predict clinical response based on our scoring system to the treatment. There were some patients who had low immune scores who did benefit and vice versa. What we're excited about is that we just got back the whole exome sequencing data on the samples, we worked with UC San Diego Moores Cancer Center to get that work done. There's some nice correlation with patients who benefit and don't in terms of what we found on the genome. We're excited to report that in the manuscript that's coming up.

What do these results mean for next steps with this research?

What this shows is that immunotherapy has a potential to be a preventative strategy, whether it's preventing a first cancer event as primary prevention, or even a second event in someone who's had a prior oral cancer who's at risk for another. This data shows us that there is safety and feasibility, and that there is probably a group of patients who would benefit from these therapies. What we could do next is add additional immune agents based on what we find in the post-treatment biopsies of patients who didn't respond to see if we can increase the response rate and the benefit rate.

I think a randomized trial would be next. Similar to the EPOCH study [NCT01483027], I think we'd be looking to do a similar study where we would randomize patients to immunotherapy, either nivolumab or nivolumab plus an additional agent vs placebo because that's how we know that the cancer-free survival is and that the response signal is real. I'm hopeful that through discussions with sponsors, we can look towards a follow-up study that randomizes patients because I do think this has some potential to improve outcomes and to decrease cancer events. If we can decrease the time to cancer events, hopefully patients will have better outcomes, they'll have better quality of life, and of course, better overall survival.

Reference:

Hanna GJ, Villa A, Shi R, et al. A phase II study of nivolumab for high-risk oral leukoplakia. Ann Oncol. 2022;33(suppl 7):S295-S322. doi:10.1016/annonc/annonc1056