Navigating Proton Therapy, Immunotherapy Sequencing, and AI in HNC

Discussions at the American College of Radiation Oncology (ACRO) 2026 Radiation Oncology Summit highlighted the rapid evolution of technology and treatment sequencing in head and neck oncology. Heath Skinner, MD, PhD, emphasized that the field is currently grappling with the integration of adaptive radiation and the comparative clinical value of proton therapy. While a phase 3 trial (NCT01893307) suggests significant benefits in reducing gastrostomy tube dependence, the phase 3 TORPEdO trial demonstrated disparate results, reinforcing that while proton therapy is a potent tool, it remains one of several standards alongside modern intensity-modulated radiation therapy (IMRT) and volumetric modulated arc therapy (VMAT).^1-3

A significant takeaway for multidisciplinary teams involves the optimal timing of checkpoint inhibitors. Data from a University of Pittsburgh Medical Center (UPMC)-led study indicate that a sequential administration of pembrolizumab (Keytruda) may provide a more robust efficacy signal than concurrent regimens, potentially establishing a new framework for future trials.⁴ Furthermore, the integration of AI-based contouring and automated VMAT planning is significantly improving clinical efficiency, though the adoption of such technologies must be predicated on demonstrated value in patient quality of life. Finally, according to Skinner, the UPMC model showcases how standardized clinical pathways and prospective peer review ensure that high-quality care is distributed across wide geographies, effectively managing complex malignancies while minimizing chronic adverse effects.

Skinner is professor and Claude Worthington Benedum Chair of Radiation Oncology at the University of Pittsburgh and chief medical officer of the Radiation Oncology Network at the UPMC Hillman Cancer Center.

CancerNetwork®: What head and neck presentations at ACRO stood out to you this year?

Skinner: A lot of the controversies and the data in head and neck [cancer] are largely revolving [around] a few key questions. Firstly, where does technology sit? There’s a few key technologies that people are looking at. One of them is the prevalence of adaptive therapy, which means automatically changing the radiation [dose] over the course of treatment based on anatomy or anatomical changes, changes in the tumor, and secondarily, utilizing advanced technologies like proton therapy to try to improve outcomes. Indeed, several recent clinical trials have informed the thought processes behind this.

Would you be able to speak about Jason K. Molotoris’s, MD, PhD, session on proton therapy in oropharyngeal cancer?

There are a couple of recent trials that we’re specifically looking at protons in oropharyngeal cancers. By way of background, typically oropharyngeal cancers, these are tumors in the base-tongue/tonsil. In many instances in the modern era, they’re HPV positive. We know that those patients tend to do better. The tumors tend to respond [and] progression-free survival in those patients, taken writ large, about 85% [or] up to 90% in patients who are low-risk.

The question becomes, can we improve, or at least maintain, those excellent outcomes while reducing toxicity, because we know head and neck cancer therapy is toxic. You have short-term toxicities, but you also have chronic toxicities, which are arguably worse, [and include] things like swallowing dysfunction and osteoradionecrosis.

At least one focus has been on utilizing proton radiotherapy, which has different beam characteristics. The goal is to reduce the dose to normal tissues. There was a trial in the US led by Stephen Frank, MD, that was published in Lancet not long ago that demonstrated [several] laudatory outcomes, both regarding [gastrostomy] tube dependence, as well as survival outcomes that were improved with proton therapy. Contrast this with a study in England, the TORPEdO trial, that had a similar design in oropharyngeal tumors that showed no statistically significant difference in any key outcomes. numerically depending upon how you look at the data, there were some instances where even the non-proton arm did better.

You have 2 somewhat disparate trials that are presented to the audience for discussion. No trial is perfect…each area where it may not fully satisfy the question at hand, and it creates some consternation within the field [regarding] which direction we go. At the end of the day, in reviewing both, my conclusions are a fewfold. Firstly, that protons remain worthy of investigation for head and neck cancer and oropharyngeal cancer. Secondly, I do not think that is the standard of care. It’s a standard of care.

If I were a patient out in the universe thinking, “Well, should I be concerned that I’m not getting protons for oropharyngeal cancer?” My answer to you is no, [and] that we have good data on both sides of the equation. We have excellent modern radiotherapy in the context of IMRT and VMAT. At least pending additional data and then the data for protons and/or pharyngeal cancer are good, it’s a very good treatment for oropharyngeal cancer, but one [cannot] call it the one and only standard of care.

A recent phase 2 trial evaluated current vs sequential pembrolizumab plus chemoradiation in head and neck cancer. Would you be able to break down the findings, and why they were consequential?

I was fortunate enough to participate in a study led by several investigators at UPMC, the goal of which was to answer a fundamental question; do we [administer] immunotherapy concurrently vs sequentially? The reason why we asked the question, and it became even more germane afterwards than during the study, is that no one has ascertained the best way to integrate immunotherapy into concurrent chemoradiation for head and neck cancer.

The earliest data in that context were in the setting of lung cancer, where they found that adjuvant therapy was [advantageous] over no immunotherapy at all––or when I say adjuvant, after the chemoradiation. The question remains, can you do it concurrently? Many studies at the time were focused on that concurrent question, and unfortunately, several of them were negative, at least for their primary outcome. Now there’s a lot of discussion in those studies if you selected for patients with high levels of PD-L1, for example, which you enriched for a population that would benefit those are ongoing discussions and questions.

Our study showed that sequential… immunotherapy performed better in the primary outcome as well as secondary outcomes than concurrent. [This] provides a framework for what we should be considering in larger trials. This was a small single institution randomized trial, but it had a dramatic signal compared [with] concurrent. With the ostensible failure of several concurrent trials, it’s worth consideration, particularly because [being the] trailblazer in this regard, i.e., locally advanced, non–small cell lung cancer, the adjuvant or sequential approach has been effective.

How have the advancement of technology improved or made your practice more efficient? As a chair of the of the radiation department at UPMC, how do you encourage your fellows to explore its use?

There are several different changes in technology that have occurred or are occurring that are going to be germane. One of them is the advent of machine learning and artificial intelligence [AI]. Everybody is talking about it at this point. One place where we have found a great deal of utility, is in the context of contouring. We use an off-the-shelf, effective AI-based contouring solution for our entire network with regards to organs-at-risk [and] our normal tissue structures. We’re starting to use them for nodal contours. We’re working… to develop contouring the primary tumor which is still in the early phases. A lot of people are doing this, and that’s one of the places where the field is going to try to make us more efficient regarding planning and [caring for] patients.

Another area where machine learning is playing and will play more of a role as time goes on is in the setting of planning. Functionally, we can generate beautiful VMAT plans in time frames that I would never have dreamed of as a resident, and that continues apace. Other areas of development include adaptive technologies, for example… and many of the technologies I’m describing, particularly the ones that are costly, it is predicated upon us as a field to demonstrate the value of these technologies.

Case in point, for adaptive as an example, it is predicated––the machines are more expensive, [same thing with protons]. It is predicated upon us as a field to demonstrate value, either in the context of symptom reduction or in the context of patient quality of life in a randomized way to ultimately integrate these technologies into practice. Unfortunately, in radiation oncology, we have not been able to do that as effectively as we all would hope. Is that because it’s harder to do trials in some instances with advanced technologies? Potentially, but we’re still going to have to do it. As we all get enamored with technology, that’s one of the reasons that draws radiation oncologists in the field.

At the end of the day, if we can’t demonstrate true value, those technologies will not move forward. One of the things I encourage my residents is look at every technology with an eye towards skepticism as it is introduced into the clinic, because ultimately, these are costly things. We need to [ensure] that we provide value to the patient, because otherwise they won’t move forward.

How does the multidisciplinary collaboration work [at a large center like UPMC]? Is there anything that [your practice excels] at that other institutions [could] adapt?

My primary focus is head and neck, and we do a wonderful job [for this disease]. We have a robust multidisciplinary clinic that includes surgeons, medical oncologists, radiology, pathology––everyone involved in the care of the patient with head and neck cancer. There is a wonderfully streamlined workflow, from diagnosis to the surgeon to us trying to get the patients in as short a period as possible. Within the context of ourradiation oncology network, there are several things that are worthy of consideration for any decent sized network.

One of them is that we participate in prospective peer review. Our entire network has a set time each week where we get together in groups…coupled with disease site and content experts, and every case that’s curative, that’s definitive, or that’s stereotactic body radiation therapy [SBRT] is presented prior to initiation of treatment, so it’s truly prospective in nature. Secondarily, our dosimetry [and] our physics resources are all shared throughout the network. While we have physicists and dosimetrists on each site, we also can distribute work amongst the entire network in a way that’s relatively seamless, and a lot of that is because we have standard operating procedures for a lot of the different things that we do.

Moreover, we participate in a standardized, clinical pathway, both in the context of treatment decision-making, as well as in the context of what the plan should look like. Finally, in addition to all those things, we also have site-specific rounds to see sites that have some increasing levels of complexity. For example, head and neck contouring is a bit more complex sometimes. We have head and neck contouring, and discussion rounds once or twice a week. We have central nervous system, stereotactic radiosurgery, and SBRT rounds to do the same thing. We have a wide geography. The goal is to bring everyone together as much as possible, to discuss what they’re doing, how they’re treating patients, and everyone can benefit from everyone else’s expertise.

Finally, something that’s just as important is that we have a well-integrated clinical research infrastructure. With a few exceptions that are highly complicated, almost every trial that we’re doing, we offer to almost every site that we have. We [accomplish] that via a combination of on-site staff as well as telemedicine to make sure that patients, regardless of where they are, can participate in the clinical research infrastructure and clinical research trials in a way that helps them, but also helps advance cancer. I would say the same thing for treatment. In general, at all our sites, one of the nice things about having this distributed network is that the treatment that a patient gets far away from the academic center in Pittsburgh is going to be [similar] as what they would get there by and large.

References

1. Frank SJ, Busse PM, Lee JJ, et al. Proton versus photon radiotherapy for patients with oropharyngeal cancer in the USA: a multicentre, randomised, open-label, non-inferiority phase 3 trial. Lancet. 2026;407(10524):174-184. doi:10.1016/S0140-6736(25)01962-2
2. Molitoris J. Proton therapy for oropharyngeal cancer. Presented at: American College of Radiation Oncology 2026 Radiation Oncology Summit; February 4-6, 2026; Orlando, FL.
3. Thomson DJ, Price JM, Tyler M, et al. Proton beam therapy for oropharyngeal cancer (TORPEdO): a phase 3, randomised controlled trial. Lancet. 2026;407(10535):1259-1275. doi: 10.1016/S0140-6736(26)00314-4
4. Zandberg DP, Vujanovic L, Clump DA, et al. Randomized phase II study of concurrent versus sequential pembrolizumab in combination with chemoradiation in locally advanced head and neck cancer. J Clin Oncol. 2025;43(23):2572-2582. doi:10.1200/JCO-24-01580