RNA Sequencing ‘Clusters’ Identify Patients for Frontline Treatment With Nivolumab Plus Cabozantinib in ccRCC

Patient selection via tissue-based RNA sequencing biomarkers for treatment with nivolumab (Opdivo) plus cabozantinib (Cabometyx) was feasible and enhanced clinical outcomes for the frontline treatment of patients with metastatic clear cell renal cell carcinoma (ccRCC), according to Scott M. Haake, MD, PhD.

Findings from the phase 2 OPTIC RCC study (NCT0536172) presented during the 2025 ESMO Congress revealed that patients (n = 25) who were identified using RNA sequencing–defined clusters who received first-line treatment with nivolumab plus cabozantinib achieved an objective response rate (ORR) of 76%, with a complete response (CR) rate of 8%. All patients experienced tumor burden reduction, and the median tumor shrinkage was 42% (range, 5%-100%).

“By selecting angiogenic tumors and treating them with an immune checkpoint inhibitor plus an anti-angiogenic agent, we can enrich for response compared with unselected cohorts,” Haake explained in an interview with OncLive®. “Follow-up is still relatively short—the median follow-up was 11.1 months—so we don’t yet have mature data on progression-free survival [PFS] or duration of response, but those are coming.”

In the interview, Haake, an assistant professor of medicine in the Division of Hematology and Oncology, Department of Medicine, at Vanderbilt University Medical Center in Nashville, Tennessee, discussed the design of OPTIC RCC, the significance of data from this study presented during ESMO, and future research directions for this research.

OncLive: What was the rationale for conducting the OPTIC RCC trial?

Haake: The clinical question that led us to design the OPTIC RCC trial was fairly straightforward. In metastatic ccRCC, we have several immunotherapy-containing doublets that are FDA approved. A handful of these are some combination of immunotherapy and TKIs, and others are pure immunotherapy strategies containing 2 immune checkpoint inhibitors, ipilimumab [Yervoy] and nivolumab. They’re all FDA approved in the front line [but] we don’t have a great way of deciding which patient should go on which strategy—who should get the immune-oncology (IO)/TKI and who should get the IO/IO agents. Our goal was to develop a biomarker that would help us make this decision.

It’s been a real struggle to develop biomarkers within the context of kidney cancer, and there really are no biomarkers today that can help us make this decision. We did have some interesting data emerge from the [phase 3] IMmotion151 trial [NCT02420821].² Specifically, there were some RNA sequencing data that suggested that some groups of tumors with similar gene expression patterns had a very angiogenic gene expression pattern. It seemed like those tumors would require a potent anti-angiogenic agent to control disease, whereas others respond preferentially to immunotherapy, and it seemed like those would be the ones that would respond best to a pure immunotherapy strategy.

We took these RNA sequencing signatures that we identified retrospectively and then prospectively developed a workflow where we could biopsy tumors, assign them to these different groups, and assign them to therapy based on these gene expression patterns.

What were the key data from OPTIC RCC were presented during ESMO?

One of the early things to emerge from the trial, pragmatically, was that primary tumors are not the same as metastatic tumors. So initially, when we started the trial, we were [performing] biopsies on both metastatic tumors and primary tumors, and if we had both, we would defer to the metastatic tumors because, you know, patients with metastatic disease—it’s the metastatic tumors that are really driving morbidity and mortality. We wanted our therapy to be targeting that disease, and then we had the primary tumor to fall back on if the metastatic tumor failed quality control.

What quickly emerged is that in patients who had gene expression data for both the primary tumor and the metastatic tumor, we saw a high discordance rate. The metastatic tumor would tell us to treat, for example, with ipilimumab plus nivolumab, and the primary tumor would be angiogenic; this would suggest that we should treat [the patient] with an IO/TKI. Because we wanted to rely on the metastatic biology to guide therapy, we amended the study to require metastatic tumors to be biopsied. That’s an important conclusion from the study—that primary and metastatic disease are different, and if you want to implement a tissue-based biomarker, you need to focus on metastatic biology when treating patients with metastatic disease.

We were able to accrue 27 patients to the angiogenic arm. These were the patients treated with cabozantinib and nivolumab. The ORR among those patients was 76%; [which] met our primary end point. We had an 8% CR rate, 68% partial response [rate], and a 24% stable disease [rate]. Importantly, no patients had progressive disease as their best response, meaning that 100% of patients had some tumor shrinkage. This was very exciting, because it showed that we were able to enrich for response to an anti-angiogenic–containing regimen when we selected these angiogenic tumors.

How did the operational logistics used during the study evolve over time?

One of the questions we frequently get is how practical this is. Is this an academic exercise that you do within a university setting but can never be rolled out to a community practice? One of the things that makes this approach realistic, and practical is getting results back in a timely manner. Early on, one of the metrics we followed was time from consent to trial enrollment to when we got a cluster assignment. Included in that metric was how fast we could get the tissue, how long it took to send it for sequencing, how long sequencing took, how long it took to run the informatics and assign a patient to an angiogenic group or an IO/IO group, and then make that treatment decision.

Early on—among the first 15 patients—the median turnaround time from consent to cluster assignment was approximately 40 days. We quickly learned how to optimize our workflow to minimize sequencing time and streamline the informatics. After those first 15 patients, the median time dropped from 40 days to about 20 days, and it plateaued there for the rest of the study. To enroll these first 27 patients, we screened about 100 patients. The first 15 were slower, and then we optimized the process.

If and when we roll this out into a larger context or larger study, if a patient needs therapy tomorrow, this may not be the right approach. But if a patient can wait about 3 to 4 weeks, then in more than 90% of cases we can get data back within that timeline. Those are some of the logistics of running this type of tissue-based biomarker study.

What has been learned from the preliminary analyses of distinct gene expression patterns in tumors?

I want to stress that all the tumors enrolled on this initial arm that we presented at ESMO were angiogenic tumors treated with cabozantinib and nivolumab. We’re still enrolling in our immunotherapy-driven cohort—patients assigned to ipilimumab and nivolumab—and that arm is ongoing. Because these tumors were intentionally selected to have similar gene expression patterns, they’re relatively homogeneous. All patients had some degree of tumor shrinkage.

Even within this homogeneous group, we asked whether there were differences between the best responders—those with complete or deep responses—and those who had tumor shrinkage but were still classified as having stable disease by RECIST 1.1 criteria. We did see some differences in gene expression patterns. These are very preliminary data and relatively small numbers, but even with those caveats, we observed some differences. For instance, we saw increased expression of genes associated with epithelial–mesenchymal transition in patients whose tumors had less robust responses, which is a known pathway associated with aggressive disease. We’ll continue to refine these studies, but it’s interesting that even within a relatively homogeneous group, we can see distinct gene expression patterns between stronger and weaker responders.

References

1. Haake SM, Beckerman, Barata P, et al. Efficacy of cabozantinib and nivolumab in cluster 1/2 metastatic clear cell renal cell carcinoma: results from OPTIC RCC, a phase II trial of a novel RNAseq-based biomarker. Ann Oncol. 2025;36(suppl 2):S1298. doi:10.1016/j.annonc.2025.08.3207
2. Motzer RJ, Powles T, Atkins MB, et al. final overall survival and molecular analysis in immotion151, a phase 3 trial comparing atezolizumab plus bevacizumab vs sunitinib in patients with previously untreated metastatic renal cell carcinoma. JAMA Oncol. 2022;8(2):275-280. doi:10.1001/jamaoncol.2021.5981