The Debate Over the Predictive Utility of PD-L1–Low Expression is Ongoing in ESCC

Nataliya Uboha, MD, PhD

Current evidence suggests that PD-L1 expression may serve as a useful predictive biomarker for immunotherapy response in esophageal squamous cell carcinoma (ESCC). However, ongoing debate persists about the predictive value of low PD-L1 expression due to challenges in reliable measurement and lack of consensus on cutoffs, potentially influencing regulatory decisions and subsequent clinical use of these agents, according to Nataliya Uboha, MD, PhD.

On September 26, 2024, the FDA’s Oncologic Drugs Advisory Committee (ODAC) voted 11 to 1, with 1 abstention, against the favorable risk-benefit profile of anti–PD-1 therapies for the frontline treatment of metastatic or unresectable ESCC in patients with a PD-L1 combined positive score (CPS) of less than 1%.^1,2

Findings from the phase 3 KEYNOTE-590 (NCT03189719), CheckMate 648 (NCT03143153), and RATIONALE-306 (NCT03783442) trials examining pembrolizumab (Keytruda), nivolumab (Opdivo), and tislelizumab-jsgr (Tevimbra) based regimens, respectively, were presented for consideration during the meeting. A statistically significant overall survival (OS) benefit was observed with the anti–PD-1–containing regimens vs comparator chemotherapy regimens across all 3 trials. However, analyses revealed that OS outcomes were not significantly different between arms among patients with a PD-L1 CPS of less than 1%.

Uboha added that toxicities associated with anti–PD-L1 therapy further complicate this assessment, raising concerns about the efficacy and risk-benefit ratio of these agents for patients with PD-L1–low tumors.

“The ODAC committee voted to recommend restricting the use of these agents to patients whose tumors have PD-L1 expression, and I’m hoping that the FDA will further refine the label going forward,” said Uboha, who is an associate professor and researcher at the University of Wisconsin School of Medicine and Public Health in Madison. “I’m [also] hoping for more of a unified label across different immunotherapy agents. It will not only make it easier for us to use these agents in clinical practice, but it will potentially simplify future trial designs [in] knowing how to use the appropriate control arm.”

In an interview with OncLive^®, Uboha expanded on the ODAC committee’s consensus that PD-L1 expression is a predictive biomarker for treatment response in esophageal cancer; discussed whether the committee will recommend restricting the FDA approval of select anti–PD-L1 agents based on PD-L1 expression; and emphasized the need for standardized testing and guidelines to facilitate their use in clinical practice.

Uboha also discusses the current indications for checkpoint inhibitor therapy in unresectable or metastatic gastric, gastroesophageal junction (GEJ), and esophageal cancers in a concurrent interview.

OncLive: What were the ODAC committee’s opinions regarding the adequacy of PD-L1 expression as a predictive biomarker, the differing risk-benefit assessments across PD-L1 expression subpopulations, and the adequacy of cumulative data to restrict checkpoint inhibitor approvals based on PD-L1 expression?

Uboha: Everybody agreed that PD-L1 is a predictive biomarker of response. Everybody also agreed that it has been a challenging biomarker to use [and] adopt. Across all members, there was strong agreement that patients with a higher PD-L1 score, particularly a PD-L1 CPS of 10% or greater, derive benefit from the addition of immunotherapy to their treatment. There was also general agreement that those who have no expression of PD-1 in their tumors do not derive much benefit.

The discussion about patients whose tumors have low PD-L1 expression, in the range of 1% to 10%, was much more nuanced. Primarily, the discussion was geared to how difficult it is to reliably identify the exact score. There’s a lot of heterogeneity in expression [and] the pathologists have commented on the heterogeneity [of] expression between the primary and metastatic sites. We don’t know what the different treatments, such as radiation, can change [regarding] expression of PD-L1 in these tumors—is expression of PD-L1 dynamic? Looking at patients with a low PD-L1 score and how they respond to immunotherapy treatment was not a primary end point in any of the studies that were discussed.

As such, the final agreement from the ODAC committee was that we do not see benefit with the use of these agents in those that have a PD-L1 CPS of less than 1%, but I don’t think the agreement was reached regarding the appropriate cutoff.

Do you anticipate that the committee will recommend restricting the approval of nivolumab and pembrolizumab to PD-L1–positive populations in esophageal cancer, similar to the FDA's decision to narrow pembrolizumab’s approval in HER2-positive tumors based on PD-L1 expression?

The committee saw the data that showed that these agents are active in [these] diseases. The discussion was whether the committee should consider limiting the approval in biomarker select patient populations for both nivolumab and pembrolizumab. Tislelizumab is not approved yet, but we’re seeing similar efficacy of these agents across different studies in this patient population. We’re also seeing consistent lack of benefit from the addition of these agents in patients with tumors lacking PD-L1 expression. I would anticipate that the committee will likely make a decision to approve tislelizumab in similar settings, since we’ve seen similar efficacy, but I also anticipate that there will be a limitation to indications of both nivolumab and pembrolizumab use based on lack of benefit in those with PD-L1–negative tumors.

It will parallel [the] already limited approval of pembrolizumab in patients whose tumors have concurrent HER2 expression. Initially, pembrolizumab had a broad approval in HER2-positive tumors, regardless of PD-L1 expression. But with accumulating data, the FDA narrowed the approval for the use of pembrolizumab to only those patients who have both HER2-positive and PD-L1–positive tumors with a PD-L1 CPS cutoff of 1%. I anticipate we will see similar narrowing of the indications in HER2-negative tumors as well.

Based on data from the phase 3 RATIONALE-305 (NCT03777657) and RATIONALE-306 trials, how is the lack of consensus regarding PD-L1 cutoffs affecting the use of anti–PD-L1 agents such as tislelizumab in clinical practice, and what steps should be taken to simplify testing and recommendations for community oncologists?

We need better treatments for these patients. These agents work in many of our patients, but I do believe we need to be more nuanced. We should offer these agents to patients where they are likely to work because they do have toxicities [and] patients can develop immune-mediated toxicities that could negatively affect their quality of life. We shouldn’t be offering these drugs to patients who are unlikely to benefit.

I also believe that we need to make testing, recommendations, and guidelines simpler for our partners in the community. Right now, there is discordance between the label, FDA approval, and ASCO as well as NCCN guidelines. We need to make it simpler. We need to have or at least attempt to make similar or uniform cutoffs for PD-L1 expression, so that these agents can be used interchangeably and so that it’s easier for our pathologists and community partners to adopt them in clinical practice.

What advice would you give to community oncologists regarding the management of immune-mediated toxicities associated with the addition of anti-PD-L1 agents to chemotherapy?

The good news is that we did not see any unexpected toxicities with the addition of these agents to chemotherapy. Our community partners use these agents a lot in the management of other diseases such as melanoma or lung cancer. Our community partners have a lot of experience with the use of these agents. For toxicity management, the NCCN is a great resource on how to manage immune-mediated toxicities. [They have] a live document that’s getting updated on a regular basis, and this is a great reference to look at if those types of toxicities are encountered in clinical practice.

What implications might the FDA approval of tislelizumab in combination with different chemotherapy backbones have for the management of ESCC?

There were 2 different sessions in this meeting. The morning session was dedicated to adenocarcinomas and the afternoon session was dedicated to ESCC. One distinction that’s important to make is that most of the ESCC will express PD-L1, and so the use of anti PD-1 agents will probably be broader in patients with ESCC

The data that was presented from the phase 3 study that looked at the activity of tislelizumab in that setting was interesting because it used a different chemotherapy backbone in this disease; it used both 5-FU–based and [capecitabine– or] paclitaxel–based chemotherapy in combination with tislelizumab. If tislelizumab is approved in combination with chemotherapy in this disease, clinicians will also have an opportunity to offer a different chemotherapy backbone to patients with ESCC, [which] is something to keep in mind.

References

1. Meeting of the Oncologic Drugs Advisory Committee. FDA. September 26, 2024. Accessed October 28, 2024. https://www.youtube.com/live/ELA3JDqtcFw
2. Immune checkpoint inhibitors in patients with metastatic or unresectable esophageal squamous cell carcinoma. Oncologic Drugs Advisory Committee Meeting. FDA. September 26, 2024. Accessed October 21, 2024. bit.ly/3UjNvS1