Doctors Debate: Is PD-L1 Testing Necessary Before Starting ICI Therapy in Gastroesophageal Cancer?


During a debate at the Gastrointestinal Cancers Symposium, Florian Lordick, MD, argued that knowledge of PD-L1 expression level was needed before starting immune checkpoint inhibitor therapy. He was challenged by Aaron James Scott, MD, who argued that knowledge of PD-L1 is not necessary before initiating immune checkpoint inhibitor therapy.

Lordick, University Cancer Center Leipzig

Florian Lordick, MD, PhD

The essentiality of knowing programmed death-ligand 1 (PD-L1) expression levels in patients with gastroesophageal cancer prior to initiating immune checkpoint inhibitors (ICIs) is an enduring question in oncology. During the Immunotherapy for Gastroesophageal Cancer session at the 2022 Gastrointestinal Cancers Symposium, 2 GI oncologists debated the topic.

Florian Lordick, MD, director and professor of oncology at the University Cancer Center Leipzig in Leipzig, Germany argued that knowledge of PD-L1 expression level was needed before starting ICI therapy. He was challenged by Aaron James Scott, MD, associate professor of medicine at the University of Arizona, in Tucson, Arizona, who argued that knowledge of PD-L1 is not necessary before initiating ICI therapy.1,2

In a vote ahead of the debate, 82.9% of the audience agreed that knowledge of PD-L1 expression was needed, while 17.1% disagreed. The number changed after both debate presentations.

Knowing of PD-L1 Level Before Initialing an ICI

Looking at PD-L1 as a biomarker, Lordick started his presentation by highlighting research that showed a 96.6% interpathologist overall agreement with testing for PD-L1 expression and a 97.2% intrapathologist overall agreement. These findings indicate that the biomarker is routinely used in the space. Moreover, he explained that GI oncologists did not always agree on HER2 as biomarker, but research has changed their minds over the years.1

“If we take PD-L1 seriously, we would select approximately half of patients with esophageal and gastric cancer for checkpoint inhibitors, or in other words, we would exclude 50%,” said Lordick during his presentation.

With the validity of the biomarker decided, Lordick explained his stance on the question of PD-L1 knowledge before ICI initiation using data from the KEYNOTE-590 (NCT03189719), CheckMate-648 (NCT03143153), and CheckMate-649 (NCT02872116) trials. Collectively, these clinical trials demonstrated that some PD-L1 expression levels impacted treatment outcomes.

“The main argument was that some of the studies, showed positive effects of immune checkpoint inhibitors such as pembrolizumab [Keytruda] and nivolumab [Opdivo] in an all-comer populations. In most of the studies, especially those that lead to drug approvals in Europe and also in the United States, subgroup analyses consistently show that there is a clear enrichment of the benefits of these drugs. In those patients who have tumors that express PD-L1, mostly measured as a combined positive score [CPS], and others also measured as a tumor proportion score [TPS], we have increasing evidence now that the subgroups without an expression of PD-L1 or very weak expression of PD-L1 do not gain substantial benefit from the use of these drugs,” Lordick told Targeted Therapies in Oncology (TTO), in an interview.

In the randomized, placebo-controlled phase 3 KEYNOTE-590 trial, the combination of pembrolizumab and chemotherapy was compared with chemotherapy alone as frontline treatment for patients with advanced esophageal cancer. Of the 749 patients randomly assigned to either of the study arms, 50% of the pembrolizumab arm and 52% of the chemotherapy-only arm had a PD-L1 CPS of ≥ 10. Results for the coprimary end points of overall survival (OS) and progression-free survival (PFS) showed that pembrolizumab was not effective in patients with a PD-L1 expression < 10 compared with those who had high PD-L1 expression (HR, 0.62; 95% CI, 0.49-0.78; P <.0001).3

“This is an example where we have a different approval situation. The EMA approved pembrolizumab for tumors expressing PD-L1 with a CPS of 10 or greater while FDA approved for all-comers. Although the FDA says this, the study is only convincingly positive for patients with tumors expressingPD-L1,” Lordick told TTO.

Further evidence of the importance of PD-L1 for use of ICI therapy in esophageal cancer was described with data from the CheckMate-648 trial , which explored the combination of nivolumab and chemotherapy or nivolumab plus ipilimumab (Yervoy) compared with chemotherapy alone in 970 patients with advanced disease.1,4

The study showed that in patients with PD-L1 TPS ≥ 1%, there was a significant improvement in the end point of OS with both ICI-containing combinations compared with chemotherapy alone. Lordick specifically highlight the comparison of nivolumab/chemotherapy versus chemotherapy alone. Patients who received the ICI combination had a median OS of 15.4 months (95% CI, 11.9-19.5) compared with 9.1 months (95% CI, 7.7-10.0) in the chemotherapy arm (HR, 0.54; 99.5% CI, 0.37-0.80; P <.0001).4

“The final, very important study for the group of patients with gastric cancer or stomach cancer, including [gastroesophageal] junction adenocarcinoma and esophageal adenocarcinoma, is called CheckMate-649,” Lordick said in the interview. “Here, the combination of nivolumab plus standard chemotherapy was compared to standard chemotherapy alone. And again, there was a compelling survival advantage of the combination with the nivolumab and chemotherapy, but it was confined to those patients who have a PD-L1 CPS of 5 or greater, while a subgroup analysis showed that if PD-L1 is not expressed, or if there is weak expression, anything below 5, there is no benefit.”

The study included a total of 2687 patients. Among patients with high PD-L1 expression, the OS improvement was significant with nivolumab/chemotherapy versus chemotherapy alone (HR, 0.71; 98.4% CI 0.59-0.86]; P < .0001). PFS was also significantly improved with the addition of ICI therapy to chemotherapy versus chemotherapy alone (HR, 0.68; 98 % CI, 0.56-0.81 ; P < .0001).5

The most important point about PD-L1 expressionis its predictive value for the benefit of immune checkpoint inhibitors,” Lordick told TTO. “We do not see a clear difference in terms of prognosis, like we see for some other markers. But for PD-L1, the important point is to enrich the population, those who may derive the benefit from drugs like nivolumab or pembrolizumab or the combination of nivolumab and ipilimumab.”

Not Knowing PD-L1 Expression Before Initiating an ICI

When deciding on whether knowledge of PD-L1 status is essential before initiating ICI therapy, Scott told the Gastrointestinal Cancer Symposium audience that the matter is complicated, and context does matter. He added that the clinical question itself is simpler than the complexities with understanding immunotherapy in gastroesophageal caner, which includes factors like disease type, histology, race, chemotherapy partner type, and other molecular factors besides PD-L1.2

Scott, Arizona Cancer Center

Aaron James Scott, MD

“All of these things are important factors in how patients respond to treatment. And so, it becomes a little challenging to use PD-L1 expression level as a way to be the end-all-be-all for deciding whether to use the immunotherapy in the first-line setting,” Scott told TTO in an interview. “Also, another important feature that is supremely relevant is that these patients have high morbidity, usually very symptomatic when they're presenting to the clinic. So, time is of the essence to initiate potentially beneficial treatments, and waiting for PD-L1 testing may be a hinderance. For example, it could take up to 2 to 3 weeks for certain types of testing. So, having that data, waiting to start treatment can be potentially detrimental, in a way, to the patient.”

Data supporting Scott’s argument were from 2 clinical trials of esophageal squamous cell carcinoma, of which, he explained, both hinted that PD-L1 status did not matter.

“I think the data for the different phase 3 trials, especially in esophageal squamous cell carcinoma [ESCC] best support my argument. When you look at the median overall survival numbers and median progression-free survival, the JUPITER-06 and ORIENT-15 trials are most striking. You see that the hazard ratio for all comers regardless of PD-L1 subgroup is as good or better compared to patients with a CPS positive score,” Scott told TTO. “So, I think those numbers are very important to recognize. In contrast, of course, we have other studies that show that the numbers are better for CPS-high subgroups. But I think those 2 phase 3 trials are very important.”

In the randomized, double-blind JUPITER-06 study (NCT03829969), 257 patients with advanced or metastatic ESCC were randomized to receive either toripalimab (Tuoyi) in combination with paclitaxel and cisplatin, or placebo maintenance. In all-comers, the median OS was 17.0 months with toripalimab and chemotherapy, compared with 11.0 months with placebo (HR, 0.58; 95% CI, 0.43-0.78; P = .00037). The improvement in PFS with toripalimab and chemotherapy compared with placebo was also significant regardless of PD-L1 expression level (HR, 0.58; 95% CI, 0.46-0.74; P < .00001).6

The global, randomized, double-blind ORIENT-15 study (NCT03748134) included 659 patients with unresectable locally advanced, recurrent or metastatic ESCC who were randomly assigned to receive either sintilimab (Tyvyt) with placebo followed by paclitaxel and cisplatin or cisplatin alone. In all patients, the median OS observed with the sintilimab combination was 16.7 months versus 13.6 months in patients treated with chemotherapy alone (HR, 0.638, 95% CI, 0.480-0.848; P = .0018). The OS benefit was also demonstrated in the subgroup of patients with a PD-L1 CPS ≥ 10 (HR, 0.638; 95% CI, 0.480-0.848; P = .0018).7

The median PFS in all patients treated with the sintilimab combination was 7.2 months versus 5.7 months among those treated with chemotherapy alone (HR, 0.558; 95% CI 0.461-0.676; P < .0001). The PFS benefit of the immunotherapy combination carried over into the high PD-L1 population (HR, 0.580; 95% CI 0.449-0.749; P < .0001).7

Aside from the clinical trial evidence, Scott says PD-L1 is not an imperfect biomarker. During his presentation, he noted that spatiotemporal heterogeneity, low positive predictive value, getting different results from different assays, and delays in the start of treatment may be key limitations to using the biomarker.

“We have learned that PD-L1 expression definitely has an impact on the tumor immune microenvironment. And in most cases, we believe that PD-L1 acts as a dampening of the immune system so that the immune system isn't able to attack the cancer cells. So, in our experience, we're now building mounting evidence that shows that PD-L1 levels on both tumor cells and immune cells, such as macrophages and lymphocytes, may be important to help guide our management strategy as a predictive biomarker,” said Scott in the interview.

“What we're learning though is that PD-L1 expression is dynamic, showing that where you biopsy and when you biopsy seems to matter. There's a phenomenon called spatialtemporal heterogeneity, where you compare 1 particular study by Xao et al that was published in Clinical Cancer Research in 2020, that showed differences between positivity rates in a primary biopsy versus a metastatic biopsy in the same patient can differ as well as pretreatment biopsies versus post treatment biopsies. PD-L1 expression level can change. So, then it becomes a question of How do we make sure that a PD-L1 level expression level does give us important positive predictive value and negative predictive value? And then: When, how, and where do we biopsy? I think the ideal answer is that we have a biomarker that can study either PD-L1 expression level directly or surrogate in the blood, so we can get around this spatialtemporal heterogeneity, perhaps,” Scott added.

No Consensus

Following both debate presentations at the Gastrointestinal Cancers Symposium, 32.9% of the audience was against known PD-L1 expression level before initiating ICI therapy versus 67.1% who still agreed that it is essential.

Regarding the presentation given by Lordick, Scott said, “I really believe that when we are developing a predictive biomarker that does, in certain circumstances, seem to offer good negative predictive value, that we have to also consider costs and adverse effects/safety to be as important. So, I really appreciated his perspective on understanding the balance that we need to strike among those things.”

Lordick also enjoyed hearing Scott’s opinion, telling TTO, “Scott did a great job to show the variety of data and to show that there are also doubts and heterogeneity among the expression patterns of PD-L1. He also made a point that in some, especially Asian Studies, and in some that investigated drugs, which are not yet approved and not yet in use either in Europe or in America, there seems to be benefit of these drugs independent of the PD-L1 expression.”

For gastroesophageal cancers, should you know a patient's PD-L1 status before starting them on an immune checkpoint inhibitor?



1. Lordick, F. Debate: Knowledge of pd-l1 expression level is essential before initiating checkpoint inhibitors: FOR. Presented at 2022 ASCO Gastrointestinal Cancers Symposium; January 20-22, 2022, San Francisco, CA.

2. Scott A. Debate: Knowledge of pd-l1 expression level is essential before initiating checkpoint inhibitors: AGAINST. Presented at 2022 ASCO Gastrointestinal Cancers Symposium; January 20-22, 2022., San Francisco, CA.

2. Sun J, Shen L, Shah, et al. Pembrolizumab plus chemotherapy versus chemotherapy alone for first-line treatment of advanced oesophageal cancer (KEYNOTE-590): a randomised, placebo-controlled, phase 3 study. Lancet. 2021;398(10302):759-771. doi: 10.1016/S0140-6736(21)01234-4

3. Chau I, Doki Y, Ajani J, et al. Nivolumab (NIVO) plus ipilimumab (IPI) or NIVO plus chemotherapy (chemo) versus chemo as first-line (1L) treatment for advanced esophageal squamous cell carcinoma (ESCC): First results of the CheckMate 648 study. J Clin Oncol. 2021;39(18). doi: 10.1200/JCO.2021.39.15_suppl.LBA4001

4. Janjigian Y, Shitara K, Moehler, et al. First-line nivolumab plus chemotherapy versus chemotherapy alone for advanced gastric, gastrooesophageal junction, and oesophageal adenocarcinoma (CheckMate 649): a randomised, open-label, phase 3 trial. Lancet. 2021;398(10294):27-40. doi: 10.1016/ S0140-6736(21)00797-2

5. Xu R, WangF, Cui C, et al. 1373MO - JUPITER-06: A randomized, double-blind, phase III study of toripalimab versus placebo in combination with first-line chemotherapy for treatment naive advanced or metastatic esophageal squamous cell carcinoma (ESCC). Ann Oncol. 2021;32 (suppl_5): S1040-S1075. doi: 10.1016/annonc/annonc708

6. Shen L, Lu Z, Wang J, et al. LBA52 - Sintilimab plus chemotherapy versus chemotherapy as first-line therapy in patients with advanced or metastatic esophageal squamous cell cancer: First results of the phase III ORIENT-15 study. Ann Oncol. 2021; 32 (suppl_5): S1283-S1346. doi: 10.1016/annonc/annonc741

Related Videos
Rohit Gosain, MD; Rahul Gosain, MD; and Pamela L. Kunz, MD, presenting slides
Rohit Gosain, MD; Rahul Gosain, MD; and Pamela L. Kunz, MD, presenting slides
Rohit Gosain, MD; Rahul Gosain, MD; and Pamela L. Kunz, MD, presenting slides
Rohit Gosain, MD; Rahul Gosain, MD; and Pamela L. Kunz, MD, presenting slides
Rohit Gosain, MD; Rahul Gosain, MD; and Pamela L. Kunz, MD, presenting slides
David Zhen, MD, an expert on gastric cancer
Related Content