Considering Role of PD-L1 CPS in Frontline Gastric Cancer Treatment

The NCCN’s guidelines recommend adding immune checkpoint inhibitors in advanced esophageal and gastric cancers in patients with a PD-L1 combined positive score (CPS) of at least 1 but only give a category 1 preference if the CPS is 5 or greater.¹ During a virtual Case-Based Roundtable event, Yelena Y. Janjigian, MD, asked participants how they approach treatment for patients with CPS between 1 and 5. Janjigian, chief of the gastrointestinal oncology service and Carroll and Milton Petrie Chair at Memorial Sloan Kettering Cancer Center, discussed the overall reliability of PD-L1 scores and why it is important to take the opportunity to use immunotherapy (IO) when it is indicated in the first line of treatment.

Register today to join a Case-Based Roundtable near you.

This article is part 2 of a 2-part series from a Case-Based Roundtable event.

DISCUSSION QUESTIONS

• How does the CPS score influence your treatment selection?​
• What threshold (e.g., CPS ≥ 5 vs 1–4) drives your use of IO?​
• Do you see a difference in real-world benefit for patients with low CPS scores (1–4)?​
• What are the limitations of CPS scoring in guiding first-line therapy, and how does it affect your decision-making in practice?​

Yelena Y. Janjigian, MD: When do you consider adding an immune checkpoint inhibitor for first-line chemotherapy? What cutoff you do you use?

Haiyun Wang, MD: I use CPS of 1.

Fayez Estefan, MD: I use a CPS of 1 to discuss IO with the patient. If Claudin18.2 is negative, I [have] a discussion with the patient about the potential benefit and the adverse events and have shared decision-making, because if you don’t offer IO in the first line, this is the only chance to use this drug, and they will probably be one of the patients that will respond. I review the data with patients with CPS of 1 to 5 and make a shared decision with the patient.

Janjigian: Do you wait until you get the [PD-L1] CPS before deciding?

Estefan: Our pathology lab does the HER2, mismatch repair [MMR], and PD-L1 for all the gastric and gastroesophageal junction adenocarcinoma without waiting for me. Usually, I have these data by the time I see the patient.

Wang: Yes, it’s similar here. We do PD-L1, HER2, and MMR in-house, so usually when [I see] the patient, we [already] have the result. If not, I wait, because we have the result in the same week. I use the CPS score to make a decision for IO and then use [CPS of] 1 as a cut off. If it’s 1 or above, I would use it. Have you used IO for PD-L1 expression less than 1%?

Janjigian: Yes, I have a lot of experience with it, because often what happens is…the patient comes in from elsewhere, the PD-L1 that is ordered at their institution is pending, and then we reorder it, and it’s back. Then we’ll find CPS 1 or 5, or sometimes positive, and at the local institution it is negative, or vice versa. What I’ve seen in practice is the discordance between several different testings of the same sample. PD-L1 testing is not that reliable. I often retest. In my practice, I had a patient with HER2 positive, PD-L1–negative disease, and I’ve never seen one of those before. When I asked them to retest it, it turns out that one of the first sites they tested was a lymph node that was completely replaced by tumor, so there is not a lot of stroma there. When they retested another sample from the same timepoint pretreatment, it had a PD-L1 CPS of 8.

I think practically, there’s a lot of that going back and forth. On the other hand, we know there is level 1 evidence from phase 3 studies…in a clinical trial [setting] where it’s a very controlled environment of excellent tumor blocks and reference radiologists with the same person reading all the slides in that setting. If you have a PD-L1–low tumor, it’s a very high probability that you’re not going to benefit as much with IO. You’re not harmed, but…the survival benefit is less. But in clinical practice, the situation is very different. The quality of the sample, the testing quality and so forth, is so variable, and so that’s why I err on the side of treatment with anti–PD-1, particularly because Claudin18.2 inhibition has so much difficulty associated with it.

Bolanle Adepoju, MD: I had a patient who transferred care towards the end of his FOLFOX [leucovorin, fluorouracil, and oxaliplatin] and he was HER2 positive so he’s on FOLFOX and trastuzumab [Herceptin], but he was never started on PD-L1 therapy, and it’s not clear why. I didn’t see PD-L1 CPS testing in his transfer records, and it was mentioned that it might be added on later. He’s on cycle 12, his oxaliplatin has fallen off, and he’s going on maintenance now. What are your thoughts about adding PD-L1 therapy? I’m getting restaging scans to see where he’s at. If it is not progressing, what are your thoughts about adding [anti–PD-L1]?

Janjigian: I would. I think if you stopped oxaliplatin, if the tumor is PD-L1 positive, and if he has residual tumor left, and if he’s still fit for therapy, I would add anti–PD-1 now, because he’s technically still in first line, [so] it’s technically still approved. If [the disease] progresses, is he going to be fit enough for trastuzumab deruxtecan [T-DXd] in the second line? He may not, and anti–PD-1 clearly improves survival in the first-line setting.

Yanyan Lou, MD: Since the PD-L1 expression is quite heterogeneous in lung adenocarcinoma, given the potential benefit, if you see PD-L1 negative, do you rebiopsy…?

Janjigian: I try to retest, or I just call the pathologist and have them comment on the quality, because sometimes people get tested and you’ll see that they actually used tumor proportion score, not CPS, or the quality of the sample is not so good.

Sheila Donnelly, MD: Have you had long-term survivors with nivolumab [Opdivo] plus chemotherapy,

either squamous or adenocarcinoma, with liver metastases?

Janjigian: In our data, unlike colon cancer, liver metastases are not a negative predictor of outcome of these agents. It’s because we have a bigger fish to fry, which is peritoneal disease. With the peritoneal disease and ascites and things like that, some people tend to do poorly. But liver metastases, per se, especially small or medium-sized liver metastases, are not a threat to us. For us, the battle is usually lost in the peritoneum.

Have you used nivolumab plus ipilimumab [Yervoy] in squamous [histology] in the first line?

Estefan: I would use chemotherapy plus single-agent IO.

Janjigian: Yes, they come in pretty sick, so you need to get a quick, brisk response. I’m always very nervous about using chemotherapy-free backbone.

What are the options for patients who are chemotherapy/IO failures? Given the high proportion of patients who do not receive more than 1 line of therapy, how important is using most effective therapy first? What do you tend to use in the second line for your patient?

Estefan: A taxane with ramucirumab [Cyramza].

Janjigian: And in squamous?

Ike Onwere, MD: I typically use the same.

Janjigian: So carboplatin/paclitaxel or paclitaxel. Ramucirumab is technically not for squamous. And then for HER2 positive?

Muhammad Masab, MD: T-DXd.

Janjigian: Have you had experience and what are your thoughts about using T-DXd?

Masab: I just had 1 patient, and they just didn’t do well with it. They were just frail. The patient deteriorated in terms of performance status after a couple of rounds.

Janjigian: Did this patient get pembrolizumab [Keytruda] and trastuzumab in the first line?

Masab: This patient was HER2 positive on the repeat biopsy, so they got chemotherapy plus IO in the first line.

Register today to join a Case-Based Roundtable near you.

_{DISCLOSURES: Janjigian previously reported research funding from Bayer, Bristol Myers Squibb, Cycle for Survival, Department of Defense, Fred’s Team, Genentech/Roche Lilly, Merck, National Cancer Institute and Rgenix; serving as a consultant or in an advisory role for Basilea Pharmaceutical, Bayer, Bristol Myers Squibb, Daiichi Sankyo, Imugene, AstraZeneca, Lilly, Merck, Merck Serono, Michael J Hennessy Associates, Paradigm Medical Communications, Seattle Genetics, Pfizer, Rgenix, AmerisourceBergen, Arcus Biosciences, Geneos, GlaxoSmithKline, Imedex, Lynx Health, Peerview, Silverback Therapeutics and Zymeworks; receiving stock options from Rgenix; and nonfinancial relationships with Clinical Care Options, Axis Medical Education and Research to Practice.}

_REFERENCE

_{1. NCCN. Clinical Practice Guidelines in Oncology. Gastric cancer, version 3.2025. https://www.nccn.org/professionals/physician_gls/pdf/gastric.pdf}