Ilson Examines Chemoimmunotherapy Regimens for Metastatic Gastroesophageal Cancers

David H. Ilson, MD, PhD

Gastrointestinal Medical Oncologist

Memorial Sloan Kettering Cancer Center

New York, NY

CASE SUMMARY

• A 60-year-old man presented with abrupt 10-lb weight loss, dyspepsia, bloating after meals, and loss of appetite, and all symptoms worsened over the past 3 months.
• Family history: Father died of gastric cancer aged 50 years.
• Medical history: overweight, hypertension
• Esophagogastroduodenoscopy showed a 2-cm protruding mass in the body of the stomach, without ulceration.
• Biopsy revealed poorly differentiated adenocarcinoma.
• Molecular testing: HER2/neu negative; mismatch repair proficient/microsatellite stable; PD-L1 combined positive score (CPS), 5
• CT of the abdomen and chest revealed a gastric polypoid mass and thickening; no ascites; and 2 lesions in the lower lobe of the left lung.
• Lung biopsy confirmed metastatic adenocarcinoma consistent with gastric primary tumor.
• PET/CT confirmed metastatic disease.
• Diagnosis: stage IV gastric adenocarcinoma
• The patient was started on FOLFOX (leucovorin calcium [folinic acid], fluorouracil, and oxaliplatin) plus nivolumab (Opdivo).

PEERS & PERSPECTIVES IN ONCOLOGY: What data support adding nivolumab to chemotherapy in patients with gastroesophageal adenocarcinoma?

ILSON: This was the landmark trial CheckMate 649 [NCT02872116], a 3-arm phase 3 study. The control arm was capecitabine/oxaliplatin or FOLFOX with or without the addition of nivolumab. The third arm, which was a nonchemotherapy arm, ipilimumab [Yervoy] plus nivolumab. This was not superior to chemotherapy and had a significant rate of toxicity, so this regimen will not move forward in adenocarcinoma. The primary end point of this study was a dual co–primary end point of overall survival [OS] and progression-free survival [PFS]. They designed the trial for patients with a CPS greater than or equal to 5.

We saw unequivocal improvements in overall survival for the patients with PD-L1 CPS 5 or higher with a nearly 3-month improvement in OS and with an HR of 0.7, which was highly statistically significant [95% CI, 0.61-0.71].¹ We also saw a significant improvement in PFS of about 2 months [HR, 0.71; 95% CI, 0.61-0.82]. It’s key to note the tail on the [Kaplan-Meier] curve here showing that patients get sustained benefit from this treatment sometimes beyond 2 years of therapy, which is unprecedented. We can see significant rates of more long-term OS with the use of a checkpoint inhibitor. PFS2 [time to second progression] in patients going on to get subsequent treatment still favored upfront use of nivolumab in first-line treatment [HR, 0.67; (95% CI, 0.58-0.77) in those with CPS of at least 5].

What data were there on antitumor response to nivolumab?

There was a significant augmentation of antitumor response with an improvement from 45% to 60% and an improvement in response duration of 7.0 to 9.6 months with the addition of nivolumab. Some of these responses were durable beyond 2, 3, and 4 years [Figure¹]. In all randomly assigned patients, there also was an augmentation of overall response rate [ORR] from 46% to 58% with nivolumab, but this is mostly driven by the CPS-positive patients.

What about OS [in responders vs nonresponders]? If we look at the CPS greater than or equal to 5, this looks at an 18-week landmark, and you see again superiority for nivolumab plus chemotherapy [median OS, 20.5 months for responders vs 11.7 for nonresponders] over chemotherapy alone [median OS, 15.5 months for responders vs 10.8 for nonresponders] if we look at the PD-L1 CPS greater than or equal to 5. There was clear superiority for adding nivolumab to chemotherapy over chemotherapy alone in all randomized patients, but again, this is driven by the PD-L1 CPS-positive patients.

What was the safety profile in CheckMate 649?

Treatment-related adverse events [TRAEs] were higher with the addition of nivolumab to chemotherapy. The rate of any TRAEs increased from 45% to 60%. For serious TRAEs, there was an increase of about 7% and about a 9% or higher rate of treatment discontinuation, but not a significant difference in treatment-related deaths. We are now becoming amateur immunologists. We are learning how to be sensitive to these toxicities to stop therapy, institute intermediate- or high-dose steroids, and refer to consultants if need be.

Could you summarize the design and outcomes of the KEYNOTE-859 trial [NCT03675737]?

KEYNOTE-859 is the more recent pembrolizumab [Keytruda] trial. This treated gastroesophageal junction and gastric adenocarcinomas. This [phase 3 trial] used either the every 3 weeks schedule of capecitabine/oxaliplatin or fluorouracil/ cisplatin. Most of the patients on this trial were treated with capecitabine/oxaliplatin on an every 3 weeks schedule. This was quite a large study that treated patients globally. This had the primary end point of OS, randomly assigning patients to a 3-week schedule of pembrolizumab plus chemotherapy every 3 weeks vs placebo plus fluoropyrimidine platinum-based chemotherapy. The FDA approved pembrolizumab in first-line treatment in gastric cancer based on these positive data. The other point to make is the more liberal PD-L1 cutoff. The more restrictive CPS of 10 from the earlier phase 3 KEYNOTE-590 study [NCT03189719] has now been liberalized to a consideration for 5 or higher, and even potentially a consideration for 1 to 9.

Here, the primary trial design was CPS of 1 or higher. They also looked at 10 or higher. You can see an OS benefit in all patients [HR, 0.79; 95% CI, 0.71-0.88] in those who had CPS greater than 1 [HR, 0.75; 95% CI, 0.66-0.85], and the greatest benefit is [seen in patients with] CPS 10 or higher [HR, 0.64; 95% CI, 0.53-0.78] for adding pembrolizumab to first-line chemotherapy every 3 weeks.²

PFS similarly was improved across the board, and the greater degree of CPS positivity leads to enhancement in PFS benefit. [Looking at the] TRAEs, typically, the chemotherapy-related AEs were not significantly augmented when compared with chemotherapy alone.

What is the significance of the phase 3 CheckMate 648 trial (NCT03143153)?

This was one of the largest global trials ever to be conducted in esophageal squamous cancers, and there was a 3-arm randomization to fluorinated pyrimidine plus cisplatin with or without nivolumab and a third arm of nonchemotherapy consisting of nivolumab/ipilimumab.

The CPS cutoff of 1 seemed to discriminate here. This was a stunning result, more than a 6-month improvement in OS with an HR of 0.60 [95% CI, 0.47-0.77].³ This clearly diminishes when we include all patients, with an HR that is still significant but dropping to 0.77 [95% CI, 0.65-0.92]. The vast majority of patients, 90% of patients, had PD-L1 CPS of 1 or higher on this trial.

PFS was also improved. Again the greatest benefit was in the PD-L1 CPS 1 or higher, which was the vast majority of patients, with an HR of 0.67 [95% CI, 0.51-0.89]. In all randomized patients, we can see a diminution of benefits. Although the HR was fairly significant, you can see that it is likely that patients who are CPS negative, less than 1, are not going to benefit here [HR, 0.82; 95% CI, 0.68-1.00].

For response and duration of response, there was a pretty stunning result. [ORR with] chemotherapy alone was 20%, which improved to 53% with the addition of nivolumab to chemotherapy. Ipilimumab/nivolumab also bested chemo-therapy, 47% to 27%. It is interesting that the duration of response seemed to trend higher in the nivolumab/ipilimumab arm. We can see that PD-L1–positive patients had significant augmentation of response and a significant improvement in duration of response, and a diminution of benefit in all randomly assigned patients, again likely indicating…that those with CPS less than 1 would have no clear benefit.

What does this mean for the use of ipilimumab/nivolumab in this setting?

For ipilimumab/nivolumab, this is intriguing. This is a non-chemotherapy regimen in metastatic esophageal squamous cancers. The main focus of the trial was PD-L1 CPS of 1 or higher, and ipilimumab/nivolumab also was superior to chemotherapy [HR, 0.63; 95% CI, 0.49-0.81]. And OS in all randomized patients was also superior [HR, 0.78; 95% CI, 0.65-0.92]. There is a slightly higher early death rate in patients getting ipilimumab/nivolumab without chemotherapy. That’s a caveat with this approach. Ipilimumab/nivolumab is an acceptable alternative nonchemotherapy regimen to treat patients with first-line esophageal squamous cancer, but this is not a group of patients I would treat if they had bulky tumors that are highly symptomatic.

The better choice here is likely to be chemotherapy plus nivolumab because you do see this early drop-off in patients. We have to be judicious in which patients we select for this regimen. Personally, I would treat patients with ipilimumab/ nivolumab if the primary [tumor] were controlled and if they had low-volume metastatic disease, but it is an option and certainly a nonchemotherapy option if clinically appropriate.

PFS was also improved. The curves separate after about 6 months. In all randomly assigned patients, again we see a diminution of benefit and clearly the greatest benefit is in those with CPS of 1 or higher, which is the vast majority of patients treated on this trial [HR, 1.03; 95% CI, 0.78-1.35].

If we look at ipilimumab/nivolumab, the augmentation of response was more modest. We saw 20% for chemotherapy vs 53% for nivolumab/chemotherapy. Here it is 20% vs 35% for ipilimumab/nivolumab, so not quite as big a bump in response. But this treatment arm, ipilimumab/nivolumab, had the longest duration of response, which is intriguing, 11.8 months compared with 5.7 months [with PD-L1 CPS 1 or higher], and a slight diminution of benefit in all randomly assigned patients [11.0 months vs 7.1 months].

What was the safety of ipilimumab/ nivolumab compared with chemotherapy?

TRAEs were manageable. If we look at serious TRAEs, the rate of grade 3 and 4 TRAEs was higher with the addition of nivolumab to chemotherapy, only about a 6% to 10% increase; 23% with ipilimumab/nivolumab compared with 13% for chemotherapy alone. We see the typical predictable toxicities. Obviously, hypothyroidism is very common. We have to monitor for this and put patients on thyroid supplements. [In terms of] gastrointestinal toxicities, we mainly worry about colitis, which we probably see in only 1% or 2% of patients, and hepatitis in 1% or 2% of patients. Pneumonitis also has to be followed. These are all organ toxicities where we generally see them in single-digit percentages. Cutaneous toxicities are common, usually low grade. We can often manage them with topical steroids. If we get more severe cutaneous toxicities, this is when I consult my dermatologists. Based on the severity of the toxicities, do we do the lower prednisone dosing or the higher dosing of 1 mg/kg? For the lower-grade toxicities, we sometimes consider intermediate steroid dosing, and I certainly have a low threshold to get my consultants involved to help manage these treatable toxicities.

_REFERENCES

_{1. Shitara K, Moehler MH, Ajani JA, et al. Nivolumab (NIVO) + chemotherapy (chemo) vs chemo as first-line (1L) treatment for advanced gastric cancer/gastroesophageal junction cancer/esophageal adenocarcinoma (GC/GEJC/EAC): 4 year (yr) follow-up of CheckMate 649. J Clin Oncol. 2024;42(suppl 3):306. doi:10.1200/JCO.2024.42.3_suppl.306}

_{2. Rha SY, Wyrwicz LS, Yanez Weber PE, et al. Pembrolizumab (pembro) + chemotherapy (chemo) for advanced HER2-negative gastric or gastroesophageal junction (G/GEJ) cancer: updated results from the KEYNOTE-859 study. J Clin Oncol. 2024;42(suppl 16):4045. doi:10.1200/JCO.2024.42.16_suppl.4045}

_{3. Chau I, Ajani JA, Kitagawa, et al. Nivolumab (NIVO) plus chemotherapy (chemo) or ipilimumab (IPI) vs chemo as first-line (1L) treatment for advanced esophageal squamous cell carcinoma (ESCC): 45-month (mo) follow-up from CheckMate 648. J Clin Oncol. 2024;42(suppl 16):4034. doi:10.1200/JCO.2024.42.16_suppl.4034}