First-Line Therapy in ES SCLC: CASPIAN

Hossein Borghaei, DO, MS: Anne, CASPIAN is another study. It was first presented in 2019 and updated in 2021. It also has some data regarding biomarker analysis. I was wondering if you could shed some light on that topic.

Anne Chiang, MD, PhD: Sure. The CASPIAN trial was similar to IMpower133. This phase 3 trial had about 800 patients with extensive-stage small cell lung cancer. They were randomized 1:1:1 for chemotherapy, chemotherapy plus durvalumab, and chemotherapy plus durvalumab plus tremelimumab, an anti-CTLA4 inhibitor. The first results presented for CASPIAN showed benefit similar to the IMpower133 trial. It was a positive trial, specifically chemotherapy plus durvalumab vs placebo. The median overall survival was 13 months vs 10.3 months. The hazard ratio was 0.73. It’s very similar to the atezolizumab data, which made everybody feel a little better. It was approved by the FDA shortly thereafter.

The differences between this trial and the previous trial included the fact that patients were allowed to be treated with cisplatin and carboplatin investigator choice. They could receive up to 6 cycles of chemotherapy. Asymptomatic brain metastases were allowed in these patients, which reflects what we see in the real world. There are a lot of patients with brain metastases that we unfortunately sometimes have to exclude from our clinical trials. It also allowed patients who had gotten PCI, or prophylactic cranial irradiation.

The most recent update by Jonathan Goldman and colleagues was in The Lancet Oncology, and it showed the results from the durvalumab-tremelimumab arm, which was not positive. There was no difference in overall survival there. However, they note in their paper that the tail of the curve may be where the story is. In other words, although there wasn’t a difference in the overall survival, if you look at 12 months out at the landmark analysis, or even 24 months out, there are a significant number of patients surviving.

If you combine both the durvalumab and the durvalumab-tremelimumab arms and look at patients who have received I/O [immuno-oncology] in combination with chemotherapy at 1 year out and perhaps even longer, about 20% of them are surviving compared with about 5% in the chemotherapy-only arm. Perhaps there’s something there about the tail of the curve. It is of note that only about 60% of the patients were able to get the full 5 planned doses of tremelimumab, so perhaps the actual survival benefit was somewhat attenuated. But there does seem to be some long-term benefit for a subset of the patients. That goes back to what Steve was talking about: How can we figure out which of those patients can benefit?

At the 2020 ESMO [European Society for Medical Oncology] Virtual Congress, Jonathan Goldman and colleagues showed a further analysis of the CASPIAN trial. They looked at patients whose progression-free survival [PFS] was greater than 12 months or less than 12 months. Of course, you’d imagine that the people who are in the greater-than-12-months group do better than the others. But they saw a difference in a number of aspects. For example, the patients who were in the greater-than-12-months PFS category actually had a median of 25 cycles of durvalumab, compared with 7 in the poor PFS category. The patients who were in the better PFS—greater than 12 months—category had an overall response rate of 96% compared with 63%. Their overall survival at 2 years was 77% compared with 11%. That’s saying if you’re going to do well, you’re going to do really well. Unfortunately, if you’re going to do poorly, you’re not going to do well at all.

The other big aspect of that abstract was that TMB, tumor mutational burden, was not helpful in determining prognosis. Going back to the CheckMate 032 trial looking at nivolumab and nivolumab-ipilimumab in TMB, we were all very excited about patients who got nivolumab-ipilimumab who had high TMB and had an overall response rate of about 46%. That didn’t pan out in the IMpower133 trial. Whether or not you used a cut point of 10 or 16 mutations per megabase, that wasn’t predictive of response. TMB is not the marker here. We’ll have an interesting discussion about what we think might be a marker. But that’s basically the results of CASPIAN and what’s happened since then.

Hossein Borghaei, DO, MS: Thank you, Anne. I appreciate it. Those are a lot of good points to come back to during our discussion.

TRANSCRIPT EDITED FOR CLARITY