Demetri Interprets Survival Outcomes for Trabectedin in Soft Tissue Sarcomas

George D. Demetri, MD, expands on the challenges with overall survival as a primary outcome measure, the potential of trabectedin in soft tissue sarcomas, and the significance of a recent trial.

George D. Demetri, MD

The path of trabectedin (Yondelis), an investigational treatment for soft tissue sarcomas, has been an interesting one, said George D. Demetri, MD, director of the Center for Sarcoma and Bone Oncology at the Dana-Farber Cancer Institute.

The drug, which Demetri described as “extraordinarily powerful in the laboratory” has been in clinical development for nearly 20 years, with a focus on leiomyosarcomas and liposarcomas for the past 12 in subtypes thought to be particularly sensitive. Numerous studies, many of them designed in the United States, have resulted in the approval of trabectedin in more than 70 countries worldwide. Despite this, data have not been of the quality required for an FDA approval in the United States thus far, said Demetri.

In late 2014, an application for trabectedin was submitted to the FDA for patients with advanced soft tissue sarcoma who received prior chemotherapy. The application was granted a priority review, with a decision deadline set for August 2015. However, with this date passed, the FDA decision is still unknown.

The final analysis of the phase III data for trabectedin, presented recently at the 2015 European Cancer Congress, found that after a 21-month follow-up, the median progression-free survival (PFS) with trabectedin was 4.2 versus 1.5 months with dacarbazine (HR, 0.55; 95% CI, 0.44-0.70; P <.001). However, the median overall survival (OS) advantage, which was the primary outcome measure for the trial, was not statistically significant, at 13.7 months with trabectedin versus 13.1 months with dacarbazine (HR, 0.93; 95% CI, 0.75-1.15; P = .492).

Grade 3 adverse events found to be higher in the trabectedin arm included ALT levels (25% vs 1%), neutropenia (21% vs 11%), anemia (14% vs 11%), and increased AST levels (12% vs 0%). Sixteen percent of patients receiving trabectedin had grade 4 neutropenia compared with 10% in the dacarbazine group.

Approximately 70% of patients in each group went on to receive additional treatment, a fact that could have impacted OS data, said Demetri. Eighteen percent of patients in the trabectedin arm and 28% of patients in the dacarbazine arm received subsequent pazopanib (Votrient).

OncLive: What about trabectedin makes it interesting as a treatment for sarcoma?

What have been the most significant findings from the trial?

In an interview with OncLive, Demetri expands on the challenges with OS as a primary outcome measure, the potential of trabectedin in soft tissue sarcomas, and the significance of the recent trial.Dr Demetri: Trabectedin is fascinating, partially because we still do not understand all of the things that it does. It is very multifunctional; it is derived from a sea squirt that grows in a colony. To protect itself since it can’t get away from things, it creates this chemical to thwart predators. This chemical is now synthetic and can be made in mass quantities. The reason our sarcoma community is interested in it is because with trabectedin, some patients get extremely good responses and durable stability of their disease, sometimes lasting 4 or 5 years. This is not trivial at all. The problem is, we do not know who those people are who are going to get the extraordinary benefit, who will get modest benefit, or who will get no benefit at all. That is why we designed this trial.We have seen a very sizable difference in terms of PFS for trabectedin over dacarbazine. As a clinically meaningful endpoint, PFS was validated by the last FDA-approved drug for soft tissue sarcoma, which was pazopanib. The important thing about that approval was that pazopanib did not have a survival benefit for patients who took it after all other chemotherapies had failed. That is an important consideration.

The FDA really learned about sarcomas in the process of evaluating pazopanib, and agreed with the experts in the sarcoma community that control of disease is a good thing for patients. Patients with sarcoma are very different than patients with other carcinomas. They can actually tolerate a great deal of disease burden in their body and still feel fairly well. However, if the disease burden gets just a little bit bigger, all of the systems can fail and the patient can die very quickly. Therefore, the important thing is to control the disease as much as possible, and hopefully keep people feeling well while their disease is managed.

Why was OS selected as the endpoint for this trial and what are the challenges with it?

Interestingly, the study did not show that the trabectedin-controlled patients lived longer. There is probably a reason for that. The study began in 2011; however, in 2012, the FDA approved pazopanib so there was a new and effective therapy that patients could go on after they were part of the trabectedin study. This may have impacted OS results. In our study, patients once randomized to one of these drugs, either trabectedin or dacarbazine, did not have the option to cross over to the other arm if their disease progressed. Therefore, the patients who were on the dacarbazine arm truly needed another therapy quicker than the trabectedin-treated patients. That could have also easily confounded the OS results.When we designed that study we, along with the FDA, decided that the goal had to be OS, knowing that it is a difficult endpoint. Obviously, OS has an appeal; it is easy to measure, patients want to live longer, and we want to help patients live longer. The FDA really wanted us to test OS, likely because we have tested this drug for several years this way. When we were testing it in the 1990s, there was a lot more debate about the value of PFS.

However, OS is difficult because it measures not just what you are doing to the patient in that particular clinical trial, but also what happens to the patient after the tumor gets worse and after the patient comes off the study. It is unclear if survival is related to whatever we do to patients during the course of the clinical trial. Because of this, OS is a controversial endpoint in sarcomas.

Fortunately, the FDA Oncologic Drugs Advisory Board, who was involved with the approval for pazopanib, really settled this debate, in my opinion. By approving pazopanib based on a PFS endpoint, they made it clear that they agree with the experts that PFS is a clinically meaningful endpoint when it is of significant magnitude. Most of us feel that 3 months is a significant magnitude PFS that can get patients to the next milestone.

What was learned from this trial?

In this trial, if the PFS benefit was only 3 months for all patients, I do not think we would be as excited about it as we are. The reason most people are so excited about this is because of that tail on the curve—the group of patients, perhaps almost as high as 20%, who can have really durable survival that is measured in years. This controlled clinical trial supports that this is not just lucky patient selection, because we see a much higher percentage of patients who were treated with trabectedin and had that durable disease control, than we saw with dacarbazine.This study basically confirmed the safety data that has already been seen for 15 years with trabectedin. The interesting thing about this drug is that it is a continuous infusion, an IV drip, which gives the patient small amounts of trabectedin over 24 hours. This is a tried and true way of delivering the trabectedin, and we already went through the learning curve of understanding that back in the early 1990s.

Generally speaking, the patients tolerated both drugs as expected. With trabectedin, the side effects were very similar to dacarbazine, such as low blood count, possible risk of infection, anemia, and fatigue. These are standard side effects for chemotherapy. There is one unique side effect for trabectedin, which is that it can irritate the liver and the liver can release some enzymes into the blood stream about 5 to 10 days after trabectedin is given. Interestingly, that tends to be when patients have some fatigue, but then that spontaneously resolves and it does not get worse with subsequent cycles.

The protocol also had a lot of safety aspects built into it. Therefore, if a patient had very high liver function elevation, their dose would be modified to keep the drug tolerable. One of the most unique aspects of trabectedin is that patients can get many cycles without any cumulative damage to the liver. We have a patient on cycle 62 of trabectedin, so there is a longstanding benefit.

Patel S, von Mehren M, Reed D, et al. Final overall survival (OS) analysis of the randomized phase 3 study of trabectedin (T) or dacarbazine (D) for the treatment of patients (pts) with advanced leiomyosarcoma (LMS) or liposarcoma (LPS). Presented at: 2015 European Cancer Congress; September 25-29; Vienna, Austria. Abstract 3403.