Shortcomings in SIRFLOX Design May Have Skewed PFS Findings

Navesh K. Sharma, DO, PhD

The addition of Yttrium-90 (Y-90) resin microspheres (SIR-Spheres) to standard frontline FOLFOX-based chemotherapy with or without bevacizumab significantly improved liver-specific progression-free survival (PFS) for patients with liver metastatic colorectal cancer (CRC); however, PFS at any site, the primary endpoint of the study, was not changed with the addition of liver-directed therapy, according to data from the phase III SIRFLOX study published in theJournal of Clinical Oncology(JCO).¹

PFS at any site was conceived as a proxy for a cure; however, the minor advantage seen in those who received selective internal radiation therapy (SIRT) along with chemotherapy was too small to be statistically significant. In the SIRT arm, patients experienced a median PFS of 10.7 months compared with 10.2 months in those who received chemotherapy alone (HR, 0.93; 95% CI, 0.77-1.12;P= .43).

However, shortcomings in the study may have confounded the main findings, as the trial seemed plagued by a large number of confounding factors and a shifting trial design. Ultimately, only 84% of patients allocated to receive SIRT actually got therapy that aligned with the protocol, including 7.7% of patients (n = 19) with bilobar disease who only received SIRT in one lobe of the liver and 7.9% of patients (n = 21) who were randomized to combination therapy but did not receive SIRT at all. Additionally, the percentage of patients with an intact primary tumor was unexpectedly high, at around 45%, which may have been impacted by the international design of the study.

Despite these shortcomings, there was, however, a substantial benefit seen for PFS in the liver, and this increase was highly statistically significant. Those in the arm that received SIRT plus chemotherapy enjoyed a median liver-specific PFS of 20.5 months compared with 12.6 months for those who received chemotherapy alone (HR, 0.69; 95% CI, 0.55-0.90;P= .002)

The SIRFLOX study sought to validate earlier findings from 2001 that showed an improvement in overall survival (OS) with SIRT plus floxuridine. In SIRFLOX, 530 patients were enrolled across the globe between October 2006 and April 2013. All patients were previously untreated for metastatic CRC, and were considered to have unresectable liver metastases when the study was initiated.

In addition to missing the primary endpoint, the study also did not show a significant increase in the number of liver resections. Of the patients in the arm that received SIRT, 38 (14.2%) went on to have a liver resection compared with 36 patients (13.7%) of those in the control arm.

SIRFLOX is the first of three randomized controlled trials in a preplanned combined analysis of SIRT’s impact on OS. SIRFLOX, FOXFIRE, and FOXFIRE Global have accrued a total of 1103 patients. Combined analysis of OS data from these three trials will be examined in 2017.

To gain further insight into the findings,Targeted Oncologyspoke with Navesh K. Sharma, DO, PhD, US lead investigator of the SIRFLOX trial and one of the coauthors on the JCO paper. Sharma is the section chief of radiation oncology and associate professor at Penn State Health St Joseph Medical Center in Reading, Pennsylvania.

TARGETED ONCOLOGY:What are the main findings from the study?

SHARMA:

The biggest message from the study is that SIRT can be given safely, because patients in both arms of the study were able to get the same chemotherapy. A lot of chemotherapy drugs are processed through the liver, and traditionally people were concerned that SIRT would damage the liver enough that chemotherapy would not be possible.

TARGETED ONCOLOGY:What are your thoughts on PFS at all sites as a primary endpoint?

SHARMA:

I was not involved in the design of this study, but that endpoint in some ways was doomed to fail from the start. Development of new lung or bone metastasis was a failure of the endpoint, but it’s not surprising that people failed elsewhere than the liver. Technically, it was a failure of the systemic chemotherapy.

TARGETED ONCOLOGY:Could the chemotherapy used in this trial have been a confounding factor?

SHARMA:

In a study described in 2001 in theAnnals of Oncology,²they were not using the chemotherapy regimen that is currently used now. No oxaliplatin was involved. In general, the OS for stage 4 disease has increased from 6 months to 2 years, which is what we normally quote patients today. It’s a huge jump.

Even more recently, the original plan for the SIRFOX study didn’t have bevacizumab, because it was not standard of care when the study was planned, so that was added on later. Which patients are ideal for SIRT? Generally, my paradigm in treating patients with SIRT is that they have to be patients with liver-predominant metastases. However, if they already have significant disease progression elsewhere, studies have shown that even in the salvage setting, SIRT helps.

TARGETED ONCOLOGY:What is the importance of liver-specific PFS?

SHARMA:

Usually the liver is the first site of metastases, and it also limits how much chemotherapy they can get. If you think about people with colorectal cancer, it’s not the colorectal cancer or the lung metastases that kill them; lots of times it’s the liver.

TARGETED ONCOLOGY:How have surgical approaches to this disease become more aggressive in recent years?

SHARMA:

Surgeons have more laparoscopic and robotic-assisted surgical processes available. The data shows that controlling the liver metastases early gives patients a much longer survival time, especially if the liver metastases can be removed, because when the liver starts failing, you have to reduce the dose of chemotherapy.

One of the secondary endpoints on the study was to have more patients go to resection. But one of the reasons it didn’t make it was that a lot of the surgeons had some discomfort, although there is more literature coming out that says it is safe to do liver resection following chemotherapy and radiation. The threshold for statistical significance for unplanned exploratory analyses was set as aPvalue of 0.01.

TARGETED ONCOLOGY:Why was this so stringent?

SHARMA:

That was one of the robust aspects of study: the statistics were very robust. When you think of how many countries and institutions were involved, you have to have very stringent mechanisms when you go into it. We were looking at surgeon preference over four continents. Some were in smaller community centers, and not everybody had a dedicated tumor board, so it was not always possible to have a surgeon who was willing to operate afterward.

In the past, it’s been found that different countries have different levels of resection. We’ll be looking at some of the numbers again over the next few meetings, such as for resections, and we might be surprised in England.

1. van Hazel AG, Heinemann V, SHarma NK, et al. SIRFLOX: randomized phase III trial comparing first-line mFOLFOX6 (plus or minus bevacizumab) versus mFOLFOX6 (plus or minus bevacizumab) plus selective internal radiation therapy in patients with metastatic colorectal cancer [published online February 22, 2016].J Clin Oncol. doi: 10.1200/JCO.2015.66.1181.