Peritoneal RFS May Be a Stronger Predictor of OS in CRC Peritoneal Metastasis

Muhammad Talha Waheed, MBBS

In a comparative analysis of recurrence-free survival (RFS) and peritoneal recurrence-free survival (P-RFS) vs overall survival (OS) after curative-intent cytoreduction, investigators found that both RFS and P-RFS can be used as surrogate end points for OS in efficacy trials evaluating patients with colorectal cancer peritoneal metastasis.¹

The retrospective study was from the Peritoneal Surface Oncology Alliance for Research and Innovative Solutions (PSOLARIS) group which included 2269 patients who had undergone complete cytoreduction. investigated whether RFS and P-RFS could predict overall survival OS in patients with colorectal cancer peritoneal metastasis. Investigators sought to determine whether RFS and P-RFS could predict OS in patients with colorectal cancer peritoneal metastasis.

Data showed that the median OS was 3.7 years (95% CI, 3.5-3.9), with over 70% of patients experiencing recurrence of their cancer, most of which were peritoneal recurrences (61.8%). The key finding of the study showed that while RFS and P-RFS were moderately linked to OS, P-RFS appears to be the better surrogate for OS.

“We saw that peritoneal recurrence-free survival was doing better compared with recurrence-free survival and can potentially be preferred over recurrence-free survival as a surrogate end point,” explained Muhammad Talha Waheed, MBBS. “[A]lthough we do lack randomized studies to evaluate these surrogate end points, these conclusions remain consistent when we look at our simulated analysis.”

In an interview with Targeted Oncology^TM, Waheed, postdoctoral research fellow at City of Hope National Center, Duarte, CA, discussed research on the reliability of using RFS as an efficacy end point for trials evaluating patients with colorectal cancer peritoneal metastasis.

3D medical illustration of colorectal cancer: © anatomy insider - stock.adobe.com

Targeted Oncology: Can you provide an overview of the PSOLARIS study and its objectives in investigating colorectal cancer peritoneal metastasis?

Waheed: The title of our study is reliability of recurrence-free survival as an efficacy end point for trials of resected colorectal cancer peritoneal metastasis and results from the PSOLARIS study group. More recently, we see an increase in the use of surrogate end points compared with the traditional end points in randomized trials. For colorectal peritoneal metastasis, this means the surrogate end points are the recurrence-free survival or peritoneal recurrence-free survival. There are currently no studies that have established their liability or validity of using these surrogate end points. The main objectives of our study were to establish the reliability of recurrence-free survival or that of peritoneal recurrence-free survival as optimal surrogate end points for trials of resected peritoneal metastasis. We also wanted to compare recurrence-free survival with that of peritoneal recurrence-free survival.

This led to the creation of PSOLARIS, which stands for Peritoneal Surface Oncology Alliance for Applied Research and Innovative Solutions. It is a group of 19 centers, 11 from the [United States], 6 from France, and 1 each from Spain and Switzerland that have come together to understand and answer these questions and to have a sizable cohort to study this question.PSOLARIS itself has not done any trials yet, but we try to look at if and why the recurrence-free survival is okay to be used as an efficacy end point in future trials of colorectal peritoneal metastasis.

What specific methodologies were employed by the PSOLARIS study group to assess the reliability of recurrence-free survival as an end point?

We initially focused on patients who were undergoing complete cytoreductive surgery for colorectal peritoneal metastasis with complete follow-up information. We initially calculated their survival times, including overall survival, recurrence-free survival, and peritoneal recurrence-free survival. Initially, we used scatter plots to look at the correlation between the surrogate end points and used Spearman rank correlations to assess the differences. To further examine whether time to recurrence could affect subsequent survival, we plotted Kaplan-Meier curves stratified by recurrence-free survival intervals to look at the survival patterns after recurrence. We did not really stop there. We went on to do our event resampling, or a simulation analysis of our patient level data to look at the correlations of hazard ratios for recurrence-free survival and the effect of those hazard ratios for overall survival.

Were there any challenges encountered in measuring recurrence-free survival in this context, and if so, how were they addressed?

To establish the reliability of surrogate end points, it is important to follow the benchmark that is set. There are a number of guidelines that suggest that even though patient level correlation is important and necessary, it is still not efficient or adequate. For robust validation, we need to validate this reliability in a meta-analysis of randomized studies. While there are a lot of trials for other disease sites, when it comes to colorectal peritoneal metastasis, there is a lack of randomized trials. It was hard for us to assess this using meta-analysis of those trials.

We did a patient level correlation initially, but to mitigate that lack of trials, what we did was a simulation analysis using our patient level data. What we essentially did was we mimicked the trial using our patient level data. We asked ourselves the question, let's say if a randomized study were to demonstrate an improvement in recurrence-free survival, would that be associated with an improved overall survival or not? What we did was we took 100 random patients from our cohort and then we took [another] 100 random patients from the cohort, and we called them as arm 1 and arm 2. Then, we calculated the hazard ratios for the surrogate end points, which were recurrence-free survival and peritoneal recurrence-free survival.

We also calculated the hazard ratios for overall survival, and then we repeated this process 100,000 times. Then, we plotted around 20 data points per point through to increment of hazard ratio for recurrence-free survival. We looked at the correlation of hazard ratio of recurrence-free survival, or that of peritoneal recurrence-free survival with overall survival. We really tried to mimic what we would see in a trial and looked at that correlation. We were able to find a high correlation of the surrogate end points with overall survival in that simulation analysis.

Do any findings regarding recurrence-free survival contribute to the understanding of the efficacy of treatments for colorectal cancer peritoneal metastasis?

While we did not look at efficacy of any treatments in the study, because all of our patients were getting the same resection, 1 of our findings suggested that if any treatment were to improve the recurrence-free survival, especially beyond 2 years, it may be associated with a better overall survival after recurrence when compared with recurrence-free survival of less than 2 years. So for patients who were regarded early within half a year or 1 year or 1 and a half years, they were not going to do better. They had lower overall survival compared with patients who are living without any recurrence for 2 years after surgery.

Were there any unexpected results or insights gained from assessing recurrence-free survival as an end point in these trials?

One interesting finding is that we found that a statistically significant recurrence-free survival, which means that an improvement in recurrence-free survival, especially in our ssimulated analysis, may not always be associated with a statistically significant overall survival. However, as the size of the effect in hazard ratio of recurrence-free survival increases, the likelihood of finding a significant overall survival also increases. What that means is that if a treatment is to decrease that recurrence-free survival by a good margin, then the statistical likelihood of finding a significant overall survival also increases.

What are the key takeaways from the PSOLARIS study group's findings regarding recurrence-free survival in this patient population?

I would like to give 3 main takeaways from our study, and that would be that recurrence-free survival and peritoneal recurrence-free survival both correlate relatively well with the clinical end point for overall survival, and they support the use of these end points in future trials of colorectal cancer peritoneal metastasis. Further, we saw that peritoneal recurrence-free survival was doing better compared with recurrence-free survival and can potentially be preferred over recurrence-free survival as a surrogate end point. And lastly, although we do lack randomized studies to evaluate these surrogate end points, these conclusions remain consistent when we look at our simulated analysis. These are the 3 main takeaways from this study that could be further explored.

What are the next steps?

We would like to see this implemented and also explored and evaluated by other groups, but more importantly, the lack of randomized trials in this field [is] the root cause of many unanswered questions. We need to come together, join forces, and produce quality randomized studies to answer those questions. With these data that we have, we are trying to answer some other big questions, including the use of adjuvant chemotherapy in colorectal burden peritoneal metastasis, and we will probably see that in future studies.

REFERENCE:

Waheed MT, Kepenekian V, Sourrouille I, et al. Reliability of recurrence-free survival as an efficacy endpoint for trials of resected colorectal cancer peritoneal metastasis: results from the PSOLARIS study group. Abstract 68.