Study strengths and weaknesses
The authors deserve to be commended for this well-designed and laborious trial, which is the first of its kind. The strength of the study is its randomized design producing level I data.
Study weaknesses include lack of reporting of BRCA status and the impact of receiving targeted therapies after the trial was over. It is well established that BRCA-mutated patients have an independent survival advantage, even when taking into account platinum sensitivity.6-8BRCA status of the study population is not specifically addressed in this paper. The authors noted in the first GOG 0213 trial publication, which assessed bevacizumab in the recurrent setting, that BRCA status has an impact on patient outcomes. Subsequently, they state that they do not report BRCA status because “…its independent effect on response to an anti-angiogenesis agent was unknown,” but it clearly would affect survival analysis if unbalanced between groups.9
Similarly, in the introduction to their study, Coleman and colleagues list availability of maintenance therapy, for instance poly ADP (adenosine diphosphate–ribose) polymerase (PARP) inhibitors, as rationale for conducting their trial. They subsequently cite this as a possible reason that the median overall survival was 3 times longer than expected. However, they provide no data on which patients received maintenance therapy, which again could have drastically affected survival outcomes.10 They do report in the supplementary information that, when stratifying those receiving bevacizumab adjuvantly during the trial, the median OS was comparable between the surgical and nonsurgical groups (58.5 months vs 61.7 months).
The authors discuss the presence of patient selection bias as a weakness in the study. Selection bias is evident in this trial (as in many surgical trials) because patients with a limited volume of disease were selected to participate over those with large-volume disease. It is reasonable to conclude that this study is likely selecting patients with less aggressive tumor biology, not only evident by low-volume disease at recurrence but also by the 20.4-month median platinum-free interval in the surgical group, which certainly affects the trial’s validity. Despite being considered PS, the disease biology in a patient with a platinum-free interval of 20.4 months is surely different from the disease biology in a patient with a 6.4-month platinum-free interval; therefore, it is difficult to generalize these data to all PS recurrent ovarian cancer patients. Similarly, other research has suggested strict selection criteria, which was not apparent in this study’s methodology.11 While the number of metastatic sites were relatively equal between the surgery and no surgery groups, there were more patients in the surgical group with extraabdominal disease, which the authors used as an exclusion criterion.
Lastly, the time to treatment commencement in each arm, which was 40 days for the surgical arm and 7 days in the nonsurgical arm, could represent a flaw in this trial. While we expect a difference in duration to account for recovery time, many centers start chemotherapy as soon as 21 days after surgery, which is almost half of the median interval in the surgical group in this trial. While the authors address this by stating that they completed a landmark analysis, no data or information about what time points they used for the analysis are provided. They simply report an interquartile range of 28 to 51 days. It is hard to know what effect this may have had on the outcome.
This is the first randomized clinical trial conducted to assess whether secondary surgical cytoreduction is beneficial in PS recurrent ovarian cancer patients. It provides compelling evidence to critically evaluate whether surgical cytoreduction is appropriate in a similar patient population. However, we would recommend using caution applying these data to patients who have different platinum-free intervals or low-volume disease limited to the pelvis.
The trial is not without flaws, as the authors point out in their discussion, but currently, it is the best evidence afforded to gynecologic oncologists. There are multiple trials currently ongoing, including DESTOP-III, which had similar PFS results as GOG 0213. If consensus is reached with these 2 trials, we believe that secondary cytoreduction will be utilized far less often in patients with recurrent ovarian cancer and a long platinum-free interval, thereby changing the current treatment paradigm for these patients.
MICHAEL D. TOBONI, MD, MPH, AND DAVID G. MUTCH, MD