publication date: Jul. 19, 2019
Conversation with The Cancer Letter
UToronto researchers: Women with cervical cancer had poorer outcomes with minimally invasive surgery
Maria C. Cusimano, BHSc, MD (left)
Department of Obstetrics & Gynecology,
University of Toronto
Sarah E. Ferguson, MD, FRCSC (right)
Associate professor of gynecologic oncology,
Department of Obstetrics and Gynecology,
University of Toronto
A retrospective study of nearly 1,000 women who underwent surgery for early-stage cervical cancer showed that minimally invasive radical hysterectomy was associated with double the rate of death and recurrence—at a population level over 10 years in Ontario, Canada.
The findings of the study, conducted by gynecologic oncologists at the University of Toronto, were similar to two high-impact studies by Ramirez et al. and Melamed et al. published in The New England Journal of Medicine in October 2018 (The Cancer Letter, Nov. 2, 2018).
“What’s important here is that there is the trend in all of them is a significant increase in death and recurrence for MIS compared to the open approach,” said Sarah E. Ferguson, an associate professor of gynecologic oncology at the Department of Obstetrics and Gynecology at the University of Toronto, and director of the Research Division of Gynecology at the Princess Margaret Cancer Center, a member institution of University Health Network.
Many prior retrospective studies showed no statistically significant differences in survival or recurrence between the two approaches, said Maria C. Cusimano, a resident physician at the Department of Obstetrics & Gynecology at the University of Toronto, and a
PhD student in clinical epidemiology at the Institute of Health Policy, Management, and Evaluation.
“Our study was done across the whole population in a single province, and we used complete high-quality population-based databases to accurately ascertain death and other health outcomes,” Cusimano said.
The Ontario study, together with the 2018 NEJM papers, provide compelling evidence of harm to cervical cancer patients over many years—stemming from the introduction of new, innovative surgical approaches and techniques that were not subjected to prospective, randomized controlled trials to demonstrate long-term oncologic safety efficacy before they were adopted as the standard of care.
“Historically, most surgical procedures that have been done—appendectomy, cholecystectomy—were never tested in a randomized controlled trial to determine the best approach before they were introduced,” Ferguson said. “When we are introducing new surgical approaches—just as when we introduce new drugs—it would be ideal to get the best level of evidence in a randomized controlled trial.”
Ferguson and Cusimano spoke with Matthew Ong, a reporter with The Cancer Letter.
What are the major takeaways from your study? Is it accurate to say that patients in the minimally invasive cohort were subjected to double the risk of death and cervical cancer recurrence, compared to patients who underwent open procedures?
Patients with stage IB cervical cancer undergoing minimally invasive radical hysterectomy had about a twofold increase in the rates of death and recurrence compared to patients who underwent open procedures.
And this was very similar to a previous randomized controlled trial that you covered back in November 2018—the LACC trial by Ramirez et al.—which demonstrated these findings for the first time.
After hearing the preliminary results of the LACC trial, Maria and I set out to look at this in our population in Ontario, Canada, to see if we would find similar results. The initial study by Ramirez et al., showing a survival difference between minimally invasive and open surgery was surprising and unexpected. Many surgeons asked whether this was really happening, or really true in their environment, in their patients, in the real world.
So, we did a retrospective, population-based cohort study over a 10-year time period. We did not randomize patients, but looked at the outcomes of women who had already received one approach or the other.
It’s real-world data. We looked at recurrence as well as survival, and we controlled for confounding factors such as for surgical experience with the procedure. This was uniqu,e compared to previous work.
Was your team or your institution involved in Dr. Ramirez’s LACC study?
Yes, we were. University Health Network was involved in that trial as a recruiting center. We had patients to consent to participate in that trial.
Does your study constitute proof of principle that surgeons, especially cancer surgeons, should determine that a new robotic or minimally invasive surgical approach is not inferior to existing standard of care surgical procedures—in this case, open surgery—before widely adopting the new approach across a specialty?
That’s a good question. Historically, most surgical procedures that have been done—appendectomy, cholecystectomy—were never tested in a randomized controlled trial to determine the best approach before they were introduced. Yet, they were adopted, and the outcomes were good.
But certainly, when we are introducing new surgical approaches—just as when we introduce new drugs—it would be ideal to get the best level of evidence in a randomized controlled trial.
But these can be challenging, especially for diseases like cervical cancer—because at least in North America, it’s uncommon.
So, it is a balance with feasibility. Ramirez et al. should be congratulated, because it was a hard study to perform. It’s given us information that’s critically important.
Could you briefly describe your methodology and how your study design was adequately powered?
MC: Our study is one of the largest published on this topic.
In fact, it’s the second largest, after the National Cancer Database study that was published in The New England Journal in Medicine, alongside the LACC trial.
We were powered to see differences between the minimally invasive and open approaches in patients with stage IB disease—those patients composed over half of our cohort.
But we were, potentially, underpowered to see statistically significant associations in patients with stage IA disease and adverse pathologic features.
And radical hysterectomies are not indicated for patients with cervical cancer beyond stage I?
Radical hysterectomy is the surgical procedure for early-stage cervical cancer. Patients with later-stage disease are not usually treated with surgery, but with radiation.
In our study, we looked at patients with early-stage disease who were treated with radical hysterectomy.
I’ve been previously informed that other retrospective studies that have been done did not conclude that oncologic outcomes stemming from minimally invasive radical hysterectomy were inferior to open procedures. Please correct me if I’m wrong.
That is true. Many prior retrospective studies showed no statistically significant differences in survival or recurrence between the two approaches.
However, generally speaking, these studies were small, with few patients, and often few instances of recurrence or death. They were often done at a single center, or a small series of centers, and may not have controlled for key confounders between patients undergoing minimally invasive and open surgery.
Our study, in contrast, was quite large. We had sufficient power and follow-up to look at oncologic outcomes and account for confounding factors. And we looked at the level of an entire population in Ontario, Canada.
When individual institutions do a retrospective study, their cohorts may be highly selected and many patients may be lost to follow-up.
Our study was done across the whole population in a single province, and we used complete high-quality population-based databases to accurately ascertain death and other health outcomes.
I have perhaps an intentional apples-vs.-oranges question: you found about double the risk of death, whereas Ramirez et al. found about a fourfold risk of death on the minimally invasive arm. The bottom line is none of this is good, but is this difference important to talk about and to think about?
I think what’s important in all these studies is the direction. It is in a sense comparing apples to oranges with regards to the methodology of these studies, but what’s important here is that the trend in all of them is a significant increase in death and recurrence for MIS, compared to the open approach.
We actually discuss this in our paper, as surgeons have questioned the differences in estimates between the Ramirez study, the Melamed study, and our study.
On one hand, there may be residual confounding in the observational studies that resulted in a different estimate. On the other hand, the Ramirez trial was stopped early due to evidence of harm.
This was the appropriate decision, but such trials are known to potentially overestimate the treatment effect at the moment they are stopped. In fact, the updated LACC trial data shows that the hazard ratios decreased slightly with an additional year of follow-up.
But as Dr. Ferguson said, regardless of what the numbers are, it’s clear that there is evidence of harm associated with MIS in all of these studies. There are reasons why each study has slightly different estimates, but the key takeaway is that they all show poorer cervical cancer outcomes with minimally invasive radical hysterectomy.
Could you describe why controlling for surgeon volume is important to your study design and the overall discussion on this matter?
SF: The immediate response to the LACC trial was, “Well, perhaps there were surgeons who weren’t experienced in minimally invasive surgery in the trial.” We know that’s not true—the LACC trial had very experienced centers.
Our study looks at this in another way. We showed that the association between MIS and poorer cervical cancer outcomes remains even after controlling for surgeon volume. For us, it was trying to investigate some of the factors that could be responsible for the difference between approaches.
This is probably speculative at this point, but what are the underlying mechanisms that are contributing to these significantly higher death and recurrence rates in the minimally invasive cohort?
It’s all speculation at this point. Some of the mechanisms that have been proposed are spillage of the tumor while manipulating the specimen or insufflating the abdomen, but it’s really unclear.
Surgeons have talked about the size of the tumor, and prior studies have compared outcomes for tumors less than 2cm to tumors more than 2cm. In contrast to that, we speculate that perhaps it could be related to whether you have a macroscopic or a microscopic tumor.
Stage IB includes everything from a very early, microscopic tumor, up to a large 4cm tumor. Perhaps spillage is more likely to occur in patients with macroscopic tumors, rather than in patients with microscopic tumors. We do not know that answer—it’s a potential theory.
We’d like to look at patient-level pathologic data in this cohort if we can, to be able to study the impact of tumor size. That’s a bigger, more complicated next step, but it would be important, as there has been quite a bit of discussion among many experts about that.
Along those lines, our study did find a statistically significant association between MIS and poorer oncologic outcomes in patients with stage IB disease, but not patients with stage IA disease. However, it’s important to note that our study was likely underpowered in the IA subgroup—so those findings are really hypothesis-generating in this area.
Exactly. Even though it hasn’t been shown, to me there is some biologic plausibility behind it. It seems unlikely that patients with microscopic tumors and no residual disease after an excisional procedure would be exposed to harm through an MIS procedure.
But we don’t know, and we have to evaluate that. Until we get more answers, we do not feel that minimally invasive radical hysterectomy is safe.
If there is spillage, is there an effort to prevent exposure of the tumor to surrounding tissue? And the second question is, what does size and stage have to do with that and the minimally invasive procedures?
Oncologists always try to prevent spillage, but it’s hard to know what exactly surgeons in these studies have done. The gynecologic oncology community has talked about potentially creating very specific, systematic procedures to prevent spillage in the future.
Even if that was done, I don’t think we would just start following such a procedure unless it was in the context of another trial.
I forgot to ask, did you have data on whether the recurrences were mostly in the pelvic cavity, or were they also in the abdominal cavity?
We couldn’t evaluate that with the data that we have. So, we used patients’ treatment patterns as a marker of recurrence.
Got it. So, what are the implications of your study for clinical practice and the ongoing debate on the role of minimally invasive surgery in cervical cancer?
I think every surgeon needs to interpret the data and discuss that with their patients. The data and the risks need to be reviewed and surgeons need to have an informed discussion with their patients.
Where I work, we are now performing open surgery for radical hysterectomies for cervical cancer. Every institution and every surgeon has to look at the data and make a decision.
Did the University Health Network make the decision to switch to only open procedures after the LACC trial last October, or just recently?
We switched to open procedures after the results of the LACC randomized study came out.
Did we miss anything?
These findings specifically apply to women with early-stage cervical cancer undergoing a radical hysterectomy. This does not apply to women undergoing hysterectomy for endometrial cancer or women undergoing hysterectomy for non-cancerous gynecological conditions.
We know that minimally invasive hysterectomy with en bloc specimen removal is safe in those circumstances. This really applies to early-stage cervical cancer and radical hysterectomy.
Yes. There has been some confusion amongst patients and non-gynecologic oncologists, so we just want to make it clear that this doesn’t apply to women having other types of surgery.