「Expert connection」Prof. Jia Wang and Prof. Xiaoyun Mao and Prof. Seema A. Khan Online discussion of ECOG-ACrin 2108 research
Dr Kahn: We did not distinguish between stable disease and responsive disease, but many women had responsive disease. As you know, the treatment for breast cancer now is associated with response rates in the 70-80% range, so，of the women who were randomized, the majority would have responsive disease. We did have a good proportion of women with HER2-positive disease as well, and HER2-positive disease responds very well to HER2-directed therapy, and these women did receive HER2-directed therapy. At the moment, we don’t have any suggestion that women who responded well to treatment did better with locoregional treatment than women who didn’t respond well. In our subset analyses, we found that women with HER2-positive disease also had no survival benefit from the use of locoregional treatment, and women with hormone receptor-positive disease similarly saw no survival benefit with the use of locoregional treatment. Among women with triple-negative breast cancer, we found that women who received continued systemic therapy who did not receive locoregional treatment seemed to better. But that was a very small number of women, so it is hard to draw any conclusions from that. But among women who generally demonstrate a response to locoregional treatment, whether HER2-positive or hormone receptor-positive, they did not experience and benefit from the locoregional treatment.
Prof.Wang： Comparing with the other two prospective studies MF 07-01 and TBCRC 013, The design of this study is more similar to TBCRC013. The screening effective patients (respone patients) for locoregional treatment after systemic treatment negative results were also obtained, while in MF07-01, when we do surgical treatment firstly, then followed by comprehensive treatment, obtained the result that surgical treatment could improve the prognosis. How do we understand this difference.
Dr Kahn: It’s a good question. The difference is actually hard to understand, because I think we would all agree that the reason women die with metastatic breast cancer is due to distant disease, not local disease. For the great majority of women, the cause of death is distant disease. In a patient whose distant disease is not responding, the idea that locoregional treatment would provide better benefit is contrary to the way we think about breast cancer biology today. It is a little counterintuitive that studies that use systemic therapy first don’t show a benefit, whereas the study that used locoregional treatment first shows a benefit. I don’t know that we are going to have that question answered for us, unless our Chinese colleagues decide to repeat MF 07-01.
If you decide to repeat the study that was done in Turkey and get similar results, that will tell us something new. But in the absence of another study replicating the results of the Turkish trial, I think we are left with a little bit of a quandary, because biologically, it is not rational that the use of locoregional treatment in women who may be non-responsive to treatment would provide a benefit. The problem with randomizing at the beginning when women have not received any systemic therapy, means we don’t know whether systemic therapy is responsive or not. The other thing to remember is that these are all relatively small trials. The way we overcome bias in randomized trials is with large numbers. Unfortunately, we had to cut back on the sample size we had originally planned to recruit 880 women, eventually recruiting 390. But the results are so null, that I think if we had actually enrolled 880 women, I doubt that would have made a difference. There is not even a hint of a benefit to suggest that a larger number would have resulted in statistically significant results.
I think that contradiction between the Tata Memorial study, E2108 on the one side, and the Turkish Federation study on the other side, is a difference that is puzzling. I would challenge our colleagues in China to try to replicate that experience. Although, I have to say that there are questions that have been raised about the design of the Turkish study and the randomization between the two arms. If you remember, there was some imbalance between arms with women who received surgery having better prognostic features than women who didn’t receive surgery at the beginning. We need to break that tie. Our Japanese colleagues have a study that will mature in two years – JCOG1015. The Japanese trial will hopefully provide additional information. Their design is similar to the E2108 design. They are using systemic therapy first and randomizing responders. I think for most of us who looked at this problem and started thinking about designing trials, that seems to be the rational approach.
The other study I would mention as a reminder (although it was very small) was the Austrian study. The Austrians have done the Posytive trial published two years ago. Although they had 40-45 women per arm, their design was similar to the Turkish study. They used surgery first, so their patients did not get an initial period of systemic therapy but were randomized at enrolment to locoregional treatment or systemic therapy. Although small, the trial also pointed in the same direction as the Indian study and E2108, not showing any benefit for the use of early surgery and radiation.
Prof.Mao：A randomized trial with incorporation of exciting correlative science ideas has big potential rewards for our patients. I think ECOG-ACRIN 2108 is just changed my assumptions. Actually I can accept the selection bias of surgery for patients. Can you give us some advices about to promote the quality of life for metastatic breast cancer?
Dr Kahn: How do we promote quality of life for women with metastatic breast cancer? There are many aspects to quality of life obviously, and we always have to balance the quality of life effects of the disease (and metastatic breast cancer is a highly symptomatic disease) against the treatment effects on quality of life. That is always the balance we have to try to reach. I think that to the extent that treatment controls metastatic disease, it improves quality of life related to the disease, so the toxicity of that treatment needs to be considered. Clearly, less toxic treatment will have a lesser impact on quality of life. I am not a medical oncologist. I have to make these decisions on a daily basis. I know that in China you also manage chemotherapy, although not in the metastatic therapy if I understand correctly. Chinese breast surgical oncologists manage chemotherapy for adjuvant settings.
I think that is the challenge. Paying attention to the toxicity of treatment while achieving good responses and controlling the symptoms of metastatic breast cancer, specifically in terms of the effects of locoregional treatment on quality of life, should be the goal. The way we set up our quality of life analyses was to anticipate that local progression of the tumor would impair quality of life to some extent (and we know it does that in women with large tumors that fungating, or where there are skin nodules, or where there is pain). There are many aspects of local disease that can impair quality of life, but when applying local treatment (surgery and radiation), that also has an impact on the quality of life. So ,in our trial, we were not focusing on palliative locoregional treatment. We were focusing on locoregional treatment that was delivered in the same way it is delivered to women who do not have metastatic disease. It was also delivered to women who had no symptoms from their tumors. In that setting, where the local tumor is not causing symptoms, it seems, at least on our initial analyses, that the symptoms related to treatment also have a significant impact.
In the asymptomatic patients, adding those symptoms related to treatment (the side effects of surgery and radiation) does have an impact on quality of life. The biggest impact we observed was at 18 months. We had expected that women would have recovered from their surgery and radiation by 18 months, so the immediate short-term impact of locoregional treatment would have resolved. What we saw was that the only time point where quality of life was different in the two arms in terms of the locoregional area was at 18 months. So, we still have to explore that further and look at the details of the quality of life responses. I think these conclusions are somewhat tentative at the moment, because the number of women who completed the quality of life surveys was not ideal. There was a fall-off in the submission of quality of life surveys as the trial progressed. Fewer women completed the quality of life surveys towards the end than at the beginning. The firm conclusion that we can reach is that locoregional treatment does not improve quality of life, at least for locoregional symptoms, but whether there is a difference between arms is bit more questionable.
Prof.Wang：In my clinical experience, the responses at the primary site and of distant tumors are sometimes very different. Where distant disease is well controlled with systemic therapy, but the primary tumor is progressing, local surgery should then be considered? (22:05～22:24)
Dr Kahn: That is a reasonable strategy. Clearly, we have all seen women with metastatic breast cancer whose locoregional treatment is causing them problems. We have also seen, as demonstrated by our data and also the other two trials that have been completed, that for the majority of women who are responding to systemic therapy, they respond at both distant sites and locally. If the local tumor is ulcerated, it will often heal. If it is large, it will often shrink. Successful systemic therapy can control the local tumor as well, in terms of symptoms at least. So women who receive effective systemic therapy often don’t have symptomatic local disease.
If the distant disease is responding, the local disease also responds. There is a small group of women though, where the distant disease seems well controlled, but the local disease is progressing. I have seen that, and I am sure you have too. In that situation, offering women local treatment would be reasonable. But we must remember that women with metastatic breast cancer are seeking treatment that they think will extend survival, so when talking to women about locoregional treatment for the primary tumor, it needs to made clear that we cannot anticipate improved survival as a consequence. What it might offer them is longer-term better control if the local tumor is progressing. But if the local tumor is well controlled with systemic therapy, I don’t think there is a survival advantage in offering locoregional treatment.
Prof.Wang：Some researchers suggest that if we resect primary tumor, the judgement whether the subsequent systematic treatment is effective may be not accurate comparing with no resection. Could you give me some information about this point.
Dr Kahn: The primary tumor is an indicator of response, so it is a site that we can follow to assess response. You are asking whether if the primary tumor is removed, can we assess response as well if it were not resected? It is true that a responding primary tumor does tell us that a treatment is working, and, in general, the response at the primary site corresponds to the response at distant sites. But as you have said, that is not always true. When we are treating women for metastatic disease, we can follow the intact primary tumor for response, but we also generally follow the distant disease response. The responding distant disease has to be assessed separately from the responding primary tumor. They often go together, but they can be different. Each disease site, distant and primary, needs to be assessed. The primary tumor response will often be concordant with the distant response, but not in every patient.
Prof. Wang: Is following the response to therapy for 4-8 months sufficient?
Dr Kahn: That is a good question. We actually struggled with how long to define that interval of initial treatment. If we let it go too long, then we are not using local therapy early in the course of disease. One of the theoretical premises that the trial was built on was the idea of reseeding – that the primary tumor site was a place where circulating tumor cells could recirculate through the tumor and become more competent to set up new sites. If we leave the primary tumor alone for a long time, then we may not see the benefit of resecting the primary tumor. If we don’t treat long enough, particularly with hormone receptor-positive disease, it can take a while to see responses. The shorter end of the limit of four months was based on the idea that most chemotherapy regimens need to run for four months in order to declare a response or not. For example, sequential-ACT goes for four months. Most of the medical oncologists in the group felt that four months was a good minimum interval. For the maximum interval, we based on endocrine therapy, because we don’t often see a good response to endocrine therapy in less than six months, and we wanted to allow some flexibility. That is how we set that interval of 4-8 months.
But the intent of the trial was really to assess the value of early local therapy, so we wanted to make it open-ended， so ，patients could be treated for as long as wanted and then be randomized. We needed the interval to make the population relatively homogeneous in terms of treatment duration. That is how we came up with the 4-8 month interval. In practice, however, obviously systemic treatment goes on longer than eight months. It goes as long as necessary. So in practice, that 4-8 month interval is not as important, particularly seeing as our results did not show a value for early local therapy. If we follow the results of the trial, then we won’t have to make these decisions of when to offer local therapy. The trial shows that early local therapy is not beneficial in women whose tumors are asymptomatic.
Prof.Mao：Your answers have changed some of my viewpoints on local therapy for metastatic breast cancer. I have a final question. I noticed that in your results, FACT-B TOI was significantly lower in patients receiving early local therapy at 18 months post randomization. How we understand this problem. Maybe you can give us more details about it. Thank you so much.
Dr Kahn: Unfortunately, we don’t have more details at the moment. We are looking more closely at the quality of life data to try and understand why that 18-month timepoint was different. As I mentioned earlier, the response rate declines during the course of the study. We had fewer women responding with their quality of life surveys, so the information is not as robust as we would have liked. What we had hypothesized was that the side effects of surgery and radiation would have gone away by that timepoint 18 months after randomization. It is possible that in some women, the surgery and radiation was a little delayed, so by 18 months, the acute effects had not subsided. We did, of course, see an improvement in local control. Women who had locoregional treatment were less likely to experience progression at the local site. Despite the lower rate of local progression, we did not see improved quality of life in terms of local symptoms. This may just be a reflection of the tail-end of the acute effects of surgery and radiation. We do need to look at that data more closely, and we hope to be able to report those findings in greater detail in our manuscript.