Steve Dunn’s Review of Surgery for Metastatic Renal Cell Cancer
When I was first diagnosed in 1989, I learned from my doctors that a solitary distant metastasis could be treated with surgery, with occasional long survivals. Review articles I read suggested that few patients have a solitary recurrence, and that surgery is most likely to be beneficial if it’s been at least a year or two since nephrectomy. It sounded like surgery benefited only a tiny proportion of patients with metastatic renal cell cancer. I took that to be well settled and didn’t look further. But in late 2001, Mike Fisher, a major supporter of CancerGuide and a member of the KIDNEY-ONC e-mail list, alerted me to something I had missed. Over the years, silently, and without fanfare, much encouraging data on the results of surgery has accumulated! My review of the major papers suggests both that the benefit for surgery is on a par with that for the best known immunotherapy, and that far more patients can benefit than is generally realized. Surgery is an important option for a significant number patients with metastatic renal cell cancer!
First, I present my own conclusions and recommendations, based on my review of eight major papers. Following my recommendations, I review my selection of papers and their methodology. Finally, I present summary tables for easy comparison of the papers, followed by the abstract of each of these papers along with my commentary. My analysis has benefited greatly from discussions with Mike Fischer, whose insightful comments have been a major influence.
Conclusions and Recommendations
How Many Patients Might be Candidates for Surgery?
Because of the retrospective nature of the data it is only possible to hazard a rough guess, but it appears that more patients might benefit than is generally realized. In the papers which included a comprehensive review of all patients treated for metastatic RCC at a given institution (or set of institutions) roughly 10% of patients had a solitary metastasis. In the papers where more than one metastasis was treated, roughly half of the patients had a solitary metastasis. Therefore, 20% is a reasonable first rough estimate. I’m expect additional patients might have been considered to be surgical candidates if their doctors had known of the benefits of surgery, but there’s no way to know how many this might be. The effects of modern imaging technology on determining who is really operable could also affect the percentage. Despite these uncertainties, that a significant percentage of metastatic kidney cancer patients can benefit from surgery should prompt more research into this area, especially since the boundaries of what is possible are only beginning to be explored. None of this is directly relevant to any one patient, since you are either operable or not depending on the particular circumstances of your case, but knowing the odds you’re operable are better than you might expect should motivate you to look into it.
When Is Surgery A Possible Option?
- All Known Metastases Must be Believed Operable: Because the evidence is that surgery is only beneficial if they can get it all, the surgeon must believe that all of the known metastases are operable. You should be aware that in the two largest studies, 30-40% of the time they didn’t get it all. I certainly suggest you discuss with your doctors what can be done before surgery to minimize the chance of surprises during surgery.
- There Are Only a Handful of Metastases (Give or take a few!): The experience in these papers goes out to roughly a handful of metastases (around five, with most patients having fewer) and the bulk of the evidence is that it doesn’t matter if there is just one or a few. At the same time, there is no data on extending the technique more than just a few metastases. The data we have suggests we haven’t reached the limit, but reason suggests that if there are many you can detect, there are probably many more that are still too small to see so that systemic treatment is needed. Since the boundary is not known, It might be reasonable to push the known limits a little especially considering the odds with current systemic therapy.
- More Metastases All Confined to One Organ: The largest paper I reviewed, Kavolius, didn’t count the number of individual tumors but rather the number of different sites of metastasis, where each organ involved counts as a different site. Patients with metastasis to one site did better than those with metastasis to multiple sites, though multiple sites was not a bar to long term survival. At any rate, if you have only one site of metastasis but there are somewhat more than is usual to resect, the question is whether surgery is appropriate despite the absolute number of tumors. We don’t know how many tumors those with only one site may have had in that organ, but because the overall limit of benefit for surgery versus number of tumors resected was not reached in the other papers reviewed, my best guess is that surgery would be beneficial in this case.
- Benefit of Surgery is Not Confined to Lung Metastasis: The prognosis may vary somewhat according to the site of disease, and disease confined to the lung probably has one of the better prognoses, but this doesn’t mean that there is no benefit to surgery in other locations. It looks like bone metastasis, and particularly brain metastasis, may have a noticeably worse prognosis with surgery, but then bone metastasis also has a smaller chance of response to IL-2 or Interferon compared to lung metastasis, and worse or not, brain metastasis is normally treated with surgery or stereotactic radiosurgery (another way of treating individual tumors.). In these papers there was strikingly little data on surgery for liver metastasis. The reasons for this are not clear, but I know of no reason to suspect that there wouldn’t be benefit to complete surgical resection of liver metastases, but it is a small stretch beyond the data. In fact, one of the RCC survivors featured in CancerGuide’s Story Section, Norma Downes, has had long term survival after a complex surgery for isolated liver metastasis!
- Metastasis at Initial Diagnosis is Not a Bar to Surgery! While this may contradict the conventional wisdom, in my view, the bulk of the evidence (See van der Poel and Tobisu below) is that it makes no difference if you have metastasis at initial diagnosis, or if you have a recurrence some time after surgery. The oldest paper I analyzed, O’Dea, claims to show that synchronous metastasis is an adverse prognostic factor, but I find their analysis very unconvincing.
- Short “Disease Free Interval” is Not a Bar to Surgery (But very short DFI might be!): Disease free interval (DFI) is the time between nephrectomy and recurrence. Short DFI was an adverse prognostic factor in most of these studies, but at the same time was not a bar to long term survival. The exception was Tobisu in which the small cohort of patients with DFI less than six months all died in less than three years. Although this observation might not hold-up given more patients, it suggests caution. The other papers compared patients with DFI less than a year or two to patients with longer DFI. Note that because of the reasonable results in patients with metastasis at the time of initial diagnosis, I believe that this situation represents a different biological/statistical category than recurrence after diagnosis, and should not be counted as zero DFI.
- If Needed, You Can Have Surgery More than Once (And Likely You Will!): If you relapse after surgery for metastatic disease, and the relapse is operable, another surgery is an option. Several of these papers reported on long term survivors who had two or three surgeries, and it doesn’t look like survival is worse after the second or third surgery than after the first. In one extraordinary case, a long term survivor had 12 separate operations and was doing well almost 8 years after the first surgery! In the two largest papers, in fact, a substantial majority of those who had an apparent complete resection later had a second operation for relapse. In my opinion, those who choose the surgical path should be prepared for the likelihood that their first surgery will not be their last!
When Is Surgery A Good Option?
In addition to considering the factors above separately, you might want to weigh adverse prognostic factors which are not an absolute bar to surgery. If there are several of these, you might decide that surgery isn’t as good an option, even if it is a possibility. Surgery, of course, can mean irreversible loss of function, or a hard long recovery. Often neither of these are likely to be true, but certainly these factors should weigh in your decision. Conversely, if surgery would relieve debilitating symptoms or remove a tumor which is immediately life-threatening, these are reasons to go with surgery.
Surgery First Or Immunotherapy First?
If you’re eligible for both surgery and IL-2 based immunotherapy which should you do first? In my opinion surgery first usually makes more sense. The results with both surgery and immunotherapy include a small chance of long term survival, and it’s not clear which is better in that respect. It looks to me like if you get a complete resection, surgery is likely to benefit you for at least a few years. If you get a response to immunotherapy, there is also a decent chance of a few years of benefit (and a documented chance of very long term remission). But the chance of getting a complete resection in most cases will exceed the chance of responding to immunotherapy. So I think the overall chance of gaining a real benefit is greater with surgery.
A second reason I think surgery is may be a better first option is that it’s better to have two chances at long term survival than just one! If you do immunotherapy first and it doesn’t work, as evidenced by growth of your cancer, the tumors might have reached the point of inoperability. But if you do surgery first, and unfortunately don’t have a complete resection, then it seems likely you can then try immunotherapy. In the best case, you may have less disease when you try immunotherapy than you had before surgery. I don’t have statistics on the percentage of operable patients who become inoperable after failed immunotherapy, or the percentage who are still healthy enough to undergo immunotherapy after a failed attempt at curative surgery, so my judgment that you’re more likely to be able to try both if you try surgery first is only an educated guess. Also, as I discussed above, you need to weigh the difficulties of the surgery, as well as any adverse prognostic factors, or immediate palliative benefit.
Preparation for Surgery
Because patients with an incomplete resection don’t appear to benefit from surgery, it seems logical to make a very careful search for additional metastases before surgery. If any are found that are inoperable, you will be spared unnecessary surgery. If additional operable metastases are found and the total number of metastases is still within reason, surely it’s better to get them out now so you have a better chance of getting them all with only one surgery. I think then, that a high quality CT scan of the chest, abdomen, and pelvis, as well as a bone scan are a reasonable starting point. You may also want to investigate adding post surgical adjuvant therapy to reduce the risk of relapse before surgery (see below).
Adding Adjuvant Therapy
The odds with surgery might be improved beyond the current experience if an effective systemic treatment to reduce the risk of recurrence were given after the surgery. Unfortunately, there is no proven adjuvant therapy for renal cell cancer, but there are more or less promising therapies for this purpose in clinical trials. Some of these trials accept patients with a history of metastatic disease who’ve been rendered disease free by surgery. If you’re going to have surgery, you might want to look into participating in such a trial! Some of these trials are vaccine trials and some vaccine trials require fresh tumor tissue to make the vaccine, so if you are interested in pursuing adjuvant therapy you should investigate it before surgery! Finally, since nothing is proven, I recommend favoring less toxic therapies over more toxic therapies if there is a choice.
Using Innovative Techniques to Overcome Inoperability
In the past few years, innovative techniques for treating individual tumors such as Radiofrequency Ablation, Body Stereotactic Radiosurgery, and others have started to become available. Some of these methods are even outpatient procedures. If you have a reasonable number of metastases but are judged to be inoperable for technical reasons, then it might be reasonable to investigate these techniques. At first glance, the principle that destroying all known metastases can result in long term survival might well hold even if the method of destruction isn’t surgery. There is very little long term data so this has to be somewhat speculative. There is certainly a risk that second order interactions, for example adverse effects of radiation on the immune system, could lessen the long term survival. The risk of incomplete destruction of tumor is probably not be as well known, and you lose the close inspection for additional metastasis that normally accompanies surgery, but given the general prognosis of metastatic kidney cancer stretching it a little seems justified. I’d say that especially if the other prognostic factors look good for surgery, it’s worth looking into. Just as there are technical considerations which dictate whether surgery is feasible in any given case, each of these techniques has it’s own limitations and risks. Just as you may need to consult with a surgeon to determine whether your tumors are technically operable, you may need to consult with experts in the various techniques to determine if they are appropriate in your case.
Review of Major Surgery Papers
Introduction and General Discussion
First I present a discussion of my selection of papers to review, and a discussion of the methodology and limitations of those papers. Then I present summary tables which allow easy comparison of the papers. Finally, I present the abstract for each of the eight selected papers followed by my commentary, which is based on a detailed reading of the complete paper. At the top of each commentary, I give the number of surgical patients. This is the number of patients who had surgery which was intended to remove all metastases and the affected kidney (if necessary), whether or not the surgery succeeded in rendering the patient apparently disease free. Most of these papers also report on other patients for whom curative surgery wasn't attempted, and these patients are not included in my count.
Selection of Papers for Review
There are numerous papers on surgery for metastatic RCC, but most cover only 10 to 20 patients, making it hard to draw conclusions. Fortunately, in recent years several major papers, each covering many more patients have been published. I review those papers, plus one important pioneering paper which did have a goodly number of patients. I believe it is best to focus on the largest and most recent reports.
Retrospective Study Design
All of the papers I've found on this subject are retrospective reviews. This means that rather than carefully selecting patients for surgery in advance based on pre-defined criteria, the authors reviewed the records of patients at a given hospital who were treated according to their doctor's best judgment. Although the papers explore the characteristics of the patients who got surgical treatment, and in some cases those treated at the same hospital who didn't, details of how patients were selected for surgery are not available. We generally don't get much of an idea of how careful a search was made for additional metastases, or how closely the patients were followed after surgery. Because of the retrospective nature of the studies these things were surely anything but consistent. All of this leaves a bit of uncertainty about whether selecting patients for surgery according to the characteristics which portended a successful outcome in past studies will do as well. One distinct advantage of the retrospective method is the ability to look far enough into the past that all of the cases have years of follow-up, so the long term outcomes can be determined, and in fact these studies have long follow-up times, approaching 20 years in some cases. This has its own disadvantage though.
Technology has improved dramatically over the last decades, especially the ability to detect metastases and recurrences with techniques like high resolution CT scans and MRI scans. I expect there have also been technical advances in surgery which render more tumors safely resectable and reduce the risk of complications. All of this makes it difficult to apply results from these studies to the present, although I would expect that these technological improvements would lead to improvement in the results through more accurate selection of patients who have limited metastasis, and greater ability to completely resect their tumors. The change in practice over time also increases the risks of comparing older papers to newer ones. Actually, because these papers study patients whose initial treatment varied by years, and even decades, there could have been changes in the results over time even within a single paper. Unfortunately, none of the papers attempt to analyze trends over time. Despite all of this, these papers show benefit for patients treated from the 1950's to the 1990's. I don't believe time or technology has changed the fundamental pattern. I do suggest weighting the more recent papers more heavily.
Last, but definitely not least, retrospective studies don't include a randomized control group, or even a control group of patients matched for important characteristics who didn't have surgery. Instead there is only the "implicit" historical control of comparing the results with surgery to the generally expected poor prognosis for patients with metastatic kidney cancer, or to patients at the same hospital who didn't receive surgery but who probably differed in many other ways. As I will discuss, the lack of a good control certainly adds uncertainty to the conclusions.
Primary Evidence for Benefit
There are two relatively consistent findings which constitute the main evidence of benefit:
- A markedly unusual number of long term survivors compared to what you would expect.
- Patients with complete resection live longer than those with partial resection.
Skeptical Arguments, Uncertainty, and Maximizing the Odds
There is definitely room for skepticism about both these observations given the uncontrolled nature of these studies. The gremlin is the possible existence of bias.
Patients who were considered eligible for surgery because they had only limited metastasis certainly have less disease then the average patient with metastatic kidney cancer, and might naturally be expected to live longer. Patients selected for surgery also had to be judged sufficiently fit. Again, this selects for healthier patients who would be expected to live longer than average.
Similarly, patients who had a complete resection may well have had less extensive disease on average than patients who didn't, or it may be that locations that are more favorable for complete resection are also more favorable for survival even without resection. As an example, lung metastasis generally has a better prognosis than bone metastasis, and it is my impression that surgery for complete excision of bone metastasis is also more difficult than that for lung metastasis. It seems plausible that a greater fraction of patients with lung metastasis would achieve a complete resection than patients with bone metastasis
While these arguments have real merit, there is good reason not to regard them as conclusive:
- First, five year survival in metastatic kidney cancer appears to be unusual (less than 5%), except with the best immunotherapy treatments. But in patients who got a complete resection, five year survivals were in the range 25-44%. I think this difference is large enough to suggest that selection bias alone is not likely to be the explanation. But because it seems likely that there is at least some selection bias, it's likely that bias explains part of the apparent benefit of surgery so that the actual benefit is likely to be somewhat less than the raw statistics suggest. Likewise the difference between complete and partial resection is probably at least somewhat explained by bias. This does not, however, mean that all or most of these differences are due to bias.
- Second, statistics are best not interpreted in a vacuum. It's important to consider whether it's biologically plausible that surgery makes a difference. Clearly if there actually are only a few metastases, and you remove them all, then there is no cancer left and the patient is cured. The question is whether this ever actually happens. Given that some patients live ten to twenty years after surgery for metastatic disease and are still without recurrence, it seems very plausible. Indeed, it seems hard to deny that at least some patients got a real benefit. If an apparent complete resection is not truly complete because metastases as yet too small to detect were not removed, it still seems plausible that reducing disease to the point of being undetectable might be expected extend life by something like the length of time it takes for the undetectable metastases to become detectable. In many cases when more metastases were detected, surgery was performed again, and it may be that the last metastases were removed at second or third surgeries in some cases. One of the papers I reviewed, van der Poel describes several patients who were disease free years after a second or third surgery (Table 2, p301). It may also be that that multiple surgeries give multiple prologations of life by again radically reducing the burden of disease. None of this proves that the surgery was responsible for all of the difference completely resected and incompletely resected patients. The bias argument may have some merit, but it also seems likely that some of the apparent benefit of surgery suggested by the difference between survival after complete resection versus incomplete resection is real.
- Third, as above, if surgery is actually effective, one would expect there to be a difference in survival between apparent complete resection and partial resection. Observing such a difference certainly doesn't prove surgery is effective, but if no such difference were observed in the majority of studies, one would surely conclude that apparent resectability is just a marker for an improved prognosis, and that the surgery itself makes no difference. Since observing such a difference (given that it was tested for) is a necessary condition for believing surgery is beneficial, that this difference is actually observed is evidence that the surgery is beneficial.
- Fourth, it is important to place the evidence in context. Metastatic renal cell cancer is a disease with very few reasonable treatment options. Moreover, the evidence for these options is subject to similar arguments. The best proven, high dose Interleukin-2, also produces a small number of long term disease-free survivors. But as with surgery, this observation can be challenged on the grounds that patients who can undergo high dose IL-2 are a highly selected group, and on the grounds that IL-2 has not been compared to no treatment in a randomized trial. Given that the best evidence for surgery (or IL-2), while highly plausible given the evidence, does have some associated uncertainty, what is a patient to do? If you choose treatment and it turns out that the purported benefits of the treatment are illusory, then you undergo a potentially difficult treatment for no benefit. If on the other hand, you forgo a treatment when it turns out that it does produce long term survivors, you are giving up the chance to be one of those long term survivors. I think given the known poor prognosis and lack of better proven alternatives, most people would prefer to risk the former error rather than the latter, and furthermore that this strategy maximizes the odds.
The Summary Tables
|Paper||Accural Dates||Number of Patients a
||Number of Patients with More than One
|Max Number of Mets Resected||General Comments|
|Kavolius||’80-’93||211||>117||At least 2|
|van der Poel||’85-95||95b||55||7|
|Tobisu||’63-’87||72||N/R||5||Very separate analysis of patients with metastasis at diagnosis
those with recurrence after nephrectomy.
|Fourquier||’60-’94||50||31||17 (!)||Lung Metastasis only excepting 5 patients|
|Jett||’70-’79||44||28||At least 5||Lung metastasis only, probably overlaps some of Kierney|
|Kierney||’70-’90||41||18||6||Excludes patients who presented with metastasis at initial
as well as certain sites of metastasis.
|O’Dea||’50-’70||28c||0||1||Solitary metastasis only|
|Althausen||’77-’96||23-30d||N/R||N/R||All patients had bone mets, an unknown number had other mets as
|Abbreviations: N/R – Not reported, UN – Unknown
Notes: (a) Number of patients in whom a total resection was intended.
(b) Excludes 6 patients in whom the primary was not resected. (c) By
my interpretation of the individual cases presented. I excluded
whose treatment did not include an attempt at total resection of both
the primary (if in place) and all known metastasis. (d) It was not
to determine the exact number of patients treated with curative
Survival and Surgical Parameters
|Paper||% Surgical CR||% Surgical CRs Who Had a Second Surgery after
|5 Year Survival in Surgical CRs||10 Year Survival in Surgical CRs||Survival Curves for Surgical CR Flatten?
|van der Poel||57%||64%||About 30%||About 18%||Yes|
Metastasis at Diagnosis: 53%
Recurrence after Nephrectomy : 73%
Metastasis at Diagnosis: 18%
Recurrence after Nephrectomy : N/R
Metastasis at Diagnosis: About 40%
Recurrence after Nephrectomy : About 25%
Metastasis at Diagnosis: About 18%
Recurrence after Nephrectomy : 0%
Metastasis at Diagnosis : Yes
Recurrence after Nephrectomy : No
|Fourquier||90%||27%||44%||4% (at 9 yrs it was about 22%. The small number of pts
at end of curve makes it unstable)
|Kierney||88%||36%||31%||N/R||No – Curve has unusual shape|
|O’Dea||89%b||N/R||UN||UN||N/R (survival curves not computed) but some very long
|Althausen||N/R||N/R||55% All patients||39% All Patients||Curve for all patients appears to flatten. No curve reported for
|Abbreviations: N/R – Not reported, UN – Unknown,
CR – Complete resection of all known disease
Notes: (a) My subjective judgment. (b) By my interpretation of the
cases presented. I excluded patients whose treatment did not include
an attempt at total resection of both the primary (if in place) and
all known metastasis.
|Paper||Total Resection Better than Partial?||Solitary Better than Multiple?||Multiple Surgeries Effective?||Longer Disease Free Interval Better?||Effect of Synchronous vs. Metachronous
|Kavolius||Yes||N/R but Solitary Site better than Multiple Sites||Yes||Yes (>12 mos vs <12mos)
In patients with surgical CR
Short DFI may have included synchronous metastasis
|UN (I had difficulty interpreting this parameter here)|
|van der Poel||Yes||No||Yes||Yes (>24mos vs <24mos)
Includes patients with incomplete resection
Short DFI included synchronous metastasis
|Synchronous equal overall|
|Tobisu||Yes||N/R||Several multiple surgery cases but not analyzed separately||Yes (<6mos vs >6mos)
In patients with surgical CR
Short DFI excluded synchronous metastasis.
|Synchronous better, particularly with surgical CR|
|Fourquier||Insufficient Data (90% had total resection)||No (If anything, multiple was better than solitary but difference
not statistically significant)
|Yes||No (They don’t state how they compared or if patients with
metastasis were counted as DFI zero.)
|Among CRs. metachronous did better but there were so few
patients, the difference was statistically insignificant
|Jett||No||No||No mention of multiple surgeries||Yes (>24mos vs <24mos)
Includes patients with incomplete resection
Short DFI probably included patients with synchronous metastasis
|Kierney||Insufficient Data (88% had total resection)||No||Yes||No (DFI considered statistically as a continuous variable, patients
with synchronous metastasis excluded from study)
|Synchronous excluded from study|
|O’Dea||Yes||N/A (Solitary Only)||N/A||N/R||UNa|
|Althausen||N/R||UNb||N/R||Yes (>24mos vs <24mos)
Includes patients with incomplete resection
Short DFI included synchronous metastasis
|Metachronous better in all patients but synchronous included
patients who didn’t have nephrectomy.
|Abbreviations: N/R – Not reported, UN – Unknown, N/A
Not Applicable, DFI – Disease Free Interval, CR – Complete Response
Notes: (a) I disagree with the author’s interpretation on this point.
Resection Of Metastatic Renal Cell Carcinoma
Kavolius JP, Mastorakos DP, Pavlovich C, Russo P, Burt ME, Brady MS
J Clin Oncol 1998;16(6):2261-6
Department of Surgery, Memorial Sloan-Kettering Cancer Center, New York, NY 10021, USA.
PURPOSE: Resection of solitary metastases from renal cell carcinoma (RCC) is associated with a 5-year survival rate of 35% to 50%. Selection criteria are not well defined.
PATIENTS AND METHODS: We retrospectively analyzed our experience with 278 patients with recurrent RCC from 1980 to 1993.
RESULTS: One hundred forty-one of 278 patients underwent a curative metastectomy for their first recurrence (44% 5-year overall survival [OS] rate), 70 patients underwent noncurative surgery (14% 5-year OS rate), and 67 patients were treated nonsurgically (11% 5-year OS rate). Favorable features for survival were a disease-free interval (DFI) greater than 12 months versus 12 months or less (55% v 9% 5-year OS rate; P < .0001), solitary versus multiple sites of metastases (54% v 29% 5-year OS rate; P < .001), and age younger than 60 years (49% v 35% 5-year OS rate; P < .05). Among 94 patients with a solitary metastasis, lung (n = 50; 54% 5-year OS rate) was more favorable than brain (n = 11; 18% 5-year OS rate; P < .05). Survival rates after curative resection of second and third metastases were not different compared with initial metastectomy (46% and 44%, respectively, v 43% 5-year OS rates; P = nonsignificant). Favorable predictors of survival by multivariate analysis included a single site of first recurrence, curative resection of first metastasis, a long DFI, a solitary site of first metastasis, and a metachronous presentation with recurrence.
CONCLUSION: Selected patients with recurrent RCC who can undergo a curative resection of their disease have a good opportunity for long-term survival, particularly those with a single site of recurrence and/or a long DFI.
Commentary on Kavolius
Surgical Patients: 211, Accrued: 1980-1993
This is the latest and greatest major metastectomy paper! It has the most patients by far, and is some of the most recent, and therefore most relevant data. The combination of most patients, most recent means this paper should be weighted heavily. Fortunately, it also has the best results of any! The long term survival curves flatten much like those from IL-2 immunotherapy, but with a lesser initial slope, and a slightly higher 10 year survival, about 20%.
They report encouraging survival, even with multiple operations. Short DFI (Less than 12 months) had an adverse impact on survival, but even so, 33% of patients with DFI less than a year who had a complete resection of metastatic disease survived at least 5 years (Note that the 9% figure in the abstract refers to all patients with short DFI, including those who didn’t have a complete resection), and there were a significant number of ten year survivors as well.
In this study, like the others, excepting Jett, complete versus incomplete resection made a definite difference which supports the idea that the treatment had an effect. Interestingly though, there was some level of long term survival in all groups, including those who didn’t have a complete resection, but it was much higher when complete resection was achieved (about 20% vs 10% at 10 years). The long term survivors in the non-resection group (and maybe some in the resection group) might represent immunotherapy successes. 1993 is in the early immunotherapy era for metastatic RCC. It does look like patients who underwent a complete resection had significantly better survival than one expects for high dose Interleukin-2 at five years and slightly better at ten years! About a third of patients didn’t get a complete resection after their first operation, and of patients who were operated on a second time after an an initial complete resection and relapse, about half didn’t get a complete resection. Clearly the high rate of incomplete resection is an area for potential improvement.
Remarkably, of the 141 patients who had a complete resection, 110 or 78% both relapsed and had a second operation. The second time around, a complete resection was achieved in only about half the patients. Given the very high percentage of patients with an initial complete resection who later needed another operation, the willingness to do more than one operation as needed may have been a key factor in the excellent overall results they achieved. I also find it very interesting that such a high percentage of relapsed patients were considered eligible for a second operation in the first place!
Rather than classifying patients as having solitary versus multiple metastases as other papers do, they classify according to whether the patient had solitary of metastasis versus multiple sites of metastasis, where each affected organ counts a separate site (My best guess is that this information is more easily and more reliably obtained from case records than the exact number of tumors resected). This makes the results for this parameter hard to compare to other studies. Because the most common single of metastasis (lung) is also one of the ones with a better prognosis and the next most common sites, such as bone or liver, typically carry a worse prognosis, it may be that multiple sites is was merely a marker for having at least one site that carries a worse prognosis. Nonetheless, multiple sites of disease was an adverse prognostic factor. It is encouraging though that even with multiple sites of involvement, there were about 30% 5 year survivors in patients who had a complete resection.
Patients with brain metastasis didn’t do as well as patients with other sites of metastasis in this study, though the five year survival was still about 20%. It’s also interesting that they don’t report any patients with liver metastasis! This is a common site of metastasis for RCC. I suspect absence of patients with liver metastasis is an indicator of the degree of patient selection since liver tumors are often considered hard to resect, but it may also be that liver metastasis tends to correlate with widespread (and inoperable) disease.
I had trouble determining how many of the patients had metastasis at initial diagnosis. They state in the abstract that this was an adverse prognostic factor, but I can’t find any further discussion of it in the paper itself, which actually reads as though such patients were excluded. I don’t think this was actually so because there were enough patients who were stage IV at diagnosis in the overall population they considered (many of whom didn’t have an operation at all) that it is not mathematically possible that they excluded all patients with synchronous presentation of operable metastasis. In any event, the data as presented, certainly would not lead me to exclude a good result with surgery for synchronous metastasis!
My conclusion: It may actually be that surgery is more effective than the best immunotherapy when all disease can be resected, particularly if DFI is more than a year! It’s also clear that there was no significant benefit unless all visible disease is resected.
Metastasectomy In Renal Cell Carcinoma: A Multicenter Retrospective Analysis
van der Poel HG, Roukema JA, Horenblas S, van Geel AN, Debruyne FM
Eur Urol 1999;35(3):197-203
Department of Surgery, Memorial Sloan-Kettering Cancer Center, New York, NY 10021, USA.
OBJECTIVE: In 60-70% of patients with renal cell carcinoma (RCC), metastases develop in the course of the disease. In the present analysis, the surgical management of metastases is described, and survival data are presented. This retrospective analysis may help in the management of future cases. Due to the retrospective nature of the data, no comparison between surgical and nonsurgical management is possible.
METHODS: Between 1985 and 1995, 152 resections of RCC metastases were performed in 101 patients at four Dutch Hospitals. Thirty-five and 6 patients had metastases resected 2 and 3 times, respectively. In most patients, the primary tumor was resected (n = 95). Resections were performed for metastases at different locations: lung n = 54, bone n = 42, lymph nodes n = 18, cerebrum n = 12 and locations in the spinal canal, thyroid, bowel, and testis. Skin excisions were excluded from the analysis. Solitary metastases were resected in 40 patients.
RESULTS: Median survival after the initial metastasectomy was 28 months. Initial tumor stage, grade, or size were not related to metastasis location or survival. The number of initially resected pulmonary metastases was of no influence on survival, however, multiple consecutive resections were related with longer survival. Patients with solitary metastases (n = 40) did not show longer survival after the first metastasectomy compared to no solitary lesions. Better survival was found for lung metastases compared to other tumor locations (p = 0.0006, log rank test) and for patients that were clinically tumor free after metastasectomy (p = 0.0230, log rank test). Additional immuno- or radiotherapy did not independently influence survival. Time interval between primary tumor resection and metastasectomy correlated positively with survival: a tumor-free interval of more than 2 years between primary tumor and metastasis was accompanied by a longer disease-specific survival after metastasectomy. Eleven patients were free of disease after metastasectomy with a median time of 47 (14-65) months. The median time of hospital admittance for metastasectomy was 9 days (4-64). Lethal complications were found in 2 patients. Long-term (> 5 years) disease-free survival was achieved in 7% of patients whereas 14% of patients were free of disease with a minimal follow-up of 45 months.
CONCLUSIONS: (1) Surgical management of metastases could be performed with short hospital stay, and low complication rates were found. (2) Disease- free survival was found in 14 and 7%, with follow-ups of at least 45 and 60 months, respectively. (3) The longest survival was found after surgery for pulmonary lesions. (4) Resection of solitary metastases did not result in longer survival compared to resection of nonsolitary lesions. (5) An interval shorter than 2 years between primary tumor and metastases was correlated with a shorter disease-specific survival.
Commentary on van der Poel
Surgical Patients: 95, Accrued: 1985-1995
This is the second largest study, and is the most recent in terms of when the patients were accrued. I’d say the overall results not quite as good as Kavolius, but the general trends were mostly the same, most definitely including the possibility of long term survival.
Patients with a complete resection did better than those with an incomplete resection, though the effect was less dramatic than in some of the other papers. Unlike Kavolius, van der Poel generally did not separate out patients who had a complete resection of metastasis from those who had only a partial resection when they analyzed other prognostic factors. I prefer the way Kavolius analyzed the data, because I expect incomplete resection would have very little to no benefit.
In this study, 40 of the 95 patients (42%) didn’t get a complete resection from the initial operation. Of the remaining 55 who did, 35 (64%) had a second operation for relapse. These results are generally similar to Kavolius, and again suggest both that surgical treatment would be improved if it were possible to more accurately predict who is really completely resectable, and that willingness to do multiple operations as needed may be an important factor in the success of surgical treatment.
At any rate, longer DFI before recurrence also mattered in their analysis. In their analysis, they compared patients who relapsed more than two years after nephrectomy to those who relapsed earlier. This two year interval is longer than most of the other papers but the difference in the survival curves was striking at first glance. Still, short DFI didn’t actually preclude long term survival, and they think the better survival in patients with longer DFI may have been due to the fact that more of them had lung metastasis rather than bone metastasis. It didn’t matter whether there were solitary or multiple metastases, and it also didn’t matter overall whether the metastasis was synchronous or developed after nephrectomy. They did multiple resections as needed. While they don’t give a statistical analysis of the survival with multiple resections, many of their long term survivors had more than one operation, confirming that multiple operations can be useful. They found lung metastasis to have a better prognosis than metastasis elsewhere. Only one liver metastasis was resected, which, again, is far below the expected number.
This study, like Kavolius extends into the age of immunotherapy which could have affected the results. In fact, they report that overall immunotherapy didn’t influence survival but no details are available about whatimmnotherapies the patients got. It’s very interesting that the shape of the survival curve for incomplete resections includes long term survival (at a low rate) like Kavolius. It’s possible that a small number of successes with immunotherapy was the cause.
Lung Metastases Of Renal Cell Carcinoma: Results Of Surgical Resection.
Fourquier P, Regnard JF, Rea S, Levi JF, Levasseur P.
Eur J Cardiothorac Surg 1997;11(1):17-21
Marie Lannelongue Hospital, Le Plessis Robinson, France
OBJECTIVE: The research was designed to evaluate the results of surgical resection of renal lung metastases.
METHODS: Between 1960 and 1994, 50 consecutive patients underwent resection for pulmonary metastases from renal cell carcinoma. Mean age was 59 years (range: 40-78 years). Mean time between nephrectomy and pulmonary resection was 3 years (range: 0-18 years). Nineteen patients had solitary metastase, 13 multiple unilateral, and 18 bilateral. Wedge excision was performed in 28 patients, segmentectomy in 3, lobectomy in 17, sleeve lobectomy in 1, pneumonectomy in 5 and biopsy in 3. Twelve patients had repeat resection for recurrent metastases.
RESULTS: The resection was complete in 45 patients. Three patients also had a complete resection of limited extra-pulmonary disease. There was one postoperative death and 3 complications. Mean follow-up was 42 months without loss of follow-up. The cause of death was always metastatic recurrent disease. Five-year survival in complete resection was 44%. Only one long survivor was observed in the case of incomplete resection in a patient who had a complete response after adjuvant immunotherapy. Five-year survival for the 12 patients with repeat resections was similar to the overall survival rate (42%).
CONCLUSIONS: Resection of renal lung metastases is a safe and effective treatment. No factor influenced the 5-year survival in this series except the complete resection. Extra-pulmonary metastases does not contra- indicate pulmonary resection. In selected patients, repeat resection for recurrent disease is warranted.
Commentary on Fourquier
Surgical Patients: 50, Accrued: 1960-1994
This study was intended to be restricted to lung metastasis but a very few patients also had metastasis elsewhere. They were willing to treat multiple metastasis even in both lungs. About 2/3 of their patients had more than one tumor. One patient had 17!
They had a high rate of complete resection (90%) but a relatively low rate of re-operation (27% of surgical CR’s) even though there was a high rate of relapse. 5 Year survival in surgical CR was good 44%. Ten year survival in surgical CRs was only 4%, but this was due to a sharp drop at the very end of the curve and this may not be reliable.
As with most of the other studies, multiple versus single metastasis didn’t make a difference.
They were unusually specific about pre-operative scanning which included a full body and brain CT scan once those technologies became available, and they also thought it was important to do open surgery where the lung could be palpated for nodules not seen on the CT scan. Interestingly, despite the changing technology, the year of resection didn’t make a statistical difference in the results. Whether this comparison had reasonable statistical power is unknown.
Pulmonary Resection Of Metastatic Renal Cell Carcinoma
Jett JR, Hollinger CG, Zinsmeister AR, Pairolero PC
Mayo Clinic, Rochester, MN
Over a ten-year period, 44 patients with known primary renal cell cancer underwent thoracotomy for pulmonary metastases. The median postthoracotomy survival for all patients was 33 months. The five-year survival was 27 percent. Postthoracotomy survival was significantly better in those patients with a disease free interval of greater than 24 months and patients with metastatic lesions greater than or equal to 3 cm. No difference in survival was detected in patients with one versus more than one lesion or in patients undergoing complete resection versus incomplete resection or biopsy only. Age, sex, grade of tumor, or location of the pulmonary metastasis had no influence on survival.
Commentary on Jett
Surgical Patients: 44, Accrued: 1970-1979
This paper is limited to lung metastasis, but patients with multiple metastasis were treated.
Alarmingly, they found no survival difference between patients who had successful total resection of all metastases and those who only had partial resection or biopsy. This suggests that the surgery actually didn’t affect the natural history of the disease! Fortunately, this conflicts with the other papers, so although I am unable to find any explanation of why this was the case here, the bulk of the evidence supports the idea that those who have a complete resection do much better.
Overall survival was a roughly linear decrease which implies an increasing hazard rate over time. Some of the other papers had survival curves which appeared to become flat or nearly so over time, implying decreasing risk with time though this was not uniformly the case. The 33 months median survival was more than you’d expect for metastatic RCC and generally consistent with the other papers. This might suggest these patients who were considered candidates for resection had less advanced metastatic RCC than average. The five year survival rate is in the range of the other major papers.
They compared patients with a DFI of more than 24 months with those whose DFI was less. It looks to me like patients with metastatic disease at diagnosis were considered to have DFI of zero. I am not able to determine how many patients were in this category. As with the other papers, shorter DFI was better. This has been observed to be true with metastatic RCC in general, so it is not a surprise here, especially if they didn’t affect the natural history of the disease. The number of metastases also didn’t make a difference.
One puzzling observation was that patients whose largest metastasis was larger than 3cm did markedly better than those with smaller metastases, and in fact represented almost all of the long term survivors. One might expect a larger metastasis to represent a faster growing, more aggressive tumor, with a worse prognosis. Perhaps this indicates instead that patients whose recurrence was not discovered early (thus the large tumor) but was still limited survived longer because it was more likely that the metastasis was truly limited in nature.
One caveat on all of these conclusions is that this paper reports on a total of only 44 patients and random variations could account for some of the observations, including the shape of the survival curve, lack of difference between complete and incomplete resection, and the observation that patients with a larger metastasis did better, which was based on only 11 patients with a > 3cm tumor.
This paper does the best job of any of reporting what testing was done before surgery to determine the extent of their metastasis. Only 1/4 of patients had a lung CT, but patients were treated between 1970 and 1979 before CT scan was routine, and sounds as though in most cases there was not a careful survey for extra-pulmonary disease. It’s not possible to compare this to the other papers though since comparable information is absent.
This paper, like O’Dea and Kierney, is from the Mayo Clinic, and the accrual period is a subset of Kierney. They say they evaluated all RCC patients who underwent lung surgery at the Mayo in this period so it appears that some patients in this study might then overlap that of Kierney . I discuss this more in my commentary on Kierney below.
Surgeon’s Role In The Management Of Solitary Renal Cell Carcinoma Metastases Occurring Subsequent To Initial Curative Nephrectomy: An Institutional Review
Kierney PC, van Heerden JA, Segura JW, Weaver AL
Ann Surg Oncol 1994;1(4):345-52
Section of Gastroenterologic and General Surgery, Mayo Clinic, Rochester, MN 55905
BACKGROUND: Solitary metastases from a primary renal cell carcinoma (RCC) occur in < 10% of patients with metastatic RCC. To date, the benefit of surgically resecting such apparently solitary lesions has not been well documented.
MATERIALS AND METHODS: Forty-one patients (25 men, 16 women) with metastatic renal cell carcinoma treated by surgical excision of solitary metastases (1970-1990) were retrospectively reviewed. They comprised 9% of patients with metastatic hypernephroma seen during this period. All patients had undergone previous curative nephrectomy with a median disease-free interval of 27 months. Patients with skeletal, spinal cord, and lymph node metastases were excluded.
RESULTS: Metastases were intrathoracic (n = 20), intracranial (n = 7), and intraabdominal or in the extrapleural chest wall soft tissue (n = 10). Three patients had metastases to the thyroid gland and one had a solitary metastasis to an index finger. Median follow-up was 3.2 years. Complete resection was possible in 36 patients (88%) with a single lesion excised in 23 of these 36 patients (64%). There was no operative mortality. Predicted survival from the date of complete resection of metastases was 77%, 59%, and 31% at 1, 3, and 5 years, respectively, with a median survival of 3.4 years. One patient is alive without evidence of recurrent tumor 93 months from the first of 12 complete surgical resections. Varying adjuvant therapy was used in 50% of the patients. An increased histological tumor grade of the metastatic lesion relative to the original RCC was the only significant prognostic indicator identified. Disease- free interval and number of resected lesions were not significantly associated with patient survival.
CONCLUSION: A small fraction of renal cell carcinoma patients are candidates for potentially curative surgical resection of solitary metastatic lesions. Excision of such lesions may contribute to prolonged survival in selected instances.
Commentary on Kierney
Surgical Patients: 41, Accrued: 1970-1990
This study is limited to patients with an initial solitary metastasis which appeared after nephrectomy. They excluded patients who presented with a solitary lesion at initial diagnosis, as well as those with bone, lymph node, or spinal metastasis (It is not clear to me why!).
The survival curve for patients who had a complete resection has an unusual shape with a period of low risk (plateau) at about a year followed by increasing risk maybe mellowing a bit by the end (5 years). This may simply represent random variation.
11 of the 41 patients had more than one operation, including one patient who was alive and disease free after his twelfth operation in nearly eight years! The number of operations didn’t make a significant difference in survival. Unlike most of the other studies, they didn’t find Disease Free Interval significant. The fact that they excluded patients who had synchronous metastasis could have had an effect here. The favorable results with synchronous metastasis reported in Tobisu and van der Poel, combined with the dismal results in recurrent disease within six months of nephrectomy reported by Tobisu make it unclear which in direction the effect would be! They had such a high rate of complete resection (88%) that it was impossible to tell whether incomplete resection was an adverse prognostic factor, as was reported elsewhere.
Their surgery patients comprised 9% of their mRCC population which is consistent with the 8.5% from O’Dea, the previous Mayo Clinic Paper.
The authors evaluated the records of all RCC patients treated at the Mayo from 1970-1990, which overlaps the accrual period for Jett, also from the Mayo. 20 of the 41 patients reported here had lung metastasis and thus could have also been reported in Jett (which reports on lung metastasis only). This would be a significant overlap! They do specifically say didn’t include any patients who had been reported in O’Dea.
The Treatment Of Renal Cell Carcinoma With Solitary Metastasis
O’Dea MJ, Zincke H, Utz DC, Bernatz PE
J Urol 1978;120(5):540-2
Mayo Clinic, Rochester, MN
Between 1950 and 1970, 44 patients (2.5 per cent) with renal cell carcinoma and a solitary metastatic lesion were treated at our clinic. Generally, treatment was aggressive, involving nephrectomy and excision of the metastatic lesion when possible. Patients presenting with the primary and metastatic lesion at the same time did not do as well as patients who presented with metastasis after nephrectomy. An operation for the metastatic lesion seemed to offer the best results in patients who presented with the solitary metastatic lesion after nephrectomy. The prognosis was uniformly poor in patients presenting with the primary and secondary lesion together, regardless of the mode of therapy. However, an operation seemed to be better marginally and did produce an occasional long-term survival.
Commentary on O’Dea
Surgical Patients: 28, Accrued: 1950-1970
This is one of the early papers on the subject of surgery for metastatic RCC and it’s cited quite frequently in more recent papers. Unlike more recent papers, survival curves are not presented, rather the data is presented in the form of a table describing the treatment and outcome for each patient. Although most of the patients had surgery as the primary treatment some actually got different treatment (Thus accounting for the 28 patients, I cite above compared to the 44 mentioned in the abstract). As with most of the other papers, there were some truly long term survivors, in this case, as long as 17 years!
The authors reviewed the records for all RCC patients who were treated at the Mayo from 1950 to 1970. Thus they are able give the percentage who had a solitary metastasis, which was 2.5%. Their 2.5% figure is quoted in many other papers as if this is the percentage of patients with metastatic RCC who have a solitary metastasis. Not so! It turns out that the denominator used to calculate their 2.5% figure includes those without any evidence of metastasis! (Oddly in their records of 1761 RCC patients treated at the Mayo, only 330 had metastatic disease!) The percentage of patients with metastatic disease who had a solitary metastasis is 44/330 = 13.3%! (Assuming conservatively that all of the solitary metastases that could be surgically treatable were in fact surgically treated, the percentage of patients with metastatic RCC who were surgical candidates as 28/330 = 8.5%). Note that Tolia and Whitmore who reviewed patients with solitary metastasis treated at Sloan Kettering between 1949 and 1969 quote a similar overall incidence of 3.2%, with the incidence among those with metastatic disease 11% which again is similar to O’Dea’s results (See “Solitary Metastasis from Renal Cell Carcinoma”, Tolia and Whitmore J. Urol. 1975;114:836-8).
The paper claims that those with the metastasis present at the initial diagnosis of RCC did worse compared to those who presented with a solitary metastasis after nephrectomy. Actually this seems to be the main point they’re trying to make, but I don’t think the paper proves it, because many more of the patients who had metastasis at diagnosis also had an incomplete resection (or even no resection!) so that the issue is confounded. By my count, total resection was attempted in only 6 of the 16 patients with synchronous metastasis, while it was attempted in 22/26 of those with recurrence after nephrectomy.
Strikingly, 14 of the 44 patients in this series had a brain lesion!
Prognostic Factors And Surgical Treatment Of Osseous Metastases Secondary To Renal Cell Carcinoma.
Althausen P, Althausen A, Jennings LC, Mankin HJ
Orthopaedic Service, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts 02114, USA.
BACKGROUND: The purpose of this study was to analyze the survival of 38 cases of metastatic renal cell carcinoma with secondary osseous metastases treated at the Orthopaedic Oncology Unit of the Massachusetts General Hospital. The survival was analyzed because it seemed to be considerably longer than any reported previously in the literature.
METHODS: Survival was analyzed with respect to age, gender, site of primary tumor, presence of pathologic fracture, disease free interval, initial presentation with metastasis, solitary versus multiple metastases, and axial versus appendicular metastases.
RESULTS: Survival for the entire group was 90% at 6 months, 84% at 1 year, 55% at 5 years, and 39% at 10 years. Age, gender, and presence of pathologic fracture had no influence on survival. Presentation without metastases, long disease free interval between nephrectomy and first metastases, appendicular skeletal location, and solitary metastases were all correlated with longer survival.
CONCLUSIONS: In the authors’ view, patients with the characteristics correlated with longer survival are appropriate candidates for aggressive surgical resection of bone metastasis.
Commentary on Althausen
Surgical Patients: 23-30, Accrued: 1977-1996
The data from this paper are very difficult to interpret, and I include it only to illustrate the point that prolonged survival at five and ten years is possible even with metastasis in locations considered to have an unfavorable prognosis such as bone. Indeed the raw survivals are extraordinarily high in this series, and they would probably be even more so if the survivals were calculated on the basis of those patients who underwent surgery with curative intent (the 23-30 I cite).
There are many difficulties in interpreting this paper. First with respect to treatment, at least 11 of their 38 patients were not treated surgically with curative intent – 8 had radiation only and 3 had operations intended to stabilize pathological fractures. The treatment of seven more is totally unspecified! Finally three more were treated surgically with curettage which involves scraping tumor off the bone. Here I’m not sure if the intent of such procedures is curative. Finally, five patients never had a nephrectomy for the primary tumor. The bottom line is I can’t determine exactly how many patients were treated surgically with curative intent. There is no record of how many patients had a complete resection. Also there is no record of whether any patients who relapsed had a second operation, and in fact we aren’t told how many patients relapsed.
With respect to disease, an unknown number of these patients also had disease elsewhere and what treatment they received for this is not known.
They claim patients with multiple metastases did worse than those with a single metastasis, but they mean multiple metastasis in a much different way than the other papers I’ve reviewed. They considered patients who relapsed to have had multiple metastasis as well as patients who had multiple metastases initially. I don’t think this is a reasonable definition of multiple metastasis and thus I don’t accept their conclusion that patients with multiple metastases did worse, or at least their conclusion isn’t relevant to whether multiple metastasis as defined in the usual sense is worse than solitary.