Conference Vancouver, B.C. June 11, 2001
Speakers
Dr Jones (part of main presentation):
So in summary in terms of recommendations PET is not sufficiently sensitive certainly compared to Sentinel Node Biopsy to evaluate Stage I and II and patients and distant metastasis in this group of patients are very rare so I think it will not be very cost effective to screen this group. PET is useful in evaluating palpable nodes in Stage III patients and may detect, may direct therapy in this group if radical surgery is contemplated. In Stage IV patients, PET is useful in the preoperative assessment of patients for whom a potentially curative surgery with metasectomy is planned. It appears to be more sensitive in Stage IV patients in detecting intra-abdominal disease and bony metastases and it does not apparently have any role in the evaluation of primary unknown melanoma.
Dr Sutcliffe: Thank you very much Amanda, we are going to open this up for discussion now. Dr Baum do you have a comment?
Dr Conti: Perhaps while he is setting up I can answer a question or make a comment actually. I just want to make sure that we all understand that there are some issues related to measurement for efficacy of PET in terms of what its role should be. I think it is important to distinguish outcomes from management change. Outcomes are highly dependant on the available therapy and as imagers we may or may not have a significant impact on outcome as it relates to treatment, on the other hand if a decision is made in terms of whether or not to treat or whether not to treat with a different modality the original one that one was originally considering you get into an issue of management change or to give additional chemotherapy or things of that nature—that is probably a more reasonable yardstick, compared to measurement of health outcomes.
Dr Sutcliffe: Thank you. Yes Richard.
Dr Baum: Maybe just a short comment on this paper by Mijnhout in Cancer, which I really do not understand, also about the conclusions you draw. We have all the publish in cancer, we should exactly know, in ’98 a prospective study and without being too long I just want to would like to present the data here that where primary staging in high risk melanoma so I agree there is no role in Stage I and II because the frequency of metastasis is just too low but as you know in Stage III, the prognosis is deteriorating dramatically. We have done it also for restaging for patients with suspected or proved recurrence for deciding as you also mentioned if there should be surgery or not. We have compared that and these were 100 prospective, consecutively studied patients with blinded evaluation of the data, that was at my time at the University of Frankfurt, and we have done this standard battery of tests and compared it to one whole body PET, 52 were in the group of primary staging and 48 were in a group for restaging. This is just one example with a hypermetabolic lymph node in the groin which was unclear and this are our results and I think this will be confirmed in many studies, especially from Switzerland I might say from Zurich there are some prospective studies which were also not mentioned by this guy Mijnhout and we have had a sensitivity, also we gave patients and single lesions so there is a direct comparison, exactly in Cancer published with a 100% sensitivity in the group of the patients for primary staging and for metastasis was 91.8 and with a very good specificity also which is much better than the conventional imaging and just another example here which shows a lesion which we thought would be in the spleen or just above the spleen and when you do image fusion with the CT scan you see it is just below the diaphragm outside the spleen and it was removed and was a 1.2cm lesion so in melanoma I think there is a role in Clark level III, IV, for staging high risk patients and also for restaging patients to decide before therapy. So I don’t understand this paper you make citation of?
Dr Sutcliffe: Thank you very much. Do you want to respond to that now Amanda?
Dr Jones: One of the things that came out from the Mijnhout review, and also I did look at your paper in Cancer as well, was it became very difficult to actually determine which patients were, in which patients metastasis were detected when those patients were clinically Stage I & II and which ones where metastasis were detected where there was a very high pre-test probability because there was suspicion, because of radiological abnormalities or symptoms, where these metastasis were being found and I found that very difficult to get out of your paper, how many of Stage I & II patients who were asymptomatic actually had metastatic disease found at diagnosis, the second.
Dr Baum: Can I answer to this question?
Dr Jones: Sure.
Dr Baum: Possibly you didn’t read it correctly, because it is clearly written that the inclusion criteria were Stage III patients. We didn’t look at Stage I and II, I mean the reason for that was that as you know the 5 to 10 years survival rates are quite good and there are not frankly many metastasis so we decided to go directly in the higher groups in Stage III.
Dr Jones: No. I think that we are talking about different things: you are talking about T stage and I am talking about clinical stage.
Dr Baum: Ok, I am talking about Clark level.
Dr Jones: Yes, I am talking about Clinical Stage I & II patients.
Dr Baum: Ok.
Dr Jones: As opposed to Clark’s levels, which we do not use any more, we use millimetres of depth of invasion.
Dr Baum: Yes, we too.
Dr Jones: So I think there is no question that it is of value in the analysis of patients at Stage III, Clinical Stage III, and that is patients presenting with palpable nodes, but in the two series, in which patients, a large series including 50 to 100 patients where systematically PET scanning has been compared with Sentinel Lymph Node Biopsy the sensitivity for detecting disease within the regional lymph nodes is very low.
Dr Baum: Actually, we were not interested in the regional lymph nodes because this is done also in our Centre by Sentinel Node and it is very clear that PET cannot detect microscopic disease and in some patients you even have to use PCR to detect it, and I think this is not, this is not the role of PET. But from a practical point of view, I mean to have a stage, a clinical stage what you are speaking about you first have to do the evaluation of the patient. So our starting point was one step earlier, I mean after resection of the primary tumour, you look at the tumour thickness right? and you clearly can say if this is more than 1.5mm the risk increases and this is Clark level III so our decision way was: If this is Clark Level III or higher then it is a high risk patient, and then let’s do a PET scan versus a conventional battery of tests, and it came out that PET is much better than conventional staging methods, especially concerning specificity, I mean if you look at our data, this group the specificity was really lousy, if I remember right it was 48% or something for conventional staging with a lot of false positives, and I think this is another argument to use PET in this group.
Dr Sutcliffe: I’m just going to bring Dr Wahl on in that discussion.
Dr Wahl: Well, I’ll confess, I didn’t read the paper in Cancer 2001, I didn’t see it yet, but I did have a role in writing the first paper in which PET was describing melanoma in 1993 and found that it worked pretty well and also worked in animals before that, and I think one paper that wasn’t referenced in that review article, I believe. If I can remember who published it was a McFarlane and was that referenced in their paper? I think if I recall correctly almost all those patients went to surgery there weren’t a lot under about 20 this was in the Journal of Clinical Oncology and in that study the sensitivity in PET was superior to that described in the paper from Indiana but I think it is quite clear that our patient population included more patients who had palpably abnormal nodes and in that group PET does really well, but I think it is absolutely clear that you will fail to detect small lesion volume melanomas with PET and that it is not comparable to Sentinel Node for sensitivity and there has been a pretty good article from Italy, showing that the detection capability is dependent on the size of lesions, it observes a 1cm - 100%, even with an older scanner, and you get down to 3 mm and you may be at the 30% level, so it is clearly size related. So I would agree with you on suggesting that it would be more appropriate in larger tumours and in more advanced disease. I do think though that the assessments based on the number of lesions are quite as important, maybe less so for melanoma but in a lot of times, you know whether it is 5 versus 10 makes no difference but the issue, zero versus one, is really quite important and in colorectal which was discussed earlier the number of lesions in the liver can make a big difference as to resectabilty. So my general attitude is if you want to use a diagnostic test, then ask whether somebody has metastatic disease you should use the more sensitive one that if you don’t want to know whether they have metastatic disease you are probably better off not ordering a test.
Dr Sutcliffe: I am going to stay with.
Dr Wahl: I am sorry just one other point, I am very sorry, but the specificity described in the paper of some 47% from Duke University if I recall correctly that was high sensitivity low specificity, in a fairly large study. Some of those specificity figures are misleadingly low because they don’t include a physical examination of the patient, I think that the studies from Zurich, and our work have shown that inflammatory lesions of the skin, surgical incisions, boils and things like this can cause false positives and it really is important not to just look at a scan but to correlate the scan results with clinical findings because otherwise you can end up with lower specificities with cutaneous lesions just misleading you completely, that could be confused with melanoma, so I think the clinical exam even with our scanners is still going to be important for cutaneous abnormalities. So I think that may in part be the cause for the lower specificities as has been reported in some studies.
Dr Jones: I think the main problem is the studies have very heterogeneous groups of patients and they have Stage I, Stage II, Stage III, Stage IV and overall they may have high sensitivities and specificities but it seems to be that mostly that’s detection of metastatic lesions in advanced stage patients so I would make a case that for Clinical Stage III and clinical Stage IV patients, PET scan is very relevant, that the melanoma has a very high uptake of FDG, higher than a lot of other tumours, and it has very unusual patterns of metastatic spread and therefore PET scan will often direct you to areas that you would never envisage were going to be affected by metastasis, but I am simply arguing that compared to Sentinel Node Biopsy in early stage patients that it is very insensitive and in early stage patients with respect to detection of metastatic disease at other sites, metastatic disease at other sites is very uncommon in patients in early stages without, you know even taking patients with deeper tumours, metastatic disease at presentation is very uncommon and so it is probably not that cost effective and the specificity seems to be lower in these early stage patients because you pick up a lot of abnormalities that don’t turn out to be related to metastatic melanoma.
Dr Sutcliffe: So Amanda it seems to me that your recommendations would be:
Dr Jones: Yes.
Dr Sutcliffe: Those are the two options that you are coming forward on.
Dr Jones: And I think thirdly in evaluation on patients presenting with bony pain or with abdominal symptoms, or undiagnosed GI blood loss, where I think PET is useful and is more sensitive than all other investigations including probably endoscopy in establishing the site of metastatic disease.
Dr Sutcliffe: Having heard those recommendations is there anybody from our guest faculty who feels those would not be appropriate indications for PET? They may be conservative relative to our discussion but.
Dr Baum: No I agree, nearly completely but I would really ask to take the story one step before the staging, before you get the stage, you have to do the whole body staging and I would include in the staging process of patients with high risk melanoma, that means with a tumour depth/thickness of more than 1.5mm, I would include today PET, because as our series showed you have, I don’t agree on that with you completely. You have in a certain number of patients who have distant metastasis detected also in the primary staging but which are not detected by other methods. So this high risk group I would include PET in the staging algorithm.
Dr Sutcliffe: On depth of invasion as a test for selecting PET in staging—Peter, do you have a comment?
Dr Conti: I think, I agree 100%, with Dr Baum here, I think if you have a situation where you have a certain probability of metastatic disease based on lesion depth. If you ask the question, what is the stage of the patient, and you are going to trigger a series of steps that is going to assess the stage. The best test that is available today is PET, period.
Dr Shreve: I would, just to make that more succinct if you are considering a head to toe CT scan, PET is the superior test compared to a head to toe CT scan, even with thin cods, and there was something said earlier that CT of the liver was more sensitive than PET for melanoma of the liver, that is just absurd. That is just totally absurd. It might be more sensitive yet the specificity is practically zero because if you call every too small a characterized lesion of the liver which for those of us who are radiologists are just as common as anything, and you say those are all metastases without actually proving them with biopsy, then may be it is more sensitive, but there is no way that I can see how CT is going to be more sensitive than PET for metastases to the liver and I would ask my colleagues if they see how.
Dr Jones: What I actually said was that it was more sensitive with respect to the numbers of lesions that is all.
Dr Shreve: CT is finding more lesions than PET?
Dr Jones: Yes. Not on a per patient basis.
Dr Shreve: No, but I mean in the liver.
Dr Jones: Yes.
Dr Shreve: Like CT will find six lesions—they must be calling non-specific lesions that aren’t proven, I can’t see any other reason.
Dr Conti: What happens frequently is that you see these lesions and the report says suspicious for metastatic disease and that is the end of it. There is no further work up, because you are not going to go in and biopsy each one of these lesions.
Dr Shreve: It is completely meaningless.
Dr Sutcliffe: I am actually going to actually hold us now on melanoma—I think we have got some very clear indications where there is total agreement and I think that some ongoing discussion that we need to have around depth of invasion in relation to staging. I think that we are pretty much in agreement on the other indications.