Treatment of primary central nervous system germinoma with chemotherapy

Bookmark and Share
Published: 29 May 2013
Views: 6384
Dr Jonathan Finlay - Children's Hospital Los Angeles, USA

Dr Jonathan Finaly presents data on treatment of CNS germinoma with chemotherapy prior to reduced-dose ventricular and local boost irradiation at the 3rd International CNS Germ Cell Tumour Symposium.

The treatment strategy of the study purpose was to evaluate the outcome, both survival and quality of survival, following a uniform treatment strategy for newly diagnosed children, adolescents and young adults with primary central nervous system germinoma of any location using four cycles of carboplatinum and etoposide, loosely based on the regimen that Jeffrey Allen had developed, there were only slight differences, I think, in the doses, followed by reduced dose ventricular field and concomitant integrated focal boost irradiation. Those are the doses for you, you’ve seen things like that before – two days of carbo at modest dose, etoposide for three days, repeated every 21 days for a total of four cycles, followed by ventricular field radiation. In our case we used IMRT, anywhere between 21.4 to 24Gy ventricular field depending on what the size of the boost was going to be. The boost was simultaneously integrated so that the duration of therapy was actually a little shorter for the patients and the boost took the total dose to the primary sites up to 30Gy. This just shows you a pretty IMRT treatment plan; when you’re doing 12 field IMRT protons can go and eat their heart out. You can see a very pretty picture here and note how important it is to include the fourth ventricle in your ventricular field, especially in this day when we’re doing ventriculostomies and the potential perhaps for seeding down there although not proven is certainly a consideration.

These are the demographics of our 21 patients, total actually 16 males 5 females, 5 Asians represents a somewhat higher incidence than we would expect for our population but typical of the Asian community. Mean age of 15.5 years with a range of 7 to 21. 20 germinomas, one patient with a germinoma with a mature teratoma, predominant mature teratoma you’ll see. 20 were localised, only one was bifocal in our group, a surprising number, and zero were metastatic both by lumbar CSF cytology evaluated in all 20 patients and MRI of spine in all 20 patients. 10 were pineal, 6 suprasellar, 1 bifocal, 3 basal ganglias, remember treated the same way, ventricular field radiation, they didn’t relapse, and one cerebellar patient who has also not relapsed.

Alpha-fetoprotein markers were obtained in all patients both in serum and at CSF diagnosis. Serum AFP was elevated in zero patients, of course, at diagnosis. We had one patient with a level that we interpreted back then as 2.9ng/ml and said, ‘Well, that’s probably OK,’ the upper limit was around 1.5 so we accepted that and as we learned, to our patient’s cost, although thankfully only in the short term, that was a bad move. Anything abnormal in the CSF much lower normal range than what you have in the serum. Beta HCG levels as shown here; note marked elevations in the serum are unusual, they tend to be higher in the CSF. We had only marked elevation in one patient, I’ll come back to that patient in a moment, but in the CSF we had three patients with elevations greater than 50 which would have put them, at least in the older North American studies, into a mixed malignant germ cell tumour category.

13 had endoscopic biopsies, sometimes the pathologists would tell me there were no more than 50 cells on the whole slide which isn’t very helpful. Mark, you have to do something about that. You mean that’s more than they need, 50 cells?! 8 partial resections and no resections.

Now this is an important slide. I don’t want to go in and again it’s published but because of concerns that had been raised that rapid responses and complete responses were important and if you don’t have them you need to intensify therapy, we didn’t intensify therapy based on response after two cycles. You’ll see that although most people, 14 had greater than 90% shrinkage after two cycles, but there were some of those, one here with only 60%, two with 89%, this was our single growing teratoma syndrome after one cycle. The tumour got larger, it was resected and there wasn’t a sign of any germinoma it was pure mature teratoma at that point. But after four cycles you can see that the response is much, much better. 18 of them had greater than 99% shrinkage. The one with the weakest shrinkage had the smallest tumour, the bifocal patient actually, it was very interesting. What I can tell you right here and now is that the degree of shrinkage had no impact upon our ultimate outcome. We had, as I’ll show you in a moment, one patient at the end of four cycles in his extent of disease evaluation whose beta HCG was rising and rising. That was the kid who had the high beta HCG to start with and we’ll come back to that. It’s always important to learn from your mistakes and I try to focus on that.

So in terms of tumour marker response to therapy all five patients with HCG elevations achieved normalisation by the end of the first cycle. So we were fooled into thinking even the kid who had a very high level, ‘Oh, look how well he’s responded. He’s obviously just a germinoma, he’s going to do fine.’ Six of seven with CSF beta HCG that were elevated achieved normalisation at the first assessment by the end of either cycle 2 or cycle 4.

So there were two patients that relapsed. Number one was the patient I told you about with the CSF AFP at a minimal elevation of 2.9. He relapsed at 15 months out with AFP and beta HCG elevations. We got him to minimal disease with Cytoxan, transplanted him and he is now… I’m not sure whether it’s four or five years out without actually receiving further RT. We recommended low dose craniospinal, he refused it and he remains NED. Our conclusion, we didn’t biopsy because this was a mixed germinoma with yolk sac tumour right from the outset and at recurrence was a yolk sac tumour.

The other patient was really more difficult. This was the patient who at diagnosis had a serum beta HCG of 144. By the time we did our full extent of disease evaluation, organ function evaluations and got him started on therapy, one week later his serum beta HCG was 323 and we got the result back a few days later. But then when we saw how well he was responding we said, ‘Let’s stick with it.’ He was the one who at the end of induction his beta HCG rose and rose. I’m not going to go through the gory details of the number of regimens and then craniospinal radiation with big doses followed by tandem transplants but, as of last follow-up about 9-10 months ago, he was alive at least 3.5 years out but I have to tell you in very poor condition. I think the combination of the full dose radiation therapy followed rapidly by tandem transplants with high dose Thiotepa produced significant brain damage. We don’t have time to go through that now but it’s a warning for everybody in that regard.

Now with follow-up now in 2013 we have a relapse free survival of 91.5% at a mean of 7 years and an event free survival of 81%, overall survival of 95.6%. So what’s happened? One guy developed about 5 years out a mediastinal seminoma which we picked up because his serum beta HCG kept increasing and we kept checking his MRI of brain and it was negative until he suddenly started getting some respiratory symptoms, then we found his mediastinal germinoma. It has been treated and he’s off therapy over a year, year and a half, now and NED. But that certainly constitutes an event, although not a relapse.

One guy died of hepatic cirrhosis and liver failure. He was morbidly obese even at the time of diagnosis, he’d had a suprasellar tumour, had, if you like, a hypothalamic obesity at presentation and despite extensive work-up we could find no etiology for his liver failure. I can tell you that he had no liver toxicity going through his chemotherapy, we just do not know why it happened about three years out from completion of therapy.

In terms of updated neuropsychological outcomes, we find no deficits in IQ, memory or executive functioning in the 19 out of the 21 that excluded the badly damaged child and the one who died, of whom assessments now at a mean of 4.5 years of follow-up from diagnosis. No effect of age at diagnosis, degree of hydrocephalus meticulously evaluated at diagnosis or tumour location, and you’ll see here that the average IQ subset compared to a non-medical normative group, of course with a mean of 100, we see verbal comprehension, perceptual reasoning, working memory and processing speed all right, not significantly different from the non-medical normative group. They show average executive functioning, average verbal memory, average non-verbal memory and non-clinical levels of anxiety, depression and social functioning.

So we conclude that induction with four cycles of carbo and etoposide produce excellent response rates and subsequent survival irrespective of the response following two cycles of the induction chemotherapy. Consolidation with reduced dose ventricular field and integrated focal booster radiation is well tolerated and results in outstanding event free survival in patients who are truly germinoma. There’s the rub, that we have to be absolutely meticulous. Stu Gorman commented, others commented on the difficulty of getting complete marker analyses at diagnosis, it’s absolutely critical because otherwise we’re going to end up undertreating some patients to their cost.

Patients with even minimal AFP elevations at diagnosis, despite a biopsy, remember all of these were biopsied and had pure germinoma, require more intensive treatment regimens. Data do suggest that biopsied germinomas with beta HCG elevations at least up to 100 can be safely considered to be true germinomas and treated as such and that’s what we’re doing in North America now. Certainly I think in Japan you actually go even higher than that, up to 200, but then you have more representative pathology to feel comfortable with.

Long-term follow-up, however, as our study suggests, is really required for patients on these studies to seek out those who have completely unexpected and sometimes unexplained late catastrophic events and those who have second malignant neoplasms that may be related to treatment or may, in fact, have a genetic origin, as is going to be discussed further in this meeting.

Thank you very much for your attention.