Classification and immunohistochemical approach of intracranial germ cell tumours

Share :
Published: 29 May 2013
Views: 6226
Dr Ivo Leuschner - University Hospital of Schleswig-Holstein, Germany

Dr Ivo Leuschner presents at the 3rd International CNS Germ Cell Tumour Symposium in Cambridge, UK on germ cell tumours and immunhistochemistry.

Good morning everybody, it’s a great honour for me to be here and to present a few data about the classification of pathology of the germ cell tumours, especially about the intracranial germ cell tumours. Maybe I can demonstrate a little bit of insight about this very heterogeneous group of tumours with which we deal. I think it’s very important to go back into the history of the classification of these tumours. In the 1970s there were two competing classifications which were completely descriptive and it was already realised that there were some pure tumours like the seminoma germinoma group and the yolk sac tumours and the teratomas but the rest was just a mixture of description without any… I don’t want to say clinical relevance but it did not reflect the biology of these tumours. These classifications stayed there for quite a long while and people tried to explain how these different types of differentiation can occur and it was obvious that there must be a primitive germ cell or a totipotential embryonal cell which can go into the different directions and somehow they merge to these mixed malignant germ cell tumours. Nevertheless it did not really explain the real different biologic behaviour of these different tumours. The great step forward was done by the Dutch people of Looijenga and Oosterhuis who published, beginning in the late 1990s and then in the next decade, these five types of germ cell tumours. You can see, I will go into this in a little bit more detail, we have three types which are relevant for the brain and for the other sites and we have some special ones here in the testicular region and in the placenta and the uterus. There are different age groups, there are different genetic backgrounds in these tumours and, what is very important, the histology is different but nevertheless there is some overlap in between these groups of tumours. This we have to keep in mind, that in some of these types the teratoma group is here and the type I and in the type II and in the type IV germ cell tumours; the yolk sac tumour can occur in the type I and the type II and the seminoma or germinoma differentiation is almost exclusively in the type II germ cell tumour and in the special type of the testicular spermatocytic seminoma.

I will go into details in some of these tumours in a minute. The other very important thing was the understanding how these tumours might develop and how the different biological behaviour is related. So this is the normal development of the germ cell in the male and the female and the very important thing is the time point of meiosis and the arrest of these germ cells for the differentiation of these cells. Looijenga and Oosterhuis also said it’s probably these embryonic or these early germ cells which are the basis for these tumours but it might explain that we have, especially in the testicular region, two types in these more adult type germ cell tumours and these infant type germ cell tumours because the meiosis is something around between 12 and 14 years of age whereas the meiosis in the girls is much earlier and so the distribution of these different types in girls and males is different and is probably different in the different sites of origin of these tumours. So it’s still a lot of open questions in these tumours but it shows the large heterogeneity in these germ cell tumours.

If we start with the type I germ cell tumours it’s usually found in the neonates and children; we can find it in the gonads, sacral retroperitoneum, almost everywhere in the body in the midline where there is movement of the germ cells, and of course in the brain. The histology is quite limited: we have the teratoma differentiation including the immaturity and the yolk sac tumour differentiation and that’s all within this group. I just want to show the histology of it: the teratoma is just a mixture of mature or immature types of tissues we see in the body, here you can see different types of epithelia mixed together, some goblet cells with more some squamous cell differentiation here and some cilium or some respiratory epithelium, you can see here some glands, sometimes in the depths there is a glial tissue so it really is a mixture of mature tissue within these tumours. Here you can see the glial tissue within the fibrous tissue very adjacent to other structures. The immaturity is basically found in the neural differentiation of these tissues and there are two grading systems. The WHO is using the one published by Norris and Thurlbeck very early in the 1960s and ‘70s. The basic difference is the one from is really looking to the whole tumour and really saying the percentage of immaturity of the whole tumour which has been embedded by the pathologist whereas the WHO grading just focusses on single or multiple spots within the tumours, it’s a description of these high power fields. It’s not very well defined what is a low power field and so on but nevertheless the biological difference between these grading systems is not so important and there are not so many differences. If there is immaturity we have a higher risk for a yolk sac tumour differentiation in these tumours.

This is a typical immature differentiation. These are structures resembling the neural tube of an embryo. You can see here the lumen, these cellular cubes, here next to it some cartilage. You also can find somewhere some epithelial immaturity but the biologic impact of this epithelial immaturity is not very well understood. It’s much more important to identify these neural tube like structure are here for immaturity in a teratoma.

The yolk sac tumour differentiation is a bit complicated because it’s not every time spotted by the pathologist because there’s a large morphological variability within these tumours. Just here are some patterns you can see within this differentiation; I just want to show you some major types. This is the most typical one, these clear cells arranged in these structures, cupular structures, this is very typically seen in the yolk sac tumour differentiation, or these structures with these little spaces in it, it is very typical and we have these protein droplets within the cytoplasm of these tumour cells.

This is a more endometrioid differentiation which is really close to an epithelial differentiation. It must be realised that this is a part of the yolk sac tumour differentiation. Or this is a very solid differentiation, the so-called differentiation within a yolk sac tumour.

We can use, of course, immunohistochemistry. You know yolk sac tumours express AFP or you can measure it in the serum or in the spinal fluid. We can demonstrate that here with immunohistochemistry these brown labelled cells are expressing AFP but you can see it’s only in a minority of cells as labelled with AFP so we have other markers. This one here is again AFP, a little bit stronger than in the previous case. Now we use today glypican-3 which labels very nicely the AFP or the yolk sac tumour differentiation in germ cell tumours. So there are almost all tumour cells positive for glypican-3.

One of the major problems is to spot these so-called microfoci of yolk sac tumours, they are very small and there are a number from 1998 from the American group saying that these microfoci of yolk sac tumour differentiation is missed in almost 30% of the cases by the initial pathologists. So there is a rule that we say that if there is any elevated AFP in the child, being maybe older than 3 or 6 months, the pathologist really has to search for the yolk sac tumour differentiation, it must be somewhere.

Here is an example, this little tiny gland is a yolk sac tumour differentiation. It’s very tiny and it really can be overlooked by if you just rush through the slides using the microscope.

We have some genetic changes in these type I germ cell tumours. The teratomas are usually diploid and the yolk sac tumours are usually un-diploid. We have some gains and some losses. I think we will hear some more about these genetic findings in the afternoon so I don’t want to go into details here.

The type II germ cell tumours, the patients are usually older. I’m not sure if these 15 years proposed by the Looijenga group is right, it really depends on the sex and on the site of the tumour. So this is roughly for the testicular tumour saying something between 12 and 15 years might be the separation between type I and II, in other sites it’s completely different. We have it in the gonads, we have it in the and of course in the brain. So this is the other type of germ cell tumours we find in the brain. We have the pure seminomas, and we can find it in all sites, and we have the group of the known seminomas including these mixed tumours, pure teratomas and this really heterogeneous group of germ cell tumours.

This type of tumour is the most undifferentiated type of germ cell tumours we have. It’s a very monomorphous type. In our registry, the KTR, the Kiel Paediatric Tumor Registry, the number is not very high and we usually see it only in the CNS but the reason is we don’t see the testicular tumours because these patients, the older patients, are usually treated not by the paediatricians in Germany but by the urologists so the number is not correct in the epidemiological sense. This is a very typical picture of a germinoma, you see these monomorphous round cells, very undifferentiated cells, and intermingled with a lot of lymphocytes, histiocytic cells, so there is no other differentiation in these pure germinomas, it’s a very monomorphous tumour. We have some markers for it, the very classical one is PLAP but I will show later to you PLAP is not a very specific marker for germinoma differentiation. The more popular one is OCT3/4, it labels the nuclei of the cells but is also not very specific for it. Again we have CD117, we have already seen it in the previous demonstration, it’s expressed in the germinoma differentiation and we’ll come later to this one. This is CD41, this is a cancer testes antigen which is also expressed in the germinoma seminoma differentiation but in a variable extent.

The teratoma differentiation is very similar to the one seen in the type I germ cell tumour so I don’t want to repeat it here. You can see the same differentiation tissue types in the type II teratomas and so I don’t want to repeat it again.

The embryonal carcinoma is a different way of differentiating these tumours. It resembles early epithelial structures in the embryo, the so-called primitive epithelial cells of the embryonic disc. If you see the morphology it’s a little bit like the yolk sac tumour but the cells don’t have this clear cytoplasm, they form more these strange structures but if you don’t have enough experience it’s difficult to separate just by morphology between yolk sac tumour differentiation and embryonal carcinoma. It’s good to have a marker for it, it’s CD34, so it labels these cells specifically in the cell membrane. But, as I said before, none of these markers is really completely specific. The embryonal carcinoma component also expresses PLAP so there is overlap to the germinoma expression.

One of the questions – is there a transition between these different types of differentiation in maybe yolk sac tumour and embryonal carcinoma – this is a tumour cell nest within the germ cell tumour, here labelled with CD30. Cytology and growth pattern and CD30 fits very nicely to an embryonal carcinoma. These cells around have the appearance of a yolk sac tumour and this is labelled by Glypican-3. So we don’t know if there is a transition between these different histological types or these tumours; there must be a similar genetic background if we have these mixtures of different types within one tumour.

The choriocarcinoma component is resembling some structures in the placenta and if you look to it it’s a different type of cell, it is a large type. We have these large multinucleated cells, different growth pattern, and they express better HCG. You can see here the red labelled cells and these smaller components very typically. But there is one problem and this is some of these problems having these slightly elevated better HCG, we have so-called better HCG positive giant cells within these tumours. These are only a bit larger cells, don’t have these large clusters like the choriocarcinoma component like this one here, and they are probably the reason for these slightly elevated serum levels for the patient.

The mixed germ cell tumours usually have a teratoma component and in the other component in a really variable extent. We probably think it must be different than these pure teratomas from the genetic background and it’s probably a different group of tumours having all these different malignant tumours.

The immaturity in the teratoma is also seen in these type II tumours so it’s not related to the type I germ cell tumours that we have mature and immature tumours.

Just one special type, these so-called malignant transformation germ cell tumours. This is somehow mixed together. As a pathologist if I speak of malignant transformation in a germ cell tumour it does not mean we have a teratoma with yolk sac tumour differentiation, it means that we have a somatic malignant component in these tumours and this is usually PNET-like differentiation like Ewing’s sarcoma like differentiation. So it’s a completely different tumour wrapped with some myosarcoma components, adenocarcinoma components. It’s very rare but we see it occasionally in these tumours and we don’t know the genetic background in these tumours. Here you can see it’s an immature teratoma with these cupula-like structures, neurotubules and here is a monomorphous small cell component like in the Ewing’s sarcoma. We don’t find the genetic changes of the Ewing’s sarcoma so it’s only the phenotype which is similar but it looks like a Ewing’s sarcoma within these germ cell tumours.

One very important point is the precursor lesion. In the testes we have this intratubular germ cell neoplasia and this is very easily detectible and this is only found in the type II germ cell tumours but not in the type I germ cell tumours. There must be one probably in the ovary and probably also in the brain but we can’t detect it yet. So here it’s very nicely seen. This is here in the testes, you can see these PLAP positive cells. These are these atypical germ cells as a precursor lesion in the testes only found in the type II tumours. Here again the genetics, it’s different from the type I tumour. I just want to point out this isochromosome 12p and this now today we use for discriminating a little bit these tumours between I and II. It’s a very tricky investigation because you have this additional short arm on chromosome 12 and you can see you need this sandwich labelling with the centromeric probe in red and these two greens really packed together and it’s quite difficult to see under the fluorescence microscope.

Just briefly, the last teratoma type we have is a type IV germ cell tumour. It’s only seen in the ovary, I don’t believe it, we also see it probably in the testes. This is the dermoid cyst of the ovary and I’m sure in young boys we have the same lesions. This is the completely mature teratoma only having these epithelial structures in here. This is a very typical case from the ovary even with teeth, hair and so on, very well structured, no immaturity, no malignant component. They are usually diploid or tetraploid peridiploid.

Just to summarise the immunohistochemistry, we have some very good markers to identify the different components, seminoma, embryonal carcinoma, teratoma and so on but there is some overlap and I just want to finish up with some examples again. This is an embryonal carcinoma which is CD40 positive but also expresses OCT3/4 in a similar pattern as seen in the germinoma differentiation. So there is some overlap in it.

We have here this picture I’ve already shown, the embryonal carcinoma with the PLAP expression which should be found only in the germinoma differentiation. Last but not least, yolk sac tumour, a very pleomorphic yolk sac tumour with expression of PLAP as we have seen should only see in the germinoma differentiation.

Finally what is one of our focus in our interests are these cancer testis antigens, the CT45. It is expressed in typical germ cells, you can see here in a testicular lesion, and this is also related to the methylation of these cells. So there are some types of epigenetics which probably will be one of the fields in the future to explain how we have these different differentiations in these germ cell tumours.

I hope I have quickly shown you a little bit about the heterogeneous group of germ cell tumours showing sometimes very similar histologic patterns with a completely different biological behaviour. I think these classifications of Oosterhuis and Looijenga really made a large step forward in the understanding of germ cell tumours. In the future we need a combination of morphology, antigen expression and molecular genetics to classify these tumours better for better treatment.

I thank you for your attention.