Diagnosis #1 - Anaplastic medulloblastoma.
Diagnosis #2 - Low-grade lipoastrocytoma.
Diagnosis #3 - Pilocytic astrocytoma with radiation induced modification.
To our knowledge, the present case illustrates the exceptional occurrence of three different brain tumors in a single patient. The clinical history and the clinico-radiological findings of our patient ruled out a possible syndrome as the cause of such an association. One alternative explanation is the possibility of second neoplasms. With specific regards to gliomas as possible sequelae of the treatment of medulloblastoma in children, Hope et al. (4) reported 16 secondary astrocytomas in a review article published in 2006 and added a new personal case. Since then, three new cases have been reported (two high grade gliomas and a pilocytic astrocytoma), all of them occurring in irradiated children (2,5,7). The main hypothesis is that adjuvant radiotherapy and/or chemotherapy are able to provoke genetic aberrations in the irradiated tissues, leading to the development of induced neoplasms.
In the present case, the cerebellar pilocytic astrocytoma could represent a secondary tumor, although it was expected to be a recurrence of the primary metastatic medulloblastoma. Nevertheless, the time from the adjuvant treatment to the appearance of this tumor (about 1 year) seems to be too short compared with the criteria used to define a secondary malignancy. Even though an exact interval of time has not been defined yet, the shorter interval reported in the literature in patients previously treated for medulloblastoma is 26 months (average: 13.1 years) (8).
Koksal et al. (5) proposed an alternative etio-pathogenetic hypothesis for medulloblastoma-related secondary tumors based on foci of glial or neuronal differentiation possibly found in medulloblastomas. Accordingly, radiation and/or chemotherapeutic drugs administration would result significantly more effective against the undifferentiated cells than against the differentiated areas; subsequently, the second malignancy would represent a "recurrence" (or a "residue") of the primary medulloblastoma, which is finally represented only by the more differentiated surviving cells. On these grounds, it can be hypothesized also that the second malignancy originates from the medulloblastoma cancer stem cells (CSCs). Astrocytomas occurring after radio/chemotherapy for medulloblastoma, indeed, would represent the differentiation along glial lineage of the multipotent and multiresistant medulloblastoma CSCs. Actually, CSCs have been demonstrated to be more radio/chemo-resistant than the other, mitotically more active tumor cells (1). Moreover, radio/chemotherapy can induce just the differentiation of the CSCs (10).
Alternatively, it has to be accepted that our patient harbored two different tumors at the time of surgery in the posterior cranial fossa: an anaplastic medulloblastoma, which was grossly removed, and a small pilocytic astrocytoma, possibly infiltrating the tentorium and/or the surrounding cerebellar tissue, which was not excised during the first operation so that it was not appreciable at the first pathological examination. The comparison between the first preoperative MRI (Figure 1) and the MRI at the moment of the recurrence (Figure 9) seems to suggest that the small subtentorial nodule in Figure 1 (black arrow) - considered as the expression of the tumor spreading - is separated from the main tumor mass and it reappears grossly unchanged more than one year later (Figure 9) (white arrow). Moreover, this neoplasm showed areas largely composed of monstrous cells, with irregular and hyperchromatic nuclei and without proliferative activity. Such features can be related on the effect of radiation therapy on a pre-existing tumor. This observation further support the hypothesis that the pilocytic astrocytoma was coexistent with the medulloblastoma rather than a radiation induced tumor.
In contrast to the cerebellar pilocytic astrocytoma, the frontal lipoastrocytoma was clearly detectable at the first MRI, and did not disappear after radio- and chemo-therapy, even showing a mild increase in size over the time. Lipoastrocytoma is a rare (only 4 cases described so far, including the present one), quite recently described variant of pediatric low-grade glioma, consisting of astrocytic cells with a diffuse lipoma-like degeneration (3). In the present case, the contiguity of this tumor with the meninges and the presence of several other subarachnoid contrast enhanced nodules in the whole neuraxis initially suggested the diagnosis of supratentorial metastasis of medulloblastoma. Surprisingly, the pathological investigation provided the diagnosis of an "incidental", synchronous astrocytic neoplasm. Owing to the presence of the pilocytic astrocytoma, one could postulate that there were two foci of low-grade glioma which followed two distinct pattern of differentiation. The adipocyte-like appearance due to the coalescence of small fat droplets into large lipid droplets actually characterizes the lipoastrocytoma, while the presence of monstrous, multinucleated cells describes the post-radiation changes of the pilocytic astrocytoma.
An alternative hypothesis can be formulated considering the "changing histology" after adjuvant therapy. In 1969, Oppenheimer (6) defined the changes found in an irradiated medulloblastoma as "giant cell glioblastoma". The author described marked cellular pleomorphism, numerous bizarre giant cells with giant or multiple nuclei, karyorrhexis, necrosis, infiltrate of phagocytic cells. Such features are noticeably different from the histopathological appearance of a lipoastrocytoma.
However, Oppenheimer's observations were made post-mortem, a few days after the completion of the radiant therapy. Consequently, we can argue that most of the described tumor cells were necrotic because the tumor had not been completely cleaned by phagocytes due to the patient's death. In our case, instead, radio/chemotherapy could have killed all the medulloblastoma typical cells, leaving behind only a lipomatous degeneration and glial cells that, as stated before, are more radio/chemoresistant than the undifferentiated medulloblastoma cells. Such a theory is supported by the description of possible adipose transformation in PNETs (9).
The explanation of the association of three different neoplasms in our patient remains speculative. Consequently, it should be considered as the result of the occurrence of three synchronous brain tumors.
Contributed by Luca Massimi, MD, Massimo Caldarelli, MD, Quintino Giorgio DíAlessandris, MD, Massimo Rollo, MD, Libero Lauriola, MD, Felice Giangaspero, MD, Concezio Di Rocco, MD