Surgery

Craniopharyngiomas project tight connexions with adjacent anatomical structures = hypothalamus, optic chiasm and nerves, meninges (pia meter and arachnoid), and more rarely arteries of the circle of Willis, sphenoid body, cavernous sinus and cerebral parenchyma that make radical resection challenging. Since the early 1950s complete resection has been attempted by experienced neurosurgeons with a mortality rate of approximately 20% initially. Progressively quality of resection and safety have improved due to the emergence of micro-surgical techniques along with the use of the ultrasonic aspirator and the laser, the benefit of pre-operative mapping by high resolution CT scan and MRI and the medical management of the post-operative period. Various approaches have been described that carry their own risks and advantages, the most popular being the subfrontal and transpterional ones. A gross total removal is made possible in 8 to 78%  the higher rate corresponding to the pediatric series. Most extensive resections have been associated with an anatomical situation, pre-chiasmatic or purely intra-sellar, with the gross solid appearance and with the size < 3 cm. Less extensive resections have been mostly described in retro-chiasmatic and multi-locular situations, in gross cystic appearance, and in tumor size > 3 cm. It is also certainly correlated with the surgeon’s experience. Post-operative mortality ranges currently between 0 and 20%, recent series reporting approximately 5%. Retro-chiasmatic sites look at higher risk. It can be due to hydro-electrolytic disorders, infection or hydrocephalus. It has been dramatically improved by the post-operative administration of cortisone. Following radical resection, local relapses have been described in 0 to 60% the majority being 25% or so. Young children seem at higher risk for relapse than older patients. In case of subtotal resection, the relapse rate raises to 60% and above Five year survival ranges between 60 and 100% in case of complete resection, and 35 to 70% in case of incomplete one. Cyst aspirations have not been considered as curative measures but can be part of a combined surgical-radiotherapeutical approach.

Radiation therapy

External irradiation has been recommended following incomplete surgical resection or sometimes following a simple biopsy. It has been widely used since the initial publication by Carpenter and Kramer although less frequently in children due to the concern for major radiation-induced complications in this age-range. Total dose usually ranges between 45 and 55 Gy, conventional fractionation, administered with megavoltage equipments through multiple beam arrangements: 2 opposed, 3, 4 orthogonal fields, rotational techniques. A relapse rate of 15 to 60% has been reported the majority being between 20 and 30%. Five year survival is in excess of 80%, which is clearly superior to partial resection alone and at least equivalent to complete resection. Our series fits well with these results (91% at 5 years) but indicates that long-term local failures can occur up to 9 years (5 and 10 year event-free survivals 78 and 56.5%). Little is known on the impact of a surgical procedure prior to radiation therapy and of quality of resection. But our series shows a significant difference between patients with or without resection, on EFS at 10 years although none of the procedures was complete macroscopically; however, some series suggest a dose-effect relationship: Sung observed on 32 patients treated by the combined approach, among 109 patients, that the relapse rate was 46.7% below 50 Gy, 16% between 55 and 57 Gy and 22% between 60 and 70 Gy. Recurrences at the highest dose-level were outside the irradiated volume. Regine evidenced a 44% risk of relapse below 54 Gy versus 16% above, but on small figures. Graham also found a significantly better local control and survival at a dose of TDF 83 and above than at lower doses. Our own series also evidences a statistical difference but didn’t really explore the high dose levels (maximum physical dose: 56 Gy). Conversely, other series have found no dose-effect relationship.

Several series emphasize the impact of timing of radiotherapy on the outcome. For Regine, there is a 78% 20-year survival when children are treated at the time of presentation versus 25% at the time of recurrence. A similar finding was not evidenced in the adult subgroup. Conversely Jose found that radiotherapy administered at the time of recurrence to a tumor dose of 50 Gy, yielded an excellent 77% 10 year-survival. Also our experience indicates a trend for a better outcome following an early administration of RT but this is not statistically significant. It also indicates that failure after radiation is not a hopeless situation when surgical salvage can be proposed: 4 out of 8 children who failed were operated on. One child died 57 months later. The other 3 survived with NED, 216, 42, and 46 months. Only one patient out of the four not operated on is still alive at 7 months.

Another debatable issue is the place of stereotactic techniques that have emerged since the early 1960s.

The stereotactic radiosurgery has likely limited applications. This relates to the fact that a safe procedure includes: small lesions (< 20 mm), mainly solid, well circumscribed on imaging and away by approximately 5 mm from the high dose region for major critical structures (i.e., brain stem, optic pathway, cranial nerves, retina). Furthermore, the tumors look to respond unconstantly to this form of radiation, which could be possibly related with the histological type: the adamantinous type, most frequently occuring in children, would respond less. Lunsford  reported in 5 patients treated for progression after prior microsurgery or radiation, that only one patient experienced a complete regression. Kobayashi  presented encouraging results on 10 patients treated at presentation (except 1) to 12–18 Gy on the 50% isodose line. Mean number of isocenters was 6. Seven patients showed marked shrinkage, 3 central necrosis without significant neurological deterioration. One recurrent tumor also benefited from this technique at the price of a visual deterioration.

The stereotactic radiotherapy that differs from the previous technique by the use of relocatable head frames that allow fractionated treatments, has also recently emerged as a possible treatment option. Children above 6 years of age and adults are treated with a frame stabilized by an acrylic oral appliance. Those below 6 years of age are treated under general anesthesia and the frame stabilized by ear plugs. Nine patients with craniopharyngiomas have been treated up to December 1993 by the Boston group, 7 of whom at the time of relapse. Tumor volume was kept below 18 cm3. It was possible to deliver 52 to 54 Gy with conventional fractionation (1.8 Gy per fraction) and excellent tolerance.

In our current approach, children with intra-cranial tumors benefit from a 3 D treatment planing based on CT simulation. A thermoplastic mask is initially constructed and then used for the entire simulation process and treatment course which greatly improves set-ups reproducibility. As far as craniopharyngiomas, we have recently implemented a high precision radiotherapeutical program based on a combination of high energy photons and protons. Initially, a high quality helicoı̈dal CT scan is performed with quick acquisition of 1-mm thick slices throughout the target volume and 3 mm in the rest of the head and neck region. A 3-mm spacing between slices generates high quality digitally reconstructed radiographs. A 4 to 6 noncoplanar beams’ arrangement is then generated. The implantation of 3 to 4 fiducial markers in the outer skull allows a submillimetric positioning accuracy of the proton part. Preliminary dosimetric investigations have shown that a 10 to 15% dose-escalation could be envisaged safely with a tight conformation of the high dose volume to the tumor and adequate sparing of the organs at risks of complication like the optic pathway and the 3rd ventricle.