paediatrics Brussels 17

Ependymoma reirradiation d T. E. M ERCHANT et al .

95

and, to a lesser extent, symptoms were apparent within 6 months of RT in the SRS group, their persistence resulted in morbidity and mortality even years after treatment. Two patients treated by using CSI experienced signs and symp- toms suggestive of necrosis. As noted, 1 patient experienced progression to necrosis in the cerebellum requiring surgery and HBOT. This case was notable because the event might have been predicted based on the short interval between the RT 1 and RT 2 treatment courses. However, because our tech- nique of reirradiation using a combination of CSI and boost treatment seeks comprehensive coverage of the neuraxis, shielding large volumes of previously irradiated tissue invites reseeding. The other case was a patient who underwent meta- stasectomy of a cervical spinal cord metastasis and developed signs of myelopathy. She was successfully treated by using HBOT. The interval between courses of radiation is recognized, along with such other clinical factors as the specific region of the brain or spinal cord, as an important variable to con- sider when offering re-treatment (6) . Although investigators showed by experimentation or in clinical studies that the spi- nal cord may tolerate reirradiation to relatively high doses (7, 8) , we restricted re-treatment to the spinal cord when admin- istering CSI to patients who had undergone previous infraten- torial irradiation that involved the upper cervical spinal cord. The addition of CSI to a dose of 39.6 Gy and to regions treated to less than 30% of the previously prescribed 54 Gy achieves a biological equivalant dose (BED) of approxi- mately 96 Gy 2 , whereas overlap of the entire cervical cord with the combined doses of 54 and 39.6 Gy would achieve a biological equivalent dose (BED) of approximately 246 Gy 2 . The former value is within the lower range and the latter is within the upper range of the previously reported cumula- tive spinal cord doses of 102–181.5 Gy 2 , for which the risk of myelopathy was estimated at 25%. In our patients, catego- rized as intermediate risk based on cumulative dose, the BED of the first course of treatment was about 129 Gy 2 (8) . Indications for reirradiation require exploration. In addi- tion, guidelines need to be established with regard to dose and volume. Patients who experience progression with com- bined local and metastatic failure after previous RT fare poorly with RT 2 and should be considered for experimental therapy or a combined-modality approach involving CSI. Pa- tients who experience progression with metastatic disease, but remain controlled at the primary site, should be consid- ered for aggressive metastasectomy and CSI, with attention given to the timing of reirradiation, normal tissue tolerances, and adequate treatment of the volume at risk. Progression at sites of metastasectomy and high-dose irradiation with dura- ble control at the primary site suggest the need for more aggressive surgery to achieve negative margins, greater cumulative doses of radiation to the metastatic site, or a com- bined-modality approach that would include agents synergis- tic with RT. Patients who experience progression with local failure require careful neuraxis surveillance for metastatic disease and aggressive local resection with definitive evidence that

1.0

(1)

0.8

(2)

0.6

0.4

Overall Survival

0.2

(3)

0.0

0

24

48

72

96

120

144

Time (months)

experimental systemic therapies instead of reirradiation for this patient population. The recent use of conformal RT for the treatment of chil- dren with EP has decreased the volume of normal tissues re- ceiving the highest doses and increased our knowledge of dose to normal tissues volumes, leaving open the possibility for reirradiation in the setting of failure after conventional treatment. We have taken advantage of this information to perform additional surgery and reirradiation by using a frac- tionated treatment approach. We attempted to present results of this series according to RT 2 type: SRS, FFRT, and CSI, with additional subgroup analysis according to RT 1 failure type. The major findings from this review are the lack of long-term disease control and morbidity of patients undergoing SRS, the durability of disease control and lack of major toxicity for patients under- going CSI, and the excellent rate of disease control for pa- tients with local recurrences re-treated by using fractionated therapy. The effectiveness of reirradiation is ultimately measured by the rate of long-term disease control and functional out- come. Although this series spans nearly 20 years, 28 of 32 pa- tients had their RT 2 administered during the past 10 years, making the results preliminary. In this preliminary series, the PFS RT2 /PFS RT1 ratio was greater than unity for 20 of 38 patients, including 4 of 6 SRS patients, 5 of 13 patients with local failure treated by using FFRT, and 8 of 12 patients with metastatic failure treated by using CSI. Exploring normal tissue tolerance with reirradiation can lead to lethal complications or tremendous morbidity. This is readily apparent from the SRS group and remains a concern for patients treated by using fractionated reirradiation regard- less of the follow-up interval. Although changes in imaging Fig. 5. Overall survival dated from the start of reirradiation accord- ing to treatment method and initial tumor pattern failure (blue (1) = 12 patients with initial metastatic failure treated with craniospinal re- irradiation; green (2) = 13 patients with local failure retreated with focal fractionated irradiation; red (3) = 5 patients with local failure treated with radiosurgery).