Preliminary experience with personalized and targeted therapy for pediatric brain tumors.

Background: A new generation of anticancer drugs has reached clinical care in common diseases, but their use in rare diseases such as pediatric brain tumors lags behind since conventional clinical trial design requires larger patient numbers.

Procedure: We designed individualized treatment protocols for pediatric patients with relapsed brain tumors, based upon the patient's treatment history. In addition, each tumor was analyzed with morphoproteomics using a panel of markers to show treatment targets, resulting in a list of potential novel drugs to be added to chemotherapy. Here, we present the concept and report the experiences of the first patients enrolled in the program.

Results: Eleven treatment protocols were designed using morphoproteomic information and given to eight patients. The histological diagnoses included: medulloblastoma (n = 3), glioblastoma multiforme (n = 2), atypical teratoid rhabdoid tumor (n = 1), choroid plexus carcinoma (n = 1), and primitive neuroectodermal tumors (n = 1). Tumor markers included p-ERK, Topoisomerase IIa, Bcl-2, VEGF-A, p-STAT3, ER-beta, p-mTOR, and p-NF-kappaBp65. The novel agents included sorafenib, bevacizumab, fulvestrant, rapamycin, bortezomib, and curcumin. The response to the first protocol was complete response: 1, partial response: 1, stable disease: 0, progressive disease: 4, and continuous complete remission: 2. The median Event-Free Survival was 0.32 year ± 0.4. For the comparison with the institutional control group, the individual response probability was calculated. The observed response was superior to the historical controls (P = 0.006 Whitman U-test).

Conclusion: This approach warrants further, systematic evaluation as proof of concept and then expansion to drug-specific hypotheses.