Proton beam radiation (PT) is getting an increasing role in the treatment strategy of complex tumour cases and especially in children. AT PSI, over 100 children were treated so far.

In this analysis we present the evaluation of 62 children treated until the end of 2007 for sarcomatous tumours. Twenty-nine girls and 33 boys were included. Median age at time of diagnosis was 8.1 yrs. (range, 0.1–19.0). The histopathologies were embryonal RMS (n = 24), Chordoma (n = 10), Ewing sarcoma (n = 6), Chondrosarcoma (n = 5), unclassified/undifferentiated RMS (n = 5), Osteosarcoma (n = 4) and miscellaneous. All, but 2 patients had localized disease at time of diagnosis. Tumour site was head and neck in 43 patients, and spine or pelvis in 19 patients. In 50 out of the 62 patients, PT was performed after biopsy or incomplete resection. Forty-four patients had received chemotherapy before or during PT. Median dose of irradiation was 54 Gy (range, 45 – 74 Gy) with 1.8–2.0 Gy fraction dose 4 – 5 times weekly.

Median FU time was 20 months (1.4 – 101). 54 children were still alive at the time of analysis. Twelve patients failed, of them 9 locally and 3 patients at distant site. Acute toxicity was exceeding grade 2 (RTOG/ EORTC) mainly for bone marrow in children with parallel chemotherapy (n = 23). In 6 children skin/mucosal reaction was exceeding grade 2. Late effects were not exceeding grade 2 in the majority of children. One serious adverse event was observed in a very young girl with a parameningeal sarcoma experiencing a lethal ischemia in the pontine area after surgery, chemotherapy and PT.

In the vast majority of patients proton therapy was well tolerated. Local control and survival rates are promising. Longer follow-up time and a greater cohort will help to provide more reliable data.

Tumour volume reduction (i.e. response), assessed following induction chemotherapy, has been identified as a prognostic factor for localized embryonal rhabdomyosarcoma (RME) in the CWS studies. In combination with other risk factors, it has been used to stratify secondary local and systemic treatment. It is however unclear whether the poor outcome of non-responders is due to insufficient local and/or systemic post-induction treatment.

We analyzed post-induction therapy of RME-patients < 21 years with unresected localized tumours (IRS-III) and poor response (NR, i.e. < 33% tumour volume reduction) treated 1980–2005 in five consecutive CWS-trials. The NR were reviewed and subclassified (Objective Response (OR; i.e.< 33%–0%) vs. Stable Disease/Progression (PD; i.e. no reduction)).

From 758 IRS-III RME-patients, 59 were NR (n=34 OR, n=25 PD). Induction for NR included dactinomycin, vincristine, alkylators ± anthracyclines in all patients. There were no significant differences in comparison of the control group and NR with regard to age, size, TN-classification, apart from site (p=0.04), and no differences regarding these parameters between OR and PD. Twenty-four NR received continued induction chemotherapy, n=32 other combinations, and n=3 no further chemotherapy following response assessment. Four patients were treated with additional high-dose chemotherapy. Fourty-two NR were irradiated with a median dose of 48Gy (control group: 45Gy). In 20 NR, the tumours were completely resected. As of 9/2008, with a median follow-up of 4.5 years (range: 0.9–12.1) for NR survivors, 34 NR are alive in CR. Reasons for the 25 deaths were: local/combined failure (n=21), systemic failure (n=1), and other reasons (n= 3). 5-yrs-OS was 71±4% for the control group, 78±15% for OR, but only 43±15% for PD (p< 0.01).

Response is an important surrogate marker of outcome, but per se associated with a poor prognosis only in tumours without any volume regression to induction chemotherapy. Ineffective local control drives mortality in these patients.