Head and neck sarcomas: prognostic factors and implications for treatment

One hundred and thirty patients with soft tissue sarcoma of the head and neck were treated at the Royal Marsden Hospital between 1944 and 1988. Pathological review was possible in 103 of these cases; only pathologically reviewed cases have been analysed. The median age at presentation was 36 years, and 53% were male. Four had neurofibromatosis type I, and one previous bilateral retinoblastoma. Six had undergone previous radiotherapy, 12 to 45 years prior to developing sarcoma. The tumours were < or = 5 cm in 78% of cases and high grade in 48%. Only one patient presented with lymph node metastases and only one with distant metastases (to lung). Malignant fibrous histiocytoma was the commonest histological type, occurring in 30 cases. The overall 5 year survival was 50% (95% CI 39-60). Local tumour was the cause of death in 63% of cases and 5 year local control was only 47% (95% CI 36-58) with local recurrence occurring as late as 15 years after treatment. The only favourable independent prognostic factor for survival was the ability to perform surgery (other than biopsy), with or without radiotherapy, as opposed to radiotherapy alone (hazard ratio 0.39; P = 0.003). Only one patient had a biopsy with no further treatment. Favourable independent prognostic factors for local control at 5 years were site (tumours of the head as opposed to the neck, hazard ratio 0.42; P = 0.02) and modality of treatment (combined surgery and radiotherapy compared to either alone, hazard ratio 0.31; P = 0.002). Patients in the combined modality and single treatment modality groups were well balanced for T stage, grade and tumour site. The patients in the combined treatment group had less extensive surgery, yet their local recurrence-free survival was longer. Unlike soft tissue sarcomas at other sites, those in the head and neck region more often cause death by local recurrence. The addition of radiotherapy to surgery may result in longer local recurrence-free survival.