The propensity of pleomorphic adenomas to recur is generally attributed to the biological nature of the tumour, and surgery close to the capsule is perceived as undesirable. At the Christie Hospital, Manchester, between 1947 and 1992, 475 tumours arising within the superficial portion of the parotid gland were treated by two surgical techniques: extracapsular dissection (380 patients) and superficial parotidectomy (95). Recurrence rates were 2 per cent in each group (median follow-up 12.5 years). Contact of the tumour with the facial nerve was recorded in 51 per cent of patients. There was no difference between the treatment groups in the incidence of permanent facial nerve injury (2 versus 1 per cent respectively). This study demonstrates that dissection in close proximity to the tumour is possible without inducing recurrence and that in practice the microinvasion of the capsule by tumour buds has limited clinical significance.

Between 1947 and 1992, 1403 patients with 1432 salivary gland tumours were treated at the Christie Hospital, Manchester. There were 1194 epithelial neoplasms: parotid, 1082 (91 per cent); submandibular, 47 (4 per cent); minor glands, 65 (5 per cent). The commonest histological diagnoses were pleomorphic adenoma (n = 776) and adenolymphoma (n = 159). A total of 244 carcinomas were seen (adenoid cystic carcinoma, n = 75). Treatment was primarily surgical, conservative where possible, and determined by tumour extent and not histology. Adjuvant radiation therapy was used in over half the definitively treated malignancies. The recurrence rate following the treatment of 551 new parotid pleomorphic adenomas was 1.6 per cent at median follow-up 12.5 (range 1-34) years, increasing to 15 per cent in the secondarily referred group (n = 170). For patients with definitively treated primary salivary carcinomas (n = 148), the disease-free survival rate at 5, 10 and 15 years was 58, 47 and 45 per cent respectively. Using multivariate analysis, clinical stage was the most important predictor of survival; the 10-year survival rate for stages I-IV was 96, 70, 47 and 19 per cent respectively.

No abstract

A clinicopathological survey of tumours of salivary glands seen in a special surgical clinic at the Christie Hospital and Holt Radium Institute is reported. Nine hundred and seventy-seven benign and malignant tumours have been seen within the past 30 years. The results of treatment are included with particular attention to those of the parotid glands which present the greatest therapeutic problem. The approach to surgical treatment of pleomorphic adenomas is described in detail and guidelines offered as to the procedure that should be used based on full exploration and assessment of each tumour. The need for adjuvant irradiation depends on the standard of surgery that is done. The relationship between carcinomas and benign adenomas is discussed and the need for surgery in all cases is stressed in order to make a definite histological diagnosis in view of the wide range of pathological abnormality encountered.

Carcinomas of the parotid gland are relatively uncommon malignancies (0.6 per 10 5 population; HMSO, 1997). They are frequently characterized by a long natural history and, consequently, their study is rendered difficult both by the time taken to accrue sufficient patients for analysis and the fact that at least 10-year follow-up is required to adequately assess treatment outcome (Spiro, 1986). Prospective randomized trials are usually impractical in such circumstances. Consensus on treatment can only be obtained from analysis of large retrospective studies in which multivariate analysis is the most appropriate method of evaluation. The purpose of this study is to identify important prognostic and treatment-related factors that influence survival in parotid cancer. For purposes of comparison, factors critical to locoregional recurrence and distant metastasis will also be examined. PATIENTS AND METHODSBetween 1952 and 1992, a total of 825 patients with previously untreated parotid neoplasms were surgically treated at the Christie Hospital, Manchester and previously summarized by the authors . A histological diagnosis of carcinoma was established in 143 patients (16%), of which 40 patients were considered incurable at diagnosis and treated palliatively. The remaining 103 formed the focus of this study. Survival data and disease status has been evaluated to 1997, such that all survivors were followed for a minimum of 5 years (median 12, range 5-32).There were 52 males (median age: 55; range 9-85 years) and 51 females (median age: 62; range 10-92 years). The median duration of symptoms was 10 (range 1-300) months. Tumours frequently presented as clinically 'benign' lumps with frank malignant features seen in only 42 (41%) patients (indurated tumour, 39; facial nerve palsy, 12; including nine with both). Cervical lymphadenopathy was uncommon at presentation (8%) and notably was not documented in patients with adenoid cystic carcinoma (Fisher's exact test; P < 0.0001).Tumours were staged retrospectively in accordance with AJCC (Fleming et al, 1997). Histological classification was reviewed in the mid-1970s and updated in line with the criteria of WHO (Seifert and Sobin, 1991). Tumour types were assigned to three grades representing different levels of biological behaviour. Mucoepidermoid and adenocarcinoma were subclassified into low-and high-grade based on histological features, while the remaining tumour types were assigned as reflected by their natural history (Table 1). This clinical grading system is similar to those proposed by the AFIP (Ellis et al, 1991) and other major institutes (Spiro, 1986;Kane et al, 1991).All patients were treated surgically by two consecutive surgeons ensuring continuity in management (late WAB Nicholson 1956-1972 ENG 1973 ENG -1992. Twenty-four (23%) underwent local extracapsular dissection , 45 (44%) formal parotidectomy with nerve identification (superficial, 34; total, 11) and 34 (33%) a more radical approach (partial nerve sacrifice, 12; total nerve sacrifice, six; extended parot...