Purpose: Tumours arising from the pulmonary apex or the posterior mediastinum may be removed en bloc in combination with total or partial vertebrectomy in the event of spinal invasion. Extra-tumour resection is a perfectly described surgical procedure recognised for its carcinological effectiveness. Raymond Roy-Camille and Bertill Steiner described en bloc spinal resection of a thoracic vertebra via a simple posterior approach. Paulson, described resection of tumours of the pulmonary apex (Pancoast Tobias) via a cervicothoracic approach. We associated these two techniques to allow en bloc resection of posterior mediastinal tumors or pulmonary apex tumours associated with spinal invasion.

Material and methods: We recommend different surgical approaches to the cervicothoracic and mid thoracic spine. For the cervicothoracic spine, an anterior approach is used with simple dislocation of the sterno-clavian joint without resection of the clavicle. The subclavian vessels and the brachial plexus are dissected and exposed. The tumour is then dissected followed by the peripheral, particularly œsophageal, attachments. The tumour is not detached from the spine to which it adheres tightly. Conventional thoracotomy is used for the thoracic level with dissection of the tumour from the adjacent soft tissue. In the event of a tumour in the posterior mediastinum, the anterior time is followed by a posterior approach. For primary pulmonary tumours, lobectomy or segmentectomy, or even pneumonectomy, is performed during the anterior time. Total or partial vertebrectomy, depending on the level of spinal involvement, is performed during the posterior approach.

Results: Thirty-six patients underwent this type of procedure. Total vertebrectomy was necessary in seven patients, partial vertebrectomy in 29. Mean follow-up has been six days to 7.2 years (mean 23.3). One patient died during the postoperative period due to a cause unrelated to the tumour. Only 35 patients were retained for analysis. Twenty-one patients (60%) died after a mean survival of 16.7 months (8–44 months). The 14 others (40%) are living at a mean 38.2 months survival (8–87 months).

Discussion: This technique requires a long learning curve and an extremely careful evaluation of tumour extension. Results obtained to date confirm the feasibility of the technique and point out its contribution in case of wide resection. Certain patients have lived more than five years after resection.