Purpose of the study: The objective was to assess the diagnostic yield of angioscanner evaluation of arterial invasion of limb tumours before surgery.

Material and methods: This was a prospective study conducted from January 2005 to May 2008 designed to assess 55 arterial segments and limb or pelvic tumours on the preoperative angioscan. Screen displays of the radiographic images were reread conjointly but radiologists using vascular analysis software. Contact of a large vessel with the tumour was classified as follows: fatty line, contact less than 90° of the vessel circumference, contact less than 180°, contact ≥ 180°. Intraoperative vessel-tumour contact was classified as follows: cleavable, non-dissectible. Agreement between the angioscan and the operative findings was analysed.

Results: Forty-five arteries were cleavable and ten non-dissectible. All arteries that were separated from the tumour by a fatty plane or were in contact over less than 90° of the circumference of the vessel were cleavable at surgery (n=37/37). The sensitivity, specificity and diagnostic accuracy and the rate of positive scan for probable arterial contact > 180° were 90, 93, 93 and 13.5% respectively.

Conclusion: The angioscanner provides a satisfactory means for predicting vascular invasion of limb and pelvic tumours and contributes pertinently to the pre-operative work-up.