Towards a 1% measurement of the Lense-Thirring effect with LARES?