Erratum

After the publication of this work [1], it was brought to our attention that the time for patient follow-up specified in the ‘Statistical methods’ section of the protocol was incorrect; follow-up will continue for six months after the end of treatment, not 12 months as stated in the original published version.