0mgm�C2 and higher that are thought to represent an active dose range and have been used across the spectrum of indications in phase II or III trial setting, the response rate may be even higher (4 out of 21 patients; 19%). Further, the disease control rate with 20MI, single-agent patupilone was 58%, and long-lasting disease stabilisation was observed never at doses as low as 6.5mgm�C2, suggesting activity throughout the dose range tested. The median TTP of 4.3 months in the 20MI arm also compares favourably with other second-line agents. It has been demonstrated in patients with mCRC that response rate and PFS are valid surrogates of overall survival (Tang et al, 2007). As the survival of mCRC patients has been shown to correlate with the number of active agents available (Grothey et al, 2004), the potential of patupilone in this disease should be further explored.

Although PRs were observed at higher doses (9.0 and 10.0mgm�C2), so too was CID, resulting in potentially more dose adjustments/interruptions. Therefore, lower doses such as 8.0mgm�C2 may provide clinical efficacy and be well tolerated, potentially providing a more favourable toxicity/efficacy profile; these could be considered for future studies in this indication. The promising activity of patupilone observed in the present trial contrasts with the lack of efficacy that was reported in patients with mCRC for another epothilone B analogue, ixabepilone (Eng et al, 2004). Compared with patupilone, ixabepilone is more water soluble, but also less cytotoxic (Fumoleau et al, 2007).

The results of the present trial of patupilone and the phase II trial of ixabepilone demonstrate that both drugs not only may differ in the activity in patients with mCRC, but also have differences in the spectrum of side effects. In the present trial, the tolerability and MTD of patupilone administered every 3 weeks was assessed using three different infusion schedules. The 5-day 16-h infusion elicited DLTs at the lowest dose tested, 6.5mgm�C2, and further exploration was stopped after the first three patients. Higher doses were achieved in the CI-1D arm; however, several DLTs were observed beginning at 7.5mgm�C2 and no tumour responses were evident. Although the MTD as defined per protocol was not reached in any of the three arms, comparison of the different schedules indicates that short-term infusion administration may be superior in terms of tolerability, toxicity and anti-tumour activity with no DLTs detected, even at the maximum dose of 10.0mgm�C2. Together with the four confirmed responses, this suggests that short-term infusion GSK-3 could be the preferred administration schedule. The standard of care has significantly changed during the conduct of this trial.