Biomodulation of irinotecan using ciclosporin: Results of PICCOLO, a randomized controlled trial in advanced colorectal cancer (aCRC).

Abstract

3566 Background: Phase I/II data showed that ciclosporin (Cs) profoundly reduces the hepatobiliary clearance of irinotecan (Ir), and suggested that 40% standard dose Ir with concurrent Cs (IrCs) may give equal efficacy to standard Ir, with less diarrhoea. PICCOLO is a multicentre randomized controlled trial in which patients (pts) with aCRC, progressing after prior therapy, were randomised to Ir or IrCs. It was designed to detect non-inferior efficacy with reduced toxicity. Methods: Eligible pts had: measurable aCRC, progressive during/after ≥1 prior regimen with fluoropyrimidines ± other drugs; no prior Ir; WHO PS 0-2. Before summer 2008, pts were allocated 1:1:1 (without prospective KRAS analysis) to Ir, IrCs or Ir/panitumumab. Thereafter, KRAS was analysed prospectively and pts with mutated or undetermined KRAS were allocated 1:1 to Ir (irinotecan 350 mg/m2 [300 mg/m2 if ≥70 yrs or PS2], q3w), or IrCs (irinotecan 140 mg/m2 [120 mg/m2 if ≥70 yrs or PS2] d1, Cs 3 mg/kg tds d0-2, q3w). The primary efficacy endpoint was % pts alive and progression-free at 12 wks, with a lower non-inferiority boundary of -10.6% (80% power, 2.5% significance). Grade ≥3 diarrhoea within 12 wks was a key secondary endpoint. PFS, OS, RR, QoL and loperamide use were also assessed. Analysis was performed on both the ITT and per-protocol populations. Results: 672 pts were randomised between Ir and IrCs. Median age was 64 yrs; 92% were PS0/1; 95% had received prior oxaliplatin; KRAS was mut in 60%, wt in 25%, unknown in 15%. ITT analysis showed 179/335 (53.4%) patients to be progression-free at 12 wks in the Ir arm, 159/337 (47.2%) in the IrCs arm. The difference was therefore -6.3% and the 95% CI crossed the prespecified non-inferiority margin (13.8%, +1.3%). Gr ≥3 diarrhoea was uncommon in both arms (Ir 12.2%; IrCs 10.1% pts), but fewer IrCs patients took loperamide within 12 wks (Ir 67.8%; IrCs 52.2%, p<0.001). Conclusions: As anticipated, IrCs was associated with less diarrhoea as assessed by loperamide use, but severe diarrhoea was uncommon in both arms. However, we failed to prove non-inferiority of IrCs compared with Ir, so cannot recommend it as a standard treatment option based on these data.