The objective of this study was to evaluate the maximum tolerated dose (MTD) and recommended dose of irinotecan administered as a 5-day schedule synchronously with 5-fluorouracil (5FU), leucovorin (LV) and preoperative pelvic radiation (45 Gy) for primary borderline/unresectable, locally advanced rectal cancer. The study used escalating doses of intravenous irinotecan (6, 8, 10, 12, 14, 16, 18, and 20 mg m(-2)) administered on days 1-5 and 29-33 followed by low dose LV (20 mg m(-2)) and 5FU (350 mg m(-2) over 1 h) in sequential cohorts. Preoperative pelvic radiotherapy using a three- or four-field technique and megavoltage photons comprised 45 Gy given in 25 fractions, 1.8 Gy per fraction. Surgery in the form of mesorectal excision was performed 6-10 weeks later. Histopathological examination of the resected specimen was performed according to techniques of Quirke, and compared with clinical staging. A distance of 1 mm or less between the peripheral extent of the tumour and the radial resection margin defined an involved circumferential resection margin (CRM). The MTD was determined as the dose causing more than a third of patients to have a dose-limiting toxicity (DLT) defined as specific grade 3 or 4 toxicities. Once the MTD was reached, a further 14 patients were treated at the dose level below the MTD. In total, 57 patients received irinotecan at the eight dose levels. The final cohort reached DLT after only four patients had been enrolled. The median age was 62 years (range 26-75), 37 male and 20 female subjects. The MTD of irinotecan in this schedule was 20 mg m(-2) when three out of four patients experienced DLT. Dose limiting grade 3 or 4 diarrhoea was reported in seven out of 57 patients, three at the 20 mg m(-2) dose level. Serious haematological toxicity (grade 3) was minimal and reported in only three patients; one grade 3 neutropaenia, one grade 4 neutropaenia and one grade 3 febrile neutropaenia and anaemia. Compliance was good with 93 and 89% of patients completing radiotherapy and chemotherapy, respectively. The remaining patients had only minor deviations from protocol therapy. Eight patients did not proceed to surgery, in six cases because they remained unresectable or had developed metastatic disease, one patient was unfit for surgery and one died as a result of complications from radiotherapy. Forty-nine patients underwent a potentially curative surgical resection. Histopathological examination of the resected specimen demonstrated pCR 12 out of 49 (24%) and 12 out of 57 (21%) overall. A histologically confirmed clear circumferential resection margin (CRM) was achieved in 39 out of 49 (80%) of those resected, and 39 out of 57 (68%) overall. In conclusion, MTD with this scheduled regimen of irinotecan is 20 mg m(-2) (days 1-5 and 29-33). The acceptable toxicity and compliance at 18 mg m(-2) recommend testing this dose in future phase III studies. The tumour downstaging and complete resection rates (negative CRM) are encouragingly high for this very locally advanced group.