In the UK, chemotherapy is the standard treatment for inoperable, locally advanced, non-metastatic pancreatic cancer. Chemoradiotherapy is also an acceptable treatment option, for which gemcitabine, fluorouracil, or capecitabine can be used as concurrent chemotherapy agents. We aimed to assess the activity, safety, and feasibility of both gemcitabine-based and capecitabine-based chemoradiotherapy after induction chemotherapy for patients with locally advanced pancreatic cancer.

In this open-label, randomised, two-arm, phase 2 trial, patients aged 18 years or older with histologically proven, locally advanced pancreatic cancer (with a tumour diameter of 7 cm or less) were recruited from 28 UK centres between Dec 24, 2009 and Oct 25, 2011. After 12 weeks of induction gemcitabine and capecitabine chemotherapy (three cycles of gemcitabine [1000 mg/m² on days 1, 8, 15 of a 28-day cycle] and capecitabine [830 mg/m² twice daily on days 1–21 of a 28-day cycle]), patients with stable or responding disease, tumour diameter of 6 cm or less, and WHO performance status 0–1 were randomly assigned to receive a further cycle of gemcitabine and capecitabine chemotherapy followed by either gemcitabine (300 mg/m² once per week) or capecitabine (830 mg/m² twice daily, Monday to Friday only), both in combination with radiation (50·4 Gy in 28 fractions). Randomisation (1:1) was done via a central computerised system and used stratified minimisation. The primary endpoint was 9-month progression-free survival, analysed by intention to treat including only those patients with valid CT assessments. This trial is registered with ISRCTN, number 96169987.

114 patients were registered and 74 were randomly allocated (38 to the gemcitabine group and 36 to the capecitabine group). After 9 months, 22 of 35 assessable patients (62·9%, 80% CI 50·6–73·9) in the capecitabine group and 18 of 35 assessable patients (51·4%, 39·4–63·4) in the gemcitabine group had not progressed. Median overall survival was 15·2 months (95% CI 13·9–19·2) in the capecitabine group and 13·4 months (95% CI 11·0–15·7) in the gemcitabine group (adjusted hazard ratio [HR] 0·39, 95% CI 0·18–0·81; p=0·012). 12-month overall survival was 79·2% (95% CI 61·1–89·5) in the capecitabine group and 64·2 (95% CI 46·4–77·5) in the gemcitabine group. Median progression-free survival was 12·0 months (95% CI 10·2–14·6) in the capecitabine group and 10·4 months (95% CI 8·9–12·5) in the gemcitabine group (adjusted HR 0·60, 95% CI 0·32–1·12; p=0·11). Eight patients in the capecitabine group had an objective response at 26 weeks, as did seven in the gemcitabine group. More patients in the gemcitabine group than in the capecitabine group had grade 3–4 haematological toxic effects (seven [18%] vs none, p=0·008) and non-haematological toxic effects (ten [26%] vs four [12%], p=0·12) during chemoradiation treatment; the most frequent events were leucopenia, neutropenia, and fatigue. Two patients in the capecitabine group progressed during the fourth cycle of induction chemotherapy. Of the 34 patients in the capecitabine group who received chemoradiotherapy, 25 (74%) received the full protocol dose of radiotherapy, compared with 26 (68%) of 38 patients in the gemcitabine group. Quality-of-life scores were not significantly different between the treatment groups.

Our results suggest that a capecitabine-based regimen might be preferable to a gemcitabine-based regimen in the context of consolidation chemoradiotherapy after a course of induction chemotherapy for locally advanced pancreatic cancer. However, these findings should be interpreted with caution because the difference in the primary endpoint was non-significant and the number of patients in the trial was small.

Cancer Research UK.