The LAP 07 trial was an international multicentre, open-label, unblinded, randomised phase 3 study of patients with locally advanced pancreatic cancer (LAPC).r The study was stratified to centre and performance status and used a 2-step randomisation process. The first step evaluated gemcitabine vs gemcitabine plus erlotinib as induction CT for 4 months. Patients with a tumour which was controlled proceeded to a second randomisation of either CT or CRT. The CRT involved 54 Gy in 30 fractions plus capecitabine (800 mg/m2 twice daily on days of RT). The RT planning involved 3D-CRT with a contrast enhanced planning CT. The planning target volume (PTV) margins used were 3cm craniocaudally and 1.5cm in the other directions.
OS was not significantly different between CT (16.5 months) vs. CRT (15.2 months) (p=0.83). However, there were differences observed in time to reinitiation of therapy for CRT (159 days) vs CT (96 days) (p=0.05) and differences in local tumour progression for CRT (32%) vs CT (46%) (p=0.0001). The reasoning for induction therapy was to spare patients with rapidly progressive disease, while also selecting those that could potentially benefit from CRT, and was supported by several retrospective studies. The trial had better median survival than other prospective studies (13 and 11 months). This suggests that induction CT selected patients with better prognosis and those excluded before second randomisation had median survival of 7.7 months. The secondary surgical resection rate was only 4% in this study.
RT compliance may have contributed to the study outcomes, with only 32% of patients in the CRT arm treated per protocol. Toxicity was not substantially different between arms, although the study did not include patient-reported toxicity.
Loehrer et al. 2011 reported on the ECOG-4201 study which compared CT alone vs CRT in patients with localised unresectable pancreatic cancer.r The prescribed dose was 50.4 Gy in 28 fractions with concurrent gemcitabine (1000 mg/m2 IV weekly). RT was delivered in two phases; phase one delivered 39.6 Gy in 22 fractions to initial large fields and phase two followed to deliver a total dose of 50.4 Gy to a reduced field (GTV + 2 cm margin).
The study was severely limited by poor accrual, enrolling only 74 of the planned 316 patients, resulting in poor statistical power. However, intention to treat analysis showed median survival was 9.2 months (CT) compared to 11.1 months (CRT) (p= 0.017). There was no difference in PFS. This may be due to scans not being performed at adequate intervals, challenging precise measurement, and that this study was primarily designed to detect OS. There were no significant differences in quality of life between the two treatment arms after week 6 of treatment.
The phase III FFCD/SFRO 2008 trial involved 119 patients with LAPC.r Patients were randomised either to induction gemcitabine alone or induction gemcitabine followed by CRT (60 Gy in 30 fractions and 5-FU infusion [300mg/m2/day, days 1-5] and cisplatin). Both arms were then followed by maintenance gemcitabine until progression or toxicity. Radiotherapy planning used a PTV expansion of the clinical target volume (CTV) + 2cm margin in all directions.
OS was significantly reduced in the CRT arm, with a median of 8.6 months compared to 13 months in the gemcitabine alone arm (p=0.03). More grade 3-4 toxicities were observed in the CRT arm. Patients in this group received fewer courses of maintenance therapy with gemcitabine. The survival difference became apparent after 8 months, suggesting that the toxicity did not directly contribute to death, but perhaps led to reduced tolerance for maintenance gemcitabine and that the RT dose used may have been too intense.
The ECOG E8282 trial was a phase III trial with 114 patients.r It showed the addition of 5-FU and mitomycin C to RT alone (59.4 Gy in 1.8 Gy fractions) increased toxicity without improving disease free survival (DFS) or OS in patients with LAPC.
Phase II trials
Mukherjee et al. reported on SCALOP, a phase II trial including 74 patients that compared capecitabine (n = 36) and gemcitabine (n = 38).r All patients received induction chemotherapy (gemcitabine and capecitabine). Patients with responding or stable disease (based on RECIST criteria), tumour diameter ≤6cm and WHO performance status 0-1 were randomised to be given a further cycle of gemcitabine and capecitabine followed by CRT (50.4Gy in 28 fractions) with either gemcitabine or capecitabine. Of the 34 patients in the capecitabine group who received CRT, 25 (74%) received the full protocol dose of RT, compared with 26 (68%) patients in the gemcitabine group. In the capecitabine vs gemcitabine group there was improved median survival (15.2 vs 13.4 months, respectively) and 12 month OS (79.2% vs 64.2%, respectively).
CALGB 89805 was a phase II trial of gemcitabine and concurrent RT (50.4Gy/28fx).r In this trial, 39 patients without disease progression were given weekly gemcitabine for five cycles. The study reported good local regional control but no survival advantage (median survival 8.2 months). Whilst this trial confirmed feasibility, it also demonstrated that treatment is still relatively toxic and gemcitabine has not been widely adopted.
Brunner et al. 2002 reported on a phase II trial of 29 patients given pre-operative CRT delivering 50.4Gy in 28 fractions and a 5.4Gy boost (in 3 fractions) with 5-FU and mitomycin C.r They reported a resectability rate of 37% (n=10), OS of 50% (resected patients) vs 6% (non-resected patients) and a median OS of 9 months. No significant acute toxicity was reported.
Borderline resectable and resectable pancreatic cancer
Murphy et al. 2019 reported on a phase II trial of total neoadjuvant therapy with FOLFIRINOX followed by individualised chemotherapy for borderline resectable pancreatic adenocarcinoma (BRPC) (n=49).r Patients were given 8 cycles of FOLFIRINOX and were then restaged. Patients with resolution of vascular involvement received short-course CRT (25 Gy delivered in 5 Gy fractions using protons) with capecitabine. Patients with persistent vascular involvement received long-course CRT with 5-FU or capecitabine. The R0 resection rate was reported as 61%.
Results are pending for the PREOPANC trial which randomised 246 patients with borderline resectable pancreatic cancer to either upfront surgery or preoperative CRT (36 Gy in 15 fractions) with gemcitabine (1000 mg/m2) preceded and followed by 2 doses of gemcitabine.r The RT dose was based on an earlier phase II dose escalation trial.r RT involved simulation with 4DCT and IV contrast. There was no elective lymph node irradiation. Full dose CT was standard.
Jang et al. 2018 reported on the first prospective randomised control trial (RCT) evaluating the oncological benefits of a neoadjuvant strategy.r In this phase II/III multi-centre RCT, 110 patients with BRPC were to be randomised to neoadjuvant CRT (54 Gy) followed by surgery or upfront surgery followed by CRT in four large-volume centres in Korea. The trial accrued 58 patients and was terminated early due to statistical significance. In the ITT analysis, the 2 year survival and median survival was significantly better in the neoadjuvant CRT than the upfront surgery group (40.7%, 21 months vs 26.1%, 12 months with a hazard ratio of 1.495 [95% confidence interval 0.66–3.36] (p= 0.028)). R0 resection rate was also significantly higher in the neoadjuvant CRTgroup than upfront surgery (51.8% vs 26.1%, P = 0.004).
In contrast Golcher et al. 2015 attempted to conduct the first prospective RCT in patients with resectable pancreatic head adenocarcinoma.r Of a planned 254 patients, the trial was ceased early due to poor accrual with 73 patients accrued. Registered patients were randomised to primary surgery or neoadjuvant CRT followed by surgery. Adjuvant CT was delivered in both arms. RT compliance was 100%. R0 resection rate was 48% in the surgery only arm and 52% in the neoadjuvant CRT arm (p=0.81). After resection, median OS was 18.9 vs. 25.0 months, respectively (p=0.79). It showed that neoadjuvant CRT is safe with respect to toxicity, perioperative morbidity and mortality.