The role of chemoradiotherapy after induction in locally advanced pancreatic cancer: lessons learnt from CONKO-007

Improving patient outcomes by converting locally advanced pancreatic cancer (LAPC) to resectable disease remains a top priority. Compared to chemotherapy alone, surgery offers a survival advantage and remains the only curative option (1). Less than 40% of the LAPC patients who receive neoadjuvant chemotherapy will proceed to surgery (2,3). In this context, chemoradiotherapy (CRT) has been proposed as an option for better local control and improved R0 resection rates. Nevertheless, across trials in localised pancreatic cancer, adding CRT to chemotherapy has not led to improved survival.

The LAP07 trial compared CRT vs. chemotherapy alone after induction and found that the CRT group had longer disease-free survival (DFS) and better local tumour control, but no overall survival (OS) benefit (4). Because of the use of less effective chemotherapy regimens and outdated radiation techniques, the potential survival advantage of CRT may have been underestimated. However, despite these concerns, neither LAP07 nor subsequent randomised trials using more contemporary chemotherapy have demonstrated a survival benefit for the addition of radiotherapy, suggesting that the above factors alone are unlikely to explain the absence of an OS effect (5,6). Furthermore, CRT did not improve OS over chemotherapy, according to a meta-analysis of 41 studies (7). In this regard, the CONKO-007 trial (8) sought to investigate the advantages of CRT following a chemotherapy induction period as opposed to continuing with the same chemotherapy regimen.

The CONKO-007 trial is a prospective, open-label, phase III trial conducted across 54 centres in Germany from 2013 to 2021, enrolling 525 patients with LAPC. After induction with either FOLFIRINOX (5-fluorouracil, leucovorin, irinotecan, oxaliplatin) or gemcitabine, 336 patients without disease progression were randomised to either to continue chemotherapy or receive 50.4 Gy of radiotherapy with concurrent gemcitabine. While resectability was assessed centrally, the decision to proceed with surgery was not randomised. Due to slow patient accrual, the authors changed OS to R0 resection rate as their primary outcome.

The R0 resection rate for all randomised patients was not significantly different between the two arms, with a rate of 25% in the CRT arm versus 18% in the chemotherapy arm (P=0.113) on intention-to-treat analysis. The rate of patients undergoing surgical exploration was comparable in both groups (P=0.91). The R0 resection rate in those patients who did proceed to surgery was significantly higher after CRT, at 69.4%, compared with 50.0% in those treated with chemotherapy alone (P=0.04). Patients in the CRT arm experienced higher rates of grade 3 or 4 haematologic toxicities, specifically leukopenia and thrombocytopenia.

CONKO-007 is a well-planned study with a pragmatic approach. The authors disclosed the change in the primary outcome, adhered to the CONSORT guidelines and comprehensively reported outcomes, including toxicity and patterns of failure. The trial complied with contemporary ethical guidelines, and the statistical analysis was appropriate.

However, the study has limitations, including the change of the primary outcome from OS to the R0 resection rate. The sample size recalculation included unblinded data from 180 early patients, which may introduce selection bias, may give false confidence in the results, and may result in an overestimation of the effect of CRT in the subgroups. Decisions about surgical exploration or the reporting of toxicity and other adverse outcomes may have been impacted by the open-label strategy. More patients were eliminated from the gemcitabine group as a result of using different induction chemotherapy regimens (47% vs. 29% in the FOLFIRINOX group; P<0.001), which could introduce confounding, as FOLFIRINOX is known to be more effective than gemcitabine (9). Interestingly the recently published PREOPANC-2 trial did not show a difference in OS between neoadjuvant FOLFIRINOX and neoadjuvant gemcitabine-based CRT in patients with resectable or borderline resectable pancreatic duct adenocarcinoma (10). Even in the case of resectable or borderline disease, the gemcitabine-based CRT has not been shown to be better than the modern chemotherapy alone in terms of survival.

The choice of surgical exploration was not randomised consequently, analyses restricted to the patients who had surgery lost the advantage of randomisation and became susceptible to selection bias. Patients who had surgical exploration after CRT may have had more favourable tumour biology that led to higher rates of complete pathologic response, or other unregistered effect modifiers compared to the patients in the chemotherapy-alone group, irrespective of any real effect of radiation.

Moreover, the relatively high conversion rate to resection in both groups (11) indicates that this cohort might not be representative of the spectrum of LAPC encountered in normal practice. For these reasons, the increased R0 rate in the subgroup of resected cases should be considered as hypothesis-generating rather than conclusive proof that CRT boosts resectability.

Furthermore, the study was underpowered for OS, and using interim data risked it being underpowered for the R0 resection rate as well. Multiple subgroup analyses were performed without adjustment for multiplicity, increasing the risk of false-positive findings. For instance, the locoregional control benefit might not stand after correction. The CRT regimen used gemcitabine, which is arguably outdated compared to capecitabine or S-1. Nab-paclitaxel/gemcitabine, which is one of the first-line treatments, was not used (12). Finally, because the study was conducted exclusively in Germany, its external validity to more diverse populations may be limited.

Nevertheless, this study adds important evidence on a critical topic. It appears that external beam radiation therapy (EBRT) should not be considered for routine use in patients with LAPC after induction chemotherapy, as it does not improve OS, the overall R0 rate, or the proportion of patients undergoing surgery. Although CRT has been utilised with the hope of improving locoregional disease control with higher R0 resection and negative margin rates, survival is primarily driven by distant recurrence (13,14). Thus, in selected cases, CRT may influence local disease behaviour and margin status, however, its use has not so far resulted in a measurable survival benefit.

The added haematologic toxicity from CRT brings to light the need for quality of life to be taken into consideration on par with the traditional oncologic endpoints.

Looking ahead, patient selection that is more nuanced will probably be very important if any role of CRT is to be determined at all. Only anatomical factors may not be enough to spot the truly chemo(radio)sensitive tumours. Radiomics (15,16) and genomics (17) research along with molecular profiling indicate that some pancreatic cancers might be markedly different with regard to their sensitivity to chemotherapy and radiation. Clinical trials that compare these highly selected subgroups of patients can investigate if certain patients get a real survival benefit from the increased local treatment. But until such data come along, randomised evidence continues to support the idea of prioritising effective systemic therapy. CONKO‑007 has thus become a part of the changing storyline about the role of CRT in pancreatic cancer, by pointing out its restrictions and by stressing the importance of systemic disease control, trial design, and patient selection. Future studies investigating the role of CRT should focus on novel locoregional treatments such as stereotactic body radiation therapy in combination with newer more potent chemotherapeutic regimes. In practical terms, clinicians should prioritise delivery of the most effective systemic chemotherapy, consider CRT only when the potential benefits and risks have been carefully weighed at an individual patient level, and actively support ongoing radio(geno)mics research aimed at identifying those groups of patients who might benefit from intensified locoregional treatment.

Supplementary

The article’s supplementary files as

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.