Arterial resections in pancreatic cancer: measure twice, cut once

Correspondence to: Monish Karunakaran (email: monish3682@gmail.com)

Department of Surgical Gastroenterology

Asian Institute of Gastroenterology

Mindspace Road, Gachibowli. Hyderabad

India 500032

BJS Open, https://doi.org/10.1093/bjsopen/zraf026, published 17 April 2025

Compared to the venous frontier, arterial involvement in pancreatic cancer has proven to be a much taller peak to scale, with dismal long-term outcomes a glaring testimony of the magnitude of the problem. The study by Hirose et al. 1 captures an impressive improvement in outcomes following arterial resection and divestment in pancreatic cancer, both in terms of major morbidity. Nevertheless, are the results prudent enough to justify or advocate pancreatectomy with arterial resections (PAR)?

Out of 584 patients, 203 (34.8%) underwent surgery. The stringent selection criteria with consequent bias needs greater acknowledgement. As the study documents significantly increased use of neoadjuvant therapy (NAT) in the latter interval (26.6% versus 95.2%) a detailed description of the exclusion criteria might help in discerning how much post-NAT case selection drove improvement in outcomes.

The current study reports an acceptable major morbidity (22.7%) and 90-day mortality (0.5%), and a significant improvement in the major morbidity in the latter era. In comparison, historical cohorts like Bachellier et al.2 (5.1%) and Loos et al.3 (8.8%) reported much higher perioperative risk, understandably because of greater inclusion of more complex resections (e.g. superior mesenteric artery (SMA) resections), which inflates the safety of PAR in the current study. Further proof to this phenomenon comes from comparing pooled perioperative mortality of arterial resections of any type4 (5%) with that of SMA resections alone5 (20%).

A 33-month median follow-up, in general, would be insufficient for conclusive 5-year survival conclusions. Five-year survival estimates using Kaplan-Meier analysis can be immature and potentially over-optimistic, which need to be interpreted with caution.

There is a significant increase in the utilization of NAT in the latter era (95 versus 26.6%), which would have been instrumental in lower pT, pN, M stages and R1 resection rates. The discussion seems to underplay the role of multiagent chemotherapy in improving the overall survival (OS) (26.0 versus 48.2 months). However, the multivariable analysis includes surgery in the latter interval as a favourable factor, while it is quite possible that preoperative chemo might have downstaged the tumours. This is not adequately adjusted in the survival analysis, and the era variable might just be a proxy for chemotherapy utilization.

Furthermore, most latter era patients received S-1 adjuvant chemo, which might be more effective, but this isn't explored in depth. The improved overall survival/post-recurrence survival but unchanged recurrence-free survival suggests post-recurrence therapies might have driven the survival gains. However, the discussion does not address this, leaving a critical gap in interpreting outcomes. It is notable that more than one-third of the patients did have an early recurrence even in the latter era. Besides, almost 20% of patients failed to receive adjuvant chemotherapy, an independent predictor of early recurrence.

A circumspective view of the PAR literature suggests that punctilious patient selection, modern NAT and R0 resection are paramount, apart from performance by highly-trained surgeons at high-volume centres. Multicentre studies should generate artery specific data. Biological and conditional resectability criteria needs to be conscientiously fulfilled prior to resections in anatomical borderline resectable pancreatic cancer / locally advanced pancreatic cancer to minimize likelihood of futile pancreatectomy6. The pragmatic pancreatic surgeon needs to leverage precision oncology to identify patients most likely to benefit from PAR, combining this with NAT to maximize survival.