Kjetil S?reide ,

a Department of Gastrointestinal Surgery, HPB Unit, Stavanger University Hospital, P.O. Box 8100, Stavanger N-4068, Norway

b Department of Clinical Medicine, University of Bergen, Bergen, Norway

Pancreatic cancer has an overall dismal prognosis compared to most other malignancies. In general, only about 15%-20% of patients are deemed upfront resectable at time of diagnosis, with a similar proportion presenting with either borderline or locally advanced disease [1] . Novel and more effective treatment regimens including FOLFIRINOX have made yet more patients become resectable, with up to 60% reported in some centers [2] .Technical advances in surgery continue to literally explore new anatomical territory [3] . More aggressive attitude towards resection of involved vessels has provided opportunity for curative attempt resections, even for a subpopulation of biological responders staged with pre-treatment unresectable disease [ 4 , 5 ]. However, while surgery is the dominant modality for a potential curative approach to pancreatic cancer, there is a subgroup of patients scheduled for surgery who ends up with an aborted resection during explorative laparotomy, also referred to as an “openclose laparotomy”. The reasons for such non-resection events are manyfold and have likely changed over time. Indeed, in one Italian study the non-resection rate remained constant at about 25%over two decades [6] . While many institutional series report nonresection rates in the same range, these figures may be influenced by patient selection, referral patterns and institutional policies towards resection and surgical aggressiveness [ 2 , 4 , 6 ].

Thus, the study from Sweden [7] in the current issue is a timely investigation into the contemporary patterns of non-resection after surgical exploration in a nationwide population. In the study, Andersson et al. [7] looked at non-resection trends during laparotomy in 1938 patients with pancreatic cancer from 2010 to 2018. The non-resection rate during surgery was 20.6%. Among the 399 openclose laparotomies, the most common cause was metastatic disease(58.6%) with the remaining events caused by locally advanced disease. The preoperative factors that were most notably related to the risk of non-resection due to metastatic disease were involuntary weight-loss before surgery and increased carbohydrate antigen 19-9 (CA19-9) levels. No cutoff level was provided in the Swedish study, but others have suggested that CA19-9 values ≥150 U/mL to be associated with computed tomography (CT)-occult metastatic disease found at time of surgical exploration [8] . A single-center,observational study from Japan found borderline resectable tumors,pre-treatment CA19-9 (with cut-off value at 260 ng/mL) and tumor size as risk factors associated with non-resectability status at surgical exploration [9] . Unfortunately, the subdivision of preoperative image-based categories of resectable, borderline or locally advanced tumors is not given in the Swedish study [7] , hence one cannot relate the preoperative risk factors to tumor imaging status in this study. Also, data on the use of neoadjuvant treatment are not included, which may potentially influence the selection to surgical exploration and thus the non-resection rate.

A Dutch nationwide study [10] covering a similar, but shorter time-period (2009-2013) had a 38.4% non-resection rate that decreased to 28.7% at the later time of the study. Occult metastatic disease diagnosed on surgical exploration remained unchanged at 18.5% during the study period. A subsequent study found that resection rates had increased and were similar across hospital types in the Netherlands [11] . Another nationwide study from Italy found that the probability of undergoing palliative/explorative surgery was inversely related to hospital volume, being 24.4% in very highvolume hospitals and 62.5% in very low-volume centers [12] .

In the past, contribution to non-resection operations may have been the result of preoperative suboptimal cross-sectional image quality or long delays (e.g.＞4 weeks) between the imaging studies performed and scheduled surgery. However, with current state of the art three-phase contrast CT, the rate of “surprise” findings in upfront resectable patients has decreased [13] . Nonetheless, the debate continues over what additional imaging modalities may be necessary to increase sensitivity for detecting advanced disease prior to explorative laparotomy [14] . Some have advocated the use of liver-specific magnetic resonance imaging (MRI) or use of positron emission tomography (PET) scans for better sensitivity. A Canadian study found that with the use of routine preoperative MRI as part of staging, an incremental 7.6% of patients were excluded from surgery with a potential reduction of up to 13.6%in futile open-close laparotomies due to liver metastases detected on MRI only [15] . While not investigating non-resectabilityperse,others have looked at the metabolic response by pre- and postchemotherapy PET scans to predict relation to survival [16] . It remains to be demonstrated if any subsequent imaging modality,such as MRI or PET, will have an additional gain in staging as there are variations in both sensitivity and costs to each modality [14] ,with PET-CT at an almost 3 times higher price compared to MRI.

Survival is affected by the non-resection rate, and is worst in patients who have metastatic disease (survival＜7 months) [7] .An argument brought forward by some for surgical exploration,is the opportunity to provide the patient with a surgical bypass if resection cannot be done. However, a study found worse survival when a bypass procedure was added over laparotomy alone if chemotherapy could be ensued [17] . Indeed, this emphasizes the need to avoid unnecessary surgery for those unlikely to be resectable or have occult metastatic disease, and the attempt of palliative intervention by least invasive means possible, sparing a surgical double bypass for a select few patients at best [18] .

Despite the progress being made in diagnosis and staging, the surgical exploration with a non-resection as the outcome still happens in about 1 in 5 laparotomies [ 6 , 7 , 10 ]. Some preoperative factors, such as weight loss and elevated CA19-9, may be indicative of non-resectability from occult diseases [7–9] . However, as more patients are subjected to neoadjuvant treatment, including FOLFIRNOX, new studies are needed to explore what preoperative factors are associated with non-resectability to the currently defined subgroups of resectable, borderline and locally advanced pancreatic cancers.

Acknowledgments

None.

CRediT authorship contribution statement

Kjetil S?reide : Conceptualization, Writing original draft, Writing review & editing.

Funding

None.

Ethical approval

Not needed.

Competing interest

No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.