Key Points

Question听 Are there differences in receiving recurrence-focused treatment and overall survival in patients with pancreatic ductal adenocarcinoma (PDAC) recurrence who received symptomatic follow-up vs routine imaging after pancreatic resection?

Findings听 In this international cross-sectional study including 333 patients from 33 centers, most patients received routine imaging after pancreatic resection. Receiving routine imaging was independently associated with receiving recurrence-focused treatment and with improved overall survival.

Meaning听 The findings of this study suggest that the use of routine imaging in the postoperative follow-up of patients with PDAC may be beneficial.

Importance听 International guidelines lack consistency in their recommendations regarding routine imaging in the follow-up after pancreatic resection for pancreatic ductal adenocarcinoma (PDAC). Consequently, follow-up strategies differ between centers worldwide.

Objective听 To compare clinical outcomes, including recurrence-focused treatment and survival, in patients with PDAC recurrence who received symptomatic follow-up or routine imaging after pancreatic resection in international centers affiliated with the European-African Hepato-Pancreato-Biliary Association (E-AHPBA).

Design, Setting, and Participants听 This was a prospective, international, cross-sectional study. Patients from a total of 33 E-AHPBA centers from 13 countries were included between 2020 and 2021. According to the predefined study protocol, patients who underwent PDAC resection and were diagnosed with disease recurrence were prospectively included. Patients were stratified according to postoperative follow-up strategy: symptomatic follow-up (ie, without routine imaging) or routine imaging.

Exposures听 Symptomatic follow-up or routine imaging in patients who underwent PDAC resection.

Main Outcomes and Measures听 Overall survival (OS) was estimated with Kaplan-Meier curves and compared using the log-rank test. To adjust for potential confounders, multivariable logistic regression was used to evaluate the association between follow-up strategy and recurrence-focused treatment. Multivariable Cox proportional hazard analysis was used to study the independent association between follow-up strategy and OS.

Results听 Overall, 333 patients (mean [SD] age, 65 [11] years; 184 male [55%]) with PDAC recurrence were included. Median (IQR) follow-up at time of analysis 2 years after inclusion of the last patient was 40 (30-58) months. Of the total cohort, 98 patients (29%) received symptomatic follow-up, and 235 patients (71%) received routine imaging. OS was 23 months (95% CI, 19-29 months) vs 28 months (95% CI, 24-30 months) in the groups who received symptomatic follow-up vs routine imaging, respectively (P鈥�=鈥�.01). Routine imaging was associated with receiving recurrence-focused treatment (adjusted odds ratio, 2.57; 95% CI, 1.22-5.41; P鈥�=鈥�.01) and prolonged OS (adjusted hazard ratio, 0.75; 95% CI, 0.56-.99; P鈥�=鈥�.04).

Conclusion and Relevance听 In this international, prospective, cross-sectional study, routine follow-up imaging after pancreatic resection for PDAC was independently associated with receiving recurrence-focused treatment and prolonged OS.

Pancreatic ductal adenocarcinoma (PDAC) is one of the most common causes of cancer-related death worldwide.¹ Pancreatic resection combined with neoadjuvant or adjuvant systemic treatment offers the best chance of long-term survival.^2,3 However, almost all patients develop local or distant PDAC recurrence within a few years, contributing to the dismal 5-year overall survival rate (OS) around 20% after pancreatic resection.^1,4-6 PDAC recurrence is typically diagnosed by imaging, which is conducted when patients develop clinical symptoms such as pain, fatigue, and weight loss. In asymptomatic patients, disease recurrence can be identified through elevation of the serum tumor marker carbohydrate antigen 19-9 (CA19-9) or by planned imaging.^7,8 It remains debated, however, whether routine use of these diagnostics leads to recurrence detection at an earlier stage and whether this increases eligibility for receiving additional recurrence-focused treatment or prolonged survival.

The European Society of Medical Oncology (ESMO) guidelines do not recommend routine use of diagnostics during postoperative follow-up, in the absence of evidence from prospective studies for a beneficial impact on survival.⁹ Additionally, routine diagnostics may cause psychological stress, which negatively impacts quality of life.¹⁰ Most country-specific guidelines adhere to the ESMO recommendations.^3,11-16 Nevertheless, management of PDAC recurrence is evolving.¹⁷ Systemic combination therapies such as 5-fluorouracil, leucovorin, irinotecan, and oxaliplatin (FOLFIRINOX) or gemcitabine/nab-paclitaxel, as well as local ablative therapies such as magnetic resonance (MR)鈥揼uided stereotactic body radiation therapy show promising results.^18-24 Hypothetically, more patients could benefit from these therapies if PDAC recurrence is detected at an early stage, when tumor burden is still low and patients have a good performance status.^25,26 This may improve survival.³ As a consequence, guidelines from countries such as France, the US, and Japan recommend the use of routine diagnostics after pancreatic resection for PDAC.^17,27-30

The lack of consensus and inconsistent recommendations in international guidelines regarding follow-up strategies highlight the need for further investigation of PDAC recurrence detection strategies and its impact on clinical outcomes. The aim of this study was to compare clinical outcomes, including recurrence-focused treatment and survival, in patients with PDAC recurrence who received either symptomatic follow-up or routine imaging in international centers affiliated with the European-African Hepato-Pancreato-Biliary Association (E-AHPBA).

According to the governing body Medical Research Ethics Committee United, this study does not fall under the scope of the Medical Research Involving Human Subject Act; therefore, informed consent was waived. This study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology () reporting guidelines.³¹

We performed a prospective, multicenter cross-sectional study. A total of 33 centers from 13 different countries (eFigure 1 in Supplement 1) responded positively to the invitation for study participation, which was disseminated among E-AHPBA centers by the E-AHPBA Scientific & Research Committee. According to the predefined study protocol, during the 1-year accrual period (September 2020-September 2021), patients were prospectively included by local investigators at the moment they were diagnosed with local and/or distant PDAC recurrence after a previous macroscopically radical PDAC resection. PDAC recurrence was pathologically proven or diagnosed based on imaging or elevated serum CA19-9 levels, preferably by consensus during a multidisciplinary meeting. Patients who underwent upfront pancreatic resection or resection after neoadjuvant treatment were included. Patients with incomplete data on postoperative follow-up strategy were excluded.

Main study outcomes were (1) overall survival (OS), defined as the time from start of neoadjuvant treatment or pancreatic resection to death or last follow-up; (2) progression-free survival (PFS), defined as the time from start of neoadjuvant treatment or pancreatic resection to PDAC recurrence diagnosis; (3) disease-free interval (DFI), defined as the time from pancreatic resection to PDAC recurrence diagnosis; and (4) postrecurrence survival (PRS), defined as the time from PDAC recurrence diagnosis to death or last follow-up. Other outcomes included patient and disease characteristics at time of recurrence diagnosis (ie, location of recurrence, presence of symptoms, performance status, and CA19-9 levels) and the type of recurrence-focused treatment.

Patients were stratified into 2 groups according to their postoperative follow-up strategy: (1) symptomatic follow-up or (2) routine imaging. Routine imaging was defined as any form of planned imaging procedures (ie, computed tomography [CT], positron emission tomography [PET]鈥揅T, or magnetic resonance imaging [MRI]) with a certain frequency. Prespecified options were every 4 to 6 weeks, 3 to 4 months, 6 months, or at another predetermined routine frequency. All other patients were included in the symptomatic follow-up group. For a sensitivity analysis, patients were stratified to 3 groups according to whether or not they had received routine serum CA19-9 measurements or routine imaging: (1) no routine diagnostic testing (ie, CA19-9 and imaging only on indication), (2) routine CA19-9 only (imaging only on indication), or (3) routine imaging with or without CA19-9 testing.

Predefined electronic case report forms were sent to the participating centers. Each center was responsible for their own data collection. Baseline data included age, sex, height, weight, Charlson Comorbidity Index (CCI) score, Eastern Cooperative Oncology Group Performance Status (ECOG-PS), American Society of Anesthesiologists (ASA) score, and serum CA19-9 levels. Body mass index was calculated using height and weight. A CCI score less than 3, an ASA score less than 3, and an ECOG-PS less than 2 were chosen as cutoff values to represent a good preoperative health status.^29,32 Data on participant race and ethnicity were not collected. The primary focus of this study was to compare clinical outcomes in patients with PDAC recurrence, and the variables of interest did not necessitate the collection of race and ethnicity data, as our research objectives were designed to be inclusive of all participants regardless of their racial or ethnic background.

Treatment characteristics included data on neoadjuvant or adjuvant systemic therapy and details of pancreatic resection. Pathology variables included tumor size, tumor differentiation, resection margin status, total number of lymph nodes resected, number of positive lymph nodes, pathological tumor, node, metastasis stage according to the 8th edition of the American Joint Committee for Cancer, and microscopic perineural and lymphovascular invasion.³³

Follow-up data included the use and interval of clinical evaluation, serum CA19-9 assessment, and any imaging procedures during the first 2 years after pancreatic resection, date of PDAC recurrence, PDAC recurrence location (ie, local only, liver only, lung only, multiple site, other isolated distant metastasis), presence of symptoms, ECOG-PS at PDAC recurrence diagnosis, secondary progression (ie, new PDAC recurrence site after initiation of recurrence treatment), administration and type of recurrence-focused treatment, vital status, and date of death or last follow-up.

Descriptive statistics were used to present patient, tumor, and treatment characteristics. In case of normal distribution, continuous variables were expressed as mean with SD and were compared using analysis of variance. Nonnormally distributed variables were expressed as median with IQR and were compared using the Mann-Whitney U test. Categorical values were presented as number and percentage and were compared using the 蠂² test. Missing baseline data were considered missing at random and were imputed with the iterative Markov chain Monte Carlo method (5 imputations, 10 iterations).³⁴

Median OS, PFS, DFI, and PRS in months with corresponding 95% CIs were estimated with Kaplan-Meier curves and compared using the log-rank test.

Multivariable logistic regression analysis was performed to assess whether follow-up strategy was independently associated with receiving recurrence-focused treatment, adjusting for potential confounders: age (continuous), sex (male vs female), CCI (鈮�3 vs <3), log CA19-9 at primary diagnosis (continuous), location primary tumor (tail vs head/neck), neoadjuvant therapy (yes vs no), tumor size (continuous), number of positive nodes (continuous), lymphovascular invasion (yes vs no), perineural invasion (yes vs no), resection margin (R1 vs R0), and adjuvant therapy (yes vs no). Outcomes were presented as odds ratios (ORs) with corresponding 95% CIs.

Multivariable Cox proportional hazard analysis was used to evaluate whether follow-up strategy was independently associated with OS, PFS, DFI, and PRS, adjusted for the same potential confounders. We performed a frailty model with a 纬-distributed random effect for center to account for potential heterogeneity across centers. Results were presented using hazard ratios (HRs) with corresponding 95% CIs.

A 2-tailed P value <.05 was considered statistically significant. Statistical analysis was performed using R, version 3.5.1 (Bell Laboratories) with the survival, ggplot2, arsenal, and mice packages.

A total of 342 patients with PDAC recurrence were prospectively included during the 1-year study period. Nine patients (3%) were excluded due to missing data on follow-up strategy. Of the remaining 333 patients (mean [SD] age, 65 [11] years; 149 female [45%]; 184 male [55%]), 98 patients (29%) received symptomatic follow-up, and 235 patients (71%) received routine imaging (eTable 1 in Supplement 1). In the symptomatic follow-up group, 24 patients (24%) received routine CA19-9 testing. In the routine-imaging group, CT was used as the main imaging modality in 217 patients (92%), PET-CT in 4 patients (2%), and MRI in 14 patients (6%). The database was locked in October 2023, 2 years after inclusion of the last patient. Median (IQR) follow-up after resection was 40 (30-58) months.

Baseline characteristics are provided in Table 1. Median (IQR) CA19-9 level at first diagnosis of PDAC was 308 (140-775) U/mL in patients with symptomatic follow-up, as compared with 131 (48-532) U/mL in patients with routine imaging (P鈥�=鈥�.02). Neoadjuvant therapy was administered in 25 patients (26%) with symptomatic follow-up and 91 patients (39%) with routine imaging (P鈥�=鈥�.02). In the symptomatic follow-up group, an R0 resection was achieved in 33 patients (34%), as compared with 130 patients (55%) in the routine-imaging group (P鈥�&濒迟;鈥�.001). Adjuvant chemotherapy was administered in 55 patients (56%) and 155 patients (66%) in the groups receiving symptomatic follow-up and routine imaging, respectively (P鈥�=鈥�.09).

Median OS was 23 months (95% CI, 19-29 months) in patients with symptomatic follow-up, as compared with 28 months (95% CI, 24-30 months) in patients with routine imaging (P鈥�=鈥�.01) (Figure, A). PDAC recurrence was diagnosed after a median PFS of 13 months (95% CI, 11-15 months) in the symptomatic follow-up group vs 12 months (95% CI, 11-13 months) in the routine-imaging group (P = .93) (Figure, B) and after a median DFI of 12 months (95% CI, 9-14 months) and 10 months (95% CI, 9-12 months), respectively (P = .64) (Figure, C). Median PRS was 6 months (95% CI, 5-10 months) and 12 months (95% CI, 9-14 months) in the symptomatic follow-up vs routine-imaging groups, respectively (Figure, D).

In the routine-imaging group, median OS was 21 months (95% CI, 19-27 months) in patients with symptomatic PDAC recurrence vs 30 months (95% CI, 26-38 months) in patients with asymptomatic PDAC recurrence (P = .003).

The sensitivity analyses demonstrated a median OS in patients with no routine diagnostic testing of 21 months (95% CI, 17-28 months) vs 30 months (95% CI, 16-46 months) in patients with routine CA19-9 measurements only vs 28 months (95% CI, 24-30 months) in patients with routine imaging with or without CA19-9 measurements (P鈥�=鈥�.03) (eFigure 2 in Supplement 1).

Recurrence characteristics are provided in Table 2. Symptoms at time of recurrence were present in 83 patients (85%) with symptomatic follow-up, as compared with 82 patients (35%) undergoing routine imaging (P鈥�&濒迟;鈥�.001). In the 15% of patients without symptoms in the symptomatic follow-up group, PDAC recurrence was detected on imaging that was performed for another reason than symptoms. Isolated local recurrence occurred in 30 patients (31%) in the symptomatic follow-up group vs 77 patients (33%) in the routine-imaging group (P鈥�=鈥�.70). Multiple-site recurrence occurred in 45 patients (46%) and in 80 patients (34%) in the symptomatic follow-up and routine-imaging groups, respectively (P鈥�=鈥�.04).

Recurrence-focused treatment was administered in 47 patients (48%) in the symptomatic follow-up group, as compared with 173 patients (74%) in the routine-imaging group (P鈥�&濒迟;鈥�.001).

Routine imaging was associated with receiving recurrence-focused treatment (adjusted OR, 2.57; 95% CI, 1.22-5.41; P鈥�=鈥�.01). Routine imaging was also associated with OS (adjusted HR, 0.75; 95% CI, 0.56-0.99; P鈥�=鈥�.04). Including center as a random effect did not significantly improve the model (P鈥�=鈥�.96, estimated variance of the random effect: P <.001), suggesting minimal heterogeneity across centers after accounting for the fixed effects. Full results of the multivariable analyses are provided in eTable 2 in Supplement 1.

Moreover, routine imaging was associated with PRS (adjusted HR, 0.59; 95% CI, 0.45-0.79; P鈥�<鈥�.001) but not with PFS (adjusted HR, 1.08; 95% CI, 0.83-1.40; P鈥�=鈥�.58) and DFI (adjusted HR, 1.09; 95% CI, 0.85-1.41; P鈥�=鈥�.49).

To our knowledge, we performed the first prospective cross-sectional study investigating the association of routine follow-up imaging in patients with PDAC recurrence after pancreatic resection with recurrence-focused treatment and survival. It was also the first international study on this topic, to our knowledge, accurately reflecting current pan-European clinical practice. Approximately two-thirds of patients received routine imaging during follow-up after pancreatic resection for PDAC in E-AHPBA鈥揳ffiliated centers, even though this is not recommended by European guidelines. Patients who received routine imaging were more likely to receive recurrence-focused treatment. Moreover, routine imaging was associated with improved survival. These results suggest that the use of routine imaging in the postoperative follow-up of patients with PDAC was beneficial.

Several retrospective studies^{7,8,19,25,26,35-40} have suggested improved outcomes when using routine diagnostics during follow-up after PDAC resection. Most studies were small single-center studies that evaluated only one of the follow-up groups of our study and, therefore, did not perform a direct comparison. A recent meta-analysis¹⁵ comparing studies that reported PDAC recurrence outcomes after different follow-up strategies demonstrated that patients with routine follow-up are more often diagnosed with asymptomatic recurrence. Moreover, patients with asymptomatic recurrence were more likely to receive recurrence-focused treatment and had improved survival.

In contrast to our findings, some retrospective cohort studies did not demonstrate survival benefits from routine imaging after PDAC resection. One study⁴¹ showed that patients with survival greater than 51 months who received annual CT scans did not have improved OS as compared with patients with fewer CT scans. Similarly, in a more recent study,⁴² no improved OS was seen in patients who received additional diagnostics at a cancer center, as compared with patients who were discharged to primary care physicians. Finally, 1 study⁴³ showed that patients who received symptomatic follow-up, as compared with patients undergoing additional imaging, additional tumor marker testing, or both, had similar survival.

Given these data, the direct impact of routine diagnostics in the follow-up after pancreatic resection on survival remains controversial. The main hypothesis of survival benefit by using routine diagnostics relies on timely detection and subsequent initiation of treatment for PDAC recurrence in an early, asymptomatic phase. More potent treatment options such as combination chemotherapies and local ablative therapies have only emerged in recent years. Patients with PDAC recurrence in the current study may have benefited from these therapies as compared with patients in earlier years, potentially resulting in improved oncological outcomes. Comparing different recurrence-focused treatment modalities was, however, beyond the scope of the current study. Interestingly, DFI and PFS did not differ in patients who received symptomatic follow-up and patients who received routine imaging. We hypothesize that the group of patients with aggressive, fast progressing recurrences are anticipated to present with symptoms irrespective of the received follow-up approach, as these recurrences mostly rely on symptom manifestation that could also occur in between the follow-up intervals. Thus, routine diagnostics mainly advance the detection of biologically less aggressive, slowly progressing recurrences, that would otherwise become symptomatic at a later point in time. In this context, routine diagnostics may preempt systemic disease progression. It is possible that in a larger study, differences in DFI and PFS would also become apparent. Our hypothesis is supported by the results of our study, showing that patients with asymptomatic PDAC recurrence after routine imaging have improved survival compared with patients with symptomatic PDAC recurrence after routine imaging.

Interestingly, when comparing the use of routine CA19-9 measurements only with the use of routine imaging with or without CA19-9 measurements, similar survival benefits over no routine diagnostic testing were found. Possibly, follow-up with routine CA 19-9 only shortens the interval to imaging, thereby increasing the number of asymptomatic recurrences. Previous studies^44-46 have suggested that CA19-9 dynamics can be used for recurrence-focused follow-up and may even precede the detection of recurrence up to 10 months. Moreover, CA19-9 measurements are less burdensome and costly than imaging and could therefore play a more important role in the timely detection of PDAC recurrence. Our findings suggest the use of personalized follow-up approaches, in which optimal type, duration, and frequency are subject to both patient and tumor characteristics.⁴ Yet, based on our findings, no recommendations regarding optimal frequency of diagnostic testing or preferred imaging modality could be made. We did not collect data on the number of imaging procedures performed. One study³⁹ has previously suggested that increasing frequency of routine postoperative follow-up from every 6 months to every 3 months increases costs without improving survival. Studies focusing on cost-effectiveness of routine follow-up in other types of cancer are inconsistent, and additional studies are therefore warranted.⁴⁷ After all, if routine imaging provides a certain survival benefit, increased costs might be justified. More definitive studies on predicting timing and location of recurrence based on patient and tumor characteristics could guide to these personalized follow-up approaches in the future.^48,49

Finally, it is possible that current recurrence-detection strategies with both imaging and CA19-9 levels are not sensitive enough to detect pancreatic cancer recurrence at an early phase.⁸ Recent studies^15,50-53 have reported new detection strategies such as circulating tumor cells and circulating tumor DNA. These biomarkers have shown to precede the first sign of recurrence on imaging by 3 to 6 months.^54,55 Moreover, advances in the combined field of radiomics and machine learning are expected to contribute to improved diagnostic tools in the upcoming years.⁵⁶

Ethical considerations persist regarding potential disadvantages of routine diagnostics. These include false-positive results, patient fear of cancer recurrence, increased costs, and negative impact on quality of life.^10,39,57-59 Despite these potential downsides, it has been suggested that patients and clinicians generally support active surveillance, possibly because it may enable symptom-directed clinical management and early recurrence detection in an era of more effective recurrence-focused treatment options.^3,15 This could be an explanation of the high routine diagnostics rate observed in the current study, in contrast to European guidelines. Because equipoise on the topic of follow-up strategies in these patients remains, the Dutch Pancreatic Cancer Group has initiated the randomized Recurrent Disease Detection After Resection of Pancreatic Adenocarcinoma Using a Standardized Surveillance Strategy (RADAR-PANC) trial,⁶⁰ which studies the impact of routine diagnostics on survival and quality of life after PDAC resection.

This study has several limitations. First, it relies on data collection from multiple centers individually, which can result in variations in data accuracy (ie, registration bias). Second, it might be argued that the findings are subjected to selection bias and confounding by indication, as patients who have a more favorable tumor biology or who have a stronger desire for recurrence-focused treatment if needed are more likely to receive routine follow-up imaging and consequent recurrence-focused treatment. There may have also been differences in patient populations between centers who do or do not routinely use imaging, resulting in the observed differences in baseline characteristics between the follow-up groups. Moreover, there could be an association between selection of routine follow-up and recurrence-focused treatment at the level of centers and care. Nevertheless, in the multivariable analysis to correct for potential confounding including center effect, routine imaging remained associated with improved OS. Residual bias, however, cannot be ruled out as other unmeasured biological and clinical differences could still confound the findings. Third, we prospectively included patients on diagnosis with PDAC recurrence and not directly at the start of their initial treatment. This may have also resulted in a selected cohort. Prospectively including patients at the start of their initial treatment may have offered a more comprehensive understanding of the true impact of routine surveillance. We did not choose for this design for the current project, however, for reasons of feasibility and length of follow-up time, among others. Fourth, differences in PRS might have been influenced by lead time bias. To increase the robustness of our findings, OS was chosen as the main survival outcome. Fifth, length-time bias may have occurred as routine diagnostics mainly advance the detection of less aggressive, slowly progressing recurrences, which generally have improved survival. Finally, the international, multicenter setting of our study supports generalizability of our findings; however, we cannot rule out that our results cannot be extrapolated to certain centers worldwide, ie, to hospitals with less access to health care resources or where diagnostics and treatment after PDAC resection are paid out of pocket by patients.

In conclusion, this international, prospective, cross-sectional study showed that patients in whom recurrence was detected during routine imaging in the follow-up after pancreatic resection for PDAC received recurrence-focused treatment more often, and that routine imaging during follow-up was associated with improved OS. These findings challenge existing international guidelines and support further research to establish optimal postoperative follow-up protocols for PDAC.

Accepted for Publication: August 29, 2024.

Published Online: November 6, 2024. doi:10.1001/jamasurg.2024.5024

Corresponding Author: Hjalmar C. van Santvoort, MD, PhD, Department of Surgery, Regional Academic Cancer Center Utrecht, UMC Utrecht Cancer Center, Heidelberglaan 100, Utrecht 3584 CX, the Netherlands (h.c.vansantvoort-2@umcutrecht.nl).

Author Contributions: Drs van Santvoort and Daamen had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Drs Andel and van Goor are considered co鈥揻irst authors. Drs Molenaar, Daamen, and van Santvoort are considered co鈥搒enior authors.

Concept and design: Andel, van Goor, Bouwense, Dorcaratto, Frigerio, Izbicki, Klug, Nikov, Verdonk, Groot, Molenaar, Daamen, van Santvoort.

Acquisition, analysis, or interpretation of data: Andel, van Goor, Augustinus, Berrevoet, Besselink, Bhojwani, Boggi, Cirkel, van Dam, Djanani, Dreyer, den Dulk, Ghorbani, Goetz, Groot Koerkamp, Gryspeerdt, Hidalgo Salinas, Intven, Izbicki, Jorba Martin, Kauffmann, Liem, Luyer, Maglione, Martin-Perez, Meerdink, Nieuwenhuijs, de Meijer, Nikov, Nunes, Pando, Radenkovic, Roeyen, Sanchez-Bueno, Serrablo, Sparrelid, Tepetes, Thakkar, Tzimas, ten Winkel, Zerbi, Groot, Molenaar, Daamen.

Drafting of the manuscript: Andel, van Goor, Besselink, Nikov, Serrablo, Tepetes, Groot, Daamen, van Santvoort.

Critical review of the manuscript for important intellectual content: Andel, van Goor, Augustinus, Berrevoet, Bhojwani, Boggi, Bouwense, Cirkel, van Dam, Djanani, Dorcaratto, Dreyer, den Dulk, Frigerio, Ghorbani, Goetz, Groot Koerkamp, Gryspeerdt, Hidalgo Salinas, Intven, Izbicki, Jorba Martin, Kauffmann, Klug, Liem, Luyer, Maglione, Martin-Perez, Meerdink, Nieuwenhuijs, de Meijer, Nikov, Nunes, Pando, Radenkovic, Roeyen, Sanchez-Bueno, Serrablo, Sparrelid, Thakkar, Tzimas, Verdonk, ten Winkel, Zerbi, Groot, Molenaar, Daamen, van Santvoort.

Statistical analysis: Andel, van Goor, Dorcaratto, Nikov, Groot, Daamen, van Santvoort.

Obtained funding: van Santvoort.

Administrative, technical, or material support: Andel, Augustinus, Besselink, van Dam, den Dulk, Ghorbani, Goetz, Hidalgo Salinas, Izbicki, Luyer, Roeyen, Sanchez-Bueno, Sparrelid, Thakkar, ten Winkel, Molenaar, Daamen, van Santvoort.

Supervision: van Goor, Boggi, Djanani, den Dulk, Frigerio, Groot Koerkamp, Izbicki, Kauffmann, Liem, Maglione, Martin-Perez, Meerdink, Nieuwenhuijs, Nikov, Thakkar, Verdonk, Groot, Molenaar, Daamen, van Santvoort.

Conflict of Interest Disclosures: Dr Intven reported receiving personal fees from Elekta outside the submitted work. Dr de Meijer reported receiving grants from ZonMW Veni, Health~Holland PPP Allowance, and NVGE Research Grant outside the submitted work. No other disclosures were reported.

Group Information: Members of the European-African Hepato-Pancreato-Biliary Association appear in Supplement 2.

Data Sharing Statement: See Supplement 3.