Minimally Invasive Pancreatoduodenectomy Is Noninferior to Open Surgery for 90‑Day Complications but Raises Important Safety and Generalizability Questions

Highlight

– In a multicenter randomized noninferiority trial (ISRCTN27483786), minimally invasive pancreatoduodenectomy (MIPD; predominantly robot-assisted) was noninferior to open pancreatoduodenectomy (OPD) for 90‑day overall complications measured by the Comprehensive Complication Index (CCI).

– MIPD shortened median time to functional recovery by ~1 day (median 7 vs 8 days) and lowered rates of postoperative pancreatic fistula (22.6% vs 35.7%) and surgical‑site infection (12.6% vs 22.7%).

– Conversion to open surgery was 8.4%. Ninety‑day mortality was numerically higher after MIPD (4.7% vs 2.0%) though confidence intervals were wide and the difference was not statistically conclusive.

Background

Pancreatoduodenectomy (Whipple procedure) remains the cornerstone operation for resectable pancreatic and periampullary neoplasms. Historically a high‑morbidity procedure, centralization to high‑volume centers and incremental improvements in perioperative care have reduced mortality and improved outcomes. Minimally invasive approaches — laparoscopic and, more recently, robot‑assisted — aim to accelerate recovery, reduce wound complications, and shorten hospital stay, but uptake has been limited by concerns about safety, a steep learning curve, and the complexity of pancreatoduodenectomy reconstruction.

Study design and methods

de Graaf et al. conducted an international, multicenter, patient‑blinded randomized noninferiority trial in 14 high‑volume centers, enrolling adults with primary resectable pancreatic and periampullary neoplasms. Patients were randomized 2:1 to MIPD (n=190; 170 robot‑assisted, 20 laparoscopic) versus OPD (n=98) and blinded to allocation until postoperative day 5. The primary end point was overall complications within 90 days quantified by the Comprehensive Complication Index (CCI; range 0–100, higher values denote greater cumulative morbidity). The prespecified noninferiority margin was −7.5 CCI points and noninferiority was tested with a one‑sided 97.5% confidence interval (P<0.025 required for noninferiority). The main secondary end point was time to functional recovery (TTFR), tested for superiority. Analyses were intention‑to‑treat.

Key findings

Primary outcome (overall complications)

– Mean CCI: MIPD 33.4 ± 27.5 versus OPD 35.3 ± 25.5 (mean difference −1.9; 95% CI −8.5 to 4.7). The trial reports P=0.002 for noninferiority, concluding MIPD is noninferior to OPD for 90‑day overall complications according to the prespecified one‑sided test.

Time to functional recovery

– Median TTFR: MIPD 7 days (95% CI 6–8) versus OPD 8 days (95% CI 7–11). The reported median improvement of approximately 1 day favors MIPD.

Key secondary safety outcomes

– Postoperative pancreatic fistula (POPF; overall rate): 22.6% after MIPD versus 35.7% after OPD. Relative risk 0.63 (95% CI 0.43–0.91), indicating a statistically and clinically meaningful reduction with MIPD.

– Surgical‑site infection (SSI): 12.6% after MIPD versus 22.7% after OPD. Relative risk 0.57 (95% CI 0.32–0.98).

– Conversion rate: 8.4% of MIPD procedures converted to open laparotomy.

– Ninety‑day mortality: 4.7% after MIPD versus 2.0% after OPD (relative risk 2.40; 95% CI 0.51–11.30). The wide confidence interval includes both no effect and clinically important harm; the trial did not demonstrate a statistically significant difference in mortality.

Interpretation and clinical relevance

This randomized trial offers high‑quality evidence that, when performed in experienced, high‑volume centers, MIPD is at least noninferior to OPD with respect to overall 90‑day complication burden as measured by CCI, and provides modest benefits in recovery time and reductions in POPF and SSI rates.

However, several points require careful interpretation before widespread adoption:

• Magnitude vs statistical designation: The mean CCI difference (−1.9 points) is small and unlikely to represent a meaningful clinical benefit by itself; the trial was designed to demonstrate noninferiority rather than superiority for overall complications.
• Statistical nuance: The authors report a 95% CI for the mean difference (−8.5 to 4.7) and a P value supporting noninferiority under a one‑sided 97.5% framework with a −7.5 margin. Readers should note that differences in reported intervals arise from the chosen one‑sided versus two‑sided approaches; the trial’s prespecified noninferiority test supported the authors’ conclusion.
• Safety signals: Reduced POPF and SSI are clinically relevant and may translate into faster recovery and fewer reinterventions. Conversely, the higher point estimate for 90‑day mortality after MIPD—albeit imprecise—warrants caution. The trial was not powered for mortality, and the wide confidence interval means chance or imbalances could explain the difference.
• Predominance of robotic approach: 170 of 190 MIPD cases were robot‑assisted. Thus the findings primarily reflect contemporary robotic MIPD rather than conventional laparoscopy; generalizability to laparoscopic MIPD is limited.
• Center and surgeon experience: Participating centers were high‑volume, and outcomes likely reflect advanced institutional expertise and structured training. Results may not be replicable at low‑volume centers or during the learning curve.

Strengths of the trial

Randomized design with patient blinding until postoperative day 5, multicenter international participation, intention‑to‑treat analysis, clinically relevant primary and secondary endpoints, and prespecified noninferiority testing strengthen the validity and applicability of the findings within similar high‑volume contexts.

Limitations and unanswered questions

– External validity: Outcomes in high‑volume expert centers may not generalize to community or low‑volume settings.

– Learning curve: The trial does not directly address outcomes during adoption phases; surgeon experience and center volume greatly influence MIPD safety and efficiency.

– Oncologic outcomes and long‑term follow‑up: The report focuses on 90‑day perioperative outcomes. Important oncologic endpoints such as margin status, lymph node harvest, time to adjuvant chemotherapy, disease‑free survival, and overall survival require longer follow‑up.

– Cost and resource utilization: Robotic platforms incur substantial capital and running costs; cost‑effectiveness analyses incorporating differences in complication profiles and length of stay are necessary.

– Mortality signal: The numerical increase in 90‑day mortality after MIPD, though not statistically significant, mandates vigilance. Identifying patient subgroups at higher risk and ensuring robust case selection are essential.

Expert commentary and implications for practice

For experienced teams with robotic platforms and structured training programs, MIPD appears to offer at least equivalent overall perioperative safety to OPD with advantages in select complications (POPF, SSI) and modestly faster functional recovery. Centers contemplating adoption should ensure formal training pathways, proctoring, credentialing, and robust auditing of outcomes. Surgeons and institutions should track outcomes prospectively, including mortality, POPF severity, reintervention rates, oncologic metrics, and cost measures.

Guideline and policy considerations

Professional societies and guideline developers may consider conditional recommendations supporting MIPD in high‑volume centers with demonstrated expertise, while discouraging rapid, unstructured implementation in low‑volume settings. Payment and accreditation policies should incentivize outcome reporting and centralization for complex pancreatic surgery.

Future research directions

– Long‑term oncologic outcomes and quality of life after MIPD versus OPD, including time to adjuvant therapy and survival endpoints.

– Cost‑effectiveness studies that incorporate capital costs of robotic platforms, conversion rates, complication‑related expenditures, and patient‑reported outcomes.

– Trials or registries evaluating outcomes during and after the learning curve and implementation studies to define minimum volume and training requirements.

– Subgroup analyses to identify patients most likely to benefit—or be harmed—by MIPD (e.g., high BMI, locally advanced tumor characteristics, vascular involvement).

Conclusion

In this randomized international trial performed in high‑volume centers, minimally invasive pancreatoduodenectomy (predominantly robot‑assisted) met the prespecified noninferiority criterion for 90‑day overall complications compared with open surgery, yielded modestly faster functional recovery, and reduced postoperative pancreatic fistula and surgical‑site infection rates. These benefits must be weighed against the limited generalizability to lower‑volume settings, the predominance of robotic procedures in the trial, and a nonconclusive signal toward higher 90‑day mortality that requires surveillance. Adoption of MIPD should be accompanied by rigorous training, outcome monitoring, and further research on long‑term oncologic outcomes and cost‑effectiveness.

Funding and trial registration

Funded by Intuitive Surgical and Fondazione Poliambulanza Istituto Ospedaliero. Trial registration: International Standard Randomised Controlled Trial Number Registry, ISRCTN27483786.

Reference

de Graaf N, Emmen AMLH, Ramera M, van Hilst J, Björnsson B, Boggi U, Bruna CL, Busch OR, Daams F, Droogh DHM, Ferrari G, Festen S, Guerra M, de Hingh I, Keck T, Groot Koerkamp B, Lips DJ, Luyer MDP, Mieog JSD, Morelli L, Molenaar IQ, de Wilde RF, Ali M, Ferrari C, Berkhof J, Maisonneuve P, Abu Hilal M, Besselink MG. Minimally Invasive versus Open Pancreatoduodenectomy for Resectable Neoplasms. NEJM Evid. 2025 Dec;4(12):EVIDoa2500045. doi: 10.1056/EVIDoa2500045. Epub 2025 Nov 25. PMID: 41288428.