Surgery · Medical Literature
An Exchange about "Minimally Invasive versus Open Pancreatoduodenectomy for Resectable Neoplasms".
Clinical relevance
This NEJM Evidence exchange addresses post-publication reader concerns about the generalizability of the DIPLOMA-2 trial—the largest patient-blinded RCT to date comparing minimally invasive pancreatoduodenectomy (MIPD) to open pancreatoduodenectomy (OPD) for resectable pancreatic head and periampullary neoplasms. Pancreatoduodenectomy is one of the most technically demanding abdominal operations, and the safety and recovery benefits of minimally invasive approaches have remained actively contested. The exchange is clinically urgent because it directly interrogates whether DIPLOMA-2 findings can inform surgical decision-making beyond the narrow stratum of very-high-volume European expert centers where the trial was conducted.
Key findings
1. Noninferiority of MIPD confirmed: In 288 patients randomized 2:1 (MIPD vs. OPD) across 14 high-volume European centers, MIPD was noninferior to OPD for the primary endpoint of 90-day overall complications as measured by the Comprehensive Complication Index (CCI; scale 0–100; prespecified noninferiority margin of 7.5 CCI points; 2.5% one-sided α, 80% power). Specific mean CCI values are not available in the public abstract. 2. Early functional recovery favored MIPD: At postoperative day 30, the MIPD group averaged 659 more daily steps (95% CI, 79–1,240; P = .03) and 22 more active daily active minutes (95% CI, 3–40; P = .03) compared with OPD, per wearable activity tracker data (Fitbit Inspire 2). 3. Autonomic recovery favored MIPD: Heart rate variability (HRV) at postoperative day 30 was 9 milliseconds higher in the MIPD group (95% CI, 2–15 ms; P = .006), indicating greater parasympathetic recovery and reduced physiological stress. 4. Recovery benefit was time-limited: Physical activity advantages attenuated by 90 days postoperatively; step counts (MIPD median 4,340 vs. OPD median 4,442; P = .90) and active minutes were no longer significantly different at day 90. 5. Robot-assisted approach predominated: 88% of MIPD cases were robot-assisted, with a 6% conversion rate to open surgery, reflecting the high-expertise context in which these results were achieved. 6. Generalizability is the central unresolved concern: The exchange article specifically addresses reader-raised questions about applicability of findings outside the trial's highly specialized setting (centers required ≥30 MIPDs and ≥30 OPDs annually; surgeons required ≥60 personal MIPD and ≥60 OPD cases), with specific quantitative estimates of generalizability not available from the public abstract.
Practice implications
Surgeons and surgical program directors at high-volume pancreatic centers with established robotic or laparoscopic pancreatoduodenectomy expertise can consider MIPD a safe alternative to OPD for upfront resectable pancreatic head and periampullary neoplasms without vascular involvement, with the expectation of noninferior 90-day complication burden and a modest but objective early recovery advantage (approximately 2–5 additional weeks of improved physical activity post-discharge). However, the authors and post-publication exchange explicitly caution against extrapolating these findings to lower-volume centers or surgeons below the established experience thresholds (≥60 personal MIPDs), given the strong influence of surgeon experience and institutional volume on outcomes. Until DIPLOMA-2's oncologic endpoints (R0 rate, adjuvant therapy initiation, and 36-month survival) and cost-effectiveness data mature, adoption should be limited to centers that can replicate the trial's rigorous volume and training standards, with prospective monitoring of conversion rates and oncologic margins.