The treatment strategies for NADAs are unclear. Tumors that are not amenable to endoscopic resection require operative intervention. PD has been the standard operation for invasive lesions that involve the ampulla of Vater or that require lymph node dissection. PPD are usually indicated for select tumors, including gastrointestinal stromal tumors, large adenomatous lesions, T1a adenocarcinoma, and small NEN, that has no risk of LN metastasis. The choice of PPD versus PD is dependent on factors, such as tumor size, location (proximity to the ampulla of Vater), risk of lymph node metastasis, and patient’s overall fitness. However, there are few reports comparing oncological results between PD and PPD, and an ideal surgical approach has not yet been established. Including total duodenectomy, supra- or infra-ampullary partial duodenectomy, and wedge resection of the duodenum, PPD may be one of the curative treatments for non-ampullary duodenal neoplasms depending on the tumor location and diagnosis [3]. When PPD is performed, handling of the accessory papilla is important. Since its damage leads to postoperative complication including pancreatic fistula, it should be preserved if possible. If a communication between the Santorini and the Wirsung duct is confirmed, it can be sutured or ligated. It is necessary to confirm it by MRCP or ERCP before surgery when the accessory papilla is planned to be ligated or sutured.

PPD has several advantages over PD in terms of operative complexity and organ preservation that leads to a lower incidence of postoperative complications and preservation of pancreatic function. There are some reports that show the usefulness of PPD, but the overall survival was not significantly different, although patients undergoing PD had a higher morbidity [4,5,6]. An important limitation of PPD is insufficient oncological resection due to the absence of LN dissection. In early non-ampullary duodenal adenocarcinoma (NADA), many reports showed that there was no incidence of LN metastasis in the NADA limited to the mucosa [7]. Otherwise, in reports that examined small numbers, high rates of LN metastasis in submucosal invasive NADA were reported, ranging from 14 to 42% [8,9,10]. In addition, there is no crucial opinion on the range of LN dissection depending on the tumor location; furthermore, the optimal range of LN dissection for the NADA in the horizontal and ascending parts of the duodenum is unknown. In a case of NADA located in the horizontal or ascending duodenum PPD may be sufficient even in advanced stages because LN dissection around the superior mesenteric artery could be performed [11, 12]. It is better to avoid unnecessary PD if the tumor does not invade the pancreas directly, especially early NADA. Although patients 1 and 3 were diagnosed with adenoma and T1a adenocarcinoma, respectively, endoscopic treatment or local resection was difficult to perform without complications due to tumor size and location. EUS could be useful for more accurate preoperative diagnosis in terms of depth of invasion, though it has not been performed on these patients. We performed PPD in both cases and negative margins were achieved. There was no risk of LN metastasis; therefore, the selected procedures were better than PD.

We performed partial duodenectomy rather than local resection because of lymph node dissection around the supra-ampullary duodenum for the second patient with NEN. Soga et al. [5] reported that the LN metastasis rate of duodenal NENs among 655 patients was 10.6% for tumor diameters of 5 mm or less, 13.9% for 6–10 mm, 24.7% for 1.1–2.0 cm, and 24.7% for tumors above 2.0 cm. In a retrospective study conducted in Japan, risk factors for metastasis were reported to be NET G2, multiple tumors, tumor size greater than 1.1 cm, and positive vascular invasion [6]. Although PD is superior to PPD in terms of lymph node dissection, PPD can be selected for a relatively small NEN if negative margins can be confirmed and there is no LN swelling on the preoperative images. Since the second patient did not undergo PD, it is necessary to follow up strictly for recurrence, including checking LNs.

The anatomical features of the duodenum, such as the narrow lumen and thin wall, make endoscopic resection of tumors difficult. To prevent complications of ESD, laparoscopic–endoscopic cooperative surgery (LECS), in which the defect of the duodenal wall by endoscopic procedure is closed by laparoscopic suture from the outside of the duodenum, was recently implemented to treat patients with NADA [13]. Although LECS requires complicated procedures and surgical instruments, it is an effective backup technique [14]. Duodenal neoplasms located at the opposite side of the ampulla of Vater, which are difficult to resect by partial duodenectomy, can be resected by the LECS that can determine the excision range and repair the defect of duodenal wall after excision. Although we did not perform LECS for our 3 patients because of tumor size and location, LECS would be useful for a certain number of cases with NADA.

We agree that PPD is an attractive and promising alternative procedure to PD for the treatment of NADNs, especially small NENs, gastrointestinal tumors, large adenomas, and early NADAs. The technical points and pitfalls of this operation have not been sufficiently discussed thus far. Further large studies to examine the safety and curability of PPD are needed in the future.