Anomalies of the right hepatic artery in periampullary cancer treatment: are pathological and clinical outcomes different? A single tertiary referral center retrospective analysis

Pancreatoduodenectomy (PD) is the gold standard of treatment of periampullary malignancies and is widely known as a complex surgical procedure, burdened by a rate of postoperative complications up to 60% and a related mortality rate over 5% in low-volume centers [1]. In this context, the high variability of the regional vascular anatomy may increase the surgical technical difficulty. Notably, anomalies of the hepatic arteries are encountered in the 25–45% of cases, and variations of the right hepatic artery (RHA) represent the majority of them (12–18% of cases) [2, 3]. This has brought several authors to analyze the incidence rate of such variants and propose a classification with the aim of providing a common nomenclature. Currently, the most widely used are those proposed by Michels et al. [2] and Hiatt et al. [3], although a more comprehensive and complex classification has been more recently introduced by Yan et al. [4]. On the counterpart, contrasting results are present in the literature on the potential impact of on an aberrant RHA on short- and long-term outcomes of patients undergoing PD. Traverso advocated a potentially higher risk of biliary fistula (BF), possibly due to ischemia of the common bile duct stump following RHA skeletonization [5]. Furthermore, some authors reported an increased risk of R1 resections and local recurrence in case of anomalies of the RHA [6, 7]. Conversely, other authors did not evidence any difference when comparing patients with modal vascularization with those with anomalies of the RHA [8‐10].

These conflicting results may find justification in the poor quality of data analyzed, mainly derived from small and non-homogeneous populations, as well as from non-referral centers for the surgical treatment of pancreatic diseases.

Based on these premises, the aim of this study was to assess the impact of anomalies of the RHA on perioperative outcomes after PD in a high-volume tertiary referral center for the surgical treatment of pancreatic diseases. In order to accomplish this purpose, patients who underwent PD for periampullary carcinomas with and without an aberrant RHA were compared. In addition, a propensity score matching (PSM) analysis was conducted to homogenize the two study populations.

From January 2017 to December 2022, 321 patients underwent PD at the Pancreatic Surgery Unit of the Fondazione Policlinico Universitario Agostino Gemelli IRCCS of Rome. Of these, 232 (72.3%) were diagnosed with a periampullary carcinoma and were therefore included in the analysis (Fig. 1). An APV was preoperatively detected at the CT scan in 38 patients (16.4%) and further confirmed intraoperatively (Fig. 2). The remaining patients presented a MPV.

After PSM, 76 patients with a MPV were compared to 38 patients with an APV. The presence of APV was classified as follows: 7 (18.4%) patients had an aRHA, 26 (68.4%) a rRHA, and 5 (13.2%) a rHA arising from the SMA. As a whole, NAT was performed in 10 (8.8%) patients. Sixty-eight patients presented severe jaundice at the diagnosis, and a biliary stent was preoperatively placed via ERCP in 63 of them (55.3%), while a trans-hepatic biliary drainage was positioned in 5 cases (4.4%). In terms of tumor location, 24 patients (21.1%) presented an ampullary cancer, 78 (68.4%) a pancreatic ductal adenocarcinoma of the head of the pancreas, 3 (2.6%) of the uncinate process, 5 (4.4%) had a distal bile duct adenocarcinoma, and 4 (3.5%) a duodenal adenocarcinoma. Clinico-demographic characteristics and their comparison between the two study cohorts are reported in Table 1.

PD is widely known as one of the most challenging procedure in general surgery, whose complexity may be significantly increased by the concomitant presence of vascular anomalies of the periampullary region. Nevertheless, no unanimous data are currently present in the literature on the potential impact of vascular variants on perioperative outcomes, and scarce and contrasting evidences are currently present in particular on the potential role of an aberrant RHA.

According to our findings, the presence of an anomalous vascular pattern did not significantly influence the incidence rate of postoperative complications. In addition, no difference was evidenced between APV and MPV patients in terms of pathological findings, here including the rate of positive resection margins.

The potential correlation between an aberrant RHA and the onset of postoperative complications has been the focus of several studies. In particular, the accidental damage of aberrant vessels, as well as the alteration of local vascularization due to the excessive skeletonization of the aberrant arteries, has been proposed as the main predisposing features to postoperative complications, in particular BF. Indeed, as reported by Northover et al. [20], the proximal extra-hepatic biliary tree is vascularized by the RHA for almost the 40% of the total, while the remaining blood flow comes from the retroduodenal, the retroportal, and the gastroduodenal arteries. In this context, the presence of a RHA originating from the SMA and running along the posterior aspect of the retroportal lamina, that needs to be skeletonized in order to guarantee oncological radicality, may hypothetically increase the risk of bile duct (and subsequent hepaticojejunostomy) ischemia, potentially leading to a higher rate of BF. This issue has already been enlightened by Traverso in 1989 [5] and assessed by several authors in recent years. However, the majority of them did not evidence any statistical difference between MPV and APV patients. Indeed, Eshuis and Rammohan showed that, despite higher intraoperative technical difficulty, the presence of an rRHA or aRHA did not affect perioperative outcomes, including BF incidence [21, 22]. More recently, Alexakis conducted a matched analysis only on patients who underwent surgery for benign diseases, reporting similar rates of complications [23]. Our findings further corroborate these evidences since no differences were detected between APV and MPV patients both in terms of postoperative complications and, in particular, BF incidence (p = 0.2 and p = 0.58, respectively). This would underline how an adaptive system would take place in case of an aberrant RHA, maintaining a constant and efficient blood supply for the biliary duct stump vascularization.

Great concern has been also expressed about the potential role that an aberrant RHA may have on the oncological radicality, especially in terms of resection margins. For instance, an aRHA or rRHA generally arise from the SMA and run along the posterior aspect of the duodeno-cephalopancreatic block through the retroportal lamina. This last is widely recognized as the most frequent site of R1 and local tumor recurrence in case of infiltration or incomplete removal [24, 25]. In this scenario, the need of preservation of the anomalous artery, such as an aRHA and rHA, might hinder a proper oncological dissection, thus paving the way to a positive retropancreatic margin and to a dismal oncological result, increasing at the same time the risk of PPH due to the necessary skeletonization of the anomalous artery. Although the oncological role of vascular arterial resection is highly debated, some author [26] advise a vascular resection of the aberrant artery when the tumor is closer than 10 mm to the artery in order to achieve an R0 resection, eventually performing a vascular reconstruction through the interposition of the gastroduodenal stump [27, 28]. According to our findings, no difference has been evidenced between the MPV and APV cohorts in terms of resection margin positivity (here including the mesopancreas margin), in line with Turrini et al. that reported no difference in terms of margin status when the variant right hepatic artery was not directly infiltrated [9]. Moreover, no arterial vascular resections were needed intraoperatively, and PPH rate was similar between the two study cohorts. These outcomes may find justification in the importance of performing such complex procedure in high-volume referral centers [18, 29]. Specifically, the preoperative assessment of tumor resectability and vascular anatomy, performed in a multidisciplinary and specialized setting, has permitted to adequately select patients for PD, excluding those with evidence of tumor involvement of major vessels, notably candidate to neoadjuvant treatment. Similarly, the surgical expertise developed over the years has permitted to achieve similar PPH rate between the MPV and APV populations (p = 0.6), with even comparable operative time detected in two study cohorts (p = 0.79) and no accidental vascular resection during surgery.

This study presents several limitations especially due to its retrospective study design as well as to the relatively small sample size of the study cohort. On the counterpart, to our knowledge, this is the first study in the literature on this topic based on a PSM-derived cohort that has significantly limited the potential selection biases, making the two populations homogenous for the majority of confounding factors. Moreover, our findings derive from a high-volume tertiary referral center for the surgical treatment of pancreatic diseases, giving an objective and real picture of clinical outcomes when PDs with APV are performed in specialized centers.

In conclusion, the present analysis has confirmed that the presence of anomalies of the RHA does not significantly modify the postoperative clinical course nor the quality of surgical dissection. It is, however, implicit the need to perform this more complex procedure in high-volume centers and in a multidisciplinary setting. Nevertheless, additional studies with larger series are necessary to further corroborate our results.