Radical modular pancreatoduodenectomy for pancreatic head cancer using a combination of multiple artery-first approaches technique.

The aim of this study was to describe and assess the efficacy of a combination of multiple artery-first approaches (CMAFA) in pancreatoduodenectomy (PD) depending on the tumor location from an embryonic point of view.Between January 2011 and December 2016, seventy-nine consecutive patients with pancreatic head cancer (PHC) underwent PD with curative intent. Patients were classified into two groups according to the surgical procedure: CMAFA-PD group (n = 38) and conventional PD (Co-PD) group (n = 41). Clinicopathlogical variables and clinical outcomes were compared among the two groups.The CMAFA technique demonstrated an improved rate of R0 resection (89.5% vs. 70.7%, P = .038) and a higher median lymph node yield (24 vs.20, P = .034). The CMAFA-PD group was associated with reduced blood loss (450 vs. 600 ml, P = .049), lower rate of blood transfusion (23.7% vs. 46.3%, P = .035), and shorter length of hospital stay (19 vs. 26 days, P < .001). The rates of 90-day mortality, major morbidity, and readmission were comparable among the two groups.This study demonstrates that CMAFA is a feasible and efficient technique with acceptable perioperative and oncological outcomes in treating patients with PHC.

Introduction

Pancreaticoduodenectomy (PD) is the standard surgical procedure for pancreatic head cancer (PHC). Since the initial report of the PD technique by Whipple and Kausch,[ some modified PD procedures at the resection stage have been developed over the years, such as superior mesenteric artery (SMA) first approach,[ no-touch isolation technique,[ uncinate process-first approach,[ and the hanging maneuver.[ Artery-first approach (AFA) is one of the important progresses in the evolvement of PD, featuring early judgment of resectability,[ optimal exposure of arterial abnormalities,[ simplification of clearing the posteromedial resection margins,[ and significant reduction of intraoperative blood loss.[ Up to now, there are main six artery-first approaches reported in the published papers, all of which have its own applicable characteristics and provide pancreatic surgeons with some options in the treatment of PHC.[ Most studies reported AFA mainly focused on the surgical exploration for the purposes of early determining resectability and usually used one of those six techniques as the initial approach for SMA dissection. But in most cases, it is not enough to use a single AFA to expose the whole length of the SMA and/or celiac axis (CA), particularly for locally advanced PHC. Therefore, a combination of different artery-first approaches based on the tumor location and extent of invasion may be pretty necessary.

Several Japanese studies demonstrated that lymphatic spread and perineural invasion of the PHC differ greatly between the dorsal and ventral primordium.[ For tumors confined to the ventral pancreas domain, lymph node metastases and perineural invasion are prone to spread toward the SMA; on the other hand, tumors confined to the dorsal pancreas domain have a propensity to spread toward common hepatic artery (CHA) and the hepatoduodenal ligament. These results suggest that a safe and effective lymph nodes and nerve plexus dissection should be designed according to the primary tumor location.

In the past several years, we have developed and implemented the radical modular PD technique with the combination of multiple artery-first approaches (CMAFA) depending on the aforementioned tumor classification. In the present study, we described the details of this technique and assess its feasibility and efficacy based on clinicopathological results.

Methods

Concept and indications of the combination of multiple artery-first approaches

Based on the previous study,[ it is relatively easy to recognize which domain the PHC is located by the line linking the portal vein/superior mesenteric vein (PV/SMV) and anterior edge of the intrapancreatic bile duct on the preoperative computed tomography (CT) images. Therefore, we classify the PHC into 3 types according to the preoperative imaging studies. Type I is greater than 90% of the tumor confined to ventral pancreatic head (Fig. 1). Type II is greater than 90% of the tumor confined to dorsal pancreatic head (Fig. 2). Type III is the tumor extended to both areas of the pancreatic head (Fig. 3). Up to now, six AFAs were described in the published studies. In order to facilitate its application in this study, we divided the six AFAs into three types: superior approach,[ posterior approach,[ and inferior approach. CMAFA-PD is performed with different combinations of AFAs mainly based on four principles: the first is a priority to the targeting of the resection area guided by the tumor location; the second is the early judgment of resectability for borderline resectable PHC; the third is reducing the intraoperative blood loss so as to ensure the safety of the operation; and the fourth is the en bloc dissection of the posteromedial pancreatic margin. Accordingly, for Type I PHC, the area of resection should be focused along the whole length of SMA, thus both inferior and posterior approaches are considered; while in the case of Type II PHC, the area of resection should be concentrated along the CHA to the origin of the CA, thus superior and posterior approaches are considered. For Type III PHC, the root of SMA is usually involved by the large tumor mass, thus posterior approach is performed first and dissection should be including along the SMA and the CHA. Therefore, superior and inferior approaches are also required. Additionally, superior and inferior approaches are performed in almost every CMAFA-PD for early ligation of the gastroduodenal artery (GDA) and inferior pancreaticoduodenal arteries (IPDA), except the circumstances that IPDA is easily dissected using posterior approach and thus the inferior approach is not necessary in type II PHC. Details of surgical approaches for different tumor types in CMAFA-PD group are shown in Table 1.

Figure 1

(A) CT scan image showing a tumor confined to the ventral pancreatic head. (B) Schematic diagram corresponding to (A). AA = abdominal aorta, Ca = cancer, CBD = common bile duct, Duo = duodenum, GDA = gastroduodenal artery, IPDA = inferior pancreaticoduodenal artery, IVC = inferior vena cava, SMA = superior mesenteric artery, SMV = superior mesenteric vein.

Figure 2

(A) CT scan image showing a tumor confined to the dorsal pancreatic head. (B) Schematic diagram corresponding to (A). AA = abdominal aorta, Ca = cancer, CBD = common bile duct, Duo = duodenum, GDA = gastroduodenal artery, IVC = inferior vena cava, SMA = superior mesenteric artery, SMV = superior mesenteric vein, WD = Wirsung duct.

Figure 3

(A) CT scan image showing a tumor extends to both ventral and dorsal pancreatic head. (B) Schematic diagram corresponding to (A). AA = abdominal aorta, Ca = cancer, CBD = common bile duct, Duo = duodenum, IVC = inferior vena cava, SMA = superior mesenteric artery, SMV = superior mesenteric vein.

Table 1

Surgical approaches for different tumor types.

Surgical technique

Superior approach

The superior approach used in this study was similar to that described by Sanjay et al.[ In the present study, there are two levels in superior approach. Level 1 is the first dissection of the hepatoduodenal ligament to expose the whole length of CHA until its origin from the CA. At this stage, GDA is divided, and the PV is isolated at the upper border of the pancreatic neck. Level 2 is the dissection along the CA down to the aorta so as to expose the root of the SMA. For type II PHC that usually spreads to the CHA/CA and the root of the SMA, the superior approach is used to start the surgery and both levels are required to be performed; while for type I PHC, only level 1 is required. The crucial surgical steps in superior approach are depicted in Figure 4A.

Figure 4

Intraoperative images demonstrating the superior artery-first approach (A), posterior artery-first approach (B), and inferior artery-first approach (C). CA = celiac axis, CBD = common bile duct, CHA = common hepatic artery, Duo = duodenum, FJA = first jejuna artery, GDA = gastroduodenal artery, HA = hepatic artery, IVC = inferior vena cava, LRV = left renal vein, PV = portal vein, SA = splenic artery, SMA = superior mesenteric artery, SMV = superior mesenteric vein.

Posterior approach

The posterior approach we used was similar to that described by Pessaux and Popescu et al.[ Using an extended Kocher's maneuver, the inferior vena cava and the distal portion of the left renal vein (LRV) can be exposed completely. The origin of the SMA is exposed where it passes in front of the distal portion of LRV. Using a vascular tape around the SMA, the right semicircle exfoliation of the nerve plexus together with the lymphatic connective tissue around the SMA is dissected cranio-caudally. Dissection along the right semicircle of the SMA is continued until the exposure of the lateral border of the SMV. At this stage, the IPDA and the replaced right hepatic artery can be identified. Then, the dissection is performed along the anterior surface of the aorta to expose the origin of the CA (Fig. 4B). The right semicircle connective tissue around the CA is then dissected, and the origin of the CHA can be exposed, if necessary. For type III PHC that occupied predominantly in the ventral pancreas, a combination of inferior approach is necessary. Similarly, for type III PHC that occupies predominantly in the dorsal pancreas, superior approach is usually required.

Inferior approach

The inferior approach to the SMA used in our institution included right/medial uncinate approach,[ inferior supracolic approach,[ and mesenteric approach.[ Medial uncinate approach as one of the inferior approaches is used most frequently in this study, which is similar to that described by Hackert et al and Shukla et al.[ The SMA is dissected in a supracolic, anterior, and retrograde fashion initiated from the caudal end of the uncinated process. In this process, IPDA and the first jejunal artery are divided and controlled (Fig. 4C). In most cases, a combination of inferior and posterior approaches could be conductive to access the full length of the SMA.

Conventional PD

In this study, Co-PD is defined that the dissection of the SMA is performed at the final stage of resection. In Co-PD, the dissection usually begins with the Kocher's maneuver so that a hand can be passed behind the pancreatic head to palpate the tumor mass. Thus, this step can roughly determine the tumor infiltration of the SMA or PV/SMV. After the irreversible divisions of pancreas, the pancreas head is pulled rightward, and the posterior surface of PV/SMV is carefully dissected from the uncinate process. Then, taping the SMV, the SMA can be divided from cranial to caudal and the right semicircle dissection of soft connective tissue around the SMA is dissected completely.

Patients

From January 2011 to December 2016, 101 consecutive patients with PHC underwent surgery with curative intent at the Department of Hepatopancreatobiliary Surgery, Second Affiliated Hospital of Harbin Medical University. Of these, a total of 79 patients who received PD, performed by a senior surgeon (Y.F. Cui), were enrolled in this study, excluding 22 patients underwent palliative surgery. In the present study, Co-PD was mainly performed in the early period (2011–2013), and our routine procedure was converted to CMAFA-PD afterward. This study only included patients with resectable or borderline resectable PHC according to the National Comprehensive Cancer Network (NCCN) guidelines and expert consensus statement in 2009.[ Excluding criteria were as follows: (1) patients with severe medical complication; (2) with evidence of distant metastasis; and (3) patients were older than 75 years. All patients in this study were discussed and treated by a multidisciplinary team. The included patients were divided into 2 groups according to the surgical procedure. The first group of 41 patients underwent conventional PD (Co-PD group), and the second group of 38 patients received PD using the CMAFA technique (CMAFA-PD group).

Preoperative assessment of the resectability of the PHC was based on the enhanced CT with thin-slice scans and magnetic resonance cholangiopancreatography. All clinical, surgical, and pathological data were recorded prospectively. A R1 resection margin was defined as the tumor cells were ≤1 mm from the margins. The postoperative pancreatic fistula was classified into 2 groups (grades B and C), according to the latest version of the International Study Group of Pancreatic Fistula definition.[ Delayed gastric emptying, postpancreatectomy hemorrhage, and biliary fistula were defined according to the International Study Group on Pancreatic Surgery definitions.[ This study was approved by the Institutional Ethics Committee of Harbin Medical University. All study participants provided written informed consent prior to study enrollment.

Statistical analysis

For continuous variables, descriptive statistics were expressed as median values with range. Continuous variables were compared by Wilcoxon rank-sum test. Categorized variables were performed using Pearson Chi-square test or Fisher's exact test. P values of less than .05 were considered as statistically significant. Data were analyzed using IBM SPSS 22.0 software (version 22.0; SPSS, Chicago, IL).

Results

Patient characteristics

A total of 79 consecutive patients with PHC underwent radical PD during the 6-year study period (CMAFA-PD = 38, Co-PD = 41). No significant differences were found between the 2 groups with respect to age, gender, BMI, and the rates of diabetes. The distribution of tumor location was similar in 2 groups. The rates of borderline tumors in 2 groups were also similar. In addition, the rates of patients who received neoadjuvent chemotherapy (gemcitabine 1000 mg/m2 on days 1 and 8 plus S-1 65 mg/m2 on days 1 through 14 of a 21-day cycle, 3 cycles of GS chemotherapy before surgery) in 2 groups were well balanced. Baseline patient demographics are presented in Table 2.

Table 2

Patient and tumor characteristics.

Pathologic analysis

There were no significant differences between the two groups in terms of tumor size, TNM stage, histologic type, perineural invasion, vessel invasion, and lymphatic permeation. Total number of dissected lymph node in CMAFA-PD group, however, was significantly higher than that of the Co-PD group. Curability was significantly higher in CMAFA-PD group. R0 resection rate was 70.7% in the Co-PD group and 89.5% in the CMAFA-PD group. Three patients with microscopically positive posterior resection margin were found in the CMAFA-PD group, 1 patient with microscopically positive margins on pancreatic stump. However, there were 5 patients with microscopically positive posterior margins, 3 patients with microscopically positive margins on pancreatic stump, and 4 patients with macroscopically positive margins on SMA in the Co-PD group. Pathologic details were shown in Table 2.

Perioperative outcomes

The perioperative outcomes are presented in Table 3. CMAFA-PD was associated with less blood loss (450 vs. 600 ml; P = .049) and lower rates of blood transfusion (23.7% vs. 46.3%; P = .035). Operative time and the number of patients with organ coresection and reoperation were comparable among groups. The rate of postoperative pancreatic fistula (POPF) was similar between the CMAFA-PD group and Co-PD group. The length of hospital stay was significantly shorter in the CMAFA-PD group, whereas the rates of the overall complication and 90-day readmission were similar between groups. One patient (2.4%) dead in Co-PD group because of portal vein thrombosis after PV reconstruction, leading to liver failure.

Table 3

Perisurgical outcomes.

Feasibility and validity of CMAFA-PD

There was no CMAFA-PD conversion to Co-PD. Twenty two patients were excluded from this study. Of these patients, 13 patients were found to show peritoneal dissemination and/or liver metastasis. We used the CMAFA technique to detect SMA involvement before reaching the “point of no return” in 6 patients who underwent palliative surgery afterward. Three patients with conventional approaches also underwent bypass surgery for major vessel invasion. Moreover, 4 patients experienced the “point of no return” in Co-PD group and ended up with macroscopically positive margins on SMA, whereas no patients experienced that point in CMAFA-PD group.

Discussion

Although the PD procedures are highly standardized in most high-volume pancreatic centers, consensus does not appear to have been reached regarding the optimal operative decision making according to the tumor location and extent of invasion. This is the first report that described a combination of multiple artery-first approaches for patients with PHC depending on the tumor location from an embryological point of view. The current study showed that CMAFA technique was associated with decreased intraoperative blood loss, blood transfusion, and length of hospital stay. Moreover, CMAFA technique was also associated with improved lymph nodal harvest and margin status.

It is necessary to establish a classification of PHC that contributes to optimize the operative decision making. Several classification systems have been published, such as the Japan Pancreas Society and the Union Internationale Contrele Cancer (UICC) classifications.[ However, both of the two classification systems mainly focused on the reliability of predicting outcome rather than guidance of operative decision making. Recently, Wang et al proposed a classification system of PHC by the relationship of the tumor to key vascular around the pancreatic head.[ However, the details of their technique for each type were vague. Although the AFA technique has been increasingly performed, there remains no consensus as to which AFA is indicated in which situation.[ At our center, we used the duct of Santorini/Wirsung, the PV/SMV, and the bile duct as landmarks in the enhanced CT with thin-slice scans to access the tumor location and extension preoperatively, which was described by Okamura et al.[ According to the finding in the imaging studies, we categorize the PHC into three types based on an embryological division of the pancreatic head. Therefore, we have the opportunity to tailor different approaches to fit the different types of the PHC during PD.

Four major advantages seem to contribute to develop the CMAFA technique to PD. The first was a priority to dissect the vessel, nerve plexus, and lymph nodes that tumors were prone to spread and extend, following oncologic principles.[ Especially for borderline resectable PHC, the presence of key vasculature involvement could be identified early enough to avoid the “point of no return”. The second was establishing a logical strategy to dissect the SMA margin first and to guarantee the maximum cancer margin. The third was the facilitation of the “no touch” resection of PV/SMV infiltrated by the PHC. In this case, the dissection of the SMA to PV/SMV was the critical part of the PD procedure. The CMAFA technique enabled the surgeon to perform an en bloc tumor resection without interfering with tumor integrity.[ The fourth was the en bloc resection of tumors located on the pancreatic body and required PD. Using the CMAFA technique, the final transection of pancreas can be performed at any desired level without unnecessary incision of the tumor. Additionally, if total pancreaticoduodenectomy (TPD) was needed to perform, our CMAFA technique facilitated the dissection along the splenic artery and splenic vein to the splenic hilum, and thus contributed to preserve spleen. Using CMAFA technique, we successfully performed 4 TPD in CMAFA-PD group with no mortality in 90 days of resection.

At present, evidence that an AFA can increase the rate of R0 resection is sparse, and evidence that an AFA improves the lymph node yield is even more so. Five studies compared Co-PD with an AFA technique,[ and only one study demonstrated significantly improved R0 rate and a higher lymph node yield.[ Similarly, results from our study showed improved margin status and lymph node yield with CMAFA-PD. Three potential factors contribute to interpret the results. First, the CMAFA technique guarantees a complete resection of “the mesopancreas triangle,” which allows an en bloc resection of the pancreatic head with retroperitoneal tissue. [ Second, the CMAFA technique facilitates the en bloc resection of the tumor-bearing region with high ligation of supplying arteries, such as GDA and IPDA. Third, we detected key vascular involvement in 6 cases by using CMAFA technique, and thus avoided unnecessary resection. On the contrary, 4 cases experienced the “point of no return” in Co-PD group, and thus received R2 resection.

A previous study compared Co-PD with a posterior AFA and showed that the AFA technique was associated with a significant reduction in estimated blood loss and duration of surgery time.[ Although the operating time in our study was comparable among the two groups, both intraoperative blood loss and length of hospital stay in CMAFA-PD group were significantly lower than in the Co-PD group. The CMAFA technique in this study was used not only to explore the SMA but also focused on the early ligation of IPDA and GDA, thus reducing the congestion into the pancreatic head. We believe that it is an important advantage of CMAFA technique, which is feasible in every type of patients with PHC. Horiguchi et al used another modified procedure of PD to first ligation of IPDA and GDA before dissecting the corresponding afferent veins.[ By doing this, bleeding from the resected surface was significantly reduced. It is reported that estimated blood loss was an important risk factor for postoperative overall complications. In this study, there was no significant difference in the rates of the overall complication between groups, although operative blood loss was significantly higher in Co-PD group.

There are several inherent limitations in this study. First, because the sample size of this study was small, and the follow-up time of the CMAFA group was short, analysis of the long-term survival was not performed. Second, the study periods of the two groups did not overlap, thus our results cannot rule out the impact of the learning curve of the surgeon over time. Third, due to the surgical records regarding the abnormal vascular were incomplete, our study did not specifically document the early recognition of arterial abnormalities during PD.

Conclusions

In conclusion, this study demonstrates that CMAFA is a feasible and efficient technique with acceptable perioperative and oncological outcomes in treating patients with PHC. Based on our current experience, we believe that CMAFA would be applicable for radical modular PD depending on the tumor location and extent of invasion to choose an optimal combination of AFAs. This novel concept needs further randomized controlled trials to confirm.

Author contributions

Conceptualization: Kai-Ming Leng, Zhi-Dong Wang, Yun-Fu Cui.

Data curation: Yun-Fu Cui.

Formal analysis: Xing-Ming Jiang, Zhi-Dong Wang.

Funding acquisition: Xing-Ming Jiang, Yun-Fu Cui.

Investigation: Hao Wang.

Methodology: Sheng Tai, Peng-Cheng Kang, Ming Wan.

Project administration: Yi Xu, Yun-Fu Cui.

Resources: Yi Xu.

Software: Peng-Cheng Kang, Hao Wang.

Supervision: Sheng Tai, Yun-Fu Cui.

Validation: Xiang-Yu Zhong, Sheng Tai, Zhi-Dong Wang, Yun-Fu Cui.

Writing – original draft: Kai-Ming Leng.

Writing – review & editing: Kai-Ming Leng, Xiang-Yu Zhong.