Comparison of Standard Distal Pancreatectomy and Splenectomy with Radical Antegrade Modular Pancreatosplenectomy

Abstract

Radical antegrade modular pancreatosplenectomy (RAMPS) has been reported to provide improved margin resection and lymph node retrieval for tumors of the body and tail of the pancreas compared with standard resection. We examined our experience with RAMPS and standard resection to determine differences in clinicopathologic outcomes. A comparison of RAMPS procedures was made to standard distal pancreatectomy and splenectomy examining various clinicopathologic variables through retrospective chart review. Twenty-six patients underwent distal pancreatectomy with or without splenectomy between November 2004 and June 2011. Twenty patients underwent standard resection and six patients underwent RAMPS procedures for a variety of histologies. As a result of the heterogeneity of diseases, which included benign lesions, margin status was not applicable in some cases and therefore was not assessed overall. Fisher’s exact test and Wilcoxon rank sum tests demonstrated a significant difference in number of lymph nodes removed with mean of 4.3 and 11.2 lymph nodes obtained for standard resection and RAMPS, respectively (P = 0.03). The RAMPS procedure for lesions of the body and tail of the pancreas retrieved significantly more lymph nodes than standard distal pancreatectomy and splenectomy. It should be the preferred surgical approach when lymph node count is important for tumor staging.

For over 100 years distal pancreatectomy has been the standard procedure for tumors in the body of the pancreas. The two goals of pancreatic resection include complete tumor resection with a margin of normal tissue and removal of regional lymph nodes. Tumors of the pancreatic head have been the focus of much innovation in achieving these two goals. Much effort has gone into developing surgical techniques that help eliminate positive margins at the uncinate or posterior border in pancreatic head resections. Extensive lymph node mapping has afforded surgeons more opportunity to resect regional nodes as well. Pancreatic body and tail tumors are reported to be just as aggressive invading locally and metastasizing through lymph nodes, but very little had been done to address improving these two markers of operative success.¹

Radical antegrade modular pancreatosplenectomy (RAMPS) addresses the apparent disparity between pancreaticoduodenectomy and standard distal pancreatectomy. Strasberg et al.¹ describes an operative technique that allows for a more complete dissection posteriorly and incorporates lymph node mapping for resection of all regional nodes.

RAMPS differs from a standard approach in that the initial dissection begins medially and the neck of the pancreas is transected early as well as the splenic vessels. The dissection continues posteriorly to the aorta at the celiac and superior mesenteric trunks. If the tumor does not break the posterior plane of the pancreatic body, the left adrenal gland is retained and the posterior plane of dissection continues left from medial exposing the left renal vein and clearing Gerota’s fascia off the left kidney. When the tumor breaks the posterior plane of the pancreas, the left adrenal is resected en bloc and the dissection continues posteriorly to the diaphragm using the retroperitoneal muscles as the posterior border, diaphragm as the superior border, and renal vein as the inferior border of the dissection plane. The rationale for this approach is to ensure a negative deep margin with complete regional lymph node dissection.

Using lymphatic mapping by O’Morchoe,² it is shown that there are two primary groups of nodes that should be considered N1 or regional (Fig. 1). The first is a ring of lymphatics fed directly by lymphatic drainage from the body and tail of the pancreas. The ring is comprised of gastrosplenic nodes incorporated in the gastrosplenic omentum, splenic nodes in the hilum of the spleen, infrapancreatic nodes deep to the pancreatic body, and gastroduodenal nodes. The second primary group lies along the aorta, in relation to the celiac and superior mesenteric arteries, and may receive direct lymph drainage from the pancreas and thus must be considered N1. Therefore, to achieve full resection of all potentially N1 nodes, a complete resection should include all the nodes of the ring described as well as the celiac lymph nodes and those anterior to and to the left of the superior mesenteric artery. Given the early ligation of the splenic vessels in the medial approach and necessity of removing the nodes in the splenic hilum, all patients undergoing RAMPS also undergo splenectomy.

Fig. 1.

Lymph node mapping of the pancreas.

In summation, RAMPS offers improved visualization of the posterior dissection plane, offers early control of the major blood vessels encountered, early division of the neck of the pancreas, better control of the splenic vessels,³ and the lymph retrieval allows for a more complete regional nodal dissection. Using this technique, improved margin resection and lymph node retrieval for tumors of the body and tail of the pancreas have been reported compared with standard resection. We examined our experience with RAMPS compared with standard resection to determine differences in clinicopathologic outcomes.

Methods

This study was approved by the Wake Baptist Health Institutional Review Board and Protocol Review Committee of the Wake Forest Comprehensive Cancer Center.

Data Collection

A prospectively created database of patients undergoing standard distal pancreatectomy or RAMPS included 26 patients between November 2004 and June 2011. The data were acquired from operative notes, anesthesia notes, pathology reports, and follow-up visits.

Procedure

A standard surgical technique was used for the traditional distal pancreatectomy. This typically consisted of either one of two approaches. Vascular control was then obtained of the splenic artery and vein, which were then serially transected with a vascular stapler. The pancreatic body was then transected with an endoscopic stapler using a blue, green, or white load depending on the pancreatic body texture. In the other method, the dissection began medially at the proposed site of transection on the pancreatic body. Control is again obtained of the splenic artery and vein and of the pancreatic body. The pancreas is transected first with an endoscopic stapler followed by stapling of the splenic vessels. In both approaches, the dissection posterior to the pancreas follows the natural plane between the retroperitoneum. RAMPS procedures were carried out as previously described and using an anterior or posterior approach (Fig. 2A-B).¹ Using RAMPS, the resection of the body and tail of the pancreas was hoped to be achieved with more complete N1 node dissection, improved visibility for the posterior margin, and superior blood vessel control.

Fig. 2.

(A) Computed tomography (CT) scan with anterior dissection plane. (B) CT scan with posterior dissection plane.

Pathologic Assessment

On gross examination, surgical specimens included a segment of distal pancreas and associated peripancreatic soft tissues. Depending on the surgery, a portion of stomach, spleen, and/or adrenal gland would be present. The pancreatic segment would be serially sectioned to expose the lesion. When applicable, the lesion would get evaluated for extension into peripancreatic soft tissues, vasculature, spleen, and/or the adrenal gland with representative sections submitted for microscopic examination. A lymph nodal evaluation would then be performed on the peripancreatic soft tissues with attention focused on the posterior pancreatic aspect near vasculature. Lymph node candidates, palpable or appreciable visually, are dissected from the adipose tissue and submitted for microscopic examination. The microscopic examination would include characterization of the lesion, a total lymph node count, and assessment for possible lymph node metastases. With some cases for malignant tumors with no lymph node candidates found on initial evaluation, the entire peripancreatic fat would be submitted for evaluation to obtain more lymph nodes.

Data Analysis

Notable clinicopathologic variables examined included: total nodes resected, operative time (OR time), estimated blood loss (EBL), length of stay (LOS), and complications. There was high heterogeneity of disease, including some benign lesions, and therefore the resection margins were not compared in this analysis. For analyzing a small sample size, Fisher’s exact test and Wilcoxon rank sum tests were used to measure association between our clinicopathologic variables and the type of surgical resection performed. To assess the difference between patients receiving RAMPS and patients not receiving RAMPS, univariate analyses were performed. Frequencies and percentages were calculated for margin status and complications by RAMPS procedure and statistical significance was assessed using Fisher’s exact tests. For total number of lymph nodes, OR time, EBL, and LOS, means and standard deviations were calculated and statistical significance was assessed using Wilcoxon rank sum tests. All analyses were performed using SAS 9.2 (Cary, NC).

Results

Twenty-six patients underwent distal pancreatectomy with or without splenectomy between November 2004 and June 20, 2011. Twenty patients underwent standard resection for the following histologies: one adenocarcinoma, five mucinous cystic neoplasms, eight neuroendocrine tumors, one serous cystadenoma, three pseudocysts, one intraductal mucinous papillary neoplasm, and one solid pseudopapillary neoplasm. Six patients underwent RAMPS procedures with the following histologies: three adenocarcinomas, one neuroendocrine tumor, one mucinous cystadenoma, and one serous cystadenoma (Table 1).

Table 1.

Resection Method by Tumor Classification

Resection Method and Tumor Classification	Number
Standard resection Adenocarcinoma Mucinous cystic neoplasms Neuroendocrine Serous cystadenoma Pseudocyst Intraductal mucinous papillary neoplasm Solid pseudopapillary neoplasm	20 1 5 8 1 3 1 1
RAMPS Adenocarcinoma Neuroendocrine Mucinous cystadenoma Serous cystadenoma	6 3 1 1 1

RAMPS, radical antegrade modular pancreatosplenectomy.

Table 2 shows clinicopathologic variables measured including total nodes resected, OR time, EBL, and LOS. As a result of the heterogeneity of diseases including benign lesions, margin status was not compared between RAMPS and standard resection. In operations involving cancers, the operative technique for labeling and pathological technique for assessment varied significantly enough during the course of this study to make comparison insignificant. Of the 20 standard resections, nine were for cancer and there was one positive margin. Of the six RAPMS operations, four were performed on cancers with zero positive margins. Four patients who underwent RAMPS procedures had greater than 14 lymph nodes resected, one had seven resected, and one had one node resected. The patient with one resected had undergone octreotide therapy for VIPoma before surgical intervention. One standard resection resulted in 20 nodes removed, which was the only traditional approach to achieve more than 12 removed. Six standard resections resulted in zero nodes removed.

Table 2.

Outcomes of Standard Resection vs RAMPS Procedure

Variable	Standard Method		RAMPS		P Value
	Mean	Standard Deviation	Mean	Standard Deviation
Total nodes	4.25	5.418	11.17	5.913	0.03
OR time (minutes)	295.26	83.823	300.00	86.948	0.97
EBL (mL)	581.25	559.186	500.00	260.768	0.79
LOS (days)	6.85	1.387	7.67	3.077	0.62

RAMPS, radical antegrade modular pancreatosplenectomy; OR, operating room; EBL, estimated blood loss; LOS, length of stay.

Table 3 shows the difference in approach for the standard resections. Exactly half were performed laparoscopically and half were open resections. The resection direction is also exactly divided in half. There were equal numbers of medial-to-lateral resections as there were lateral to medial.

Table 3.

Standard Technique Approach

Variation	Number
Incision Open Laparoscopic	20 10 10
Dissection Medial to lateral Lateral to medial	20 10 10
Splenectomy Spleen-preserving Spleen-resecting	4 16

Fisher’s exact and Wilcoxon rank sum tests showed no difference in OR time (295.26 vs 300.00 minutes; P = 0.97), EBL (581.25 vs 500.00 cc; P = 0.79), or length of hospital stay (6.85 vs 7.67 days; P = 0.62) for standard resection versus RAMPS, seen in Table 2. A significant difference in number of lymph nodes removed (4.3 vs 11.2; P = 0.03) was found for standard resection and RAMPS, respectively.

Complications also were not significant (11 of 20 vs three of six; P = 1.00) for standard resection versus RAMPS, seen in Table 4. Complications were categorized and tabulated (Table 5) using the Clavien grading system (Table 6). There were a total of 14 complications with the standard approach versus five total complications with RAMPS. This resulted in 10 Grade I, three Grade II and one Disability Grade for Standard Resection, and four Grade I and one Grade IV complications with RAMPS.

Table 4.

Complications by Procedure

Complication	Standard Resection		RAMPS		P Value 1.00
	Number	Percent of Total	Number	Percent of Total
Yes	12	60	3	50
No	8	40	3	50

RAMP, radical antegrade modular pancreatosplenectomy.

Table 5.

Complications by Type

Complication	Standard Resection		RAMPS
	Number	Percent of Total Patients	Number	Percent of Total Patients
Vomiting	2	10	2	33
Bleeding	1	5	1	17
Pancreatic leak	6	30	0	0
Diabetes mellitus Type I	1	5	0	0
Diarrhea	2	10	1	17
Pneumonia	1	5	0	0
Surgical Site Infection	1	5	1	17
Total	14/20		5/6

RAMPS, radical antegrade modular pancreatosplenectomy.

Table 6.

Complications by Clavien Grading System

Complication	Standard Resection		RAMPS
	Number	Percent of Total Patients	Number	Percent of Total Patients
Grade I	10	50	4	66
Grade II	3	15	0	0
Grade III	0	0	1	17
Grade IV	0	0	0	0
Grade V	0	0	0	0
Disability	1	5	0	0
Total	14/20		5/6

RAMPS, radical antegrade modular pancreatosplenectomy.

Conclusion

The RAMPS procedure for lesions of the body and tail of the pancreas retrieved significantly more lymph nodes than standard distal pancreatectomy and splenectomy. Margin status was not assessed in this diverse pancreatic resection cohort as a result of the benign nature of some neoplasms. The lymph node resection in RAMPS was aided in large part by O’Morchoe’s work in discovering the lymphatic drainage of the human pancreas.² Strasberg et al. has published data showing their experience with RAMPS over time.^{1, 4} Their study includes 47 patients who were followed and showed a 35 per cent 5-year survival rate. Contrasted with a literature review of standard distal pancreatectomies, this is much improved over the reported range of 10 to 19 per cent.⁴

This study demonstrated increased lymph node yield with RAMPS (mean of 11.2) compared with standard distal pancreatectomy (mean of 4.3) with a P value of 0.03. However, of the 26 operations, 23 margins were negative. All three margins that were not negative were in the standard distal pancreatectomy group. One was a positive retroperitoneal margin in a neuroendocrine tumor, another was a serous microcystic adenoma with equivocal margins, and the last one was a pancreatic intraepithelial neoplasm, which had evidence of PanIN-1 at the margin. An argument could be made the last two cases really were not positive margins because the lesions were not invasive malignancies.

The standard distal pancreatectomies were performed either open or laparoscopic using a medial-to-lateral or lateral-to-medial approach. RAMPS offers improved visualization of the posterior dissection plane, early control of the major blood vessels encountered, and, with early division of the neck of the pancreas, better control of the splenic vessels. RAMPS procedures were performed according the procedure described by Strasberg in 2003. The division of the neck of the pancreas is performed immediately after incision and access to the pancreas. Celiac lymph node dissection is then made and a determination of dissection anterior or posterior to the adrenal gland is made based on whether the tumor breaks the posterior plane of the pancreatic body or tail. This determines the posterior extent of the dissection. The dissection then proceeds laterally to the spleen, mobilizing the tail of the pancreas along the way.¹

Despite the more aggressive dissection in the RAMPS procedure, there was no noted difference in perioperative outcomes. There was no statistical difference in EBL, LOS, or OR time. This is in contrast to the work performed by Bonenkamp examining D2 lymph node dissection in gastric cancer whose radical approach showed significantly higher rates of complications than did those in the standard group (43 vs 25%, P < 0.001), more postoperative deaths (10 vs 4%, P = 0.004), and longer hospital stays (median, 16 vs 14 days; P < 0.001).⁵

Success of a surgical approach to pancreatic cancer is measured primarily using margin status and lymph node yield. As described previously, margin status was not assessed in this study because not all resections were performed on malignant disease. However, our comparison of RAMPS with standard distal pancreatectomy showed a statistically significant increase in lymph node yield using RAMPS. This finding is important and supports the use of RAMPS as the treatment of choice for neoplasms in the body and tail of the pancreas.

Lymph node resection plays a key role in determining the prognosis of pancreatic cancer. Nodal status and number of nodes resected are two major factors considered when assessing the value of a lymphadenectomy. As described previously, the technique for lymph node resection in RAMPS takes the lymphatic drainage into account, allowing for complete N1 nodal resection. Therefore, this approach inherently offers a more complete dissection of potentially involved lymphatic tissue. Further supporting this are data suggesting that increased yield in pancreatic lymph node resection improves survival prognosis.^{1, 4, 6}

Huebner et al. showed that in lymphadenectomies in which 11 or more nodes were examined, there is increased accuracy of survival prognosis in node-negative patients. Furthermore, it was shown that if less than 11 lymph nodes were examined, the risk of missing one or more positive nodes was between 10 and 41 per cent. Their results showed that patients with suboptimal node removal (less than 11) had worse survival outcomes and suggests that the low yield was responsible for missed metastatic disease.⁶ Our data showed an average of 11.17 nodes resected in the RAMPS group compared with 4.25 nodes removed in the standard resection group.

The anatomic location of lymph node dissection in RAMPS is also superior to the standard approach. Schwarz et al.⁷ showed that in tumors of the pancreatic head, half of lymph node involvement was on the posterior aspect of the pancreas with one-third lying adjacent to the superior mesenteric artery with one-fifth being paraaortic. The approach taken in the RAMPS lymph node dissection is intended to resect a more posterior fascial plane of tissue to include sentinel nodes, as mapped by O’Morchoe, specifically including superior mesenteric and periaortic nodes. Schwarz goes on to conclude that increased node removal and examination influences the quality of staging and overall survival. Specifically their data show that 15 total nodes removed or approximately 10 negative nodes optimized operative benefit.⁷

To conclude, lymph node yield and status is a significant factor in pancreatic cancer prognosis and staging. House et al.,⁸ using data on 696 patients undergoing pancreatic resection, concluded that inadequate surgical lymphadenectomy gives an N0 patient similar outcomes as patients with one node positive out of 12 assessed. This finding supports the rationale that a more thorough anatomic dissection will not only increase lymph node yield, but in resecting all N1 sites, allow for comprehensive pathologic assessment of all primary sites of metastasis.

There were limitations to our study. The sample size is relatively small with only 26 patient experiences to analyze. This is attributable in large part to the fact that all operations were performed by the same surgeon, which is a strength countering the small size. This ensured intraoperative consistency of technique and approach that many other multicenter trials were unable to achieve. The long specimen collection interval also introduces another confounding factor. Pathologic assessment was not performed by the same pathologist. Thus, the pathologic assessment, including lymph node analysis and margin status in particular, may have not been as controlled as if one pathologist evaluated all 26 specimens. Also, the retrospective nature of this study limited our ability to record consistent data regarding complications and operative/pathologic data.

Our study supports the use of RAMPS as the preferred approach to neoplasms of the body and tail of the pancreas. Based on sound anatomic and lymphatic mapping studies, RAMPS resection is actually not a radical procedure, but merely combines a partial pancreatectomy with a complete regional lymphadenectomy. There was no increase in perioperative morbidity and mortality and the increased nodal yield may contribute to improved staging of disease. We aim to continue studying our experience with RAMPS and evaluate it for future improvements including the potential for a laparoscopic approach with or without robotic assistance.