Multivisceral Ex Vivo Surgery for Tumors Involving Celiac and Superior Mesenteric Arteries

Authors


Tomoaki Kato, Tk2388@columbia.edu

Abstract

Abdominal tumors involving both roots of the celiac and superior mesenteric artery are deemed unresectable by conventional surgical methods. We performed three cases of multivisceral ex vivo surgery involving temporary removal of the entire abdominal viscera followed by vascular reconstruction, ex vivo tumor resection and autotransplantation of excised organs. We achieved a complete tumor resection with negative margins in all cases. All patients have survived with no tumor recurrence to date at 17-, 27- and 38-month follow-up. Postoperative complications included diarrhea, sphincter of Oddi dysfunction and arterial stenosis; all responded to directed treatments. Multivisceral ex vivo surgery applying techniques of deceased donor multivisceral transplantation is feasible in achieving local control of otherwise unresectable abdominal tumors. This surgery is best suitable for locally invasive tumors unresectable because of location and vascular involvement.

Abbreviations: 
IRB

Institutional Review Board

SMA

superior mesenteric artery

SMV

superior mesenteric vein

We report three cases of novel multivisceral ex vivo surgery. Previous ex vivo surgeries were performed on single organs (kidney, liver or intestine) with autotransplantation of the particular organ (1,4,5,7). Intestinal ex vivo surgery is indicated when tumor only involves the root of the mesentery sparing the hepatic and celiac artery. In this series, we report cases of temporary removal of entire abdominal viscera to remove extensive tumor that involves both celiac artery and the superior mesenteric artery (SMA). We describe the technical, clinical and oncological feasibility of the procedure.

Multidisciplinary Team Discussion, Informed Consent and Institutional Review Board Review

Each case was thoroughly reviewed by a multidisciplinary team that considered the extent of disease, metastatic potential of the tumor and the estimated survival without surgery. Evaluation included imaging studies to rule out extraabdominal disease spread, oncologic review of prior treatment, cardiopulmonary risk assessment, nutritional and performance status and psychosocial assessment. Alternative surgical approaches were discussed including multivisceral allotransplant, and conventional resection with positive margins followed by adjuvant medical intervention. After considering short- and long-term risks and benefits of the procedure, we concluded that multivisceral ex vivo surgery was a reasonable option. All three patients (or their families) were provided the details of the above discussion, the risks and potential benefits of the procedure and alternative approaches. Written informed consent for the procedure was obtained from the patient (or guardian). Written permission was also obtained for the publication of this report. The Institutional Review Board (IRB) of the Columbia University Medical Center reviewed the patient's permission and full IRB review was deemed unnecessary for this report.

Case 1

The patient was a 63-year-old female with a leiomyosarcoma involving the abdominal aorta and the roots of celiac and mesenteric arteries. The tumor size was 7.4 × 6.4 cm at first presentation. The mass was considered unresectable by conventional surgical techniques because of its location and vascular involvement. The patient underwent chemotherapy and radiation therapy with partial response. The tumor decreased in size but was still deemed unresectable. After a thorough discussion of the surgical risks and benefits (acknowledging unknown risk because this was the first attempt of its kind), we performed our initial case. The final decision to proceed was made at the time of exploration. After division of the gastrointestinal tract at the gastroesophageal junction and splenic flexure of the colon we were able to mobilize the entire abdominal viscera from the retroperitoneum. Given the location, size of the tumor and vascular elements involved the team was confident that the viscera could be salvaged after ex vivo tumor resection. After dissection and encircling of the abdominal aorta above and below the organ bloc, we continued with the planned operation. The organs were eviscerated en bloc including the liver, stomach, pancreaticoduodenal complex, spleen and small and large intestine to the distal transverse colon. Veno-venous bypass was not needed. After the abdominal organs were removed, the portions of resected abdominal aorta and the vena cava were replaced by synthetic grafts (Figures 1 and 2). The left and the right renal arteries were temporarily disconnected with the kidneys remaining in situ and were reimplanted to the synthetic vascular graft (Figures 1 and 2). The organ block was moved to the back table, flushed with preservation solution and chilled. The tumor was dissected from the organ block in a cold basin (Figure 3). The common hepatic artery and the SMA were reconstructed with a bifurcated synthetic graft (Figure 1). The splenic artery and the left gastric artery were ligated. The organ block was reimplanted to the reconstructed grafts of both the vena cava and the aorta (Figure 1). The organ block included the stomach, pancreas, spleen, liver and intestine as in multivisceral allotransplantation. Total evisceration time was 90 min. The blood supply to the stomach left unaltered was from the gastroduodenal artery via the SMA and the right gastric artery via the hepatic artery. No gastric venous congestion was observed. Continuity of the gastrointestinal tract was reestablished using standard surgical techniques (Figure 1). A pyloroplasty was performed. The surgical specimen confirmed tumor-free margins. A temporary colostomy was created to permit continual surveillance of the autograft and was taken down 2 months later. The patient's immediate postoperative course was uneventful and she was discharged home on the 21st postoperative day with nutritional support via tube feeding. On the eighth postoperative month she developed biliary sludge because of sphincter of Oddi dysfunction that was treated with endoscopic sphincterotomy. Liver function tests normalized after the procedure. She remains well with no evidence of tumor recurrence at her 38-month postoperative follow-up.

Figure 1.

Summary of surgical procedure of Case 1. The tumor involved the abdominal aorta and the roots of the celiac artery and the SMA (A, B). It also involved the right and left renal arteries. The abdominal aorta and the vena cava were clamped above and below the organ block (C) and the entire abdominal viscera were moved to the back table (D). The tumor was removed at the back table and the hepatic artery and the SMA was reconstructed using Y-shaped synthetic vascular graft (E, F, G). Synthetic tube graft was placed to reconstructed resected portion of both abdominal aorta and the vena cava (I). Organ block was reimplated to the synthetic vascular graft (J). Continuity of the gastrointestinal tract was reestablished using standard surgical techniques (K).

Figure 2.

“Empty” abdominal cavity after total removal of the abdominal organs. The abdominal aorta and the vena cava were replaced with synthetic grafts (A). Renal arteries were reconnected to the synthetic graft (B).

Figure 3.

All abdominal organs were moved to the basin on the back bench. The tumor on the abdominal aorta encasing the celiac artery and the SMA was dissected. Multivisceral ex vivo surgery enabled the surgical team to access the tumor located deep in the abdomen involving major vascular pedicles in the bloodless field with easy exposure.

Case 2

The patient was a 7-year-old girl who presented with an extensive inflammatory myofibroblastic tumor of the pancreas involving both celiac artery and the SMA (Figure 4). The tumor occupied most of the pancreas. The portal vein, superior mesenteric vein (SMV) and splenic vein were encased by the tumor and were thrombosed including the mesosplenic junction. She developed portal hypertension and ascites. The patient had a prior attempt at resection approximately a year prior including partial pancreatectomy and distal gastrectomy. The tumor recurred in the pancreas 6 months after the original surgery. The tumor was deemed unresectable by conventional surgical methods because of its location. She received medical therapy with antiinflammatory agents including celecoxib, steroids and subsequently chemotherapy with etoposide with no response. We proposed multivisceral ex vivo procedure. The preprocedure imaging revealed that the remaining stomach, pancreas and the spleen would likely be sacrificed to achieve complete resection of the tumor. We were hoping to salvage pancreatic islets to avoid brittle diabetes, but extensive infiltration of the organ was found intraoperatively. In case reimplantation of the liver was not possible secondary to vascular tumor invasion, we prepared the father of the child as a potential living donor. The father underwent the standard living donor evaluation and was on standby at the time of his daughter's procedure. Although both the celiac artery and the SMA were encased by the tumor, the abdominal aorta was tumor free. We were able to preserve 1–2 mm of vascular cuff at the origins of the celiac artery and the SMA (Figure 5). The liver and the intestine were separated from the organ block. The tumor involved the right and left bifurcation of the hepatic artery and an ex vivo reconstruction was performed at the back table (Figure 5). The liver was autotransplanted first using inflow provided by the left renal vein to the portal vein (renoportal anastomosis using a jugular vein interposition graft; Figure 5). The hepatic artery was anastomosed to a synthetic graft placed as an interpositional graft to the root of celiac artery (Figure 5). After the liver autograft was reperfused, we proceeded with autotransplantation of the intestine (Figure 5). The SMA was anastomosed using synthetic graft placed on the root of SMA. The SMV was anastomosed to the side of jugular vein graft (connecting left renal vein to the portal vein). Approximately 15 cm of the proximal jejunum was sacrificed because of tumor infiltration of the mesentery. Cold preservation times for the liver and the intestine were 163 min and 218 min, respectively. Surgical margins were clear of tumor. She was discharged home on an insulin pump with nutritional support via jejunal tube feeds and pancreatic enzyme replacement on the 32nd postoperative day. A temporary ileostomy for bowel surveillance was taken down 3 months after the resection. She required lysis of adhesions 7 months later but remains tumor-free at 27-month postoperative follow-up.

Figure 4.

The tumor was occupying most part of pancreas (Case 2) encasing the celiac artery (arrow) (B) and the SMA (arrow; C). Upper border of the tumor was at the hepatic hilum (A) and the lower border of the tumor was at the root of mesentery (D).

Figure 5.

Summary of surgical procedure of Case 2. The tumor was occupying almost entire pancreas. The celiac artery and the SMA were encased (A). Entire abdominal viscera were moved to the back table (B). The liver and the intestine were separated from the tumor and the hepatic artery was reconstructed at the back table (F, G). The stomach, pancreas and spleen were sacrificed (C). Synthetic vascular grafts were placed on the roots of the celiac artery and the SMA (E). The liver was reimplanted using inflow from the left renal vein with interpositional jugular vein graft (H). The intestine was reimplanted connecting autograft SMV to the side of jugular vein graft (I). Biliary reconstruction and GI reconstruction was done using standard surgical techniques (J, K).

Case 3

The patient was an 8-year-old girl with a Kaposiform hemangioendothelioma. She was followed with a 2-year history of a pancreatic head mass with a presumed diagnosis of chronic pancreatitis. She subsequently developed biliary stenosis requiring repeated endoscopic stenting of the common bile duct. She later developed colonic obstruction requiring surgical resection. The surgical specimen demonstrated Kaposiform hemangioendothelioma. She was referred for further management. Imaging studies revealed a pancreatic head mass and extensive portomesenteric venous thromboses. The mass was encasing the celiac artery and the SMA. The portal vein, SMV and splenic vein were thrombosed as in Case 2. We informed the family regarding treatment options and postoperative anticipated consequences and proceeded with an exploration with an anticipated multivisceral ex vivo procedure. Intraoperatively, we confirmed that the pancreas mass was indeed a Kaposiform hemangioendothelioma and performed a procedure similar to the one described in detail in Case 2 (including total abdominal evisceration, vascular reconstruction and autotransplant of the liver followed by intestine, with sacrifice of the pancreas, spleen and stomach). In this case, the autograft did not include the colon because of a previous right hemicolectomy. Cold preservation times for the liver and the intestine were 128 min and 195 min, respectively. The patient developed hepatic artery stenosis that was treated successfully with balloon dilatation and she was discharged home with an insulin pump with nutritional support via jejunal tube feeds and pancreatic enzyme replacement on the 44th postoperative day. She is doing well at her 17-month follow-up.

Discussion

Ex vivo surgery and autotransplantation is a technical byproduct of organ transplantation. The first ex vivo repair and autotransplantation of the kidney was reported in a patient with iatrogenic ureteral injury (1). Subsequently, the technique was used in renovascular disease (2,3). The extended use of ex vivo surgery was applied in patients with hepatic tumor that was considered unresectable using conventional techniques (4). Ex vivo liver surgery is feasible secondary to advances in organ preservation and liver transplant, evolving in a manner similar to ex vivo repair of the kidney in concert with kidney transplantation. Ex vivo hepatectomy has been reserved for large primary or secondary hepatic malignancies with extensive vascular involvement (5). Though use of ex vivo resection for malignant hepatic tumors has been controversial, recent advances in technique and organ preservation suggests the approach is appropriate for selected cases (6).

The technology of intestinal transplantation has opened a new horizon in ex vivo surgery (7). Under conventional surgical methods, a tumor that involves the mesenteric vascular pedicle is considered unresectable. An en bloc removal of such a tumor together with the intestine, ex vivo resection and autotransplantation of the intestine has been performed applying techniques of intestinal transplantation (8). Alternatively, the resection in such cases can also be achieved with intestinal allotransplantation; however, it carries significant morbidity and mortality because of a high rate rejection (9–11). Avoiding allotransplantation by performing autoransplantation is far more desirable. In some cases of the intestinal ex vivo procedure we have reported previously, we removed the pancreatic head and/or part of the liver. However, in those cases, the hepatic artery was intact and the liver or pancreas was not completely removed from the body because the tumor did not involve the celiac artery.

Extending the idea of the intestinal ex vivo procedure, we performed a multivisceral ex vivo procedure. With this approach, tumor that involves both roots of the celiac artery and the SMA may be removed. Theoretically, a multivisceral ex vivo procedure is the ultimate form of abdominal ex vivo surgery that potentially enables a surgeon to operate on tumors with visceral vascular involvement if the tumor carries a favorable prognosis after complete resection. Multivisceral allotransplantation may also be used for locally invasive tumors; however, as in intestinal allotransplantation, multivisceral allotransplantation involves significant risk associated with organ rejection (12,13).

One of the most difficult surgical decision points surrounding these cases was to determine whether or not organs can be reimplantated after the tumor resection at the back table. It is difficult to assign specific inclusion and exclusion criteria because they have yet to be established. We do believe this technique can be applied for locally invasive tumors with low metastatic potential such as sarcomas. In a multivisceral ex vivo procedure, one at least has to be able to reimplant the liver and intestine. Preoperative assessment suggesting that these organs cannot be salvaged would constitute an exclusion criterion for the procedure. Both liver and intestine autografts can be partial if the remaining size and/or length are adequate. Therefore if the vascular involvement of those organs is limited to their relatively proximal branch level, theoretically, this procedure is feasible using partial hepatectomy or partial enterectomy. Alternatively, if the tumor involves more than the tertiary branches of these vital organs, candidacy for the procedure is likely excluded because vascular reconstruction may not be possible.

Each case had its unique challenges. In Case 1, because the tumor involved the abdominal aorta and the more proximal portions of the celiac artery and the SMA, the biggest surgical challenge was whether or not the aorta could be mobilized safely particularly after radiation injury. Preoperative assessment demonstrated that the distal hepatic artery and the distal SMA were clear from the tumor suggesting that reimplantability of these organs was not a limiting factor in the decision to proceed. Considerable amount of time was spent at the beginning of surgery to ensure feasibility of the procedure. We mobilized the organs completely from the retroperiotoneum to the level that entire abdominal viscera were only connected to the body with abdominal aorta and the vena cava. In Case 2, the tumor involved the more peripheral branches of the celiac artery and the SMA. We anticipated sacrificing the pancreas and the stomach because of the tumor location and vessels involvement. We were confident that the segment of SMA below the pancreas could be separated from the tumor; however, because the tumor involved the hepatoduodenal ligament extensively, there was uncertainty if the hepatic artery could be reconstructed. Given this concern we prepared the father of the child as a standby living donor, in case reimplantation of her native liver was impossible.

The greatest risk of ex vivo procedures is vascular thrombosis as in any transplant surgery. Vascular thrombosis is not a common occurrence in multivisceral transplantation because the vascular anastomoses are relatively large in size. In our procedures, synthetic vascular grafts were used to replace vessels that were involved by tumor. The vascular anastomoses in the two pediatric cases (Cases 2 and 3) were more complex compared to standard multivisceral allotransplantation because the liver and intestine were autotransplanted separately. Although we avoided vascular complications, the risk and consequences of vascular thrombosis may be higher than that seen in allotransplant surgery.

Reconnecting liver and intestine autografts at the back table and en bloc reimplantation of liver and intestine may be a consideration in Cases 2 and 3. We opted not to perform en bloc reimplantation because the portal vein would be susceptible to kinking and the anhepatic phase would be prolonged. Although the renoportal anastomosis added some complexity to the procedure, it stabilizes the portal vein to avoid kinking and provides the necessary inflow to the liver autograft. Long-term safety and efficacy of a renoportal anastomosis was known from the author's own experience (14).

Another potential complication is ischemia reperfusion injury to the organs that were autotransplanted. Relatively shorter cold preservation times in this procedure as compared to standard allotransplantation makes this complication less likely. All three patients fared well in terms of recovery from ischemic reperfusion injury. Liver function tests normalized in 3–7 days. As in intestinal and multivisceral allotransplant, the denervation of the bowel causes dysmotility and diarrhea. This complication is manageable and usually improves with time as in multivisceral or intestinal allotransplantation recipients.

Total gastrectomy and total pancreatectomy were necessarily in Cases 2 and 3. The impact of these procedures could be significant on long-term outcome. Total pancreatectomy would expect to result in maldigestion and diabetes. Pancreatic enzyme supplementation can address the exocrine deficiency adequately but must be adjusted with growth. The loss of glucose regulatory and counterregulatory hormones has more profound consequences causing brittle type of diabetes mellitus. At this point in time both patients are able to tolerate oral diabetic diets with supplemental tube feeding at nighttime without any intravenous fluid requirement. They are both using an insulin pump with relatively well-controlled blood sugar level. We did not include detailed nutritional assessment data but both have normal serum albumin levels and are in the 25–50th percentile for weight by Z score. They are both attending school full time. We continue to monitor the long-term nutritional outcome as a measure of success of the procedure.

In oncologic surgery, a tumor is deemed inoperable, either because of the extent of the tumor obviating surgical intervention or the location of the tumor making resection technically impossible. Other considerations include the ability of the patient to tolerate a radical procedure and the biological aggressiveness of the tumor. If tumor is too extensive and the surgical removal of the tumor is not expected to change the outcome, then ex vivo resection should not be considered even when the tumor could be removed technically. In addition, the role of this procedure is questionable for highly malignant tumors with a tendency to metastasize early or aggressively. We believe that this procedure is best suited to locally invasive tumors for which a tumor-free surgical margin is expected to benefit the patient's long-term outcome. Despite significant risks involved in the procedure, all three patients have survived the perioperative period and have demonstrated no recurrence of tumor to date. With further demonstration of operative safety and long-term disease free survival, ex vivo resection may become first line surgical therapy for young patients with these anatomically challenging tumors.

In summary, we performed multivisceral ex vivo procedures in three patients with locally invasive tumors. All had tumors deemed unresectable with conventional surgical techniques because of location and vascular involvement. Surgery was performed successfully and to date the patients are thriving without tumor recurrence albeit with short to midterm postoperative follow-up. This technique appears both feasible and effective in the proper clinical situation when performed by a multidisciplinary team experienced in multivisceral transplantation, expanding the role of transplant surgery in specific oncological problem.

Acknowledgment

No funding was provided for this study.

Disclosure

This manuscript was not prepared in any part by commercial organization.

The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.

Ancillary