Intestinal Autotransplantation for Locally Advanced or Locally Recurrent Colon Cancer Invading Superior Mesenteric Artery

Abstract

Objective:

To examine the outcomes of intestinal autotransplantation (IATx) in patients with locally advanced (LACC) or locally recurrent (LRCC) colon cancer invading the superior mesenteric artery (SMA).

Background:

SMA involvement in LACC or LRCC is deemed unresectable and is associated with a poor prognosis. Combined extended resections of multiple organs together with SMA, followed by IATx may offer favorable clinical outcomes. However, data on its safety and efficacy are scarce.

Methods:

This retrospective cohort study included patients undergoing IATx between May 2018 and December 2022 in intestinal transplant programs at 2 university-affiliated hospitals in China. Patients with LACC or LRCC concomitantly with SMA contact of more than 180 degrees were included. Patients with a locoregional peritoneal, pelvic, or distal metastasis were excluded.

Results:

Ten patients underwent either IATx combined with pancreaticoduodenectomy (n = 8) or IATx alone (n = 2). Eight patients (80%) were males, and the median age was 55 years (range, 32–71 years). The Kaplan-Meier estimates for recurrence-free survival and overall survival at 3 years after IATx were 68% and 80%, respectively. No perioperative deaths occurred. All 10 patients experienced postoperative complications, including Clavien-Dindo grade I (n = 1), grade II (n = 4), grade IIIa (n = 1), grade IIIb (n = 3), and grade IVa (n = 1), which comprised acute venous thromboses, upper gastrointestinal hemorrhage, anastomotic leak, gastropareses, and significant pleural effusions. With an average follow-up of 23.9 months, 8 patients (80%) were currently alive without evidence of disease.

Conclusion:

Extended resection for LACC or LRCC invading SMA can be performed safely and is associated with prolonged survival.

Colon cancer is the third most common malignancy and the second leading cause of cancer-related deaths worldwide.¹ Despite considerable advances in the management of colon cancer, locally advanced (LACC) or locally recurrent (LRCC) colon cancer remains difficult to treat.^2–4 At diagnosis, ~5% to 12% of patients have involvement of adjacent organs (T4 tumor) or locally advanced disease.⁵ For 70% of all patients with colon cancer, the standard therapy is invariably surgical resection with curative intent.⁶ Among them, 30% to 40% will develop relapse after curative resection for primary colon cancer, with isolated locoregional recurrence ranging from 4% to 13%, usually on the sites of perianastomosis, mesentery, retroperitoneum, and peritoneum.^7–9 Because of the prevalence of colon cancer and its frequent intra-abdominal recurrent pattern, patients with LACC or LRCC represent an important group for which an effective strategy should be available.¹⁰

LACC and LRCC present technical challenges as these lesions often involve adjacent organs and structures. As right colon cancer directly invades the duodenum and pancreas, extended en bloc resection of the involved structures to achieve microscopic negative margins (R0) may result in favorable clinical outcomes.^11,12 Although rare, involvement of the superior mesenteric artery (SMA) in LACC or LRCC is traditionally deemed unresectable and is associated with a dismal prognosis. Currently, there are no effective surgical or nonsurgical treatments for such patients. Combined resections of multivisceral organs with the SMA remain poorly delineated, probably because of the surgical risks of irreversible intestinal ischemic damage. In recent years, the development of ex vivo surgery and intestinal autotransplantation (IATx) by our team and others may render such a challenging disease potentially treatable.^13–15 Hereby we described 10 patients with LACC and LRCC invading the SMA undergoing either combined pancreaticoduodenectomy (PD) and IATx or IATx alone. The surgical indications for IATx, perioperative complications, and short and mid-term follow-ups are described in this report. Our efforts illustrate that aggressive multivisceral resection including the SMA is feasible and is associated with prolonged survival.

METHODS

Patients

In this cohort study, a retrospective analysis of prospectively collected data from 2 academic centers (First Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou and First Affiliated Hospital, the Fourth Military Medical University, Xian, China) was conducted for patients who had undergone IATx between November 2018 and December 2022. The clinical, pathologic, and operative records, and follow-up data on these patients were reviewed. The Institutional Review Board and Ethics Committee approved the study protocol at each site. All patients provided written informed consent before the procedure, which was performed in accordance with the Declaration of Helsinki.

The surgical indication for IATx was LACC or LRCC concomitant with SMA contact of more than 180 degrees but with no evidence of distal metastasis. We also did not consider patients with pathologically confirmed peritoneal or pelvic metastasis. In addition, patients with recurrent disease should have a disease-free interval of at least 6 months. The patients who failed standard treatment and did not benefit from systemic chemotherapy were also considered for IATx. Patients with a present or prior distal metastasis of other malignant neoplasms were excluded from this study. All potential candidates were evaluated by a multidisciplinary team comprising colorectal and transplant surgeons, oncologists, pathologists, and radiologists to determine their resectability. All the patients underwent ¹⁸F-fluorodeoxyglucose positron emission tomography to exclude a distal metastasis. A computed tomography (CT) angiography is routinely performed to assess the anatomy and pathology of the mesenteric vascular tree. LACC and LRCC were downsized with systemic chemotherapy where appropriate before IATx.

Surgical Procedures

The surgical procedure has been described previously in detail.^13,16 In brief, the abdominal cavity was thoroughly examined for evidence of metastatic disease, particularly in the liver, peritoneal surfaces, and pelvic cavity. Lesions suspected of metastasis were biopsied and sent for frozen section analysis. Extensive kocherization of the duodenum and a subsequent medial rotation of the pancreatic head was performed to expose the origin of the SMA and to assess the extent of the disease. Upon confirming disease involvement, we decided to proceed with either PD-IATx or IATx alone.

A suitable segment of the intestine with reasonably sized SMA and superior mesenteric vein (SMV) for vascular anastomosis was initially evaluated and measured for future segmental grafting. We then identified and isolated the SMA and SMV just below the inferior border of the disease of the mesenteric root. The mesentery was subsequently divided into an inverted “V” shape with the tip of the “V” directed towards the isolated SMA and SMV. After the separation of the bile duct, stomach, and head of the pancreas, the SMA and SMV supplying the segmental graft were transected and the bowel autograft was removed and flushed with cold University of Wisconsin solution through the artery on the back table and subsequently kept chilled until use. In the setting of a short mesentery, a back-table reconstruction of the graft’s proximal SMA and SMV was performed using cryopreserved blood-type identical iliac allograft vessels.

After standard PD with extended lymphadenectomy, a bowel autograft was brought into the operative field and revascularized either through the infrarenal aorta or the retained SMA stump. The graft SMV was anastomosed to the portal vein or the vena cava. Upon establishment of the blood supply for the bowel graft, the pancreatic remnant was anastomosed to the posterior gastric wall between the lesser and greater curvature. The continuity of the gastrointestinal (GI) tract was reconstructed using a 45 cm Roux-en-Y limb to complete choledochoenterostomy, gastroenterostomy, colocolostomy, ileocolostomy, or ileostomy.

Postoperative Management

After the procedure, intravenous albumin was administered to maintain blood levels over 30 g/L. Prophylactic antibiotics were administered for at least 72 hours postoperatively. Ultrasonographic color Doppler imaging was used to evaluate blood flow in the SMA and SMV daily for 3 days postoperatively. When ultrasonography was inadequate for the evaluation of suspected vascular thrombosis, CT angiography was used to confirm the clinical diagnosis. All the patients were initially maintained on total parenteral nutrition. Enteral or oral feeding was resumed within a week after the operation and advanced as tolerated. Total parenteral nutrition was gradually discontinued when patients were able to provide sufficient enteral nutrition.

After discharge, the patients were followed up as outpatients once a month in the first year, every 3 months in the second year, and every 6 months thereafter. Chest, abdominal, and pelvic CT scans were obtained every 3 months in the first year and thereafter every 6 months thereafter.

Statistical Analysis

The Kaplan-Meier estimates were calculated using MedCalc for Windows, version 12.5 (MedCalc Software). Continuous variables are reported as medians and ranges and categorical variables are reported as the counts and percentages of the patient population.

RESULTS

Patient Characteristics

During the study period, 26 patients were referred to our institution because of advanced primary or recurrent colon cancer invading the SMA. Of these patients, 12 (46.2%) showed LACC or LRCC with no evidence of distant metastasis and were potential candidates for IATx. Two patients received palliative surgery because of pelvic and/or peritoneal metastases and were excluded from IATx. The remaining 10 patients met all requirements for IATx. The demographic and clinicopathological characteristics of the 10 patients with LACC (n = 2) and LRCC (n = 8) are listed in Table 1. There were 8 men and 2 women with a median age of 55 years (range, 32–71 years). The median Body Mass Index was 20.3 (range, 16.4–28.4). Primary cancers were located in the cecum (n = 3), ascending colon (n = 5), and transverse colon (n = 2). All patients had primary tumors equal to or greater than stage T2 (T2 in 2, T3 in 5, and T4 in 3). Seven patients (70%) had moderately differentiated tumors and 3 (30%) had poorly differentiated tumors. Positive lymph nodes were noted in 4 patients, vascular invasion in 3, and neural invasion in 2. The median preoperative serum carcinoembryonic antigen level was 5.1 ng/mL (range, 1.5–27.3). Four patients (40%) received 5-fluorouracil–based adjuvant chemotherapy after primary resection. The median time from primary resection to disease recurrence was 21.6 months (range, 8.2–57.6).

TABLE 1.

Clinical, Demographic, and Oncologic Characteristics of Patients and Primary Right Colon Cancer

Characteristics	Patients (n = 10)
Age (yr), median (range)	55 (32–71)
Sex
Male	8
Female	2
BMI, median (range)	20.3 (16.4–28.4)
Primary location (%)
Right colon	5
Transverse colon	2
Cecum	3
Tumor stage
T2	2
T3	5
T4	3
Tumor grade
Moderate	7
Poor	3
Nodal status
Positive	4
Negative	6
Vascular invasion
Yes	3
No	7
Neural invasion
Yes	2
No	8
Adjuvant chemotherapy for primary disease
Yes	4
No	6
CEA level at time of primary resection (ng/mL), median (range)	5.1 (1.5–27.3)
Disease-free interval to recurrence (mo), median (range)	21.6 (8.2–57.6)

BMI indicates Body Mass Index; CEA, carcinoembryonic antigen.

Figure 1 illustrates representative preoperative CT images of SMA and SMV involvement in the LACC and LRCC. Eight patients underwent extensive oncological therapy before IATx (Table 2). KRAS mutations were present in 4 patients (40%), dMMR in 2 (20%), and co-occurrent KRAS/dMMR mutations in 1 (10%). The median number of chemotherapy cycles was 16 (range, 12–21). Targeted chemotherapy, including bevacizumab and cetuximab, was administered to 8 patients (80%). Of these, 3 patients received additional programmed cell death protein 1 inhibitors. Among 8 patients who received systemic chemotherapy, 5 had a sustained response, 1 had a partial radiologic response, and 2 presented with progressive disease. Two patients did not receive systemic chemotherapy due to bowel obstruction. The median time from the primary diagnosis or recurrence to IATx was 6.0 months (range, 1.5–23.3).

FIGURE 1.

Representative preoperative CT images. A, Axial contrast-enhanced CT image of case 2 with LACC demonstrating a hypodense mass in the ascending colon invading into the mesentery root and SMA (red arrow), B with SMV stricture and its distal branch dilatation (red arrow). C, Coronal reformatted CT image revealing >180 degrees encasement of SMA and SMV (red arrow). D, Axial contrast-enhanced CT image of case 10 with LRCC demonstrating a mesenteric recurrent lesion infiltrating SMA (red arrow); E coronal reformatted CT image showing SMV involvement (red arrow). F, >180 degrees encasement of SMA and its branches (red arrow).

TABLE 2.

Details of Systemic Therapy and Response on CT in 10 Patients With Locally Advanced or Locally Recurrent Disease Undergoing IATx

Case	Gene Status	Targeted therapy	Response on CT	HTG	TRG	INT (mo)
1	pMMR	Bevacizumab, cetuximab	SD	G2	G2	5.5
2	pMMR	None	NA	G2	NA	1.5
3	pMMR, mKRAS	Bevacizumab	PD	G3+mucinous	G3	4.8
4	pMMR	Bevacizumab, cetuximab	SD	G3+mucinous	G2	23.3
5	pMMR	Bevacizumab	SD	G3	G2	4.6
6	pMMR, mKRAS	PD1+bevacizumab+regorafenib	PR	G2	G0	18.3
7	dMMR, mKRAS	PD1+bevacizumab+regorafenib	PD	G3+mucinous	G2	11.6
8	dMMR	PD1+cetuximab	SD	G2	G2	4.4
9	pMMR, mKRAS	Bevacizumab	SD	G2	G3	7.4
10	pMMR	None	NA	G2	NA	1.6

HTG indicates histologic tumor grading; INT, interval time; NA, not assessable; PD, progressive disease; PR, partial response; SD, stable disease; TRG, tumor regression grading.

Of the 10 patients, 8 underwent PD-IATx and 2 underwent IATx alone (Table 3). The length of the bowel autograft varied between 340 cm and 510 cm. The median total operative time was 502 minutes (range, 320–612) and transfusion requirements ranged from 2 to 7 units of packed red blood cells. The median warm ischemia time from removal of the graft out of ice until reperfusion was 36 minutes (range, 25–51), and the median cold ischemia time from cold perfusion of the graft until reperfusion was 110 minutes (range, 44–185), respectively. The graft artery was anastomosed to the SMA stump (n = 2) or the infrarenal aorta (n = 8). The grafted vein was anastomosed to the inferior vena cava (n = 8), portal vein, or SMV (n = 2). A cryopreserved blood-type identical interposition iliac allograft was used for vascular reconstruction in 3 patients. Eight patients with PD-IATx underwent a pancreaticogastrostomy. The cecum and right ascending colon were transplanted together with a bowel graft in 5 patients. The right colon was removed from the bowel graft in the other 5 due to poor blood supply after reperfusion. Among 10 patients, 8 (80%) underwent a primary colocolostomy or ileocolostomy and 2 (20%) underwent a temporary loop ileostomy. The median number of days in the intensive care unit was 2 (range, 1–3) days. The median length of postoperative hospital stay was 21 days (range, 10–36).

TABLE 3.

Operative Details and Postoperative Complications of 10 Patients With LACC or LRCC Undergoing IATx

Characteristics	Outcomes
Type of graft
IATx alone	2
PD-IATx	8
Graft length (cm), median (range)	450 (340–510)
Blood transfusions (units), median (range)	4.5 (2–7)
Warm ischemia times (min), median (range)	36 (25–51)
Cold ischemia times (min), median (range)	110 (44–185)
Total operating times (min), median (range)	502 (320–612)
PV drainage (%)
PV	2 (20)
IVC	8 (80)
Lymph nodal involvement
Negative	5
Positive	5
Surgical margin status
Negative	10
Positive	0
ICU stay (d)	2 (1–3)
Hospital stay (d)	21 (10–36)
CD classification complications (%)
I	1 (10)
II	4 (40)
IIIA	1 (10)
IIIB	3 (30)
IVA	1 (10)
Disease recurrence (%)
Yes	2
None	8
Follow-up times (mo)	23.9 (10.1–92.0)

CD indicates Clavien-Dindo; ICU, intensive care unit; IVC, inferior vena cava; PV, portal vein.

All 10 patients underwent resection, with pathologically clear margins. Three patients presented mucinous-type adenocarcinomas. Histopathological examination of the resected specimen showed malignant infiltration of the SMV in all cases (100%), SMA in 4 (40%), lymph node metastases in 5 (50%), and neural invasion in 3 (30%). Figure 2 representatively shows microscopic colon adenocarcinoma infiltrating the SMA.

FIGURE 2.

Representative micrographs showing: (A) colon adenocarcinoma invading into the tunica adventitia of SMA (black arrow) and (B) PV (black arrow; H&E stain, ×20). H&E indicates hematoxylin and eosin; PV, portal vein; SMA, superior mesenteric artery.

Up to date, 6 patients have not received any chemotherapy after IATx and 4 patients received chemotherapy because of disease recurrent (n = 2) or the lack of chemotherapy before the procedure (n = 2).

Postoperative Complications

No perioperative deaths occurred. All patients experienced postoperative complications (Table 3), including Clavien-Dindo grade I (n = 1), grade II (n = 4), grade IIIa (n = 1), grade IIIb (n = 3), and grade IVa (n = 1), which included acute venous thromboses, upper GI hemorrhage, anastomotic leak, and significant pleural effusions. Of note, immediately after the procedure, 2 patients developed acute venous thrombosis of the graft with open thrombectomy to successful revascularization in 1 patient but lost the graft in the other. Four months after the procedure, the latter patient underwent intestinal allotransplantation and was doing well at the last follow-up visit. Two patients presented with acute upper GI bleeding from the cut surface of the pancreas, 2 and 3 weeks after the procedure, respectively. Using an upper endoscope, clips were applied after dilute epinephrine was injected into the cut surface of the pancreas and the patient had no further bleeding afterwards. Upper endoscopy failed to locate the bleeding site in another case. Immediate relaparotomy showed a ruptured small artery in the upper border of the cut surface of the pancreas, and the patient recovered well and was discharged home 28 days after the IATx.

Delayed gastric emptying occurred in 4 patients (40%). After conservative management, including a nasogastric tube and oral prokinetics, all 4 patients responded well with an average time of nasogastric tube removal of 18 days and an average time of tolerance of solid food of 27 days. In addition, 6 patients (60%) required antidiarrheal medications at discharge. Most of the patients had 4 to 7 stools per day during the first half of the year, which decreased to 2 to 3 times per day after 6 months. Diarrhea episodes were controlled with diet modification and antidiarrheal medications as needed. All patients had a well-functioning autograft and did not require any intravenous fluid hydration after hospital discharge.

The median survival was 23.9 months, with the Kaplan-Meier estimates of a 68% 3-year recurrent-free rate and 80% 3-year survival rate after IATx (Fig. 3). Eight patients (80%) survived without evidence of disease. Two (20%) had tumor recurrence after 7.7 and 29.7 months, respectively, one (case 3) with multiple lung metastases and the other (case 2) with a retroperitoneal parapsoas metastasis. The first patient died of a pulmonary infection 34 months after IATx. The second patient underwent local resection and is currently alive at the time of this writing.

FIGURE 3.

Kaplan-Meier estimates of overall and recurrence-free survival in patients who underwent combined PD and IATx or IATx alone.

DISCUSSION

The role of extended surgical resections for LACC or LRCC infiltrating the SMA remains unclear. This study describes a combination of PD-IATx or IATx alone with SMA resection for patients with right colon cancer with patients with LACC or LRCC infiltrating the SMA. Using this strategy, radical resection (R0) was achieved in all 10 patients, who had an 80% overall actuarial survival and a recurrent-free survival of 68% at 3 years with an acceptable level of morbidity. Our results suggest that this aggressive approach is a viable option for patients with unresectable locally advanced or recurrent lesions that invade the SMA.

One of the significant and valuable results was that SMA resection for LACC or LRCC is technically feasible with acceptable morbidity and mortality. Our 10 patients experienced a median survival of 23.9 months and an estimated 3-year survival of 80%, indicating that this strategy is valuable in otherwise unresectable locally later-stage patients. Although this study involved a small number of selected patients, it represents the largest series reported to date on the surgical management of LACC or LRCC invading the SMA. Traditionally, involvement of the SMA by a variety of malignant neoplasms, although rare, has been considered inoperable and associated with a poor outcome. Our team initially reported favorable outcomes of IATx in a series of patients who underwent radical resection of heterogeneous neoplasms originating from the pancreas, duodenum, or retroperitoneum invading the SMA.^14,15 With increasing experience, we further expanded this approach to unresectable LACC or LRCC encasing the SMA. In the setting of LACC invading the duodenum and pancreas, en bloc combined PD and right hemicolectomy are currently recommended and 5-year survival after curative resection ranges from 49% to 66.3%.^12,17,18 However, the value of resection for locoregionally recurrent colon cancer remains controversial.^19,20 Recent pooled data indicate that R0 resection was associated with a 3-year survival rate of 58% and a 5-year survival rate of 52%. Otherwise, R1 resection carries a poor outcome with 3-year or 5-year survival rates of 27% and 11%, respectively. After macroscopically incomplete resection (R2), 3-year survival was only 11% with no 5-year survivors.²¹ Our study again underlines the significant effect of complete resection on survival, suggesting that patients with limited intra-abdominal mesentery recurrences should be considered for aggressive surgical management.

We acknowledge that in our limited experience, this approach carries a considerable risk of perioperative complications. In our series, the major postoperative complication rate (Clavien-Dindo grades III–V) was 50%, reflecting an aggressive approach to achieving optimal oncological clearance. Current overall morbidity rates associated with PD for benign and malignant pancreatic neoplasms range from 30% to 60%, and the mortality rates have been reported as low as 2% to 4% in high-volume centers.^22,23 Similarly, in a systemic review, 33 of 63 patients (52.4%) undergoing PD for LACC experienced one or more postoperative complications including pancreatic fistula, delayed gastric emptying, and infection.¹² Of special note, our study suggests several issues that are important in considering SMA resection for LACC or LRCC. First, vascular thrombosis is a devastating postoperative complication specifically related to IATx, possibly leading to irreversible bowel damage. If highly suspicious radiologically, prompt second-look surgery is warranted to prevent irreversible bowel damage. Second, appropriate management of the remnant pancreatic stump is the key determinant of the outcome after PD-IATx. The adequacy of blood supply at the cut surface of the pancreas should be evaluated routinely, and if deemed inadequate, more of the pancreas should be removed or even a total pancreatectomy may be performed to prevent anastomosis leakage. We favor pancreaticogastrostomy instead of pancreaticojejunostomy as a possible maneuver to reduce the incidence of pancreatic fistulas. The proximity of the pancreas to the posterior gastric wall allows for potentially less tension on the anastomosis. Third, diarrhea is a common adverse effect early after the procedure, which adversely affects nutrition and initiation of adjuvant chemotherapy. Preservation of the right colon, particularly the ileocecum can decrease the severity of diarrhea and promote nutritional autonomy. A novel method to decrease intestinal preservation injury and improve graft function may be useful.^24,25 Understanding the mechanisms of repair and regeneration in the autonomic nervous system after intestinal denervation is required.

The recurrence rate was 20% in our cohort of patients, a figure that compared favorably to that of patients with LACC or LRCC undergoing multivisceral R0 resection. In a systemic review, Khalili et al¹² showed that of 106 patients undergoing PD for LACC, 27 (25.5%) developed recurrence. In another systematic review, Chesney et al²¹ showed that R0 resection for locoregionally recurrent colon cancer was achieved in 191 patients with a pooled recurrence rate of 25%. However, the latter study did not discern recurrence between right and left colon cancer. Recurrence may reflect aggressive tumor biology, where viable cells have already escaped the limits of resectability.²⁶ For nonresectable liver metastases from colon cancer, the Fong Clinical Risk Score and the Oslo Score based on response to systemic therapy, tumor size, number of lesions, preoperative carcinoembryonic antigen levels, and interval from primary resection to recurrence are the key surrogates to select candidates for liver transplantation.^27–29 The number of patients in this study is too small to make appropriate recommendations regarding the appropriate candidate for IATx. Tumor biology assessments, including molecular testing, may facilitate risk stratification to better identify candidates who are likely to benefit from IATx.

Similar to other studies,^30,31 our results show that patients with mucinous colon cancer have a poor response to systemic chemotherapy. Currently, this unique type of colon cancer is treated with the same approaches used for other forms of colon cancer. However, chemotherapy may not be as effective for mucinous adenocarcinoma as for other types of colon cancer. Despite earlier recurrence in one patient with mucinous adenocarcinoma after IATx, 2 patients currently had no evidence of local or systemic relapse with short-term survival benefits. Radical resection combined with effective systemic therapy warrants further evaluation for patients with otherwise unresectable local lesions invading the SMA.

Our study had certain limitations. These limitations include its retrospective nature and uncontrolled experience. Further, the patients treated comprised a heterogeneous group, and the numbers were small with follow-up of <2 years for half of the patients in the cohort. Thus, our findings need to be confirmed in a multicenter study with long-term follow-up.

CONCLUSIONS

Our results indicate that IATx allows patients with unresectable LACC or LRCC involving the SMA to undergo complete resection, resulting in favorable outcomes. Meticulous preoperative assessment and preparation are critical for better outcomes in this technically challenging procedure. Selecting appropriate patients, establishing surgical indications for IATx, and developing effective chemotherapy for colon cancer will further improve the clinical outcomes.