En Bloc Right Hemicolectomy with Pancreaticoduodenectomy for Locally Advanced Right-Sided Colon Cancer


 Background: En bloc right hemicolectomy with pancreaticoduodenectomy (RHCPD) is the optimum treatment to achieve the adequate margin of resection (R0) for locally advanced right-sided colon cancer (LARCC). Information regarding the indications and outcomes of this procedure is limited.Method: In this retrospective study, patients who underwent RHCPD for LARCC during October 2010 to May 2019 were identified. The overall survival (OS), disease-free survival (DFS), mortality, postsurgical complications, gene mutational analysis, and prognostic factors were evaluated. Survival was estimated using Kaplan-Meir method.Results: Nineteen patients who underwent RHCPD were included in the study. The OS was 88.2%, 65.9%, and 57.6% at 1, 3, and 5 years. The DFS was 71.6%, 56.4%, and 56.4% at 1, 3, and 5 years. The median operative time was 320 minutes (range: 222-410 minutes), and the median operative blood loss was 268 mL (range: 100-600 mL). The OS was significantly better among patients with well-differentiated tumor (P=0.03), N0 stage (P=0.01), and high microsatellite instability (MSI) (P=0.047) and in patients who received chemotherapy (P=0.027). The major postoperative complications occurred in 8 patients (42%), with pancreatic fistula (PF) being the most common. On the basis of the univariate analysis, poorly differentiated tumor, regional lymph node dissemination, MSI status, and no perioperative chemotherapy were the significant predictors of poor survival (P<0.05). By multivariate analysis, only lymph node status was the significant factor (hazard ratio [HR]: 79.045; P=0.021).Conclusions: This study suggests that RHCPD is feasible and can achieve complete tumor clearance with favorable outcome, particularly in patients with lymph node-negative status.


Background
Colorectal cancer (CRC) is the third most common malignancy globally with a high mortality rate at the advanced stages [1][2][3]. The rate of adjacent organ invasion in CRC is 5.2-23.6% [4,5]. Locally advanced CRC represent about 5-22% of all CRC, and carcinoma of the right colon invading adjacent viscera is rare (11%-28%) [5][6][7]. Surgical resection is considered as the standard curative treatment for CRC [3]. Incomplete resection and separation of colon cancer from adherent organs are considered leading to tumor recurrence and a poor prognosis [8][9][10]. En bloc resection is the curative resection for locally advanced CRC that has invaded adjacent structures without distant metastasis. Locally advanced right-sided colon cancer (LARCC) can involve the duodenum, pancreas, and other organs. In this situation, extended or multivisceral resection is necessary to achieve the adequate margin of resection (R0).
En bloc right colectomy with pancreaticoduodenectomy (RHCPD) is the treatment of choice for LARCC to achieve R0 resection [11,12] and was rst reported in 1953 [13]. Despite the complexity of RCD, several studies have reported acceptable morbidity and mortality rate [7,11,14]. However, few studies reported the long-term outcome and clinicopathology ndings of patients undergoing RHCPD for LARCC [5,12,[15][16][17][18]. In China, there are very limited studies describing the long-term survival and histologic ndings of LARCC treated with RHCPD [5,12,19]. In addition, there is no study to reveal the potential relationship between the clinical evidence with histologic-genetic status and prognosis. Hence, we report a retrospective study aimed to assess the feasibility of RHCPD in achieving complete tumor clearance with favorable outcome and to identify the in uencing factors that affect the prognosis in patients undergoing en bloc RHCPD for LARCC with malignant in ltration into adjacent organs. Pathologic features and gene expression were also studied in these patients.

Patient characteristics
In this retrospective study, patients with primary right-sided colon cancer who underwent radical right colectomy (RC) during October 2010 and May 2019, at the Beijing Cancer Hospital, were included. The patient demographics, duration of surgery, estimated blood loss, tumor pathology, chemotherapy received, morbidity, mortality, and long-term survival were collected. Patients with (1) histologically proven colon carcinoma; (2) no evidence of metastasis disease in preoperative imaging; (3) biopsy-proven T4 cancer either to duodenum or pancreas; and (4) patients who were capable of completing the radiologic follow-up at our institution were included. All the patients underwent potential curative resections. Patients with distant metastasis, local recurrent tumor, or secondary involvement of the duodenum and/or pancreatic head rather than direct in ltrate were excluded. This study was approved by the institutional ethical committee of the Beijing Cancer Hospital, and written informed consent was obtained from all the patients included in the study. All the procedures were performed in accordance with the 1964 Helsinki declaration and its later amendments.

Diagnosis and indications for RHCPD
Local tumor in ltration was evaluated using preoperative computed tomography (CT). The preoperative carcinoembryonic antigen (CEA) and cancer antigen 19 − 9 (CA19-9) levels were routinely tested in all the patients. Preoperative colonoscopy and histopathologic examination of the tumor were performed for the con rmation of RHCPD. The indications for RHCPD were as follows: (1) preoperative histologic con rmation of colon carcinoma; (2) colon cancer that could not be dissociated from the duodenum and/or pancreas; (3) radical resection feasible on preoperative evaluation without distant metastasis; and (4) patients without severe comorbidity and tolerable to a radical multivisceral excision [5,16,17].

Patients characteristics
Between October 2010 and May 2019, 2269 patients with primary right-sided colon cancer underwent radical RC at the Beijing Cancer Hospital. Among them, 19 patients (12 men and 7 women) underwent RHCPD for LARCC with direct in ltration into the duodenum and/or pancreas. The median age of the patients was 60 years (range, 35-75 years). The tumors were located in the ascending colon (3 patients) and hepatic exure (16 patients). In ltration of the tumor into the duodenum was observed in all the patients, pancreas in 4 patients, the liver in 2 patients, and the SMV in 1 patient. The median preoperative CEA was 5.73 ng/mL (range: 0.934-18.53 ng/mL), and CA19-9 was 43.19 IU/mL (range: 0.50-228.40 IU/mL). During presentation, 10 patients had anemia, and 8 had abdominal pain. Other prominent symptoms were history of signi cant weight loss (n = 7), abdominal distension (n = 5), and vomiting (n = 5). Preoperative chemotherapy of the patients is shown in Table 1.
Surgical procedure for RHCPD Resection A Cattell-Brasch maneuver was initially performed followed by an extended Kocher maneuver to mobilize the duodenum fully [14]. The resectability of LARCC and the amount of in ltration into the duodenum and/or pancreas was evaluated after complete mobilization of the right colon and duodenum without dissecting the adherent organs from the cancer. RHCPD was performed after establishment of R0 resection via the standard procedures.
Pancreaticoduodenectomy (PD) was undertaken with a pyloric antrectomy in all patients [14,16]. If the portal vein and/or superior mesenteric vein (SMV) was involved, resection of the mesentericoportal vein and an end-to-end anastomosis were carried out.

Reconstruction
Reconstruction was performed in accordance with the modi ed Child's reconstruction method, with an end-to-side pancreaticogastrostomy or pancreaticojejunostomy, depending on the situation of the pancreatic duct. The stent of the pancreatic duct was routinely used in pancreatojejunostomy. Reconstruction of the bowel was performed via stapled side-to-side anastomosis of the ileum and transverse colon. After reconstruction, rubber drains were inserted near the biliary and pancreatic anastomoses, and the abdominal wall wounds were closed [5,14,16].

Pathology and gene testing
The tumor stage was assessed as per the classi cation of American Joint Committee of Cancer (AJCC) [20]. Postoperative complications evaluated includes the presence of pancreatic stula (PF), delayed gastric emptying (DGE), and intraabdominal abscess. Postoperative PF was categorized according to the International Study Group on Pancreatic Fistula De nition [21,22]. DGE was de ned according to the consensus de nition of DGE after pancreatic surgery suggested by the International Study Group of Pancreatic Surgery [23]. Intraabdominal abscess was de ned as a pocket of infected uid and pus located inside the abdominal cavity.
Formalin-xed, para n-embedded blocks was used to determine tumor cellularity. Tumors were macrodissected to remove contaminating normal tissue, resulting in samples containing > 20% neoplastic cells. Sample preparation, library construction, exome capture, next-generation sequencing

Follow-up
All patients were followed up at 3-month intervals during the rst 2 postoperative years, then at 6-month intervals during the next 3 years, and at least annually thereafter. During every follow-up, physical examination, CEA, CA19-9 measurement, chest radiography, and abdominal ultrasound or CT were performed. The patients were followed up as per a standard protocol [24]. Patients were followed up via annual colonoscopy examination at outpatient clinics.

Primary and secondary outcomes
The overall survival (OS) was considered as the primary outcome. The disease-free survival (DFS), 30-day postoperative mortality, postsurgical complications, gene testing, and prognostic factors evaluation were the secondary outcomes.

Statistical analysis
Numerical data were presented as the median (range). The survival rate was analyzed by the Kaplan-Meier (KM) method. Prognostic variables were examined using univariate analysis, and variables with P value of < 0.1 were used in the multivariate Cox proportional hazards model to identify independent predictors. P < 0.05 was considered as statistical signi cance. Hazard ratio (HR) was reported with 95% con dence interval (CI). All data were analyzed using SPSS v. 16.0 software (SPSS, Chicago, IL).

Postoperative complications and treatment
The treatments and outcomes data of the patients are listed in Table 2. The median operative time was 320 minutes (range: 222-410 minutes), and the median operative blood loss was 268 mL (range: 100-600 mL). Blood was transfused intraoperatively in 11 patients with an average amount of

Pathologic ndings and genetic testing
The pathologic ndings and genetic testing results are listed in Table 3. Tumors were classi ed as well-differentiated adenocarcinoma (3 patients), moderately differentiated adenocarcinoma (12 patients), and poorly differentiated adenocarcinoma (4 patients) based on histologic ndings. According to the AJCC classi cation system, 15 patients were N0 staged, 1 as N1b, 2 as N2b, and 1 as N2a. All of the tumors had clear resection margins (R0). In mutation testing, 11 patients were K-Ras mutant, 1 patient was B-Raf V600E mutant, 2 were Her-2 mutant, and none of the patients were N-Ras mutant as identi ed by using NGS. A total of 8 patients were identi ed as MSI-high status.

Prognostic factors for overall survival
On the basis of the univariate analysis, tumor differentiation, N stage, MSI status, K-Ras, B-Ras, Her-2, and perioperative chemotherapy were the signi cant prognostic factors (P < 0.05). In multivariate analysis, only N + was a signi cant independent predictor (HR: 79.04; P = 0.021) of poor survival (Table 4).

Discussion
Right-sided CRC invading duodenum and/or pancreas is a rare condition [6,23,25], and only a few studies have reported adjacent-organ resection [4,6,7,[14][15][16][25][26][27][28][29]. In our study, among 2269 patients with primary right-sided CRC who underwent radical RC screened, only 19 patients (12 men and 7 women) underwent en bloc RHCPD for LARCC with direct in ltration into the duodenum and/or pancreas. The patients who underwent duodenal resection with correction by direct suture or pedicled ileal ap were excluded because of the poor outcome and high rate of morbidity and mortality [8].
LARCC once con rmed during the operation, all adhesions between tumor and adjacent organs should be considered as malignant invasion due to 33-84% malignant invasion on pathologic examination [4,15] and should not be separated as there exist a risk of tumor recurrence rate of 90-100% [9,10]. In our study, en bloc resection was performed, and adhesions were veri ed as malignant only after histopathologic examination.
Right-sided CRC invading duodenum and/or pancreas was considered to have poor outcomes and unresectable in the earlier days. However, recent studies have reported a promising prognosis, with a 5-year survival rate ranging from 21-55% in patients with LARCC invading adjacent organs undergoing en bloc multivisceral resection [4,5,12,15,19]. Similarly, in our study, all the patients who underwent curative RHCPD achieved good outcomes, with 1-, 3-, and 5-year OS rate of 88.2%, 65.9%, and 57.6%, whereas DFS rates of 71.6%, 56.4%, and 56.4% respectively. Comparatively higher OS rate in our study is probably because of the fact that even though all the patients were staged T4b, regional lymph node dissemination of the cancer may not be that advanced because only 5 out of 19 patients were lymph node positive (N2a and N1b). In addition, some colon cancers exhibits locally aggressive invasion instead of distant spread [4,15,27,29]. According to the previous reports, 25-60% of right colon carcinoma that invaded the adjacent duodenum or pancreas do not have lymph node metastasis [4,6,15,27,29]. Furthermore, Saiura et al reported signi cantly longer survival in patients with node-negative status than node-positive patients [15]. Similarly, in our study, the survival of patients with node-positive had short survival (< 3 years) at the time of last follow-up Meanwhile, the 3-and 5-year OS of 14 patients with N0 were 88.9% and 77.8%, respectively, among them 4 patients survived for > 5 years. In subsequent univariate and multivariate analysis, regional lymph node dissemination was signi cant prognostic factor for poor survival.
Furthermore, based on the KM curve analysis, OS was signi cantly better in patients with well-or moderately differentiated tumor compared with patients with poorly differentiated tumor. This is probably because of the fact that the histologic type of tumor may affect the lymph node metastasis and prognosis in patients with LARCC as reported by a retrospective study conducted by Saiura et al 2008. The rate of lymph node metastasis was signi cantly higher in well-differentiated adenocarcinoma than mucinous or poorly differentiated adenocarcinoma in LARCC (P = 0.015) [15]. In our study, only 1 patient (6.7%) with moderately differentiated adenocarcinoma had node-positive status, whereas all the patients with poorly differentiation adenocarcinoma had node-positive status, and the 3-year OS rates of the 2 groups (well-or moderately differentiated tumor vs. poorly differentiation adenocarcinoma) were 82.1% versus 0%. Molecular markers such as K-Ras, N-Ras, B-Raf, HER2, and MSI play a signi cant role in the disease prognosis in CRC, and hence, analysis of these biomarkers helps in facilitating proper treatment to the needy patients [30]. In our study, all the patients in MSI-H status survived for > 3 years, while 3year OS of patients in MSS status was only 35.0% (P = 0.047). OS did not differ signi cantly between K-Ras mutant and wild-type (P = 0.888), BRAF V600E mutant and wild-type tumors (P = 0.771) nor Her-2 (P = 0.635). Hence, only MSI was the signi cant prognostic factor affecting survival.
In FOxTROT trial, preoperative chemotherapy has resulted in signi cant downstaging of tumor in patients with locally advanced colon cancer compared to postoperative chemotherapy (P = 0.04) [31]. Another retrospective study by Arredondo et al. has also shown tumor downstaging (62.5%), R0 resection (100%), and a promising prognosis (median OS of 31 months) in locally advanced colon cancer patients treated with preoperative chemotherapy [32]. Similarly, in our study, the 3-year OS rates was greater in the preoperative and postoperative chemotherapy group (100% and 77.8%) compared with no perioperative chemotherapy group (25.0%). However, these nding need to be con rmed by considering studies with large sample size.
There are several limitations associated with the present study. First, the number of LARCC patients with invasion of duodenum and/or pancreas is low, and hence, the number of participants is small in this study. Large-scale studies may produce more reliable results. Second, the chemotherapy regimens varied among patients. In the era of advanced chemotherapy, administering the same regimen for a long-term study seems formidable.
However, the present study also has several strengths such as this study gathered the largest number of patients, and all the clinical information and follow-up were proved to be accurate. Moreover, histologic-genetic examination was performed in detail, and we were able to build a bridge between the preoperative clinical evidence with histologic-genetic ndings and prognosis.

Conclusions
En bloc RHCPD may result in long-term survival in patients having LARCC with in ltrated adjacent organs. This aggressive approach may help improve the prognosis, particularly in patients with node-negative status. This study also evaluated the prognostic factors for OS and the role of mutational status of several genes on disease prognosis. Long-term and larger-scale studies may produce more reliable results. This study was approved by the institutional ethical committee of the Beijing Cancer Hospital, and written informed consent was obtained from all the patients included in the study. All the procedures were performed in accordance with the 1964 Helsinki declaration and its later amendments.

Consent for publication
Written informed consent for publication was obtained from all participants.

Availability of data and material
All data generated or analysed during this study are included in this published article.

Competing interests
The authors declare that they have no competing interests. XLY proposed the study, collected and analyzed the data, and wrote the rst draft. All authors contributed to the design and interpretation of the study and to further drafts. BCX is the guarantor. All authors have read and approved the manuscript.

Patients Pathologic Findings
Genetic Testing Overall Survival and Disease-Free Survival in Overall Patients