TRIANGLE operation, combined with adequate adjuvant chemotherapy, can improve the prognosis of pancreatic head cancer: A retrospective study

Abstract

BACKGROUND

The TRIANGLE operation involves the removal of all tissues within the triangle bounded by the portal vein-superior mesenteric vein, celiac axis-common hepatic artery, and superior mesenteric artery to improve patient prognosis. Although previously promising in patients with locally advanced pancreatic ductal adenocarcinoma (PDAC), data are limited regarding the long-term oncological outcomes of the TRIANGLE operation among resectable PDAC patients undergoing pancreaticoduodenectomy (PD).

AIM

To evaluate the safety of the TRIANGLE operation during PD and the prognosis in patients with resectable PDAC.

METHODS

This retrospective cohort study included patients who underwent PD for pancreatic head cancer between January 2017 and April 2023, with or without the TRIANGLE operation. Patients were divided into the PD_TRIANGLE and PD_non-TRIANGLE groups. Surgical and survival outcomes were compared between the two groups. Adequate adjuvant chemotherapy was defined as adjuvant chemotherapy ≥ 6 months.

RESULTS

The PD_TRIANGLE and PD_non-TRIANGLE groups included 52 and 55 patients, respectively. There were no significant differences in the baseline characteristics or perioperative indexes between the two groups. Furthermore, the recurrence rate was lower in the PD_TRIANGLE group than in the PD_non-TRIANGLE group (48.1% vs 81.8%, P < 0.001), and the local recurrence rate of PDAC decreased from 37.8% to 16.0%. Multivariate Cox regression analysis revealed that PD_TRIANGLE (HR = 0.424; 95%CI: 0.256-0.702; P = 0.001), adequate adjuvant chemotherapy ≥ 6 months (HR = 0.370; 95%CI: 0.222-0.618; P < 0.001) and margin status (HR = 2.255; 95%CI: 1.252-4.064; P = 0.007) were found to be independent factors for the recurrence rate.

CONCLUSION

The TRIANGLE operation is safe for PDAC patients undergoing PD. Moreover, it reduces the local recurrence rate of PDAC and may improve survival in patients who receive adequate adjuvant chemotherapy.

Key Words: TRIANGLE operation; Pancreatic ductal adenocarcinoma; Heidelberg triangle; Adjuvant chemotherapy; Prognosis; Pancreaticoduodenectomy

Core Tip: Although the TRIANGLE operation has shown efficacy in treating locally advanced pancreatic ductal adenocarcinoma (PDAC) patients, long-term oncological data from patients with resectable PDAC who underwent pancreaticoduodenectomy (PD) are limited. This study demonstrated the safety and efficacy of the TRIANGLE operation in reducing local recurrence of resectable PDAC and suggested survival benefits for patients receiving adequate adjuvant chemotherapy.

INTRODUCTION

Pancreatic ductal adenocarcinoma (PDAC) is one of the most common fatal malignancies and has a low 5-year survival rate of 11%[1]. While approximately 20% of pancreatic cancer patients are eligible for curative resection, local recurrence and metastasis are common, particularly in patients with ductal adenocarcinoma of the pancreatic head[2]. In addition to the TNM stage, the status of the surgical margin is closely related to overall survival (OS) and is a prognostic factor[3]. However, the rate of positive surgical margins remains high, with the retroperitoneal margin being the most commonly affected site. Local recurrence typically occurs after pancreaticoduodenectomy (PD)[4-7]. Therefore, it is crucial to meticulously dissect this area to improve patient prognosis.

Gockel et al[8] first proposed the concept of the mesopancreas based on the mesorectum, emphasizing that resection of the mesopancreas could achieve a negative resection margin of the retroperitoneum. Adham and Singhirunnusorn provided further knowledge about the mesopancreas by proposing anatomical boundaries such as the celiac axis (CA)/common hepatic artery (CHA), superior mesenteric artery (SMA) and portal vein (PV)/superior mesenteric vein (SMV)[9]. These authors also introduced total mesopancreas excision (TMpE), which improves posterior clearance and achieves R0 resection. Similarly, Hackert et al[10] proposed the TRIANGLE operation and applied it to patients with locally advanced pancreatic cancer. This procedure is performed on patients who are in stable condition after induction therapy and who have no viable tumor cells in the tissue surrounding the artery, as confirmed by frozen sectioning. The TRIANGLE operation aims to reduce the percentage of patients with a positive resection margin. It involves dissection between the SMA, SMV/PV, and CA/CHA, as well as the location of pancreatic head nerve plexus I (PLphI) and pancreatic head nerve plexus II (PLphII) proposed by the Japan Pancreas Society (JPS)[11]. Perineural invasion (PNI) occurs in approximately 80%-100% of pancreatic cancers[12]. Therefore, the use of the TRIANGLE operation holds potential prognostic value for removing these nerve plexuses, lymph nodes, and adipose tissue in the mesopancreas.

While initial findings from a pilot study involving 15 patients with locally advanced PDAC suggest acceptable morbidity associated with the TRIANGLE procedure[10], comprehensive data from a larger cohort need to be established. Moreover, while the safety of the TRIANGLE operation has been the subject of several studies[13,14], research focusing on the impact of combining this procedure with PD on survival outcomes in patients with resectable pancreatic head cancer is still limited.

The aim of the present study was to assess the safety of the TRIANGLE procedure and its impact on the survival of patients with resectable ductal adenocarcinoma of the pancreatic head in a single-center cohort.

MATERIALS AND METHODS

This retrospective cohort study received approval from the institutional review board of Huadong Hospital affiliated with Fudan University, and the protocol was approved and registered on ClinicalTrials.gov (ID: NCT05703581) before statistical analysis (January 30, 2023). This study was also approved by the Ethics Committee of Huadong Hospital Affiliated to Fudan University (No. 20170014), and all patients provided informed consent before the operation. This study adhered to the STROBE guidelines.

Study design and patient selection

This study included adult patients diagnosed with pancreatic head cancer who underwent PD between January 2017 and April 2023 at the Department of General Surgery, Huadong Hospital affiliated with Fudan University. All patients had complete clinicopathological and prognostic data and were followed up for a minimum of 6 months. The exclusion criteria were as follows: (1) Died due to non-neoplastic causes; (2) Incomplete follow-up data; and (3) Received chemotherapy prior to surgery. Regarding the extent of lymph node dissection, based on preoperative planning, from January 2017 to May 2019, all patients with resectable pancreatic head cancer underwent standard lymphadenectomy according to the International Study Group on Pancreatic Surgery (ISGPS) consensus[15] and were categorized into the PD_non-TRIANGLE group. Starting in June 2019, all patients with resectable pancreatic head cancer underwent the TRIANGLE operation in addition to standard lymphadenectomy and were classified into the PD_TRIANGLE group.

Treatments in both groups

The adjuvant chemotherapy regimens included gemcitabine-based chemotherapy and 5-fluorouracil-based chemotherapy. Regardless of the regimen, a continuous adjuvant chemotherapy course lasting longer than 6 months was defined as adequate for patients who underwent PD[16].

Our institution successfully performed PD using a standard approach. Following the removal of the pancreatic head, we employed a modified version of Child’s method to accomplish intestinal reconstruction. The reconstruction procedure involved performing a pancreatojejunostomy based on the Blumgart method, utilizing a duct-to-mucosa, end-to-side technique, and placing a pancreatic drainage tube. A continuous suture technique was used for biliary-enteric anastomosis. Gastrojejunostomy was achieved using interrupted 3-0 polypropylene monofilament sutures.

The PD_non-TRIANGLE group underwent standard lymphadenectomy during the PD operation according to the ISGPS consensus[15], which included the removal of suprapyloric and infrapyloric lymph nodes (No. 5, 6), the anterior superior region of the CHA (No. 8a), the anterior and posterior pancreatoduodenal lymph nodes (No. 13a, 13b, 17a, 17b), the right side of the hepatoduodenal ligament (12b1, 12b2, 12c), and the right side of the SMA from its origin to the inferior pancreaticoduodenal artery (14a, 14b). However, Heidelberg triangle dissection was not performed in this group.

The PD_TRIANGLE group, on the other hand, underwent standard lymphadenectomy along with Heidelberg triangle dissection during the PD operation. The dissection of the Heidelberg triangle involved removing all vascular, lymphoid, and neural tissues within the triangular area surrounded by the PV-SMV, CA-CHA and SMA. The aforementioned blood vessels within this area were skeletonized; however, for the SMA, we cleared only the right 180 degrees, preserving the left 180 degrees (Figure 1)[10]. The patients had their standard lymph nodes dissected according to the ISGPS consensus, as well as the posterior lymph node of the CHA (No. 8p) and the lymph node around the CA (No. 9) within the Heidelberg Triangle. A clear intraoperative photograph of the Heidelberg triangle dissection is shown in Figure 2.

Figure 1 The preoperative imaging of Heidelberg triangle. A: Preoperative contrast-enhanced computed tomography image of the abdomen. The orange arrowhead points to vessels around the Heidelberg triangle. The portal vein-superior mesenteric vein, celiac axis-common hepatic artery (CHA), and superior mesenteric artery represent the three boundary vessels of the Heidelberg triangle. The dorsal pancreatic artery (DPA) in the figure originates from the CHA and passes through the triangle, reflecting one of the numerous pathways of DPA. The left gastric vein (LGV), splenic vein (SPV), and inferior mesenteric vein (IMV) typically pass anteriorly to the triangle; B: Frontal view of the Heidelberg triangle generated by three-dimensional reconstruction. The black arrowhead points to vessels around the Heidelberg triangle. The image reveals the DPA coursing within the triangle and the LGV, SPV, and IMV traversing anteriorly to the triangle. GDA: Gastroduodenal artery; PV: Portal vein; SMV: Superior mesenteric vein; CA: Celiac axis; CHA: Common hepatic artery; SMA: Superior mesenteric artery; DPA: Dorsal pancreatic artery; LGV: Left gastric vein; SPV: Splenic vein; IMV: Inferior mesenteric vein.

Figure 2 Intraoperative image of the Heidelberg triangle dissection. A: Intraoperative image before Heidelberg triangle dissection. At this stage, the superior mesenteric artery (SMA) is located on the dorsal side of the pancreatic remnant, and the portal vein (PV)-superior mesenteric vein (SMV), common hepatic artery (CHA), and pancreas are visible. The blue shading illustrates the boundaries of the Heidelberg triangle, encompassing vascular, lymphatic, and neural fiber tissues within the triangle; B: Intraoperative image after Heidelberg triangle dissection. The black arrowhead points to vessels and tissues around the Heidelberg triangle. The black triangle shows the range of Heidelberg triangle dissection. After thorough clearance, the precise boundaries of the triangle and its three bordering vessels, namely, the PV-SMV, CHA, and SMA, are clearly visible. SMV: Superior mesenteric vein; SMA: Superior mesenteric artery; CHA: Common hepatic artery; PV: Portal vein.

Definitions and data collection

The following baseline data were obtained from the hospital laboratory information system: age, sex, body mass index (BMI), American Society of Anesthesiologists (ASA) score, main pancreatic duct size, carbohydrate antigen 19-9 (CA19-9), carcinoembryonic antigen (CEA), albumin, and serum bilirubin. The following pathological data were obtained from the pathology database: tumor grade, PNI, perivascular invasion (PVI), number of examined lymph nodes, and margin status. R0 resection was confirmed microscopically when tumor cells were absent from the margin within a 1 mm distance, and R1 resection was identified when tumor cells were present within the 1 mm boundary. The TNM stage was determined according to the eighth edition of the American Joint Committee on Cancer (AJCC) staging system[17]. Postoperative complications, such as postoperative pancreatic fistula (POPF)[18], bile leakage[19], chyle leak[20], delayed gastric emptying (DGE)[21], and postpancreatectomy hemorrhage (PPH)[22], were classified according to the ISGPS definitions. Clinically relevant complications (grades B and C according to ISGPS) were also recorded. Diarrhea was defined as having more than three bowel movements per day and receiving regular use of antidiarrheal medication during the second postoperative week[23]. The occurrence of diarrhea during both the postoperative in-hospital period and the subsequent follow-up period was recorded. Recurrence-free survival (RFS) was calculated from the date of treatment to the first radiographic evidence of recurrence, and postoperative imaging was obtained as per the National Comprehensive Cancer Network (NCCN) guidelines[24]. OS was calculated from the date of treatment to the date of death or the last follow-up. Locoregional recurrence was defined as radiographic or pathological evidence of recurrent disease in the remnant pancreas and retroperitoneum along the SMV/PV, SMA, or CA/CHA. Distant recurrence was defined as a tumor that spread outside the locoregional area, including but not limited to the peritoneum, lungs, and liver. Simultaneous local and distant recurrence was defined as the presence of both local and distant recurrences at the same time. According to the radiological resectability criteria of the NCCN guidelines, all pancreatic head cancer patients enrolled in our study were classified as resectable[24].

Statistical analysis

Continuous variables with a normal distribution are presented as the mean ± SD and were analyzed using Student’s t test. Continuous variables that were not normally distributed are presented as medians with interquartile ranges and were analyzed using the Mann-Whitney U test. Categorical variables were analyzed by the chi-square test or Fisher’s exact test, depending on the situation.

For survival analyses, patients who were lost to follow-up were censored at the time of the last normal imaging or clinical visit for RFS and the last interaction within the institution’s medical record for OS. Kaplan-Meier survival curves were generated for RFS and OS, and survival was compared using the log-rank test. Univariate and multivariate analyses with Cox proportional hazard models were performed to evaluate significant predictors of RFS. Variables included in the multivariable model were selected based on their significance in the univariate comparison of the two groups. The HR and its 95%CI were calculated using Cox proportional hazard analysis. All tests were two-tailed, and the statistical significance was set at 0.05. Statistical analyses were performed using SPSS V.26 software (IBM, United States), and Kaplan-Meier survival curves were generated using the survival and survminer package in R statistical software (v. 4.2.2).

RESULTS

Cohort characteristics

We retrospectively analyzed 127 patients with pancreatic head ductal adenocarcinoma who underwent PD between January 2017 and April 2023 at the Department of General Surgery, Huadong Hospital, Fudan University. All operations were performed by the same surgical team. We excluded four patients who died from nonsurgical causes, four patients with incomplete follow-up data and twelve patients who received chemotherapy prior to surgery. The final analysis included 107 patients with PDAC who met the inclusion criteria. These patients were divided into the PD_TRIANGLE (n = 52) and PD_non-TRIANGLE (n = 55) groups based on whether they underwent Heidelberg triangle clearance (Figure 3). There were no statistically significant differences in the baseline data, such as age, sex, BMI, ASA score, CA19-9, or CEA, between the two groups. Similarly, there were no statistically significant differences in clinicopathological indexes, including the PNI, PVI, TNM stage, or resection margin status, between the two groups. The longest follow-up period was 56 months, with median follow-up times of 37.0 months and 40.0 months for the PD_TRIANGLE group and the PD_non-TRIANGLE group, respectively. Notably, the median number of examined lymph nodes was significantly greater in the PD_TRIANGLE group than in the PD_non-TRIANGLE group (20 vs 14, P = 0.001), indicating that Heidelberg triangle dissection enabled the acquisition of a greater number of lymph nodes (Table 1).

Figure 3 Flow diagram of the study cohort. Pancreaticoduodenectomy (PD)_TRIANGLE, patients underwent standard lymphadenectomy along with Heidelberg triangle dissection during pancreaticoduodenectomy; PD_non-TRIANGLE, patients underwent standard lymphadenectomy during pancreaticoduodenectomy without Heidelberg triangle dissection. PD: Pancreaticoduodenectomy.

Table 1 Demographic and clinicopathologic data.

Variables	PDTRIANGLE (n = 52)	PDnon-TRIANGLE (n = 55)	P value
Age, median (IQR; yr)	69 (60-74)	67 (61-74)	0.589
Sex			0.154
Male	26 (50.0)	35 (63.6)
Female	26 (50.0)	20 (36.4)
BMI (mean ± SD, kg/m²)	22.86 ± 3.19	22.15 ± 2.79	0.224
ASA score			0.387
1	10 (19.6)	15 (27.3)
2	37 (72.6)	33 (60.0)
3	4 (7.8)	7 (12.7)
Main pancreatic duct size (mm)			0.208
≤ 3	18 (34.6)	19 (34.5)
> 3	34 (65.4)	36 (65.5)
CA19-9, median (IQR; IU/mL)	94.4 (37.3-358.5)	108.0 (57.6-213.2)	0.711
CEA, median (IQR; ng/mL)	3.4 (2.1-6.5)	3.9 (2.3-5.7)	0.521
Albumin (mean ± SD, g/L)	39.7 ± 3.9	40.7 ± 4.5	0.227
Serum bilirubin, median (IQR; μmol/L)	29.3 (8.3-186.4)	18.8 (9.4-114.9)	0.660
AJCC stage (8th ed)			0.291
I/IIa	28 (53.8)	24 (43.6)
IIb/III	24 (46.2)	31 (56.4)
Grading			0.385
1	2 (3.9)	0 (0.0)
2	39 (75.0)	45 (81.8)
3	11 (21.1)	10 (18.2)
PNI	44 (84.6)	44 (80.0)	0.532
PVI	28 (53.9)	27 (49.1)	0.623
Number of ELN, median (IQR)	20 (14-26)	14 (8-20)	0.001a
Margin status			0.579
R0	42 (80.8)	42 (76.4)
R1	10 (19.2)	13 (23.6)

^aP < 0.01.

Values in parentheses are percentages unless indicated otherwise. PD: Pancreaticoduodenectomy; IQR: interquartile range; BMI: Body mass index; ASA: American Society of Anesthesiologists; CA19-9: Carbohydrate antigen 19-9; CEA: Carcinoembryonic antigen; AJCC: American Joint Committee on Cancer; PNI: Perineural invasion; PVI: Perivascular invasion; ELN: Examined lymph nodes.

Perioperative and recurrent data

There were no significant differences in operative time (386 minutes vs 390 minutes, P = 0.527) or intraoperative bleeding volume (300 mL vs 300 mL, P = 0.863) between the two groups, suggesting that Heidelberg triangle dissection was safe in PD. The postoperative length of stay (14 d vs 15 d, P = 0.177) and the proportion of patients requiring ICU admission for ≥ 2 d (25.0% vs 30.9%, P = 0.496) were similar between the two groups. Although extended lymphadenectomy was involved in Heidelberg triangle dissection, there was no significant difference in complication rates, including POPF (B/C), bile leakage (B/C), chyle leak (B/C), DGE (B/C), PPH (B/C), intra-abdominal infection, or diarrhea, between the two groups. Moreover, there was no significant difference in the proportion of patients requiring reoperation between the two groups (1.9% vs 5.5%, P = 0.651; Table 2).

Table 2 Surgical and recurrence data.

Variables	PDTRIANGLE (n = 52)	PDnon-TRIANGLE (n = 55)	P value
Operative time, median (IQR; min)	386 (324-423)	390 (345-425)	0.527
Intraoperative blood loss, median (IQR; mL)	300 (200-500)	300 (200-500)	0.863
LOS, median (IQR)	14 (11-19)	15 (12-22)	0.177
ICU stay ≥ 2 d	13 (25.0)	17 (30.9)	0.496
POPF (B/C)	9 (17.3)	13 (23.6)	0.418
Bile leakage (B/C)	0 (0.0)	3 (5.5)	0.262
Chyle leak (B/C)	0 (0.0)	2 (3.6)	0.496
DGE (B/C)	3 (5.8)	5 (9.1)	0.775
PPH (B/C)	0 (0.0)	2 (3.6)	0.496
Intra-abdominal infection	10 (19.2)	9 (16.4)	0.698
Diarrhea	14 (26.9)	9 (16.4)	0.184
Reoperation	1 (1.9)	3 (5.5)	0.651
90-d mortality	0 (0.0)	0 (0.0)	1.000
Adjuvant chemotherapy, ≥ 6 months	30 (57.7)	27 (49.1)	0.373
Chemotherapy regimens1			0.887
Gemcitabine-based	24 (80.0)	22 (81.5)
5-fluorouracil-based	6 (20.0)	5 (18.5)
Recurrence			< 0.001a
Yes	25 (48.1)	45 (81.8)
No	27 (51.9)	10 (18.2)

^aP < 0.001.

¹Variables calculated on the total of patients who received adjuvant chemotherapy lasting longer than 6 months Values in parentheses are percentages unless indicated otherwise.

PD: Pancreaticoduodenectomy; IQR: Interquartile range; ICU: Intensive care unit; LOS: Length of stay; POPF: Postoperative pancreatic fistula; DGE: Delayed gastric emptying; PPH: Postpancreatectomy hemorrhage.

Additionally, while the two groups were balanced in terms of receiving adjuvant chemotherapy longer than 6 months (57.7% vs 49.1%, P = 0.373) and chemotherapy regimens (P = 0.887), the recurrence rate was lower after Heidelberg triangle dissection (48.1% vs 81.8%, P < 0.001). Notably, clearance of the Heidelberg triangle significantly reduced the local recurrence rate of PDAC patients from 37.8% to 16.0% and reduced the combined local and distant recurrence rates from 26.7% to 12.0% (Figure 4).

Figure 4 The recurrence patterns in both groups. A: Recurrence pattern in the Pancreaticoduodenectomy (PD)_TRIANGLE group. The proportions of local, distant, and combined local and distant recurrences were 16%, 72%, and 12%, respectively; B: Recurrence pattern in the PD_non-TRIANGLE group. The percentages of local, distant, and combined local and distant recurrences were 37.8%, 35.5%, and 26.7%, respectively. PD: Pancreaticoduodenectomy.

Prognostic factors for recurrence

The median RFS was 30.0 months in the PD_TRIANGLE group and 12.6 months in the PD_non-TRIANGLE group (P < 0.001). Similar results were observed for the median OS of both groups (44.1 months vs 28.3 months, P = 0.029; Figure 5). To identify independent risk factors for postoperative recurrence, Cox univariate and multivariate regression analyses were performed for each variable. Multivariate Cox regression analysis revealed that PD_TRIANGLE (HR = 0.424; 95%CI: 0.256-0.702; P = 0.001), adequate adjuvant chemotherapy ≥ 6 months (HR = 0.370; 95%CI: 0.222-0.618; P < 0.001) and margin status (HR = 2.255; 95%CI: 1.252-4.064; P = 0.007) were found to be independent factors associated with the recurrence rate (Table 3). Further analysis was also conducted to explore the factors affecting the recurrence pattern. Moreover, the PD_TRIANGLE (HR = 0.175; 95%CI: 0.059-0.521; P = 0.002) was found to be an independent risk factor for a lower local recurrence rate. Moreover, adequate adjuvant chemotherapy ≥ 6 months (HR = 0.332; 95%CI: 0.164-0.672; P = 0.002) and CA19-9 > 37.0 (HR = 5.684; 95%CI: 1.354-23.868; P = 0.018) were found to be independent factors for the distant recurrence rate. Finally, both the PD_TRIANGLE (HR=0.188; 95%CI: 0.052-0.683; P = 0.011) and the AJCC stage (8^th ed; HR = 3.608; 95%CI: 1.057-12.308; P = 0.040) were found to be independent factors for simultaneous local and distant recurrence (Table 3).

Figure 5 Kaplan-Meier curves of both groups. A: Kaplan-Meier curve of recurrence-free survival (RFS) between the Pancreaticoduodenectomy (PD)_TRIANGLE and PD_non-TRIANGLE groups. The median RFS was notably longer in the PD_TRIANGLE group at 30.0 months than in the PD_non-TRIANGLE group at 12.6 months (P < 0.001); B: Kaplan-Meier curve of overall survival (OS) between the PD_TRIANGLE and PD_non-TRIANGLE groups. The median OS was significantly better in the PD_TRIANGLE group than in the PD_non-TRIANGLE group (44.1 months vs 28.3 months, P = 0.029). PD: Pancreaticoduodenectomy; RFS: Recurrence-free survival.

Table 3 Univariate and multivariate analyses of independent prognostic variables for different recurrence patterns.

Variables	Total (n = 70)			Local (n = 21)			Distant (n = 34)			Local + distant (n = 15)
	P1	HR (95%CI)	P2	P1	HR (95%CI)	P2	P1	HR (95%CI)	P2	P1	HR (95%CI)	P2
Age (yr), > 70/≤ 70	0.206	-		0.092	-		0.864	-		0.470	-
Sex, male/female	0.572	-		0.304	-		0.866	-		0.707	-
PDTRIANGLE, +/-	< 0.001c	0.424 (0.256-0.702)	0.001b	0.002b	0.175 (0.059-0.521)	0.002b	0.649	-		0.011a	0.188 (0.052-0.683)	0.011a
Grading, 2-3/1
2	0.663	-		0.997	-		0.997	-		0.240	-
3	0.813	-		0.997	-		0.997	-		0.292	-
Margin status, R1/R0	0.031a	2.255 (1.252-4.064)	0.007b	0.249	-		0.346	-		0.066	-
AJCC stage (8th ed), I-IIa/IIb-III	0.057	-		0.801	-		0.077	-		0.038a	3.608 (1.057-12.308)	0.040a
PNI, +/-	0.427	-		0.887	-		0.478	-		0.512	-
PVI, +/-	0.064	-		0.694	-		0.036a	1.862 (0.940-3.691)	0.075	0.833
CA19-9, > 37.0/≤ 37.0	0.098	-		0.141	-		0.019a	5.684 (1.354-23.868)	0.018a	0.152	-
CEA, > 5.0/≤ 5.0	0.823	-		0.068	-		0.289	-		0.612	-
Adjuvant chemotherapy, ≥ 6 months/< 6 months	0.001b	0.370 (0.222-0.618)	< 0.001c	0.527	-		0.002b	0.332 (0.164-0.672)	0.002b	0.127	-

^aP < 0.05.

^bP < 0.01.

^cP < 0.001.

P₁: Univariate P value; P₂: Multivariate P value; PD: Pancreaticoduodenectomy; CA19-9: Carbohydrate antigen 19-9; CEA: Carcinoembryonic antigen; AJCC: American Joint Committee on Cancer; PNI: Perineural invasion; PVI: Perivascular invasion.

Subgroup analysiss

The data were divided into four groups based on the two independent prognostic factors derived from the Cox analysis: Group 1: No clearance of the Heidelberg triangle and no adjuvant chemotherapy; Group 2: No clearance of the Heidelberg triangle but receiving adjuvant chemotherapy (≥ 6 months of adjuvant chemotherapy); Group 3: Clearance of the Heidelberg triangle but no adjuvant chemotherapy; and Group 4: Clearance of the Heidelberg triangle and receiving adjuvant chemotherapy (≥ 6 months of adjuvant chemotherapy). The analysis revealed that the median RFS (mRFS) was significantly greater in group 4 than in group 2 (30.0 months vs 16.4 months, P = 0.027), and the mRFS was significantly greater in group 3 than in group 1 (24.6 months vs 9.8 months, P = 0.013). The OS of the four groups was further analyzed, revealing that the median OS (mOS) of group 4 was longer than that of group 2 (not reached vs 32.2 months, P = 0.040), and the mOS of group 4 was longer than that of group 3 (not reached vs 24.0 months, P < 0.001). These differences were statistically significant (Figure 6; Table 4).

Figure 6 Kaplan-Meier curves of the four groups. A: Kaplan-Meier curve of recurrence-free survival (RFS) between the four groups. The median RFS (mRFS) for groups 1, 2, 3, and 4 were 9.8, 16.4, 24.6, and 30.0 months, respectively. From group 1 to group 4, the mRFS progressively increased. The mRFS of group 1 was significantly shorter than that of groups 2, 3, and 4; group 4 exhibited a significantly longer mRFS than group 2 did; B: Kaplan-Meier curve of overall survival (OS) between the four groups. The median OS (mOS) for groups 1, 2, 3, and 4 were 23.0, 32.2, 24.0 months, and not reached, respectively. The mOS of groups 2 and 4 was significantly longer than that of groups 1 and 3, and the mOS of group 4 was significantly longer than that of group 2. Group 1: No clearance of the Heidelberg triangle and not receiving ≥ 6 months of adjuvant chemotherapy; Group 2: No clearance of the Heidelberg triangle but receiving ≥ 6 months of adjuvant chemotherapy; Group 3: Clearance of the Heidelberg triangle but not receiving ≥ 6 months of adjuvant chemotherapy; Group 4: Clearance of the Heidelberg triangle and receiving ≥ 6 months of adjuvant chemotherapy. RFS: Recurrence-free survival; OS: Overall survival.

Table 4 P values for survival comparisons between the four groups.

Group comparison	P value of mRFS	P value of mOS
Group 1 vs group 2	0.002b	< 0.001c
Group 1 vs group 3	0.013a	0.488
Group 1 vs group 4	< 0.001c	< 0.001c
Group 2 vs group 3	0.663	0.031a
Group 2 vs group 4	0.027a	0.040a
Group 3 vs group 4	0.147	< 0.001c

^aP < 0.05.

^bP < 0.01.

^cP < 0.001.

Group 1: No clearance of the Heidelberg triangle and not receiving ≥ 6 months of adjuvant chemotherapy; Group 2: No clearance of the Heidelberg triangle but receiving ≥ 6 months of adjuvant chemotherapy; Group 3: Clearance of the Heidelberg triangle but not receiving ≥ 6 months of adjuvant chemotherapy; Group 4: Clearance of the Heidelberg triangle and receiving ≥ 6 months of adjuvant chemotherapy;mRFS: Median recurrence-free survival; mOS: Median overall survival.

DISCUSSION

The extent of lymphadenectomy in PD patients remains a topic of debate. Previous studies have suggested that there is no significant difference in the long-term prognosis between standard lymphadenectomy and extended lymphadenectomy for PDAC, although the latter may result in longer operation times and increased intraoperative bleeding[25-29]. In 2017, Hackert et al[10] introduced the TRIANGLE operation for locally advanced PDAC, which is a modified form of extended lymphadenectomy encompassing the same range as the TMpE proposed by Adham and Singhirunnusorn[9] in 2012. This procedure has been shown to improve the margin status of PDAC without increasing the complication rate[23]. However, studies on the long-term prognosis of the TRIANGLE operation are currently lacking. The objective of this study was to investigate the effect of combining PD with the TRIANGLE operation on the long-term prognosis of PDAC patients.

Previous studies examining the early postoperative outcomes of the TRIANGLE operation have demonstrated that it yields a greater number of examined lymph nodes than standard lymphadenectomy does, with a lower rate of R1 (direct) resection, although no significant difference has been observed in R0 rates[23]. Similarly, our study revealed that the PD_TRIANGLE group had a greater lymph node yield than the PD_non-TRIANGLE group did, while the R0 rate did not significantly differ. Additionally, we observed that the TRIANGLE operation did not increase the operative time, intraoperative bleeding volume, or postoperative complication rate, suggesting that it is safer than extended lymphadenectomy, as discussed in previous studies[25-29]. Regarding postoperative quality of life, extended lymphadenectomy has been associated with increased rates of postoperative diarrhea[30]. Our results revealed a greater incidence of diarrhea in the TRIANGLE operation group (26.9% vs 16.4%, P = 0.184), but this difference was not statistically significant. This increase in diarrhea rates might be attributed to the resection of autonomic nerves around the SMA and CA. Furthermore, in the present study, the TRIANGLE operation involved dissection of only the right 180° of the SMA while preserving the nerves on the left 180°; hence, the incidence of diarrhea was not very high[25]. Notably, patients who experienced diarrhea responded well to oral antidiarrheal medication, implying a limited impact on overall quality of life.

Previous studies investigating factors influencing postoperative recurrence in patients with PDAC have shown the PNI to be an independent prognostic factor for OS and disease-free survival (DFS), while adjuvant chemotherapy has been shown to reduce the rate of local and distant recurrence after surgery, resulting in an improvement in OS[31-33]. Despite both study groups receiving equal proportions of adjuvant chemotherapy, our findings showed a lower recurrence rate when the Heidelberg triangle was cleared. Notably, in patients who experienced recurrence, clearing the Heidelberg triangle reduced the local recurrence rate of PDAC from 37.8% to 16.0% and the combined local and distant rate from 26.7% to 12.0%. These findings suggested that the TRIANGLE operation reduces the local recurrence rate and alters the recurrence pattern of PDAC. Cox regression analysis of the recurrence pattern further confirmed this phenomenon. The PLphI and PLphII, as mentioned in the JPS guidelines, are located within the Heidelberg triangle[11], and the TRIANGLE operation involves clearance of the corresponding nerve plexus, thereby reducing the local recurrence rate. This may provide a plausible explanation for the observed phenomenon. A recent clinical study demonstrated that selective extended dissection, which involves removal of the extrapancreatic nerve plexus that may be invaded by tumors, including the Heidelberg triangle area, can significantly prolong the DFS of patients with pancreatic head cancer and reduce the rate of local recurrence. This finding supports the rationale behind this approach[34]. In the present study, survival analysis revealed mRFS values of 30.0 and 12.6 months for the PD_TRIANGLE and PD_non-TRIANGLE groups, respectively, which were significantly different. Similar results were obtained when analyzing the mOS in both groups, suggesting a potential prognostic benefit attributed to the TRIANGLE operation or adjuvant chemotherapy.

Previous studies have demonstrated that the completeness of adjuvant chemotherapy is more critical than the timing of chemotherapy initiation[35]. Complete adjuvant chemotherapy (> 6 cycles) has been shown to improve OS and DFS after curative surgery for PDAC, while early termination of chemotherapy has been associated with shorter OS[36-38]. Typically, chemotherapy regimens for pancreatic cancer that last 6 cycles span 6 months or longer[38]. Thus, in this study, we used 6 months rather than 6 cycles as the criterion for complete chemotherapy. Through multivariate Cox regression analysis, we identified adjuvant chemotherapy lasting ≥ 6 months as an independent factor associated with lower rates of distant recurrence. The results of the ESPAC-3 study showed that completion of all six cycles of planned adjuvant chemotherapy rather than early initiation was an independent prognostic factor after resection for pancreatic adenocarcinoma[38]. Similarly, a previous study by Hirono et al[16] revealed that noncompletion of postoperative adjuvant therapy was a risk factor for distant recurrence in resectable PDAC patients who underwent PD. These findings emphasize the importance of completing an adequate course of adjuvant chemotherapy. Subgroup analysis based on the combination of the TRIANGLE operation and chemotherapy lasting ≥ 6 months revealed that for patients who underwent a sufficient course of adjuvant chemotherapy, previous implementation of the TRIANGLE operation significantly prolonged mRFS. In addition, not only may the TRIANGLE operation itself prolong RFS, but this effect is also more pronounced when accompanied by an adequate course of adjuvant chemotherapy. The results indicate that the TRIANGLE operation extends mRFS but should be combined with an adequate course of adjuvant chemotherapy. Furthermore, sufficient courses of adjuvant chemotherapy were found to significantly prolong OS in the PD_TRIANGLE and PD_non-TRIANGLE groups. The results highlight the importance of combining an appropriate extent of lymphadenectomy with an adequate course of adjuvant chemotherapy, consistent with the concept of comprehensive treatment of pancreatic cancer.

This study has several limitations. First, the study spans a long period, during which definitions of some clinicopathological indicators changed, and chemotherapy regimens evolved based on yearly guidelines, which may introduce confounding results. Second, this study was conducted at a single center, which may limit the generalizability of the findings. Future multicenter studies are necessary to validate the results.

CONCLUSION

The results of this study suggest that the TRIANGLE operation is a safe and feasible approach for treating PDAC during PD. More importantly, the TRIANGLE operation reduces the local recurrence rate of PDAC and may also prolong RFS and OS when combined with adequate postoperative adjuvant chemotherapy. However, additional studies with larger cohorts are needed to confirm the conclusions drawn from this study.

ARTICLE HIGHLIGHTS

Research background

Data on the long-term oncological outcomes of the TRIANGLE operation in resectable pancreatic ductal adenocarcinoma (PDAC) patients undergoing pancreaticoduodenectomy (PD) are limited.

Research motivation

The TRIANGLE operation shows efficacy postneoadjuvant therapy in locally advanced PDAC, but its role in resectable PDAC is unclear. This study explored the safety and prognostic impact of this approach in patients with resectable PDAC.

Research objectives

To assess the safety of the TRIANGLE operation during PD and its prognostic relevance for resectable PDAC recurrence and survival.

Research methods

This retrospective cohort study included patients who underwent PD for pancreatic head cancer between January 2017 and April 2023, with or without the TRIANGLE operation. Patients were divided into the PD_TRIANGLE and PD_non-TRIANGLE groups. Surgical and survival outcomes were compared between the two groups. Adequate adjuvant chemotherapy was defined as adjuvant chemotherapy ≥ 6 months.

Research results

The study included 52 patients in the PD_TRIANGLE group and 55 in the PD_non-TRIANGLE group, with no significant differences in baseline or perioperative outcomes. The PD_TRIANGLE group had a lower recurrence rate (48.1% vs 81.8%, P < 0.001) and a decrease in local PDAC recurrence from 37.8% to 16.0%. Multivariate Cox regression analysis revealed that PD_TRIANGLE, adequate adjuvant chemotherapy (≥ 6 months), and margin status were independent predictors of the recurrence rate.

Research conclusions

The TRIANGLE operation during PD is safe, reduces local recurrence of PDAC, and potentially enhances recurrence-free survival and overall survival with sufficient adjuvant chemotherapy.

Research perspectives

Further research with a larger cohort is essential to validate the survival outcomes of the TRIANGLE operation.

ACKNOWLEDGEMENTS

The authors appreciate all the team members for their help. We appreciate Hekai Shi for his help.