Most total pancreatectomies for ductal adenocarcinoma potentially can be replaced by Whipple over the splenic artery: a before and after study

Abstract

Introduction:

Recently, more and more total pancreatectomy (TP) has been performed for central-located pancreatic ductal cell adenocarcinoma (PDCA), which abuts or involves both gastroduodenal and splenic arteries and demands transaction of both of them for complete resection. Spiked by Warshaw’s procedure (spleen-preserving distal pancreatectomy with excision of splenic vessels), the authors developed a new procedure “Whipple over the splenic artery (WOTSA)” to replace TP by leftward extension of pancreatic parenchyma transaction line and preservation of pancreatic tail and spleen after excision of the splenic artery. This uncontrolled before and after study assesses the safety and efficacy of a new technique “Whipple over the splenic artery (WOTSA)” as a treatment for pancreatectomy for ductal adenocarcinoma (PDAC), which traditionally requires TP for a complete excision.

Methods:

The study group comprised 40 consecutive patients who underwent WOTSA for PDAC between August 2019 and September 2022. Their clinicopathological characteristics and survival were compared with those of a historical control group comprising 30 consecutive patients who underwent TP between January 2016 and July 2019.

Results:

None of the 40 patients in the WOTSA group required reoperation due to infarction of the pancreas and/or spleen remnant. Diabetes mellitus (DM) medication after WOTSA were none in 19, oral hypoglycemic agents in 19, and insulin preparations in 2 patients. Compared with TP, patients who underwent WOTSA exhibited similar rates of major operative complications, clear pancreatic parenchyma transaction margin, and a number of harvested positive lymph nodes, but a higher rate of adjuvant chemotherapy completion and a trend toward better median disease-free survival (14 vs. 7.5 months, P=0.023).

Conclusions:

Compared to TP, WOTSA can be safely performed and have much better postoperative glycemic status without cost of higher operative risk or impaired surgical radicality. These findings indicate that most TPs for PDAC potentially can be replaced by WOTSAs.

Introduction

Highlights

• Total pancreatectomy (TP) is required for complete excision of pancreatic cancer invading or abutting both gastroduodenal and splenic arteries.

• Whipple over the splenic artery (WOTSA) with the leftward extension of pancreatic parenchyma transaction line and preservation of pancreatic tail and spleen after excision of splenic artery can be safely performed.

• Compared to TP, WOTSA had similar operative risk, rate of clear Pancreatic parenchyma transection margin (PPTM), number of harvested positive lymph nodes, but better postoperative glycemic status, higher rate of completed adjuvant chemotherapy, and better disease-free survival.

• Most TPs for ductal adenocarcinoma can potentially be replaced by WOTSA.

Margin status is one of the most frequently reported predictors for recurrence after pancreatectomy for ductal adenocarcinoma (PDAC)^1–4. In patients undergoing pancreaticoduodenectomy (PD) for PDAC, the surgical margins include the pancreatic parenchyma transection margin (PPTM i.e. neck/body), anterior, superior mesenteric vein groove, medial (uncinate/superior mesenteric artery), and posterior/retroperitoneal margins. Of these, the PPTM is the most technically amenable because an additional resection can potentially convert a positive intraoperative frozen section (IOFS-R1) to a negative on permanent section (PS-R0)⁵. However, the literature review showed up to 22% patients with a positive IOFS at PPTM did not have additional resection^6,7. Moreover, up to 40% of additional resections failed to convert a positive IOFS (IOFS-R1) to PS-R0^6,7. Together, more than half (53%⁶, 57%⁷) of patients with IOFS-R1 at the PPTM eventually had PS-R1. Postulated reasons for these absent or failed additional resections include the following: (1) difficult to perform an additional resection wide enough to get PS-R0 and (2) reluctance to convert the operation into total pancreatectomy (TP). Steven et al. ⁸ described a procedure “Whipple at the splenic artery (WATSA)” and denoted that the maximal left extent of the PPTM in a PD is just to the right of the point that the splenic artery (SA) contacts the superior border of the pancreatic body (Fig. 1A, line B). Likewise, the maximal left extent of an additional resection in response to a positive IOFS at PPTM is similar to that in WATSA, and we denominated it as the “WATSA point”. More leftward extension of the PPTM over the WATSA point demands transection of the SA (Fig. 1A, line C), which traditionally will be followed by a conversion of the operation into TP because of the concern of insufficient blood supply to the pancreas and spleen remnants⁸. This hypothesis is consistent with Zhang’s report that TP was required to convert a positive IOFS of the PPTM to PS-R0 in nearly one-third of patients during PD for PDAC⁹.

Figure 1.

(A) Pancreatic parenchyma transection line. Hatched line A, at the neck just. above the superior mesenteric-portal vein in the conventional Whipple procedure; hatched line B, just to the right of the point where the splenic artery contacts the superior border of the pancreas in “Whipple at the splenic artery (WATSA)”; hatched line C, ≥2 cm left to the “WATSA” point with excision of the proximal segment of the splenic artery and vein. (B) After the excision of the proximal segment of the splenic artery and vein, the pancreas and spleen remnants receive blood supply from the left gastroepiploic and short gastric vessels.

The rationale to obtain a clear PPTM by TP is to improve patients’ survival. However, Schmid et al.¹⁰ reported that TP improves survival over subtotal pancreatectomy only in patients with isolated PPTM-positive PDAC. In other words, conversion of PD into TP as a response to a positive IOFS at PPTM in PDAC with positive margins at other site(s) will result in patient’s poor postoperative glycemic status without any benefit in survival. However, in face of a positive IOFS at PPTM during PD for PDAC, it would be difficult or even impossible to ascertain the margin status at other sites and whether to convert the operation into TP or not.

But, it may not be necessary to convert the operation into TP after transection of the splenic artery and leftward extension of the PPTM (Fig. 1A, line C). As shown in Warshaw’s procedure, that is spleen-preserving distal pancreatectomy with excision of splenic vessels has been widely and safely performed for decades^11,12. Theoretically, after the excision of the proximal segment of the splenic artery and transact the pancreatic parenchyma at the site left to WATSA point, the retrograde blood flow from left gastroepiploic and short gastric vessels could be sufficient for not only the spleen but also the pancreas remnants (Fig. 1B). We denominated this procedure as “Whipple over the splenic artery (WOTSA)”.

To address whether WOTSA is a better procedure in comparison to TP in patients with PDAC, which demands TP for complete excision, we compare clinicopathological characteristics and survival between 40 consecutive cases treated with WOTSA and a historic control group of 30 consecutive patients treated with TP.

Patients and methods

Choice of operative method among PD, WATSA, WOTSA, and TP

During PD for PDAC, the PPTM is performed at either the neck (Fig. 1A, line A) when it looks healthy or at the WATSA point (Fig. 1A, line B) when the tumor involves the neck. The PPTM was routinely sent for IOFS. In case of a negative IOFS at the neck, conventional PD was performed. With a positive or equivocal IOFS at the neck, an additional resection was performed at the WATSA point (Fig. 1, line B) with the PPTM sent for the second IOFS. In case of a negative IOFS at the WATSA point, WATSA was performed. For a positive or equivocal IOFS at WATSA point, TP (before) or WOTSA (after August 2019) was performed. TP or WOTSA was also performed as a priori decision when preoperative images showed tumor involvement or abutment of both the gastroduodenal and splenic arteries (Fig. 2A).

Figure 2.

Whipple over the splenic artery (WOTSA) technique. (A) Preoperative. contrasted computed tomography image shows pancreatic head-neck tumor invading the gastroduodenal and splenic arteries; (B) Planned transection sites of the splenic artery and pancreatic parenchyma are to the left of the point where the splenic artery contacts the upper border of the pancreas (WATSA point); (C). The splenic artery is divided first at its root and then at the site of the planned pancreatic parenchyma transection line; (D) The pancreatic parenchyma and splenic vein are divided as a whole with a stapler at the planned transection line; (E). After dissection, the specimen is tethered only by the superior mesenteric and portal veins; (F) The superior mesenteric and portal veins are divided and reconstructed; (G). Specimen after WOTSA; (H, I) Contrasted computed tomography performed 1 week after the surgery shows good enhancement of the pancreas and spleen remnant after the excision of the proximal segment of the splenic vessels.

Patients

Since WOTSA is a new technique, we performed it after obtaining approval from Ethics Committee and the work has been reported in line with the STROCSS criteria¹³. Supplemental Digital Content 1, http://links.lww.com/JS9/C655. Since August 2019, we used WOTSA for PDAC, which traditionally would need a TP for a complete resection due to a tumor involving or abutting both gastroduodenal and splenic arteries or a positive IOFS of PPTM at WATSA point. This study was designed as a prospective uncontrolled before and after study. The study group consisted of 40 consecutive patients receiving WOTSA for PDAC between August 2019 and September 2022 to ensure at least 12 months’ follow-up. Informed written consents were obtained from all studied patients for publication of this case report and accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal on request. The historical TP group consisted of 30 consecutive patients receiving TP for PDAC between January 2016 and July 2019 (Fig. 3 Flowchart).

Figure 3.

Classification of patients who underwent surgical resection for ductal. adenocarcinoma at the head, neck, or uncinated process of the pancreas by operative methods. Cohorts were defined as follows: (1) conventional pancreaticoduodenectomy (PD) with pancreatic parenchyma transected at neck; (2) Whipple at the splenic artery (WATSA) with pancreatic parenchyma transected just to right of the point that splenic artery contacts the upper border of pancreas (WATSA point); (3) Whipple over the splenic artery (WOTSA) with pancreatic parenchyma transected 2 cm or more left to WATSA point; and (4). Total pancreatectomy with splenectomy (TP).

Technique of WOTSA

Most parts of the WOTSA were performed as WATSA described by Steven et al. ⁶ but with the pancreatic parenchyma transacted at site of 2 cm or more left to the WATSA point. The SA was transected first at its root and then at site of 1 cm left to the planned pancreatic parenchyma transaction line (Fig. 2B and C). The pancreatic parenchyma and splenic vein were divided as a whole with a stapler (Fig. 2D). The PPTM was routinely sent for IOFS. If IOFS turned out to be equivocal or positive, additional resection would be made. Once the pancreas and splenic vein were divided, the anterior and right borders of the superior mesenteric artery were cleared (Fig. 2E). The superior mesenteric and portal veins were cross-clamped, divided 0.5 cm from the tumor, and reconstructed (Fig. 2F). Pancreatic stump was closed without main duct reconstruction in the initial 4 cases because of concern of ischemia-related poor wound healing and another 2 cases with MPD orifice unrecognizable on cut surface of the pancreatic stump. The stapled cut edge on the pancreatic stump was removed to expose orifice of the main pancreatic duct, and then an end-to-side one-layer pancreaticojejunostomy was completed with a trans-anastomotic internal stent in the pancreatic duct using a suitable pediatric feeding tube as previously described¹⁴ in the remained 34 cases.

Postoperative follow-up

Patients who underwent PD were entered into a registry approved by the Institutional Review Board (IRB). Demographic, clinical and pathological data were also recorded. A pathologic composite margin analysis was performed according to a standardized protocol described by Esposito et al. ¹⁵ and a positive microscopic margin was defined as cancer cells at or within 1 mm of the margin of resection. Postoperative pancreatic fistula (POPF), delayed gastric emptying (DGE), and post-pancreatectomy hemorrhage (PPH) were defined and graded according to the International Study Group of Pancreatic Surgery (ISGPS)^16–18. Postoperative complications were graded according to the Clavien–Dindo classification, as minor (grade 2 or lower) or major (grade 3A or higher) and reported as 90-day outcomes¹⁹.

Measurement of the remnant pancreatic volume

Contrasted CT images acquired before and one week after the surgery were analyzed using 3D Slicer software (v. 4.10.2, www.slicer.org), a free, open-source software for medical image analysis. Free-hand regions of interest were drawn slice-by-slice on contrast-enhanced computed tomography (CT) images by one radiologist (Ban-Bin Chen) to include all normal pancreatic tissue preoperatively. The areas with pancreatic cancer, postoperative residual pancreatic tissue, splenic vein, and dilated pancreatic duct (≥3 mm) were excluded. The integral area of each manually segmented slice of the pancreas, defined as the functional pancreatic volume, was automatically calculated by the Slicer software. The percentage of the functional residual pancreatic volume was calculated as follows: (preoperative functional volume—postoperative functional volume)/preoperative pancreatic functional volume.

Postoperative chemotherapy and follow-up

The adjuvant chemotherapy regimen was selected based on the doctor’s suggestion and the patients’ preference. The adjuvant chemotherapy regimen comprised the administration of oral TS-1, Gemcitabine, Abraxane, and FOLFIRINOX. All patients underwent laboratory evaluation of the CEA and CA19-9 serum levels every 2 or 3 months and MDCT levels every 3 months for 2 years, and every 6 months thereafter. The type of diabetes mellitus (DM) treatment and the timing of its initiation were determined by the surgeon or DM specialist.

Evaluation of perioperative changes in glycemic status

The evaluation of perioperative changes in glycemic status was performed with blood tests 1 day before and 3 and 6 months after the operation and diabetes medications.

Statistical analysis

Statistical analysis was performed using IBM SPSS Statistics version 21×86 software. Continuous variables were summarized using means and standard deviations or medians and interquartile ranges, whereas categorical variables were summarized using frequencies and percentages. Differences between operative groups were analyzed using either a t-test or Wilcoxon rank-sum test for continuous variables and the χ² test or Fisher’s exact test for categorical variables. Overall survival (OS) was measured from the date of surgery until death. The median OS was estimated using the Kaplan–Meier method with 95% CIs. All statistical tests were two-sided, and differences were considered statistically significant at P less than 0.05.

Results

Patients

370 (41.6%) of the 890 pancreatectomies performed at NTUH between January 2016 and September 2022 were for ductal adenocarcinoma at the head, neck, or uncinated process of the pancreas. Operative methods included 253 (68.4%) conventional PDs, 47 (12.7%) WATSAs, 40 (10.8%) WOTSAs, and 30 (8.1%) TPs. Of the 30 TPs, 18 (60%) were performed as a priori decision because of tumor involvement of both GDA and SA shown on preoperative images and 12 (40%) as a deliberate reaction to a positive intraoperative frozen section (IOFS) of PPTM at the WATSA point. Of the 40 WOTSA, 17 (42.5%) were performed as a priori decision and 23 (57.5%) after a positive IOFS at the WATSA point (Fig. 3).

Clinicopathological characteristics

Comparison of clinicopathological characteristics between WOTSA and TP group patients were listed in Table 1. There were no significant inter-group differences in demographic data.

Table 1.

Demography of studied patients receiving WATSA, WOTSA, and total pancreatectomy for ductal adenocarcinoma.

	WOTSA, N=40	TP, N=30	P
Age (year)a	66.35 (9.8)	64.7 (8.4)	0.455
Male sex (M:F)	18:22	12:18	0.676
BMI (kg/m2) a	23.5 (3.2)	22.4 (3.0)	0.100
Albumin (g/dl)a	4.2 (0.3)	4.2 (0.4)	0.723
Lymphocyte count (103/μl)a	1.62 (0.55)	1.88 (0.61)	0.999
Malnutrition universal screening toolb	1 (0-2)	1 (0 – 2)	0.806
Comorbidity, n (%)
Hypertension	18 (43)	11 (37)	0.484
Coronary artery disease	4 (10)	4 (13)	0.664
Liver cirrhosis	1 (2.5)	1 (3)	0.836
Other malignance	5 (12.5)	2 (7)	0.421
Preoperative jaundice	17 (42.5)	16 (53)	0.369
Preoperative biliary drainage, n (%)
No	29 (72.5)	16 (53)	0.098
Yes	11 (28)	14 (47)
ERBD	8 (80)	12 (40)	0.420
PTBD	3 (20)	2 (7)
Clinical stage, n (%)			0.723
Resectable	32 (80)	25 (83)	0.103
No SMPV abutment	0	2 (8)
SMPV abutment <180°	32 (83)	23 (92)
Borderline resectable	8 (20)	5 (17)	0.317
SMPV encasement >180°	2 (25)	3 (60)
CHA abutment reconstruct	5 (62.5)	1 (20)
SMPV encasement and CHA abutment	1 (12.5)	1 (20)

CHA, Common Hepatic Artery; ERBD, Endoscopic Retrograde Biliary Drainage; PTBD, Percutaneous Transhepatic Biliary Drainage; SMPV, Superior Mesenteric-Portal Vein.

^aValues are mean (s.d.)

^bMedian (IQR.).

Operative procedure

The mean operative time and amount of intraoperative blood loss of the 40 WOTSA group patients were 293±60.0 min and 680±555.6 ml, respectively. There were no significant differences in the rate of portal vein and/or common hepatic artery, level of mesopancreas dissection, operative time, and amount of intraoperative blood loss between the TP and WOTSA groups (Table 2).

Table 2.

Pathological outcomes and perioperative change of glycemic status.

	WOTSA, N=40	TP, N=30	P
Tumor differentiation			0.619
Well, n (%)	9 (22.5)	4 (13)
Moderately, n (%)	26 (65)	22 (74)
Poorly, n (%)	5 (12.5)	4 (13)
Tumor size, cm, Mean±SD	3.5±1.2	4.1±1.8	0.147
T stage, n (%)			0.085
I	1 (2.5)	3 (10)
II	31 (77.5)	16 (53.3)
III	8 (20)	11 (36.7)
Peri-neural invasion, n (%)	36 (90)	28 (93)	0.622
Lymphovascular invasion, n (%)	23 (57.5)	16 (53)	0.728
Lymph node, n (mean±SD)
Resected	29±11	33±9	0.075
Positive	2.5±2.9	3±3	0.936
N stage, n (%)			0.641
0	12 (30)	8 (26.7)
1	17 (42.5)	16 (53.3)
2	11 (27.5)	6 (20)
AJCC8th tumor stage, n (%)			0.208
I A/B	12 (30)	4 (13.3)
II A/B	17 (42.5)	18 (60)
III	11 (27.5)	8 (26.7)
Negative PPTM, n (%)	40 (100)	30 (100)	0.999
Positive margin(s) # at site(s) other than PPTM, n (%)	18 (45)	13 (43)	0.890
Preoperative glycemic status, n (%)			0.934
No DM	19 (47.5)	13 (43.3)
New-onset DM	12 (30)	10 (33.3)
Long-standing DM	9 (22.5)	7 (23.3)
Pancreas functional residual volume ratio (%), median (IQR)	22.6 (16.6–29.2)	0	<0.001
Perioperative change of glycemic status, n (%)			<0.001
No or resolution of DM	26 (65)	0
Persistent /aggravated DM	14 (35)	30 (100)
DM medication at time of 6 months after operation, n (%)			<0.001
None	19 (47.5)	0
OHA	19 (47.5)	0
Insulin preparations	2 (5)	30 (100)
Adjuvant chemotherapy, n (%)			0.028
Complete	29 (72.5)	14 (46.7)
Incomplete/no	11 (27.5)	16 (53.3)

DM, Diabetes mellitus; PPTM, pancreatic parenchyma transection margin; OHA, Oral Hypoglycemic Agents; TP, total pancreatectomy; WOTSA, Whipple over the splenic artery.

Negative margin*, R0 >1 mm; Positive margin ^# , including R1 ≤ mm and R1-direct.

Operative safety

Contrast-enhanced abdominal CT performed 1 week after WOTSA showed good enhancement of the pancreas and spleen remnant in all 40 patients in WOTSA group (Supplemental Fig. 1, Supplemental Digital Content 2, http://links.lww.com/JS9/C656) and thereafter none of them had reoperation because of infarction of the pancreas and/or spleen remnant. There was one reoperation for the removal of the inadvertently stappled nasogastric tube by a stapler used for gastrojejunostomy anastomosis. Of the 40 WOTSA group patients, there were two biochemical, two grade B POPFs, five DGE (graded as A in two and B in three). The median length of hospital stay after WOTSA was 18.5 days. There was no readmission or death within 90 days after the operation. There were no significant differences in the rate and severity of the DGE and major complications between TP and WOTSA groups (Table 2). However, patients in TP group had significantly longer postoperative hospital stays and more readmissions, including 2 for hypoglycemic episodes and 2 for intra-abdominal abscess.

Pathologic data

There were no significant inter-group differences in tumor size and grade, rate of peri-neural/lymphovascular invasion, number of harvested and positive lymph nodes, and the American Joint Committee on Cancer eighth edition tumor stage (Table 2). All 40 patients in the WOTSA group had a PS-R0 at the PPTM. However, 18 (45%) of them had positive margins at other site(s), including 9 (22.5%) at the uncinate process, 8 (20%) at the anterior or/and posterior surface, and 1 (2.5%) at SMPV. Thirteen (43.3%) of 30 TP group patients had positive margins, 9 (30%) at the uncinated process, 3 (10%) at the SMPV groove, and 1 (3.3%) at the anterior surface.

Perioperative glycemic changes

There was no significant inter-group difference in the preoperative glycemic status. The median pancreatic functional residual volume after WOTSA was 22.6 (range, 9.9–56.6%). DM medication at 6 months after WOTSA was absent in 19 patients, OHA in 19, and insulin in 2. In contrast, all TP group patients needed insulin preparation for their DM control (Table 2).

Adjuvant chemotherapy

Significantly more WOTSA group patients had complete adjuvant chemotherapy [WOTSA, 29 (72.5%) vs. 14 (46.7%), P=0.028, Table 2].

Survival analysis

For the entire study group, the median follow-up time for DFS and OS analyses was 9 [interquartile range (IQR) 6–16] and 18 (IQR 12–31) months; for living patients, it was 28 (IQR 13–35) months. On unadjusted analysis, the WOTSA group patients displayed improved median DFS (14 months) compared with the TP cohort (7.5 months, P=0.023, Fig. 4A). Although not statistically significant, there was a trend toward better median overall survival in WOTSA (not yet reached) than TP group patients (17 months, P=0.059, Fig. 4B). A multivariate survival analysis based on Cox proportional hazards model showed no or incomplete adjuvant chemotherapy was a statistically significant poor predictor of both DFS [hazard ratio 2.81 (95% CI, 1.47–5.36); P=0.002, Table 3] and OS [hazard ratio 3.01 (95% CI, 1.61–5.64); P=0.001, Table 4] and lymph node metastases was the other poor predictor of OS [hazard ratio 1.88 (95% CI, 1.17–3.01); P=0.009, Table 4]

Figure 4.

(A) Disease-free and (B) overall survival. Kaplan–Meier curves for patients treated with Whipple over the splenic artery (WOTSA) compared with those treated with total pancreatectomy (TP).

Table 3.

Cox proportional hazards analyses associated with disease-free survival.

Disease-free survival
	Univariate		Multivariate
Variable	HR (95% CI)	P	HR (95% CI)	P
Age ≥75	0.634 (0.348–1.155)	0.128
Male	1.010 (0.573–1.781)	0.972
Borderline resection	1.050 (0.492–2.244)	0.900
T3 stage	1.501 (0.835–2.698)	0.183
N stage	1.657 (1.034–2.654)	0.037	1.560 (0.978–2.490	0.062
High histology grade (moderate, poor)	1.217 (0.570–2.600)	0.604
LVI	1.193 (0.678–2.098)	0.539
PNI	1.382 (0.496–3.852)	0.517
R1 resection	1.526 (0.870–2.678)	0.143
Major complication	2.621 (1.098–6.256)	0.052	1.154 (0.442–3.010)	0.770
No completed chemotherapy	3.051 (1.722–5.406)	0.000	2.808 (1.472–5.355)	0.002
Surgical method: TP	1.872 (1.064–3.295)	0.029	1.330 (0.726–2.439)	0.356

Factors in univariate analysis with P<0.1 were included in multivariate analysis.

HR, hazard ratio; TP, total pancreatectomy LVI, Lymphovascular Invasion; PNI, Perineural Invasion.

Table 4.

Cox proportional hazards analyses associated with overall survival.

Overall survival
	Univariate		Multivariate
Variable	HR (95% CI)	P	HR (95% CI)	P
Age ≥75	0.760 (0.397–1.455)	0.402
Male	0.941 (0.511–1.734)	0.846
Borderline resection	0.755 (0.295–1.930)	0.543
T3 stage	1.217 (0.641–2.312)	0.553
N stage	1.995 (1.240–3.209)	0.005	1.875 (1.169–3.005)	0.009
High histology grade (moderate, poor)	0.858 (0.394–1.868)	0.703
LVI	1.144 (0.626–2.090)	0.662
PNI	1.113 (0.394–3.145)	0.837
R1 resection	1.053 (0.564–1.968)	0.871
Major complication	2.183 (0.848–5.623)	0.141
No complete chemotherapy	3.078 (1.671–5.672)	0.001	3.011 (1.608–5.639)	0.001
Surgical method: TP	1.796 (0.958–3.366)	0.066	1.250 (0.641–2.438)	0.660

Factors in univariate analysis with P<0.1 were included in multivariate analysis.

HR, hazard ratio; TP, total pancreatectomy LVI, Lymphovascular Invasion; PNI, Perineural Invasion.

Discussion

We compared clinic-pathological data between 40 WOTSA and 30 TP patients and found no significant differences in operative risk, surgical margins, number of harvested positive lymph nodes. But WOTSA group patients had better postoperative glycemic status, higher rate of completed adjuvant C/T, and a trend toward better patients’ DFS.

Desaki et al. ²⁰ reported 18 successful PDs with resection of the splenic artery (PDSAR) for pancreatic cancer involving both splenic and gastroduodenal arteries shown on preoperative images. In addition to Desaki’s indication for PDSAR, we expand indications for this technique further to patients with a positive IOFS at the WATSA point. Therefore, we denominate it as WOTSA to emphasize its use in the extension of the PPTM leftward over the WATSA point. A retrospective review of 158 patients who underwent Warshaw’s procedure showed that three (1.9%) patients required reoperation due to splenic infarction¹². However, both pancreas and spleen remnants were safely preserved after WOTSA in all our 40 and Desaki’s 18 patients. Compared to Warshaw’s procedure, a longer splenic artery remnant is left after WOTSA and receives additional blood supply from not only the left gastroepiploic/short gastric vessels but also the posterior epiploic artery²¹, which may account for the successful preservation of pancreas/spleen remnants in all of the Desaki’s and our patients.

With recent improvements in the management of brittle diabetes mellitus, an increasing number of studies advocate TP to obtain a clear PPTM^22–24. However, a recent systemic review of 1536 patients showed that the overall quality of life after TP is impaired, particularly by the impact of diabetes-related morbidity and diarrhea^25,26. Although the functional residual volume of the pancreas remnant is low after WOTSA (median, 22.8%), patients’ glycemic status after WOTSA are surprisingly good. Compared to WOTSA, TP group patients had longer length of hospital stay after operation and more readmission after discharge because of DM-related problems. At 6 months after WOTSA, only 2 (5%) patients needed insulin preparations for their DM control. Moreover, adjuvant chemotherapy (the only positive predictor of DFS and OS) was significantly more commonly completed after WOTSA than TP. We attributed the excellent postoperative glycemic status after WOTSA to regional differences in islet distribution in the human pancreas. Wang et al. ²⁷ reported islet distribution/density is twofold higher in the tail compared to that in the head and body regions which may account for the much better postoperative glycemic status in WOTSA group patients in spite of small residual volume of functional pancreatic parenchyma.

As stated before, up to 40% of additional resections failed to convert a positive IOPFS (IOPFS-R1) to PS-R0^6,7. However, a wide additional resection with pancreatic parenchyma transacted at site 2 cm or more left to the WATSA point resulted in PS-R0 at PPTM in all 40 WOTSA group patients. Theoretically, compared to WOTSA, TP will harvest additional lymph nodes distributed along the pancreatic tail and splenic hilum. However, both our study and others showed that compared with that of the WOTSA, TP has a higher number of harvested but equal number of positive lymph nodes^28,29. Therefore, in oncologic consideration, WOTSA and TP have similar efficacy in both surgical margins and lymph node dissection.

Kooby et al. ⁶ reported that conversion of FS-R1 to PS-R0 neck margin by additional resection or even total pancreatectomy failed to prolong patients’ median OS [FS-R1 to PS-R0 (n=72), 11.9 months; PS-R1, 13.7 months, P=0.98]. In contrast, Zhang and colleagues reported improved survival after revision of pancreatic neck margins based on intraoperative FS analysis [designated as complete resection—non-en-block (CR-NEB)] [CR-NEB (n=159), 24 months; incomplete resection (n=78), 19 months, P=0.020]. Kooby and colleagues attributed failure of conversion of FS-R1 to PS-R0 based on intraoperative FS in improving survival to FS-R1-related aggressive tumor biology reflected by larger tumor size (3.3±1.1 cm), more LN metastasis (76%) and peri-neural invasion (81%). Since all our studied patients had FS-R1 aggressive tumors because they had WOTSA or TP because of a positive IOFS at the WATSA point or tumor involvement of both GDA and SA. Pathology also showed similar aggressive characteristics, tumor size of 3.7±1.4 cm, LN metastasis (71%) and peri-neural invasion (90%) as described by Kooby. But median OS [25 (IQR) months] for our entire study cohort (n=70) was more consistent with Zhang’s report (CR-NEB, 24 months). We believe, in addition to aggressive biology, tumors’ location could also account for a positive IOFS at PPTM. Small tumors abut both gastroduodenal and splenic arteries tend to have a positive IOFS because of the splenic-artery-related limitation of leftward extension of the PPTM rather than aggressive tumor biology. This hypothesis is consistent with Johnson’s report that 4434 (83%) of 5521 patients receiving TP for PDAC had tumors <5 cm in size²⁹. In consideration of 12–15 cm long normal adult pancreas, most TPs were performed because of tumors’ location instead of extensive tumor involvement of the whole or nearly whole pancreatic parenchyma.

Our and other studies showed that up to 40–50% of TPs for PDAC were performed in patients with positive margin(s) at sites other than the PPTM, which will abolish the survival benefits of conversion of a positive IOFS at neck/body margin to negative on permanent section as reported by Schmid et al. ¹¹ However, in clinical practice, margin status at sites other than PPTM and tumor biology such as, peri-neural invasion and LN metastasis) can only be determined postoperatively. In a recent review article, Datta et al. ³⁰ questioned the value of IOFS of the PPTM during PD for PDAC. We believe these also account for the reluctance to convert the operation into total pancreatectomy (TP) and subsequent no or failed additional resection as a response to a positive IOFS at PPTM. Considering the high success rate in converting FS-R1 to PS-R0 and good postoperative glycemic status, tolerance to adjuvant chemotherapy, and patients’ survival, we advocate a routine IOFS at PPTM and WOTSA as a response to a positive result during PD for PDAC.

This study is an uncontrolled before-after study with a limited number of patients. We may overestimate the effect of WOTSA on patients’ survival because it might be derived from the time difference. Besides, this study was performed in a single institution, which will also limit its generalizability. But the criteria to define the status of surgical margin, harvested lymph nodes, and perioperative glycemic change were the same during the studied period. Nevertheless, a prospective randomized trial is required to reach a more definitive conclusion.

In conclusion, our study showed WOTSA can be safely performed and had similar surgical efficacy in PPTM clearance and LN dissection. Compared to TP, WOTSA group patients had better postoperative glycemic status, more completed adjuvant chemotherapy, and a trend toward better DFS. These excellent results denote that most TPs for PDAC can and should be replaced by WOTSA.

Ethical approval

Research Ethics Committee A, National Taiwan University Hospital, No. 7, Chung-Shan South Road, Taipei, Taiwan 100, R.O.C. Reference number: 202004002RINA State for judgement: approval.

Consent

We obtained written and signed consent to publish a case report from all the studied patients. All the authors provide assurance that alterations do not distort scientific meaning.

Source of funding

All the listed authors did not receive any source of funding to support this study.

Author contribution

T.-C.K.: data collection, study design, and data analysis or interpretation. C.-H.W.: data collection, data analysis or interpretation. B.-B.C.: measurement of pancreatic functional residual volume. Y.-J.L.: sketch of the graphic abstract.

Conflicts of interest disclosure

All listed authors have no conflicts of interest to declare and nothing to disclose with regard to commercial support.

Research registration unique identifying number (UIN)

The study was registered on ClinicalTrials.gov.

Name of the registry: Most Total Pancreatectomies for Ductal Adenocarcinoma Can Be Replaced by Whipple Over the Splenic Artery: A Before and After Study.

Hyperlink to the registration: https://register.clinicaltrials.gov/prs/app/action/SelectProtocol?sid=S000DKMC&selectaction=Edit&uid=U00002 FP&ts=2&cx=-fh3jcr Unique identifying number of the study: 202004002RINA.

Guarantor

Yu-Wen Tien.

Data availability statement

Due to the sensitive nature of privacy of the patients in this study, all the authors were assured raw data would remain confidential and would not be shared publicly. However, the data base might be available to researchers upon reasonable demands, subject to approval by the responsible ethics committee.

Provenance and peer review

The submitted paper has not been invited or not commissioned, externally peer-reviewed.