The American Journal of Surgery (2015) 209, 564-569

Midwest Surgical Association

Extended distal pancreatectomy for pancreatic adenocarcinoma with splenic vein thrombosis and/or adjacent organ invasion Alexandra M. Roch, M.D., M.S., Harjot Singh, M.D., Alexandra P. Turner, M.D., Eugene P. Ceppa, M.D., Michael G. House, M.D., Nicholas J. Zyromski, M.D., Attila Nakeeb, M.D., Christian Max Schmidt, M.D., Ph.D., M.B.A.* Department of Surgery, Indiana University School of Medicine, Indianapolis, IN, USA KEYWORDS: Adenocarcinoma of the pancreatic body/tail; Distal pancreatectomy; Extended resection; Outcomes

Abstract BACKGROUND: Patients with adenocarcinoma of the pancreatic body/tail and associated vascular thrombosis or adjacent organ invasion are suboptimal candidates for resection. We hypothesized that extended distal pancreatectomy (EDP) for locally advanced adenocarcinoma is associated with a survival benefit. METHODS: We retrospectively reviewed a prospectively collected database of patients who underwent distal pancreatectomy (DP) for adenocarcinoma at a single academic institution (1996 to 2011) with greater than or equal to 2 years of follow-up. RESULTS: Among 680 DP patients, 93 were indicated for pancreatic adenocarcinoma. Splenic vein thrombosis (n 5 26) did not significantly affect morbidity, mortality, or survival. Standard DP was performed in 70 patients and 23 underwent EDP with no difference in morbidity/mortality. Patients with EDP had a survival comparable with patients with standard DP (disease-free survival 18 vs 12 months 5 .8; overall survival 23 vs 17 months, P 5.6). There was no difference in survival between EDP patients with versus without pathologic invasion of adjacent organs, but a trend favored those without. CONCLUSION: EDP is safe and should be considered in fit patients with locally advanced adenocarcinoma. Ó 2015 Elsevier Inc. All rights reserved.

Pancreatic adenocarcinoma is the fourth leading cause of cancer mortality in the United States, with an annual death

The authors declare no conflicts of interest. This article has been accepted for Plenary Session Oral Presentation at 2014 Midwest Surgical Association Annual Meeting, August 3–6, 2014, Mackinac Island, Michigan. * Corresponding author. Tel.: 11-317-278-8349; fax: 11-317-278-4897. E-mail address: [email protected] Manuscript received July 22, 2014; revised manuscript October 14, 2014 0002-9610/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjsurg.2014.10.017

rate approximating the incidence (37,390 and 43,920).1 Adenocarcinoma of the pancreatic body/tail is associated with a poor prognosis in part due to a delay in occurrence of specific symptoms until the tumor grows large and often invades adjacent organs. At diagnosis, pancreatic body/tail adenocarcinoma shows regional infiltration in 35% to 40% of patients.2 Patients with vascular thrombosis or invasion of adjacent organs may be suboptimal candidates for resection. A negative margin resection, however, may be the best chance for improvement in survival. Patients with

A.M. Roch et al.

Outcomes of extended distal pancreatectomy for adenocarcinoma

locally advanced unresectable tumors have a median survival of 9 months,1 which is not clearly improved by palliative systemic therapies.3 In an effort to increase resectability, pancreatic surgeons have attempted to push the limits of resection in patients with advanced disease with en bloc resection of the pancreas and adjacent organs and vessels. Pioneered in 1973 by Fortner,4 extended distal pancreatectomy with en bloc venous/arterial resection for vascular involvement has since then been documented. Some studies focusing on distal pancreatectomy with major vascular resections (portal vein/ superior mesenteric vein confluence, hepatic artery, superior mesenteric artery) have reported that extended/multivisceral resection for pancreatic cancer has acceptable morbidity, with similar long-term prognosis when compared with standard resection, and have thus emphasized the benefits of this procedure.5 However, other studies have also recognized that the use of multivisceral resection in the context of major pancreatic surgery increases morbidity.6 The long-term benefit of extended resection for locally advanced adenocarcinoma of the pancreatic body/tail remains unclear. We hypothesized that extended distal pancreatectomy is justified in cases of locally advanced pancreatic adenocarcinoma with adjacent organ infiltration/invasion.

Methods Patient’s selection From 1996 to 2011, data on all patients who underwent distal pancreatectomy at a single university-based medical center were prospectively collected in a database. These data were supplemented by retrospective review of electronic medical records. Patients with incomplete pathological data or whose final pathological diagnosis was not consistent with pancreatic ductal adenocarcinoma were excluded. Patients were also excluded if they were followed less than 2 years after surgery or if follow-up documentation was not available. The type of operation was divided into 2 distinct groups: standard distal pancreatectomy and extended resection. Extended resection was defined as en bloc resection of at least one additional adjacent organ that is not normally removed during the course of the operation but that is required to completely extirpate all local–regional macroscopic disease. Additional organs removed included adrenal gland, stomach, kidney, colon, or small bowel. Splenectomy was performed with standard distal pancreatectomy for adenocarcinoma, and was thus not considered an additional organ. Extended resection was performed with the clinical suspicion (based on preoperative studies and intraoperative assessment) that adjacent organs were invaded by tumor. The intent was to extirpate all local–regional tumor. Patients who underwent an extended resection were compared with patients with a standard distal pancreatectomy. Data were compiled and reported in strict compliance with patient confidentiality guidelines as defined by the Indiana University Institutional Review Board.

565

Parameters assessed Demographic data (age, sex), surgical procedure (type of procedure, mortality, complications, length of hospital stay), pathological tumor characteristics (TNM staging7, grade, adjacent organ invasion, splenic vein thrombosis, resection margins), and long-term outcomes (recurrence, disease-free survival [DFS] and overall survival [OS]) were assessed. A complication was considered as any event occurring within 30 days of surgery that modified the normal postoperative course and/or required medical treatment, including radiological drainage or other intervention. Postoperative mortality included in-hospital mortality or death within 30 days of surgical resection. Target events included death, local and distal recurrence (metastasis). Survival was calculated from the day of surgery. Follow-up information was regularly obtained from outpatient clinic visits.

Pathology Only patients with a pathological diagnosis of pancreatic ductal adenocarcinoma were included in this study. Pancreatic tumors other than ductal adenocarcinoma were excluded. Tumor stage was determined according to the 6th edition of the tumor node metastasis classification of the American Joint Committee on Cancer Staging.7 Both pancreatic transection margin and radial margin were analyzed by the pathologist. If one margin (or both) was positive for adenocarcinoma, the patient was considered as part of the ‘‘positive margin’’ group. When resection of adjacent organs was performed, close histological sectioning of those organs was used to determine the presence or absence of pathologic invasion to organs or structures of the extended resection.

Statistical analysis Data were compiled using Microsoft Excel 2011 (Redmond, WA) and analyzed with GraphPad Prism (GraphPad Software, Inc, La Jolla, CA). Descriptive statistics of continuous data included median, mean, standard error deviation, and range, whereas categorical variables were presented as percentages. Subgroup comparisons on continuous and categorical data were performed with the Student t-test and Fisher’s exact test, respectively. Actuarial DFS and OS curves were assessed according to the Kaplan–Meier method and compared with the log-rank test. Statistical significance was considered at P value less than .05.

Results Patients’ population Between 1997 and 2011, 680 patients who underwent distal pancreatectomy at Indiana University Hospital were identified from a prospectively maintained database. Of

566 them, 93 were indicated for adenocarcinoma of the pancreatic body/tail. Our population included 37 men and 56 women (sex ratio 5 .66), with a mean age of 65.4 years (range 37 to 84). Median follow-up of patients alive at the time of this report was 53.5 months (range 31 to 139).

Impact of splenic vein thrombosis Splenic vein thrombosis was present in 26 patients (28%), with pathological invasion of the vein in 13 only. Eleven patients (42.3%) with splenic vein thrombosis underwent extended distal pancreatectomy, whereas 15 (57.7%) underwent standard resection. Splenic vein thrombosis did not significantly affect length of hospital stay, morbidity, and mortality following distal pancreatectomy (10.3 vs 8.3 days, P 5 .22; 69.2% vs 61.2%, P 5 .63; 0% vs 1.5%, P 5 1). There was no impact of splenic vein thrombosis on long-term outcomes: DFS (7.4 vs 12.8 months, P 5 .13); OS (12.6 vs 18 months, P 5 .38); 1-, 3-, and 5-year survival rates (55.6% vs 64.2%, 34.6% vs 26.9%, and 11.5% vs 13.4%, P 5 .64, .46, and 1, respectively). Recurrence was not significantly increased in case of splenic vein thrombosis (30.8% vs 38.8%, P 5 .63), whether it was local recurrence (11.5% vs 13.4%, P 5 .63) or distant metastases (19.2% vs 25.4%, P 5 .6). Pathologic invasion of the vein did not impact short- and long-term outcomes.

The American Journal of Surgery, Vol 209, No 3, March 2015 procedure and not the extended resection. The remaining 2 patients had a standard hospital stay but were readmitted for delayed gastric emptying and small bowel obstruction on internal hernia. Long-term outcomes of recurrence (overall 43.5% vs 34.3%, P 5 .46; local recurrence 17.4% vs 11.4%, P 5 .48; distant metastases 26.1% vs 22.9%, P 5.78) and median survival (DFS 13.2 vs 10.4 months, P 5 .52; OS 20.3 vs 16 months, P 5 .74) did not differ between patients with extended resection and patients with standard resection. These results are summarized in Table 1 and Fig. 1. Of the 22/93 (23.7%) specimen with positive margin, 11 involved the retroperitoneal margin, 9 the pancreatic transection margin, and 2 the peripheral anterior margin. R0 resection (microscopically negative, n 5 71) was associated with longer median DFS and OS versus R1 (microscopic residual tumor) (15.1 vs 7.1 months and 22.8 vs 7.3 months, P 5 .003 and .0005), suggesting a bigger impact of the resection margin status over the type of resection itself. Among the 23 patients with extended distal pancreatectomy, 16 (69.6%) had resection of 1 adjacent organ, 5 (21.7%) of 2, and 2 (8.7%) of 3. The most commonly removed organs were the stomach (n 5 12), adrenal glands (n 5 7), colon (n 5 7), kidney (n 5 3), and small bowel (n 5 3). The type and number of organs resected did not significantly impact postoperative morbidity (P 5 .76 and .78, respectively).

Impact of adjacent organ invasion Comparison of standard distal pancreatectomy and extended distal pancreatectomy Twenty-three patients (24.7%) underwent extended distal pancreatectomy. When patients with extended resection (n 5 23) were compared to patients with standard resection (n 5 70), there was no difference in tumor histological characteristics, apart from the rate of T3/T4 tumors and pathological spleen invasion, each of which was higher in the case of extended resection. The average blood loss in the extended distal pancreatectomy group was 903 mL (range 100 to 3,500). Conversely, in the standard resection group, the average blood loss was 1,324 mL (range 100 to 6000). Median blood loss was 550 and 800 mL in the extended resection and standard resection groups, respectively (P 5 .26). Morbidity and mortality were comparable between the 2 groups (69.6% vs 61.4% and 0% vs 1.4%, P 5.62 and 1), although the mean hospital stay was significantly longer in case of extended resection (11.4 vs 8 days, P 5 .035). Among the 16 patients with extended distal pancreatectomy who had a postoperative complication, 12 of the 16 (75%) patients had a pancreatic fistula with secondary intra-abdominal abscess requiring interventional radiology drainage, or bleeding requiring radiological embolization or reintervention. Of the remaining 4 patients, 2 had a medical complication (new atrial fibrillation and Clostridium difficile colitis) requiring longer hospital stay for adequate treatment. Eight-eight percent (14/16) of patients with complications following an extended distal pancreatectomy had a longer hospital stay secondary to the pancreatic

Of the 23 patients with extended distal pancreatectomy, 11 (47.8%) had confirmed pathological invasion of adjacent organs on the surgical specimen. Margin status (P 5 1) and recurrence rate (P 5 .31 and .053 for local recurrence and P 5 1 for distant recurrence) were not significantly different in the extended resection subgroup with pathologic evidence of adjacent organ invasion versus patients with standard resection. There was also no difference in DFS (7.2 vs 10.4 months, P 5 .08) and OS (12.3 vs 16 months, P 5 .11) in patients with pathologic invasion of other organs compared with standard distal pancreatectomy patients. When we compared patients with documented pathologic invasion (n 5 11) with the entire cohort without pathologic invasion (n 5 82), patients without pathological invasion had a trend toward longer DFS and OS, although neither reached statistical significance (12.8 vs 7.2 months, P 5 .056; 16.7 vs 12.3 months, P 5 .072, respectively). Among extended resection patients, there was a trend toward decreased DFS and OS when proven pathologic evidence of adjacent organ invasion, although it did not reach statistical significance (7.2 vs 24 months and 12.3 vs 34.7 months, P 5 .07 and .05, respectively).

Comments More than 80% of pancreatic body/tail cancers are diagnosed at an advanced stage.1 Although surgical resection

A.M. Roch et al. Table 1

Outcomes of extended distal pancreatectomy for adenocarcinoma

567

Comparison of patients with extended distal pancreatectomy and standard distal pancreatectomy

Characteristics Demographic Age (mean, in years) Sex (men) Systemic treatment Neoadjuvant chemoradiation Pathology T3/T4 tumor Lymph node status Perineural invasion Lymphovascular invasion Tumor grade Low grade Moderate grade High grade Resection margins (positive R1 margins) Spleen invasion Postoperative outcomes Postoperative complications 30-day mortality Length of hospital stay (mean in days) Long-term outcomes 1-year survival 3-year survival 5-year survival Recurrence Local recurrence Distant metastases

Extended distal pancreatectomy (n 5 23)

Standard distal pancreatectomy (n 5 70)

61.8 10 (43.5%)

66.6 27 (38.6%)

.08 .81

2 (2.9%)

2 (8.7%)

.26

P value

17 16 16 9

(73.9%) (69.6%) (69.6%) (39.1%)

31 38 48 27

(44.3%) (54.3%) (68.6%) (38.6%)

.03* 0.1 1 1

8 13 8 5

(34.8%) (56.5%) (34.8%) (21.7%)

8 34 28 17

(11.4%) (48.6%) (40%) (24.3%)

.12

1

5 (21.7%)

3 (4.3%)

.03*

16 (69.6%) 0 11.4

43 (61.4%) 1 (1.4%) 8

.62 1 .04*

16 8 3 10 4 6

42 19 9 24 8 16

.47 .6 1 .46 .48 .78

(69.6%) (34.8%) (13%) (43.5%) (17.4%) (26.1%)

(60%) (27.1%) (12.9%) (34.3%) (11.4%) (22.9%)

Data are expressed as number (percent). *Significant P value.

is the best chance for long-term survival, surgeons are reluctant to perform a potentially morbid procedure in patients with vascular thrombosis and/or other organ invasion. We analyzed a subset of 93 patients who underwent a distal pancreatectomy with an intent to achieve a margin negative resection. Twenty-three of them received an extended resection to achieve complete removal of local– regional macroscopic tumor. In this report, we found that

Figure 1 (A) Disease-free survival and (B) overall survival. pancreatectomy (n 5 23).

extended distal pancreatectomy can be performed safely with comparable postoperative morbidity, mortality, and survival when compared with standard resection, with no impact of the type or number of organs resected. Patients with splenic vein thrombosis, with or without splenic vein invasion also fared equally well. A study by Shoup et al8 in 2003 compared patients with locally advanced tumor undergoing distal pancreatectomy

5 standard distal pancreatectomy (n 5 70);

5 extended distal

568 versus no resection because the tumor was deemed unresectable. They found a significantly longer median survival of 15.9 months following tumor resection compared with 5.8 months in patients who were not resected, suggesting that adjacent organ infiltration should not always be considered a poor prognosis indicator and contraindication to resection. Our study also showed that there was no difference in recurrence (local and distant) rate and survival (DFS and OS) between patients with pathological proof of adjacent organ invasion versus patients with standard distal pancreatectomy, and versus patients with extended resection but no adjacent organ invasion on pathology of the surgical specimen. However, patients with R0 resection had significantly longer DFS and OS. These results suggest that the resection margin status is more predictive of outcome than the type of resection itself,9 and confirm that an R0 resection should be the only goal when performing a distal pancreatectomy for adenocarcinoma, regardless of the extension of visceral resection. This type of multivisceral resections is already the standard of care for gastric and colon cancers with improved long-term outcomes and acceptable morbidity and mortality.10,11 To our knowledge, this study is the largest single institution series of patients with extended distal pancreatectomy for adenocarcinoma of the pancreatic body/tail, although it includes only 93 patients. This study, however, presents several limitations. It is a retrospective, single institution study, spanning over more than 15 years. This long study period can explain the abnormally high rate of positive margin in the standard group. Indeed, 10 of the 17 patients with positive margins in the standard group were operated on between 1996 and 2002, and 7 of the 10 patients had a positive pancreatic transection margin because no frozen section of this margin was used intraoperatively. Second, because of the poor resectability rate, the number of patients undergoing distal pancreatectomy for adenocarcinoma is low (93 patients in 14 years). That small number of patients, especially in the splenic vein thrombosis group (n 5 26), can have led to an error of second kind, which may have resulted in failure to assert statistical significance (seems more likely in the DFS comparison, as P value of .13 approximates statistical significance). Because of the regional referral status of Indiana University Hospital, most patients received their systemic treatment (adjuvant chemotherapy or radiation) at an outside hospital, with incomplete data on the presence, type, and duration of these treatments available at the time of this report. True pathologic invasion into adjacent organs or vascular structures is difficult to assess preoperatively on cross-sectional imaging or even intraoperatively. Although true pathologic invasion of adjacent organs had a trend of worse long-term outcomes, margin status appeared to be the main determinant. With long-term outcomes comparable with standard resection and significantly better outcomes when compared with no resection, extended distal

The American Journal of Surgery, Vol 209, No 3, March 2015 pancreatectomy should be performed in fit patients where an R0 resection can be achieved. Like every technically demanding and highly morbid surgical procedure, extended distal pancreatectomy should be performed in high-volume centers. Moreover, in those high-volume academic centers, multidisciplinary teams can assist in multivisceral resections and thus limit the postoperative mortality and morbidity. Given the low resectability rate of body/tail adenocarcinoma, larger multicenter series will be necessary to corroborate our findings.

References 1. Howlader N, Noone AM, Krapcho M et al. SEER Cancer Statistics Review, 1975-2010. National Cancer Institute. Bethesda, MD. Available at: http://seer.cancer.gov/statfacts/html/pancreas.html. Accessed December, 2014. 2. Brennan MF, Moccia RD, Klimstra D. Management of adenocarcinoma of the body and tail of the pancreas. Ann Surg 1996;223:506–11. 3. Sasson AR, Hoffman JP, Ross EA, et al. En bloc resection for locally advanced cancer of the pancreas: is it worthwhile? J Gastrointest Surg 2002;6:147–57. 4. Fortner JG. Regional resection of cancer of the pancreas: a new surgical approach. Surgery 1973;73:307–20. 5. Hirano S, Kondo S, Hara T, et al. Distal pancreatectomy with en bloc celiac axis resection for locally advanced pancreatic body cancer: long-term results. Ann Surg 2007;246:46–51. 6. Hartwig W, Hackert T, Hinz U, et al. Multivisceral resection for pancreatic malignancies: risk analysis and long-term outcomes. Ann Surg 2009;250:81–7. 7. Sobin LH, Wittekind C, International Union Against Cancer (UICC). TNM Classification of Malignant Tumors. 6th ed. New York: WileyLiss; 2002. 8. Shoup M, Colon KC, Klimstra D, et al. Is extended resection for adenocarcinoma of the body or tail of the pancreas justified? J Gastrointest Surg 2003;7:946–52. 9. Wagner M, Redaelli C, Lietz M, et al. Curative resection is the single most important factor determining outcome in patients with pancreatic adenocarcinoma. Br J Surg 2004;91:586–94. 10. Nakafusa Y, Tanaka T, Tanaka M, et al. Comparison of multivisceral resection and standard operation for locally advanced colorectal cancer. Dis Colon Rectum 2004;47:2055–63. 11. Brar SS, Seevaratnam R, Cardoso R, et al. Multivisceral resection for gastric cancer: a systematic review. Gastric Cancer 2012;15:100–7.

Discussion Dr Gerard V. Aranha (Maywood, IL). What is your blood loss in the extended versus the standard? What were the complications that kept the extended group in the hospital, 3 days longer than the standard group? Where should these operations be done; at a high volume center or can they be done in any hospital? Do your patients with distal pancreatic adenocarcinoma have laparoscopy? And how many were excluded based on your findings? Japanese have performed celiac axis resection and say that they have got good results. What are your thoughts on that? I think you told us that you don’t do that anyway. And, finally, how many patients had neoadjuvant therapy? Give us

A.M. Roch et al.

Outcomes of extended distal pancreatectomy for adenocarcinoma

your algorithm for the treatment of patients with distal pancreatic adenocarcinoma. Dr Schmidt: The mean blood loss was 900 CC’s. Standard distal pancreatectomy was 1300 CC’s. If you look at the medians, they’re closer. There is a large amount of variability in the distal pancreatectomy group just because there’s a lot of patients in that group compared to the extended. Looking at the median, 550 milliliters in extended; 800 milliliters in the standard. Most of the blood loss was associated with the pancreatectomy part of the procedure and not associated with resection of adjacent organs. The complications that kept the extended distal pancreatectomy patients extra days in the hospital will come as no surprise to you. They had a higher rate of fistula. Actually, in looking at this, it was statistically significant with a 52% incidence of pancreatic fistula in the extended group and 25% incidence in the standard group. Should extended distal pancreatectomy be performed in selected high volume centers? I would say yes with rare exception. Do all patients with distal pancreatectomy have laparoscopy and how many were excluded? I don’t have the exact number of how many were excluded, but based on recent study we had done looking at the impact of timing of CT scan on these patients, with a timely CT scan, about 10% percent of patients will have occult metastasis. In this area, we do routinely do laparoscopy, although in some of the patients in the 90s, it was not routinely done. Indeed, celiac axis resection and the so-called Appleby procedure is employed in Indiana University. It’s really done in highly selective cases, patients who have usually gone through neoadjuvant treatment and haven’t fallen out of that protocol for whatever reason. Based on the literature, it seems reasonable in high volume centers, highly selected cases. The outcomes are actually pretty remarkable, but you have to remember that it’s a pretty select group that has usually gone through neoadjuvant chemoradiation and probably selecting out a superior biology. These outcomes, however, are counterbalanced by the significantly higher morbidity and mortality. So I think these patients have to be appropriately counseled, and we have to look at whether it’s really worth it on an individual basis. How many

569

patients had neoadjuvant therapy? Just two, in the standard two in the extended distal pancreatectomy group. The large amount of neoadjuvant is in the major vascular section group. As far as an algorithm, it’s kind of evolving. Patients with distal pancreatectomy with just locally advanced tumors that involve the SMA, SMV portal veins and celiac axis go through a neoadjuvant chemoradiation approach. The current favored approach if the patient is able to tolerate the regimen, is a FOLFIRINOX regimen. Gem/ Abraxane is used for those who are less able to tolerate an aggressive chemotherapy regimen. Based on review of their data, however, there may be two other groups that we consider a neoadjuvant approach, in those with splenic vein involvement and those with adjacent organ involvement. Dr Aaron Sasson (Omaha, NE). Quick question. The ability to predict adjacent organ invasion can be difficult preoperatively. Did you find that splenic vein thrombosis was a surrogate marker for adjacent organ invasion, ie, how many patients did not have adjacent organ invasion when they had splenic vein thrombosis or vice versa? Dr Schmidt: That’s an excellent question. I will be happy to look at it and get back to. I don’t know the answer to that question. Dr Abdelkader Hawasli (St. Clair Shores, MI). On your patients who had negative margin intraoperatively, how many of these were positive and how did you handle these afterward? Dr Schmidt: I believe we had one that ended up being positive afterwards. It was negative on frozen section and the pathologist, the group of pathologists that reviewed it afterwards thought it was an error in the original reading. Dr Hawasli: What did you do with that patient? Dr Schmidt: We proceeded with adjuvant chemoradiation and counseled the patient about potentially reresection and then they declined. Dr Jeffrey Bender (Oklahoma City, OK). Less than 20% of your distals were for adenocarcinoma. What the heck were all the rest of these for? Dr. Schmidt: The largest two other groups would be IPMN and pancreatitis.

©2015 Elsevier

or adjacent organ invasion.

Patients with adenocarcinoma of the pancreatic body/tail and associated vascular thrombosis or adjacent organ invasion are suboptimal candidates for r...
242KB Sizes 0 Downloads 10 Views