Abstract
Background:
The association between a positive surgical margin and local recurrence following resection of pancreatic adenocarcinoma (PDAC) has been reported. Assessment of the location of the positive margin and the specific site of local recurrence has not been well described.
Methods:
A prospectively maintained database was queried for patients who underwent R0/R1 pancreaticoduodenectomy for PDAC between 2000-2015. The pancreatic, posterior, gastric/duodenal, anterior peritoneal, and bile duct margins were routinely assessed. Postoperative imaging was reviewed for site of first recurrence, and local recurrence was defined as located in the remnant pancreas, surgical bed, or retroperitoneal site outside of the surgical bed.
Results:
During the study period, 891 patients underwent pancreaticoduodenectomy, and 390 patients had an initial local recurrence with or without distant metastases. The 5-year cumulative incidence of local recurrence by site was: remnant pancreas, 4%(95%CI:3-5%); surgical bed, 35%(95%CI:32-39%); and other regional retroperitoneal site, 4%(95%CI:3-6%). A positive posterior margin (HR:1.50 [95%CI:1.17-1.91], p=0.001) and positive lymph nodes (HR:1.36 [95%CI:1.06-1.75], p=0.017) were associated with surgical bed recurrence on univariate analysis, and posterior positive margin remained significant on multivariate analysis (HR:1.40 [95%CI:1.09-1.81], p=0.009). There were 197 patients with an isolated local recurrence, and a positive posterior margin was associated with surgical bed recurrence in this subgroup (HR:1.51 [95%CI:1.08-2.10], p=0.016).
Conclusion:
In this study, the primary association between site of margin positivity and site of local recurrence was between the posterior margin and surgical bed recurrence. Given this association, and the limited ability to modify this margin intraoperatively, preoperative assessment should be emphasized.
Introduction
Pancreatic ductal adenocarcinoma (PDAC) is predicted to be the second leading cause of cancer-related mortality in the United States by the year 2030.1 Only 15-20% of patients with PDAC present with resectable disease, and the majority of these patients will develop a recurrence following resection. Most recurrences occur at distant sites, while approximately 40-45% of patients will initially recur locally in the region of operative resection.2-4 Within the group of patients who experience local recurrence, approximately half will have an isolated local recurrence, and the other half will fail at both local and distant sites.2-4
Pathologic variables including R1 resection,2,5-7 specifically a positive posterior/superior mesenteric artery (SMA) margin,2,8 and positive nodal disease2,8 have been associated with an increased risk of developing a local recurrence following resection for PDAC. Many reports focus on the association between a positive margin and survival, and a variety of approaches to assess the margin intraoperatively have been recommended. Whether extension of surgical margins in the face of a positive margin on frozen section, or the use of adjuvant radiation therapy, improve outcomes is controversial.9-21
While a positive surgical margin is a documented risk factor for local recurrence, an assessment of the specific location of the positive margin (transected pancreatic, posterior, bile duct, gastric/duodenal and anterior peritoneal) and the exact site of local recurrence (surgical bed, remnant pancreas or other regional retroperitoneal site) has not been well described. In this study, we sought to evaluate which surgical margins are associated with specific sites of local recurrence following pancreaticoduodenectomy (PD). Additionally, we assessed other clinical and pathologic variables for an association with the site of local recurrence.
Methods
A prospectively maintained database was queried for all patients who underwent PD between January 2000 and December 2015 and had a final pathologic diagnosis of PDAC. Approval for this study was obtained from the Institutional Review Board (IRB) at Memorial Sloan Kettering (MSK). Patients were included if they underwent R0/R1 resection, had adequate postoperative cross-sectional imaging, and had follow-up greater than six months if they remained without evidence of disease. Patients who had metastatic disease at diagnosis or underwent R2 resection were excluded. Demographic, clinical, and pathologic information was obtained from the database and the electronic medical record.
The following margins were routinely assessed and reported at the time of pathologic evaluation of the PD specimen: pancreatic transection margin, bile duct margin, gastric/duodenal margin, posterior margin and the anterior peritoneal surface. Specific sites of posterior margin were reported separately as previously defined: uncinate/SMA, vascular groove and posterior/deep surface.22,23 Consistent with the AJCC Cancer Staging Manual 8th Edition, tumor within <1mm of the surgical margin was considered positive.24 However, R1 resections were further subdivided based on the degree of involvement to R1 (direct) and R1 (within 1mm).
Cross-sectional postoperative imaging was reviewed for site of first recurrence. Postoperative imaging was obtained per the NCCN guidelines;25 computed tomography (CT) scans were routinely used to monitor for recurrence postoperatively, however, additional imaging was obtained at the discretion of the treating physician. Patients were classified into one of four groups based on site of first recurrence: distant recurrence only, isolated local recurrence, simultaneous local and distant recurrences, or no recurrence with a minimum follow-up of six months. Local recurrence was defined by the first cross-sectional imaging study where tumor growth was noted in the retroperitoneum. This was correlated with laboratory values, biopsies or changes in treatment, and confirmed by documentation of the attending medical or surgical oncologist. Local recurrences were divided into three sites: the remnant pancreas, surgical bed, or other regional retroperitoneal site outside of the surgical bed. Recurrence in the surgical bed was defined radiographically as tumor growth adjacent to the remnant pancreas, surgical clips or the superior mesenteric artery or celiac axis. Regional retroperitoneal sites included retroperitoneal lymph nodes (ie. paraaortic or retrocaval), and tumor growth adjacent to or invading other retroperitoneal structures. Patients who had a distant recurrence within 90 days of local recurrence were classified as having both local and distant recurrences as their first site of recurrence.
Statistical Analysis
Disease and treatment characteristics were summarized using frequency and percentage for categorical variables, and median and interquartile range (IQR) for continuous variables. Overall (OS) was calculated from date of surgery until date of death or date of last follow up and estimated using Kaplan-Meier methods. Among patients who developed a recurrence at the time of the study, OS was calculated from the time of recurrence to the date of death or date of last follow-up and compared between subgroups using the log-rank test.
Cumulative incidences of patterns of first recurrence (distant only, isolated local, concurrent local and distant) and death without evidence of recurrence, as well as time to specific site of local recurrence (surgical bed, remnant pancreas and other regional retroperitoneal site), were estimated using competing risks methods from the date of operation. Associations between pathologic variables and time to specific site of local with or without distant recurrence, as well as isolated local recurrence, were evaluated using Gray’s test and the Fine and Gray competing risks method. Additionally, given the terminology used to describe PD margins was standardized during our study period, we conducted a sensitivity analysis of patients resected between 2008 and 2015.
Variables associated with time to surgical bed recurrence on univariate analysis (p<0.10) were entered in a multivariable model. A multivariable model was not constructed for recurrences in the remnant pancreas or other retroperitoneal site, as the number of recurrences observed were small (n=38 and n=39, respectively). Likewise, sites of positive posterior margin were not included as the number of events in these groups was small.
All statistical analyses were performed using SAS Version 9.3 (SAS Institute, INC., Cary, NC, USA) or R Version 3.5.1 (R Foundation for Statistical Computing, Vienna, Austria) using ‘cmprsk’ package. All p-values were two-sided, and p-values <0.05 were considered statistically significant.
Results
Patient Cohort
During the study period, there were 891 patients who underwent R0/R1 PD and had greater than six months of imaging follow up. Demographic, clinical, and pathologic characteristics of the cohort are shown in Table 1. Overall median follow-up was 25 months, and the median follow-up among survivors was 55 months. At the time of analysis, initial isolated distant recurrence occurred in 358 patients, 197 patients had developed an isolated local recurrence, and simultaneous local and distant recurrences were seen in 193 patients. There were 143 patients who remained free of disease for at least six months following resection (Figure 1). Distant recurrence was the most common site of initial recurrence, with a cumulative incidence of 29.1% [95%CI: 26.1-32.1%] and 40.0% [95%CI: 36.7-43.3%] of patients having experienced distant relapse at one and five years, respectively (Figure 2a). Cumulative incidence of an isolated local recurrence was 8.4% [95%CI: 6.7-10.4%] at one year and increased to 22.3% [95%CI: 19.6-25.1%] at five years. The cumulative incidence of concurrent local and distant recurrences was 14.2% [95%CI: 12.0-16.5%] at one year and 21.5% [95%CI: 18.9-24.3%] at five years after surgery.
Table 1.
Clinical and pathologic characteristics of patients who underwent pancreaticoduodenectomy.
| Variable | Overall (n = 891) |
Local Failure (n=390) |
Local Recurrence Only (n=197) |
Local and Distant Recurrence (n=193) |
|---|---|---|---|---|
| Age, years | 68 (61-75) | 68 (61-75) | 70 (63-76) | 66 (60-73) |
| BMI | 26 (23-30) | 26 (23-29) | 26 (23-29) | 26 (23-30) |
| Gender | ||||
| Female | 438 (49) | 194 (50) | 96 (49) | 98 (51) |
| Male | 453 (51) | 196 (50) | 101 (51) | 95 (49) |
| Diabetes | ||||
| Yes | 673 (76) | 103 (26) | 53 (27) | 50 (26) |
| No | 218 (24) | 287 (74) | 144 (73) | 143 (74) |
| Smoking History | ||||
| Yes | 406 (46) | 219 (56) | 89 (45) | 82 (42) |
| No | 485 (54) | 171 (44) | 108 (55) | 111 (58) |
| Preoperative Biliary Drainage | ||||
| No | 343 (38) | 158 (41) | 79 (40) | 79 (41) |
| Yes | 548 (62) | 232 (59) | 118 (60) | 114 (59) |
| Neoadjuvant Treatment | ||||
| None | 768 (87) | 332 (85) | 161 (82) | 171 (89) |
| Neoadjuvant Chemotherapy | 75 (8) | 37 (10) | 25 (13) | 12 (6) |
| Neoadjuvant Chemotherapy + XRT | 48 (5) | 21 (5) | 11 (5) | 10 (5) |
| Vascular Resection | ||||
| No | 733 (82) | 304 (78) | 145 (74) | 159 (82) |
| Yes | 158 (18) | 86 (22) | 52 (26) | 34 (18) |
| Adjuvant Treatment* | ||||
| None | 201 (23) | 86 (22) | 36 (18) | 50 (26) |
| Adjuvant Chemotherapy | 435 (50) | 224 (58) | 124 (64) | 100 (52) |
| Adjuvant Chemotherapy + XRT | 241 (27) | 76 (20) | 35 (18) | 41 (22) |
| Tumor size, cm | 3.0 (2.2-3.5) | 3.0 (2.3-3.5) | 3.0 (2.3-3.5) | 3.0 (2.4-3.6) |
| Differentiation** | ||||
| Well | 19 (2) | 7 (2) | 6 (3) | 1 (0.5) |
| Moderate | 574 (65) | 258 (67) | 142 (73) | 116 (60.5) |
| Poor | 285 (33) | 120 (31) | 46 (24) | 74 (39) |
| Extrapancreatic Invasion | ||||
| Yes | 864 (97) | 379 (97) | 190 (96) | 189 (98) |
| No | 37 (3) | 11 (3) | 7 (4) | 4 (2) |
| Lymphovascular Invasionǂ | ||||
| Yes | 578 (66) | 260 (68) | 124 (64) | 136 (72) |
| No | 296 (34) | 123 (32) | 71 (36) | 52 (28) |
| Perineural Invasion# | ||||
| Yes | 814 (93) | 359 (94) | 179 (93) | 180 (95) |
| No | 66 (7) | 23 (6) | 13 (7) | 10 (5) |
| Lymph Node Status | ||||
| Negative | 262 (29) | 109 (28) | 64 (32) | 45 (23) |
| Positive | 629 (71) | 281 (72) | 133 (68) | 148 (77) |
| Margin Status | ||||
| R0 | 553 (62) | 232 (59) | 117 (59) | 115 (60) |
| R1 | 338 (38) | 158 (41) | 80 (41) | 78 (40) |
| R1 (direct) | 247 (28) | 108 (28) | 58 (29) | 50 (26) |
| R1 (within 1mm) | 91 (10) | 50 (13) | 22 (11) | 28 (14) |
| Pancreatic Margin | ||||
| Positive | 136 (15) | 60 (15) | 26 (13) | 34 (18) |
| Negative | 755 (85) | 330 (85) | 171 (87) | 159 (82) |
| Posterior Margin | ||||
| Positive | 211 (24) | 106 (27) | 56 (28) | 50 (26) |
| Uncinate/SMA | 112 (13) | 61 (16) | 29 (15) | 32 (16) |
| Vascular Groove | 11 (1) | 5 (1) | 2 (1) | 3 (2) |
| Posterior/Deep Surface | 78 (9) | 35 (9) | 20 (10) | 15 (8) |
| Multiple | 10 (1) | 5 (1) | 5 (2) | 0 (0) |
| Negative | 680 (76) | 284 (73) | 141 (72) | 143 (74) |
| Gastric/duodenal Margin | ||||
| Positive | 4 (0.5) | 1 (0.5) | 0 (0) | 1 (1) |
| Negative | 887 (99.5) | 389 (99.5) | 197 (100) | 192 (99) |
| Bile Duct Margin | ||||
| Positive | 26 (3) | 13 (3) | 9 (5) | 4 (2) |
| Negative | 865 (97) | 377 (97) | 188 (95) | 189 (98) |
| Anterior Peritoneal Margin | ||||
| Positive | 54 (6) | 24 (6) | 12 (6) | 12 (6) |
| Negative | 836 (94) | 366 (94) | 185 (94) | 181 (94) |
Variables are expressed as n (%) or median (IQR)
missing in 13 patients overall and 4 patients with local failure
missing in 13 patients overall and 5 patients with local failure
missing in 17 patients overall and 7 patients with local failure
missing in 11 patients overall and 8 patients with local failure
Figure 1.
Site of recurrence in all patients (n=891) who underwent pancreaticoduodenectomy between 2000-2015.
Figure 2.
Cumulative incidence of recurrence (a) and overall survival (b) of all patients based on site of recurrence. Cumulative incidence of recurrence of specific site of local recurrence in patients with local failure (c).
The median overall survival of the whole cohort was 26.7 [95%CI: 24.7-29.2] months. Patients who developed concurrent local and distant recurrences experienced decreased OS from the time of recurrence when compared to patients who developed either isolated distant or locally recurrent disease (median [95%CI], 7.2 [6.4-8.3] months vs 12.0 [10.0-13.4] months vs 15.2 [13.1-17.3] months, p<0.0001, Figure 2b).
Local Recurrence
Clinical and pathologic data for patients who experienced local recurrence with or without distant metastases is shown in Table 1. Within this group, R1 resection occurred in 41% (n=158) of patients. A positive posterior margin was the most common site of margin positivity (27%, n=106) followed by a positive pancreatic margin in 15% (n=60). A positive uncinate/SMA margin was the most common site of positive posterior margin (n=66, 17%). In patients with local failure, 72% had positive lymph nodes, whereas 28% had lymph node negative disease. In patients who had an R1 resection, 79.1% (125/158) also had lymph node positive disease and in patients who had an R0 resection, 67.2% (156/232) had lymph node positive disease. Neoadjuvant therapy was administered to 15% (n=58) of patients who experienced local recurrence and 78% (n=300) received adjuvant therapy. In this group with local failure, 20% of individuals (n=76) received adjuvant chemotherapy and radiation. Of these, 62% (n=47) had an R1 resection, and 41% (n=31) had a positive posterior margin
The majority of local recurrences occurred in the surgical bed (n=313) while 38 patients recurred in the remnant pancreas, and 39 had a recurrence at a regional retroperitoneal site other than the surgical bed. A surgical bed recurrence was the most common, with a cumulative incidence of 19% [95%CI: 16-21%] at one year and 36% [95%CI: 32-39%] at five years. The cumulative incidences of recurrence in both the remnant pancreas and other retroperitoneal site was approximately 2% at one year and 4% at five years (Figure 2c). When comparing the specific site of local recurrence among those with and without simultaneous distant metastases, a similar number of patients developed recurrence in the remnant pancreas (n=18 vs n=20) and the surgical bed (n=150 vs n=163), yet a greater number of patients with simultaneous local and distant metastases had a recurrence at other retroperitoneal sites (n=25 vs n=14).
The cumulative incidence of recurrence was then evaluated for each site of local recurrence and specific positive surgical margins were compared. An R1 resection was significantly associated with time to local recurrence, and specifically time to recurrence in the surgical bed (HR: 1.30 [95%CI: 1.04-1.62], p=0.023). There was no significant difference between R1 (direct) and R1 (within 1mm) and the risk of surgical bed recurrence (HR: 1.39 [95%CI: 0.97-2.00], p=0.074).
More specifically, a positive posterior margin was associated with a risk of surgical bed recurrence (HR: 1.50 [95%CI: 1.17-1.91], p=0.001; Table 2a). The cumulative incidence of surgical bed recurrence at one and five years was 24.6% [95%CI: 19.0-30.6%] and 44.8% [95%CI: 37.9-51.5%] in patients with a positive posterior margin as compared to 16.9% [95%CI: 14.2-19.9%] and 33% [95%CI: 29.1-36.2%] in those with a negative margin (Figure 3a). Again, there was no difference between R1 (direct) and R1 (within 1mm) and the risk of surgical bed recurrence (HR: 0.95 [95%CI: 0.63-1.43], p=0.802). The posterior margin was subsequently subdivided into three specific margins: uncinate/SMA, vascular groove and posterior/deep surface. Our results suggested that patients with a positive uncinate/SMA margin or multiple positive posterior margins had an increased incidence of surgical bed recurrence (1-year CIR: uncinate: 28.6% [20.5-37.2%]; multiple: 30.0% [5.6-60.5%], p=0.007; Figure 4). No other positive margin was significantly associated with any specific site of local recurrence.
Table 2.
Univariate association between clinical and pathologic variables and time to local +/− distant recurrence (a) and time to isolated local recurrence (b).
| a | Surgical Bed Recurrence |
Remnant Pancreas Recurrence | Other Retroperitoneal Site Recurrence |
|||
|---|---|---|---|---|---|---|
| HR (95% CI) | p-value | HR (95% CI) | p-value | HR (95% CI) | p-value | |
| R1 Resection | 1.30 (1.04-1.62) | 0.023 | 0.58 (0.28-1.18) | 0.133 | 0.91 (0.47-1.75) | 0.785 |
| Positive Posterior Margin | 1.50 (1.17-1.91) | 0.001 | 0.60 (0.25-1.43) | 0.249 | 0.70 (0.31-1.59) | 0.394 |
| Positive Pancreatic Margin | 1.06 (0.78-1.43) | 0.707 | 1.25 (0.55-2.83) | 0.591 | 0.46 (0.14-1.48) | 0.192 |
| Positive Bile Duct Margin | 0.85 (0.42-1.72) | 0.654 | 1.89 (0.46-7.79) | 0.380 | 2.88 (0.89-9.32) | 0.077 |
| Positive Anterior Peritoneal Margin | 1.02 (0.65-1.59) | 0.913 | 0.41 (0.06-3.04) | 0.385 | 1.29 (0.40-4.18) | 0.669 |
| Positive Lymph Nodes | 1.36 (1.06-1.75) | 0.017 | 0.45 (0.24-0.85) | 0.014 | 0.83 (0.43-1.61) | 0.582 |
| Tumor Size | 1.09 (0.99-1.19) | 0.068 | 0.76 (0.58-1.00) | 0.050 | 0.90 (0.69-1.18) | 0.450 |
| Poor Tumor Differentiation | 0.96 (0.75-1.24) | 0.766 | 0.73 (0.36-1.51) | 0.399 | 1.53 (0.80-2.91) | 0.194 |
| Lymphovascular Invasion | 1.26 (1.00-1.60) | 0.054 | 0.67 (0.35-1.28) | 0.228 | 0.99 (0.51-1.93) | 0.982 |
| Perineural Invasion | 1.61 (0.97-2.65) | 0.063 | 0.67 (0.24-1.88) | 0.450 | 0.98 (0.30-3.16) | 0.974 |
| Preoperative Biliary Drainage | 0.94 (0.75-1.17) | 0.558 | 0.85 (0.45-1.61) | 0.613 | 0.90 (0.48-1.70) | 0.741 |
| Neoadjuvant Chemotherapy + Radiation* | 0.82 (0.43-1.57) | 0.554 | 1.19 (0.27-5.34) | 0.817 | 0.94 (0.22-3.95) | 0.931 |
| Vascular Resection | 1.40 (1.08-1.81) | 0.012 | 1.24 (0.57-2.70) | 0.589 | 1.20 (0.55-2.61) | 0.648 |
| Adjuvant Chemotherapy + Radiation* | 0.57 (0.43-0.75) | <.001 | 0.32 (0.11-0.92) | 0.034 | 0.51 (0.21-1.25) | 0.139 |
| b | Surgical Bed Recurrence |
Remnant Pancreas Recurrence | Other Retroperitoneal Site Recurrence |
|||
| HR (95% CI) | p-value | HR (95% CI) | p-value | HR (95% CI) | p-value | |
| R1 Resection | 1.24 (0.91-1.69) | 0.178 | 0.54 (0.20-1.48) | 0.231 | 1.23 (0.43-3.55) | 0.700 |
| Positive Posterior Margin | 1.51 (1.08-2.11) | 0.016 | 0.57 (0.17-1.93) | 0.362 | 0.88 (0.24-3.17) | 0.847 |
| Positive Pancreatic Margin | 0.81 (0.51-1.28) | 0.359 | 1.38 (0.46-4.11) | 0.562 | 0.43 (0.06-3.23) | 0.409 |
| Positive Bile Duct Margin | 1.11 (0.44-2.81) | 0.824 | 3.82 (0.90-16.2) | 0.070 | 5.65 (1.29-24.8) | 0.022 |
| Positive Anterior Peritoneal Margin | 1.23 (0.69-2.18) | 0.486 | NE | NE | NE | NE |
| Positive Lymph Nodes | 1.01 (0.72-1.41) | 0.946 | 0.41 (0.17-0.98) | 0.045 | 0.55 (0.19-1.59) | 0.272 |
| Tumor Size | 1.03 (0.89-1.21) | 0.670 | 0.83 (0.60-1.16) | 0.290 | 0.76 (0.40-1.42) | 0.390 |
| Poor Tumor Differentiation | 0.64 (0.44-0.92) | 0.017 | 0.52 (0.17-1.55) | 0.240 | 0.83 (0.26-2.64) | 0.751 |
| Lymphovascular Invasion | 1.03 (0.75-1.42) | 0.865 | 0.46 (0.19-1.13) | 0.091 | 0.68 (0.24-1.97) | 0.481 |
| Perineural Invasion | 1.24 (0.65-2.35) | 0.512 | 0.70 (0.16-3.00) | 0.629 | 1.06 (0.14-8.02) | 0.952 |
| Preoperative Biliary Drainage | 0.94 (0.69-1.28) | 0.695 | 0.93 (0.38-2.26) | 0.865 | 0.83 (0.29-2.40) | 0.738 |
| Neoadjuvant Chemotherapy + Radiation* | 0.62 (0.27-1.42) | 0.257 | 1.60 (0.23-11.39) | 0.638 | 0.38 (0.04-3.33) | 0.381 |
| Vascular Resection | 1.72 (1.22-2.42) | 0.002 | 1.16 (0.39-3.46) | 0.795 | 2.58 (0.87-7.66) | 0.088 |
| Adjuvant Chemotherapy + Radiation* | 0.53 (0.36-0.78) | <.001 | 0.16 (0.02-1.24) | 0.080 | 0.18 (0.02-1.40) | 0.101 |
comparison group is patients who received chemotherapy alone; NE: not able to estimate as no recurrences in the remnant pancreas or other retroperitoneal site were observed
Figure 3.
Cumulative incidence of site of local recurrence based on positive or negative posterior margin (a) and nodal status (b).
Figure 4.
Association of posterior margin positivity and surgical bed recurrence.
We evaluated other pathologic variables and their relationship to time to recurrence in the surgical bed. Lymph node positivity was associated with an increased risk of surgical bed recurrence on univariate analysis (HR: 1.36 [95%CI: 1.06-1.75], p=0.017, Figure 3b, Table 2a). Other variables including lymphovascular invasion (HR: 1.26 [95%CI: 1.00-1.60], p=0.054), perineural invasion (HR: 1.61 [95%CI: 0.97-2.65], p=0.063) and tumor size (HR: 1.09 [95%CI: 0.99-1.19], p=0.068) trended toward significance. In the multivariate model, a positive posterior margin remained significantly associated with time to surgical bed recurrence after adjusting for lymph node positivity, tumor size, lymphovascular invasion, and perineural invasion (HR: 1.40 [95%CI: 1.09-1.81], p=0.009, Table 3).
Table 3.
Univariate and multivariate analysis for pathologic variables and cumulative incidence of recurrence in the surgical bed in patients with local recurrence with or without distant metastases.
| Univariate Analysis | Multivariate Analysis | |||
|---|---|---|---|---|
| HR (95% CI) | p-value | HR (95% CI) | p-value | |
| Positive Posterior Margin | 1.50 (1.17-1.91) | 0.001 | 1.41 (1.09-1.81) | 0.009 |
| Positive Pancreatic Margin | 1.06 (0.78-1.43) | 0.707 | ||
| Positive Bile Duct Margin | 0.85 (0.42-1.72) | 0.654 | ||
| Positive Anterior Peritoneal Margin | 1.02 (0.65-1.59) | 0.913 | ||
| Positive Lymph Nodes | 1.36 (1.06-1.75) | 0.017 | 1.13 (0.85-1.49) | 0.400 |
| Tumor Size | 1.09 (0.99-1.19) | 0.068 | 1.06 (0.97-1.17) | 0.210 |
| Poor Tumor Differentiation | 0.96 (0.75-1.24) | 0.766 | ||
| Lymphovascular Invasion | 1.26 (1.00-1.60) | 0.054 | 1.08 (0.84-1.39) | 0.550 |
| Perineural Invasion | 1.61 (0.97-2.65) | 0.063 | 1.35 (0.81-2.24) | 0.250 |
We investigated clinical variables associated with time to local failure in our cohort but did not include them in the multivariate model as the decision for neoadjuvant or adjuvant therapy was based on radiologic or pathologic tumor characteristics, and therefore, could result in a selection bias. Vascular resection was associated with an increased risk of surgical bed recurrence (HR: 1.40 [95%CI: 1.08-1.81], p=0.012). Administration of adjuvant chemotherapy and radiation as compared to chemotherapy alone was associated with a decreased risk of surgical bed recurrence (HR: 0.57 [95%CI: 0.43-0.75], p<0.001), while the use of neoadjuvant chemotherapy and radiation was not (HR: 0.82 [95%CI: 0.43-1.57], p=0.554). We then looked at the effect of radiation on time to local recurrence in the setting of a positive posterior margin. The receipt of adjuvant chemotherapy and radiation in individuals with a positive posterior margin on final pathology was associated with a decreased risk of surgical bed recurrence as compared to patients with a positive posterior margin who did not receive adjuvant therapy or received chemotherapy alone (HR: 0.60 [95%CI: 0.39-0.92], p=0.019).
No clinical or pathologic variables, including margin status, were associated with recurrence at another regional retroperitoneal site. The presence of positive lymph nodes (HR: 0.45 [95%CI: 0.24-85], p=0.014) was inversely associated with the risk of remnant pancreas recurrence, and tumor size approached significance (HR: 0.76 [95%CI: 0.58-1.00], p=0.050). A positive pancreatic margin was not associated with a risk of pancreatic recurrence (HR: 1.25 [95%CI: 0.55-2.83], p=0.591).
Given that the definitions of PD margins were standardized during our study period, we conducted a sensitivity analysis of sites of positive margin and the site of local recurrence of patients resected between 2008 and 2015. The results were similar between the entire cohort and individuals resected after 2008 for factors associated with recurrence in the surgical bed (Supplemental Table 1).
Isolated Local Recurrence
Finally, we evaluated time to isolated local recurrence as the first site of local recurrence. There were 197 patients who developed an isolated local recurrence during follow up. In this group, there were 163 patients who experienced a surgical bed recurrence, 20 patients who recurred in the remnant pancreas, and 14 patients who recurred at another retroperitoneal site. Again, there was a significant association between positive posterior margin and the risk of surgical bed recurrence (HR: 1.51 [95%CI: 1.08-2.11], p=0.016, Table 2b) on univariate analysis. Lymph node positive disease was not associated with time to surgical bed recurrence, but poor tumor differentiation was inversely associated with this finding (HR: 0.64 [95%CI: 0.44-0.92], p=0.017). Vascular resection was significantly associated with an increased risk (HR: 1.72 [95%CI: 1.22-2.42], p=0.002), and the use of adjuvant radiation was associated with a decreased risk, of surgical bed recurrence (HR: 0.53 [95%CI: 0.36-0.78], p<0.001).
Discussion
Previous data have shown that multiple factors are associated with local recurrence following pancreatectomy. One of the most well-established risk factors is a positive surgical margin. Multiple studies have demonstrated that an R1 resection,2,5-7 and more specifically, a positive posterior/superior mesenteric margin,2,8 is associated with local recurrence. This association between margin status and outcome has resulted in multiple recommendations for intraoperative margin assessment, with evaluation of the pancreatic transection and bile duct margins being common. Most studies however, have not evaluated the associations between specific sites of margin positivity and specific sites of local recurrence.
In the current study, we investigated the relationship between specific margin status and detailed site of local recurrence. We demonstrated that a positive posterior margin was independently associated with an increased risk of recurrence in the surgical bed. In particular, the finding of residual tumor at the uncinate/SMA margin was associated with the greatest cumulative incidence of surgical bed recurrence at 1-year postoperatively. In all patients who failed locally, the presence of nodal disease was also associated with time to surgical bed recurrence, and other tumor factors including tumor size, lymphovascular invasion, and perineural invasion trended toward significance on univariate analysis. Neither a positive pancreatic transection margin, nor a positive bile duct margin, nor a positive anterior pancreatic margin were associated with time to any site of local recurrence. Furthermore, no site of margin positivity was significantly associated with recurrence in the remnant pancreas or at another retroperitoneal site.
The most common site of local recurrence was the surgical bed, and a positive posterior margin was significantly associated with time to surgical bed recurrence after adjusting for other potential confounders. Specifically, the uncinate/SMA margin was associated with the greatest incidence of recurrence at this site. This is important, as when the operation is performed correctly, this margin is difficult to extend given its adjacency to the SMA. Arterial resections during pancreatic resection are associated with signficiant morbidity.26-28 Some groups have proposed extensive artery-sparing resections, such as the “triangle operation” following neoadjuvant therapy29 or total meso-pancreatic excision,30,31 although these are not routinely performed and long-term oncologic outcomes not well studied. The pancreatic, bile duct, and gastric/duodenal margins are margins that could technically be extended intraoperatively, yet they were not associated with any site of local recurrence. Thus, if local recurrence is to be avoided, it is likely of greater benefit to preoperatively assess the plane between the uncinate process and superior mesenteric artery on cross-sectional imaging than it is to perform intraoperative frozen section on the pancreatic and bile duct margins.
Prior retrospective studies have demonstrated that neoadjuvant and adjuvant therapy, and specifically radiation, are associated with a decreased risk of local recurrence and improved OS following pancreatectomy, however, the use of radiation therapy remains controversial.16-18,20,21,32-36 We found that use of adjuvant radiation as compared to chemotherapy alone was associated with a decreased risk of surgical bed recurrence. Furthermore, looking at the subgroup of patients with a positive posterior margin, those who received radiation had a decreased incidence of surgical bed recurrence as compared to patients who did not receive adjuvant therapy or underwent chemotherapy alone. We did not see these associations with the use of neoadjuvant therapy, but this is perhaps limited by the small number of patients who received neoadjuvant therapy in our cohort. We did not include these variables in a multivariate model, as administration of chemotherapy and/or radiation are based on the results of radiographic or pathologic factors, and therefore, this could introduce a selection bias into the analysis.
A number of prior studies have failed to demonstrate an association between a positive surgical margin and local recurrence. Not all studies address specific surgical margins that are positive in the setting of an R1 resection, and this study along with others,2,8 demonstrate that this distinction is important. Additonally, the definition of R1 resection varies between studies, and a positive margin has been defined by the presence of tumor cells located between 0-2mm from the surgical margin.5,37-39 In this study, we evaluated R1 resections using the AJCC 8th Edition definition of tumor >1mm away from the resection edge,24 as well as evaluated R1 (direct) and R1 (within 1mm) resections. We noted that overall R1 resection, as well as R1 at the posterior margin were associated with surgical bed recurrence but did not demonstrate any significant difference between R1 (direct) and R1 (within 1mm) in our cohort.
We initially evaluated time to local recurrence by including all patients who failed locally, whether or not distant metastases were present, as previous data has suggested that margin status is not associated with recurrence at distant sites.24 We subsequently assessed the group of patients with an isolated local recurrence, which similarly demonstrated that a positive posterior margin was associated with time to surgical bed recurrence. Positive nodal disease was not associated with recurrence in this group, yet poor tumor differentiation was associated with a decreased risk of local recurrence. This supports the idea of underlying differences in tumor biology between patients who develop isolated local recurrences and those who develop widespread disease, such that patients with poorly differentiated tumors tend to have early systemic dissemination of cancer cells leading to distant recurrence.
There are several limitations to this study. Given its retrospective nature, we were unable to account for change in reporting of pathologic margins over time. Specific posterior margins are now reported separately and additionally, the definition of R1 resection has changed over time. The definition of R0/R1 resection has been shown to impact the results of prior studies,5,37,38 and this could alter associations between margin status and site of local recurrence. Given these modifications occurred within our study period, we attempted to control for these variables by doing multiple subgroup analyses, yet this remains a limitation in the setting of a retrospective study. Additionally, we were not able to control for the specific chemotherapy and/or radiation regimens used during the study period. Given the heterogenity of this data, while there is an association of decreased risk of surgical bed recurrence with use of radiation, further prospective studies would need to confirm this. The majority of patients with a positive posterior margin also had nodal involvement, and the co-occurrence of other pathologic variables and positive margin has been previously demonstrated.10,40 Lastly, some of the subgroups are small in size, and this should be considered when interpreting results.
In conclusion, the primary association between positive surgical margin and site of local recurrence in the current study was between a positive posterior margin and surgical bed recurrence. Specifically, a positive uncinate/SMA margin was associated with surgical bed recurrence. Lymph node positive disease was associated with time to surgical bed recurrence on univariate analysis, and other tumor factors trended toward significance. Given that the posterior margin is difficult to modify intraoperatively, attention should be paid to the relationship between tumor edge and SMA on preoperative imaging. No other margin site was associated with local recurrence, and therefore, intraoperative assessment of these margins is of questionable value.
Supplementary Material
Synopsis.
In this study of patients who underwent pancreaticoduodenectomy for pancreatic adenocarcinoma, a positive posterior margin was associated with local recurrence, and specifically a recurrence in the surgical bed.
Acknowledgments
Disclosures: Murray F. Brennan, MD - NIH/NCI Cancer Center Support Grant P30 CA008748
Footnotes
Publisher's Disclaimer: This Author Accepted Manuscript is a PDF file of an unedited peer-reviewed manuscript that has been accepted for publication but has not been copyedited or corrected. The official version of record that is published in the journal is kept up to date and so may therefore differ from this version.
References
- 1.Rahib L, Smith BD, Aizenberg R, Rosenzweig AB, Fleshman JM, Matrisian LM. Projecting cancer incidence and deaths to 2030: the unexpected burden of thyroid, liver, and pancreas cancers in the United States. Cancer Res. 2014;74(11):2913–2921. [DOI] [PubMed] [Google Scholar]
- 2.Gnerlich JL, Luka SR, Deshpande AD, et al. Microscopic margins and patterns of treatment failure in resected pancreatic adenocarcinoma. Arch Surg. 2012;147(8):753–760. [DOI] [PubMed] [Google Scholar]
- 3.Groot VP, Rezaee N, Wu W, et al. Patterns, Timing, and Predictors of Recurrence Following Pancreatectomy for Pancreatic Ductal Adenocarcinoma. Ann Surg. 2018;267(5):936–945. [DOI] [PubMed] [Google Scholar]
- 4.Van den Broeck A, Sergeant G, Ectors N, Van Steenbergen W, Aerts R, Topal B. Patterns of recurrence after curative resection of pancreatic ductal adenocarcinoma. Eur J Surg Oncol. 2009;35(6):600–604. [DOI] [PubMed] [Google Scholar]
- 5.Sugiura T, Uesaka K, Mihara K, et al. Margin status, recurrence pattern, and prognosis after resection of pancreatic cancer. Surgery. 2013;154(5):1078–1086. [DOI] [PubMed] [Google Scholar]
- 6.You Y, Choi DW, Heo JS, et al. Clinical significance of revised microscopic positive resection margin status in ductal adenocarcinoma of pancreatic head. Ann Surg Treat Res. 2019;96(1):19–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Walston S, Salloum J, Grieco C, et al. Identifying Clinical Factors Which Predict for Early Failure Patterns Following Resection for Pancreatic Adenocarcinoma in Patients Who Received Adjuvant Chemotherapy Without Chemoradiation. Am J Clin Oncol. 2018;41(12):1185–1192. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Ghaneh P, Kleeff J, Halloran CM, et al. The Impact of Positive Resection Margins on Survival and Recurrence Following Resection and Adjuvant Chemotherapy for Pancreatic Ductal Adenocarcinoma. Ann Surg. 2019;269(3):520–529. [DOI] [PubMed] [Google Scholar]
- 9.Dillhoff M, Yates R, Wall K, et al. Intraoperative assessment of pancreatic neck margin at the time of pancreaticoduodenectomy increases likelihood of margin-negative resection in patients with pancreatic cancer. J Gastrointest Surg. 2009;13(5):825–830. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Hernandez J, Mullinax J, Clark W, et al. Survival after pancreaticoduodenectomy is not improved by extending resections to achieve negative margins. Ann Surg. 2009;250(1):76–80. [DOI] [PubMed] [Google Scholar]
- 11.Kooby DA, Lad NL, Squires MH 3rd, et al. Value of intraoperative neck margin analysis during Whipple for pancreatic adenocarcinoma: a multicenter analysis of 1399 patients. Ann Surg. 2014;260(3):494–501; discussion 501-493. [DOI] [PubMed] [Google Scholar]
- 12.Lad NL, Squires MH, Maithel SK, et al. Is it time to stop checking frozen section neck margins during pancreaticoduodenectomy? Ann Surg Oncol. 2013;20(11):3626–3633. [DOI] [PubMed] [Google Scholar]
- 13.Pang TC, Wilson O, Argueta MA, et al. Frozen section of the pancreatic neck margin in pancreatoduodenectomy for pancreatic adenocarcinoma is of limited utility. Pathology. 2014;46(3):188–192. [DOI] [PubMed] [Google Scholar]
- 14.Fatima J, Schnelldorfer T, Barton J, et al. Pancreatoduodenectomy for ductal adenocarcinoma: implications of positive margin on survival. Arch Surg. 2010;145(2):167–172. [DOI] [PubMed] [Google Scholar]
- 15.Nitschke P, Volk A, Welsch T, et al. Impact of Intraoperative Re-resection to Achieve R0 Status on Survival in Patients With Pancreatic Cancer: A Single-center Experience With 483 Patients. Ann Surg. 2017;265(6):1219–1225. [DOI] [PubMed] [Google Scholar]
- 16.Barbier L, Turrini O, Gregoire E, Viret F, Le Treut YP, Delpero JR. Pancreatic head resectable adenocarcinoma: preoperative chemoradiation improves local control but does not affect survival. HPB (Oxford). 2011;13(1):64–69. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Greer SE, Pipas JM, Sutton JE, et al. Effect of neoadjuvant therapy on local recurrence after resection of pancreatic adenocarcinoma. J Am Coll Surg. 2008;206(3):451–457. [DOI] [PubMed] [Google Scholar]
- 18.Neoptolemos JP, Dunn JA, Stocken DD, et al. Adjuvant chemoradiotherapy and chemotherapy in resectable pancreatic cancer: a randomised controlled trial. Lancet. 2001;358(9293):1576–1585. [DOI] [PubMed] [Google Scholar]
- 19.Neoptolemos JP, Stocken DD, Friess H, et al. A randomized trial of chemoradiotherapy and chemotherapy after resection of pancreatic cancer. N Engl J Med. 2004;350(12):1200–1210. [DOI] [PubMed] [Google Scholar]
- 20.Papavasiliou P, Hoffman JP, Cohen SJ, Meyer JE, Watson JC, Chun YS. Impact of preoperative therapy on patterns of recurrence in pancreatic cancer. HPB (Oxford). 2014;16(1):34–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Schorn S, Demir IE, Samm N, et al. Meta-analysis of the impact of neoadjuvant therapy on patterns of recurrence in pancreatic ductal adenocarcinoma. BJS Open. 2018;2(2):52–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Adsay NV, Basturk O, Saka B, et al. Whipple made simple for surgical pathologists: orientation, dissection, and sampling of pancreaticoduodenectomy specimens for a more practical and accurate evaluation of pancreatic, distal common bile duct, and ampullary tumors. Am J Surg Pathol. 2014;38(4):480–493. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Esposito I, Kleeff J, Bergmann F, et al. Most pancreatic cancer resections are R1 resections. Ann Surg Oncol. 2008;15(6):1651–1660. [DOI] [PubMed] [Google Scholar]
- 24.Amin MB, Edge S, Greene F, Byrd DR, Brookland RK, Washington MK, Gershenwald JE, Compton CC, Hess KR, Sullivan DC, Jessup JM, Brierley JD, Gaspar LE, Schilsky RL, Balch CM, ed AJCC Cancer Staging Manual (8th edition). Springer International Publishing: American Joint Commission on Cancer; 2017. [Google Scholar]
- 25.National Comprehensive Cancer Network. Pancreatic Adenocarcinoma (Version 3.2019). https://wwwnccnorg/professionals/physician_gls/pdf/pancreaticpdf. [DOI] [PubMed]
- 26.Bockhorn M, Burdelski C, Bogoevski D, Sgourakis G, Yekebas EF, Izbicki JR. Arterial en bloc resection for pancreatic carcinoma. Br J Surg. 2011;98(1):86–92. [DOI] [PubMed] [Google Scholar]
- 27.Mollberg N, Rahbari NN, Koch M, et al. Arterial resection during pancreatectomy for pancreatic cancer: a systematic review and meta-analysis. Ann Surg. 2011;254(6):882–893. [DOI] [PubMed] [Google Scholar]
- 28.Zettervall SL, Ju T, Holzmacher JL, et al. Arterial, but Not Venous, Reconstruction Increases 30-Day Morbidity and Mortality in Pancreaticoduodenectomy. J Gastrointest Surg. 2019. [DOI] [PubMed] [Google Scholar]
- 29.Hackert T, Strobel O, Michalski CW, et al. The TRIANGLE operation - radical surgery after neoadjuvant treatment for advanced pancreatic cancer: a single arm observational study. HPB (Oxford). 2017;19(11):1001–1007. [DOI] [PubMed] [Google Scholar]
- 30.Adham M, Singhirunnusorn J. Surgical technique and results of total mesopancreas excision (TMpE) in pancreatic tumors. Eur J Surg Oncol. 2012;38(4):340–345. [DOI] [PubMed] [Google Scholar]
- 31.Kawabata Y, Tanaka T, Nishi T, Monma H, Yano S, Tajima Y. Appraisal of a total meso-pancreatoduodenum excision with pancreaticoduodenectomy for pancreatic head carcinoma. Eur J Surg Oncol. 2012;38(7):574–579. [DOI] [PubMed] [Google Scholar]
- 32.Hsu CC, Herman JM, Corsini MM, et al. Adjuvant chemoradiation for pancreatic adenocarcinoma: the Johns Hopkins Hospital-Mayo Clinic collaborative study. Ann Surg Oncol. 2010;17(4):981–990. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Merrell KW, Haddock MG, Quevedo JF, et al. Predictors of Locoregional Failure and Impact on Overall Survival in Patients With Resected Exocrine Pancreatic Cancer. Int J Radiat Oncol Biol Phys. 2016;94(3):561–570. [DOI] [PubMed] [Google Scholar]
- 34.Rutter CE, Park HS, Corso CD, et al. Addition of radiotherapy to adjuvant chemotherapy is associated with improved overall survival in resected pancreatic adenocarcinoma: An analysis of the National Cancer Data Base. Cancer. 2015;121(23):4141–4149. [DOI] [PubMed] [Google Scholar]
- 35.Morganti AG, Falconi M, van Stiphout RG, et al. Multi-institutional pooled analysis on adjuvant chemoradiation in pancreatic cancer. Int J Radiat Oncol Biol Phys. 2014;90(4):911–917. [DOI] [PubMed] [Google Scholar]
- 36.Parikh AA, Maiga A, Bentrem D, et al. Adjuvant Therapy in Pancreas Cancer: Does It Influence Patterns of Recurrence? J Am Coll Surg. 2016;222(4):448–456. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Osipov A, Nissen N, Rutgers J, et al. Redefining the Positive Margin in Pancreatic Cancer: Impact on Patterns of Failure, Long-Term Survival and Adjuvant Therapy. Ann Surg Oncol. 2017;24(12):3674–3682. [DOI] [PubMed] [Google Scholar]
- 38.Strobel O, Hank T, Hinz U, et al. Pancreatic Cancer Surgery: The New R-status Counts. Ann Surg. 2017;265(3):565–573. [DOI] [PubMed] [Google Scholar]
- 39.van Roessel S, Kasumova GG, Tabatabaie O, et al. Pathological Margin Clearance and Survival After Pancreaticoduodenectomy in a US and European Pancreatic Center. Ann Surg Oncol. 2018;25(6):1760–1767. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Raut CP, Tseng JF, Sun CC, et al. Impact of resection status on pattern of failure and survival after pancreaticoduodenectomy for pancreatic adenocarcinoma. Ann Surg. 2007;246(1):52–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.