Abstract
BACKGROUND:
Level 1 data demonstrate that adjuvant chemotherapy (ACT) improves survival after surgical resection of pancreatic ductal adenocarcinoma (PDAC), (adjuvant gemcitabine, CONKO-001 study; adjuvant 5-FU, ESPAC3 study). The role of adjuvant chemoradiation therapy (ACRT) remains controversial. What is less clear is whether adjuvant therapy influences patterns of recurrence. The purpose of this study was to perform the first multicenter study analyzing patterns of recurrence after adjuvant therapy for PDAC.
STUDY DESIGN:
Patients undergoing resection for PDAC from 8 medical centers over a 10-year period were analyzed. Demographics, tumor characteristics, operative treatment, type of adjuvant therapy, recurrence pattern, and survival were reviewed. Using Cox-proportional hazards multivariate (MV) regression, the impact of ACT and ACRT on overall survival (OS), local recurrence (LR), and distant recurrence (DR) was investigated.
RESULTS:
There were 1,130 patients who were divided into those having surgery alone (n = 392), ACT (n = 291), or ACRT (n = 447). Median follow-up was 18 months. Compared with patients undergoing surgery alone, ACT, but not ACRT, demonstrated a significant OS advantage on MV analysis. Patients receiving ACT had significantly fewer recurrences (LR and DR); those receiving ACRT had significantly less LR but not DR. On subset MV analysis, ACT and ACRT resulted in less LR in patients with lymphnode (LN) positive and margin negative disease. No improvements in LR, DR, or OS were seen in margin positive patients with either ACT or ACRT.
CONCLUSIONS:
This is the first analysis demonstrating differences in recurrence patterns in PDAC patients based on type of adjuvant therapy. Adjuvant chemotherapy provided an OS advantage likely related to its effect on reducing both LR and DR. Adjuvant chemoradiation therapy appears to decrease LR, but not DR, and therefore has less impact on OS. Future investigations and treatment protocols should consider additional ACT rather than ACRT in the treatment of PDAC.
With an estimated 48,960 new cases in 2015, pancreatic adenocarcinoma (PDAC) represents 3% of all cancers in the United States. Currently, PDAC represents 7% of all cancer deaths, making it the fourth most common cause of cancer death in the US.1 Based on demographic shifts in the US population, and with the continued increase in the incidence of PDAC, cancer-related deaths due to pancreas cancer are projected to increase dramatically to become the second leading cause of cancer-related deaths before 2030.2 Although surgical resection offers a potential cure, most patients present with locally advanced or metastatic disease, and only 10% to 15% of patients are candidates for potentially curative resection. Even in patients who have had successful resection, recurrence is common and occurs in up to 50% to 90% of patients.3,4 These recurrences can occur loco-regionally around the resection bed and adjacent lymphatic tissue and at distant sites including in the liver, peritoneum, lungs, and extra-regional lymph nodes. Unlike more indolent cancers, most patients who have recurrence will ultimately die of their disease.4,5 The rationale for adjuvant therapy is based on the high incidence of tumor recurrence both locally and at distant sites, presumably because of the presence of micrometastatic disease after surgical resection.
Adjuvant strategies include the use of systemic chemotherapy alone or in conjunction with chemoradiation therapy. The role of adjuvant chemotherapy (ACT) has been well established by several large randomized trials including the Charité Onkologie (CONKO-001),6 the Japanese Study Group of Adjuvant Therapy for Pancreas Cancer (JSAP-2),7 and the European Study Group for Pancreatic Cancer (ESPAC) 1 and 3 trials,8 which all showed a significant survival advantage with ACT for resected PDAC. However, no randomized trial has been able to convincingly support the role of adjuvant chemoradiation therapy (ACRT) for overall survival (OS), and its benefit remains controversial.6,9–11 Large single-institution studies and other retrospective studies, including our earlier report from the Central Pancreatic Consortium,12–14 have suggested a possible benefit with the use of ACRT in specific high-risk patients, such as those with lymph node (LN) positive or margin positive disease.
The rationale for the use of ACT is its ability to treat occult metastatic disease, thereby preventing distant recurrence.15–19 Meanwhile, ACRT is directed toward the loco-regional area and therefore would be more efficacious in treating loco-regional recurrences rather than distant disease.18,19 Because most patients ultimately die of distant recurrence, this may explain some of the differences in survival benefit between the 2 adjuvant strategies. To date, however, the influence of adjuvant therapies on the patterns of recurrence after surgical resection for PDAC has been poorly studied.
The primary aim of this study was to understand the association of ACT and ACRT with patterns of recurrence and to determine factors influencing loco-regional and distant recurrences and their impact on OS after resection for PDAC. We sought to determine if ACT and ACRT lead to differences in recurrence patterns and thereby influence OS.
METHODS
Data source and patient acquisition
This study is an institutional review board-approved, multi-institutional retrospective review of prospectively maintained databases from 8 academic medical centers comprising the Central Pancreatic Consortium. The goal of the Central Pancreatic Consortium is to study important biologic and clinical questions in regard to pancreatic neoplasms and pancreatic surgery. As such, all institutions are high volume centers with expertise in the multi-disciplinary management of pancreatic cancer.
The study cohort consisted of all patients diagnosed with PDAC who underwent successful surgical resection between January 2000 and December 2010. Patients were excluded from the analysis if they were found to have M1 disease or R2 (grossly positive margin) resections at the time of surgery, if they had a pathologic diagnosis other than PDAC, if they died within 90 days of resection, if they received neoadjuvant therapy, or if their adjuvant therapy status or pattern of recurrence was unknown.
Demographic, clinical characteristics, and definitions
Clinical variables collected included demographic data including age at resection; sex; tumor characteristics including size, grade, stage, lymph nodes removed, margin status; and lymph node status and treatment parameters including the use of preoperative biliary stenting, type of operation, need for blood transfusion, vein resection, and the use of ACT alone, ACRT, or surgery alone. The use of adjuvant therapy was defined as initiation of treatment within 90 days of resection and before any signs of recurrence. The majority of patients who received ACRT also received ACT as part of their regimen. Because the Central Pancreatic Consortium consists of major tertiary referral centers, many of these patients received adjuvant therapy at outside institutions, so details about specific treatment regimens were not available. However, the majority of patients receiving ACT received 6 months of adjuvant gemcitabine chemotherapy, and the majority of patients receiving ACRT received gemcitabine chemotherapy before and after infusion fluorouracil and radiation therapy (50.4 Gy).
Outcomes variables
Data were collected regarding OS, as well as local recurrence (LR) and distant recurrence (DR) in all 3 cohorts: patients undergoing surgery alone, those receiving ACT, and those receiving ACRT. The time to recurrence and/or death was defined from the time of resection. Local recurrence was defined by radiographic and or pathologic evidence for recurrent disease in the operative and regional bed; distant recurrence included any recurrence outside of the local-regional area including hepatic, peritoneal, lung, and extra-regional lymph nodes. Patients with both LR and DR were included in both groups for purposes of analysis. Documentation of recurrence was determined by radiographic evidence, pathologic confirmation, and/or tumor marker elevation at the discretion of the treating physician.
Statistical analysis
Univariate parametric and nonparametric analyses of demographic and clinical characteristics were performed using analysis of variance (ANOVA), chi-square, Kruskal-Wallis H test, and Wilcoxon-Rank sum tests, respectively. Summary data are reported as mean ± standard deviation (SD) for normally distributed data and median and interquartile range (IQR) for non-normally distributed data. Kaplan-Meier analysis was used to determine OS, local recurrence-free survival (LRFS), and distant recurrence-free survival (DRFS), and to perform univariate analysis of the impact of surgery, ACT only, and ACRT on these survival parameters. Log-rank test was used to calculate statistical significance of the comparison. For multivariable analysis, separate Cox proportional hazards regression models were constructed to identify a priori chosen independent demographic, tumor-related, and treatment factors associated with OS, LRFS, and DRFS. Hazard ratios and 95% confidence intervals were calculated to evaluate the strength of association between each variable and survival. Statistical significance was defined as a p value < 0.05. Analyses were performed using Stata 13.1 statistical software (Stata Corporation).
RESULTS
Cohort
A total of 1,130 patients who underwent curative resection for PDAC between January 2000 and December 2010 and had complete data available were included in the cohort. A total of 392 patients (35%) underwent surgery alone, 291 patients (26%) underwent surgery followed by ACT, and 447 patients (39%) underwent surgical resection followed by ACRT (in addition to chemotherapy). Median follow-up was 18 months.
Demographics
Demographic data for these patient cohorts are shown in Table 1, including the number of patients in each treatment arm by institution. The median age for the entire cohort was 65 years (IQR 57 to 73 years); patients receiving ACRT were younger than those in the other groups (p < 0.001). There was a higher percentage of men in the surgery alone group as compared with the ACT and ACRT groups (57% vs 46% and 50%, respectively, p = 0.025). Treatment patterns among the 8 institutions also varied significantly (p < 0.001). A higher percentage of patients undergoing ACT and ACRT underwent preoperative biliary stenting when compared with the surgery alone group (59% and 60% vs 43%, p < 0.001). The type of surgical resection also differed among the groups (p = 0.019).
Table 1.
Demographic Data
| Variable | Surgery alone (n = 392) | Adjuvant chemotherapy only (n = 291) | Adjuvant chemoradiation (n = 447) | p Value |
|---|---|---|---|---|
| Age at operation, y, median (IQR) | 68 (58, 75) | 67 (58, 74) | 63 (55, 70) | <0.001 |
| Male sex, n (%) | 223 (57) | 135 (46) | 225 (50) | 0.025 |
| Institution, n (%) | <0.001 | |||
| 1 | 9 (16) | 32 (24) | 45 (33) | |
| 2 | 55 (29) | 11 (20) | 35 (64) | |
| 3 | 72 (74) | 50 (27) | 82 (44) | |
| 4 | 110 (32) | 9 (9) | 16 (17) | |
| 5 | 8 (15) | 86 (25) | 146 (43) | |
| 6 | 46 (33) | 18 (35) | 26 (50) | |
| 7 | 58 (43) | 41 (30) | 51 (37) | |
| 8 | 34 (27) | 44 (36) | 46 (37) | |
| Preoperative biliary stent, n (%) | 161 (43) | 167 (59) | 263 (60) | <0.001 |
| Operation type, n (%) | 0.019 | |||
| Standard PD | 265 (67) | 172 (59) | 309 (69) | |
| PPPD | 89 (23) | 86 (30) | 109 (24) | |
| LP | 31 (8) | 30 (10) | 28 (6) | |
| TP | 7 (2) | 3 (1) | 1 (0.2) |
IQR, interquartile range; LP, left (distal) pancreatectomy; PPPD, pylorus preserving pancreaticoduodenectomy; TP, total pancreatectomy.
Oncologic and treatment characteristics
Oncologic and treatment characteristics are shown in Table 2. Mean tumor size was smaller in those undergoing surgery alone (2.9 ± 1.4 cm vs 3.3 ± 2.1 cm and 3.1 ± 1.3 cm, p = 0.01). The majority of tumors were grade 2 (moderately differentiated) and did not differ among the 3 groups. Although the majority of patients in all 3 groups had stage II disease, those receiving ACT had fewer stage I tumors (p = 0.03). In addition, patients receiving ACT had the highest percentage of LN positive disease (70%) as compared with the ACRT group (58%) and surgery alone group (42%) (p < 0.001 overall; p < 0.001 for surgery alone vs ACT and ACRT; p = 0.001 for ACT vs ACRT).
Table 2.
Oncologic and Treatment Data
| Variable | Surgery alone (n = 392) | Adjuvant chemotherapy only (n = 291) | Adjuvant chemoradiation (n = 447) | p Value |
|---|---|---|---|---|
| Tumor size, cm, mean ± SD | 2.9 ± 1.4 | 3.3 ± 2.1 | 3.1 ± 1.3 | 0.01 |
| Tumor grade, n (%)* | 0.61 | |||
| 1 | 38 (11) | 38 (13) | 41 (9) | |
| 2 | 219 (63) | 175 (61) | 276 (63) | |
| 3 | 90 (26) | 76 (26) | 120 (28) | |
| Tumor stage, n (%) | 0.03 | |||
| I | 57 (15) | 21 (7) | 63 (14) | |
| II | 333 (85) | 268 (92) | 380 (85) | |
| III | 2 (0.5) | 2 (1) | 4 (1) | |
| Lymph nodes, positive, n (%) | 166 (42) | 203 (70) | 261 (58) | <0.001† |
| Vein resection, yes, n (%) | 44 (11) | 50 (17) | 70 (16) | 0.062 |
| Margin positive, yes, n (%) | 64 (16) | 61 (21) | 124 (28) | <0.001‡ |
| Lymph nodes resected, median (IQR) | 14 (8, 21) | 18 (11, 27) | 15 (9, 22) | <0.001 |
| Blood transfusion, yes, n (%) | 114 (29) | 75 (26) | 126 (28) | 0.623 |
| Length of stay, d, median (IQR) | 11 (8, 18) | 9 (7, 13) | 9 (7,12) | <0.001 |
Grade 1, well differentiated; 2, moderately differentiated; 3, poorly differentiated.
p < 0.001, surgery alone vs adjuvant chemotherapy and adjuvant chemoradiation; p = 0.001, adjuvant chemotherapy vs adjuvant chemoradiation.
p = 0.12, surgery alone vs adjuvant chemotherapy; p < 0.001, surgery alone vs adjuvant chemoradiation; p = 0.03 adjuvant chemotherapy vs adjuvant chemoradiation. IQR, interquartile range.
Patients undergoing venous resection did not differ between groups. The R1 (microscopically positive) resection margin was significantly higher in the ACRT group vs both surgery alone (28% vs 16%, p < 0.001) and ACT (28% vs 21%, p = 0.03) groups. But margin positive disease did not significantly differ between the surgery alone and ACT groups (p = 0.12). The median number of LNs resected for all patients was 15 (IQR 9 to 23), ranging from 14 (IQR 8 to 21) in the surgery alone group to 18 (IQR 11 to 27) in the ACT group (p < 0.001). Although complete data on blood loss was not available, blood transfusion rates were not different among the 3 groups. Complete complication data were not available; however, the median hospital length of stay was significantly higher in the surgery alone group (11 days, IQR 8 to 18 days) vs the ACT group (9 days, IQR 7 to 13 days) and the ACRT group (9 days IQR 7 to 12days), p<0.001.
Recurrence and survival
Median OS for the entire cohort was 25.9 months. Unadjusted median OS for the surgery alone group (25.1 months), the ACT group (26.5 months) and the ACRT group (26.8 months) were not significantly different (p = 0.84). The overall LR rate for the entire group was 22% and did not differ significantly among the groups (20% for surgery alone, 21% for ACT, and 23% for ACRT, p = 0.49). For those who suffered local recurrence, unadjusted median LRFS, however, was significantly higher in the ACRT group (24.8 months) as compared with the ACT group (21.2 months) and the surgery alone group (20.0 months). The overall DR rate for the entire group was 41% and was highest in the ACRT group (51%) when compared with the surgery alone group (30%) and ACT group (41%), p < 0.001. For those who had distant recurrence, however, unadjusted median DRFS did not significantly differ among the groups (11.7 months for surgery alone, 13 months for ACT, and 14.2 months for ACT, p = 0.44).
Multivariable analysis of overall survival
In order to adjust for competing risk factors, we performed multivariable Cox-proportional hazards models to assess the effect of ACT and ACRT on OS (Table 3). After adjusting for age, tumor size, grade, pre-operative biliary stent, venous resection, margin status, LN status, blood transfusion, and LOS, the use of ACT was independently associated with a significant OS advantage compared with surgery alone (hazard ratio [HR] 0.71, 95% CI 0.57 to 0.89). The ACRT, however, did not result in an OS advantage vs surgery alone (HR 0.84, 95% CI 0.69 to 1.02) or vs ACT alone. Preoperative biliary stenting was also associated with a survival advantage; increasing age, venous resection, blood transfusion, R1 margin, positive LNs, increasing tumor grade, and increasing length of stay were associated with decreased OS (Table 3).
Table 3.
Cox Proportional Multivariate Hazards Regression for Overall Survival
| Variable | Hazard ratio | 95% CI | p Value |
|---|---|---|---|
| Age | 1.01 | 1.00–1.02 | 0.041 |
| Sex (ref, female) | 1.06 | 0.90–1.24 | 0.453 |
| Tumor size | 1.00 | 0.99–1.01 | 0.552 |
| Tumor grade | |||
| 1 (reference) | |||
| 2 | 1.42 | 1.06–1.91 | 0.019 |
| 3 | 1.72 | 1.25–2.36 | 0.001 |
| Preoperative stent | 0.70 | 0.60–0.83 | 0.001 |
| Venous resection | 1.46 | 1.18–1.82 | <0.001 |
| Blood transfusion | 1.27 | 1.06–1.53 | 0.01 |
| Hospital length of stay | 1.01 | 1.00–1.03 | 0.028 |
| Lymph node status | |||
| Negative (reference) | |||
| Positive | 1.60 | 1.35–1.90 | <0.001 |
| No. of lymph nodes resected | 1.00 | 0.99–1.01 | 0.44 |
| Margin status | |||
| Negative (reference) | |||
| Positive | 1.56 | 1.29–1.88 | <0.001 |
| Adjuvant therapy | |||
| None (reference) | |||
| Chemotherapy only | 0.71 | 0.57–0.89 | 0.004 |
| Chemoradiation | 0.84 | 0.69–1.02 | 0.072 |
When stratifying for LN negative disease, increased OS was associated with the use of ACT as compared with surgery alone (HR 0.54, 95% CI 0.35 to 0.83) but not ACRT (HR 0.85, 95% CI 0.63 to 1.16). When stratifying for LN positive disease, however, neither ACT (HR 0.77, 95% CI 0.58 to 1.02) nor ACRT (HR 0.79, 95% CI 0.61 to 1.03) resulted in a significant OS advantage compared with surgery alone (Table 4).
Table 4.
Stratified Cox-Proportional Hazards Regression for Overall Survival, Local Recurrence-Free Survival, and Distant Recurrence-Free Survival
| Variable | LRFS | DRFS | OS |
|---|---|---|---|
| ACT (n = 291) | 0.59 (0.40–0.86)* | 0.74 (0.55–0.98)* | 0.71 (0.57–0.90)* |
| ACRT (n = 447) | 0.51 (0.37–0.72)* | 0.80 (0.63–1.03) | 0.84 (0.69–1.01) |
| LN negative | |||
| ACT (n = 87) | 0.57 (0.28–1.14) | 0.65 (0.39–1.09) | 0.54 (0.35–0.84)* |
| ACRT (n = 185) | 0.55 (0.32–0.94)* | 0.80 (0.56–1.14) | 0.85 (0.63–1.16) |
| LN positive | |||
| ACT (n = 204) | 0.58 (0.36–0.94)* | 0.73 (0.50–1.08) | 0.77 (0.57–1.02) |
| ACRT (n = 262) | 0.49 (0.31–0.76)* | 0.81 (0.56–1.17) | 0.79 (0.61–1.03) |
| Margin negative | |||
| ACT (n = 230) | 0.58 (0.38–0.89)* | 0.78 (0.57–1.09) | 0.68 (0.53–0.88)* |
| ACRT (n = 322) | 0.44 (0.30–0.65)* | 0.73 (0.54–0.97)* | 0.77 (0.62–0.96)* |
| Margin positive | |||
| ACT (n = 61) | 0.67 (0.25–1.76) | 0.55 (0.28–1.09) | 0.81 (0.48–1.38) |
| ACRT (n = 125) | 0.99 (0.41–2.37) | 0.95 (0.53–1.70) | 1.00 (0.65–1.55) |
Data shown as hazard ratio (95% CI). Surgery alone = reference.
Significant.
ACRT, adjuvant chemoradiation; ACT, adjuvant chemotherapy; DRFS, distant recurrence-free survival; LRFS, local recurrence-free survival; OS, overall survival.
For patients with margin negative disease, both ACT (HR 0.0.68, 95% CI 0.53 to 0.88) and ACRT (HR 0.77, 95% CI 0.62 to 0.96) were independently associated with improved OS compared with surgery alone. In margin positive disease, however, neither ACT (HR 0.81, 95% CI 0.0.48 to 1.38) nor ACRT (HR 1.00, 95% CI 0.65 to 1.55) provided any significant advantage in OS compared with surgery alone (Table 4).
Multivariable analysis of local recurrence-free survival
When adjusting for possible confounding variables associated with LRFS with Cox-proportional hazards modeling, both the use of ACT (HR 0.59, 95% CI 0.40 to 0.86) and ACRT (HR 0.51. 95% CI 0.37 to 0.72) were independently associated with significantly increased LRFS, as compared with surgery alone (Table 5). Lymph node positive disease was independently associated with a significantly decreased LRFS (HR 1.72, 95% CI 1.27 to 2.32). The presence of positive margins, however, was not associated with any significant differences in LRFS (Table 5). When stratifying for LN negative disease, ACRT (HR 0.55, 95% CI 0.32 to 0.94), but not ACT (HR 0.57, 95% CI 0.28 to 1.14), was associated with a significantly increased LRFS when compared with surgery alone. In patients with LN positive disease, both ACT (HR 0.58, 95% CI 0.36 to 0.94) and ACRT (HR 0.49, 95% CI 0.31 to 0.76) were independently associated with improved LRFS. Similarly, in patients with margin negative disease, both the use of ACT (HR 0.58, 95% CI 0.38 to 0.89) and ACRT (HR 0.44, 95% CI 0.30 to 0.65) were associated with increased LRFS. However, in patients with margin positive disease, neither ACT (HR 0.67, 95% CI 0.26 to 1.76) nor ACRT (HR 0.99, 95% CI 0.99 to 1.07) provided a significant benefit in LRFS (Table 4).
Table 5.
Cox Proportional Multivariate Hazards Regression for Local Recurrence-Free Survival
| Variable | Hazard ratio | 95% CI | p Value |
|---|---|---|---|
| Age | 0.99 | 0.98–1.01 | 0.926 |
| Sex (reference, female) | 1.10 | 0.84–1.45 | 0.479 |
| Tumor size | 1.00 | 0.99–1.01 | 0.254 |
| Tumor grade | |||
| 1 (reference) | |||
| 2 | 1.07 | 0.68–1.68 | 0.778 |
| 3 | 1.18 | 0.71–1.96 | 0.529 |
| Preoperative stent | 0.92 | 0.69–1.22 | 0.565 |
| Venous resection | 1.29 | 0.91–1.85 | 0.156 |
| Blood transfusion | 1.34 | 0.98–1.82 | 0.067 |
| Hospital length of stay | 0.99 | 0.97–1.01 | 0.421 |
| Lymph node status | |||
| Negative (reference) | |||
| Positive | 1.72 | 1.27–2.32 | <0.001 |
| No. of lymph nodes resected | 1.00 | 0.99–1.02 | 0.609 |
| Margin status | |||
| Negative (reference) | |||
| Positive | 1.18 | 0.84–1.64 | 0.334 |
| Adjuvant therapy | |||
| None (reference) | |||
| Chemotherapy only | 0.59 | 0.40–0.86 | 0.007 |
| Chemoradiation | 0.51 | 0.37–0.72 | <0.001 |
Multivariable analysis of distant recurrence-free survival
Cox proportional multivariate analysis of variables associated with DRFS is shown in Table 6. When compared with surgery alone, the use of ACT significantly improved DRFS (HR 0.74, 95% CI 0.55 to 0.98); ACRT did not (HR 0.80, 95% CI 0.62 to 1.03). An R1 resection margin was independently associated with a decreased DRFS (HR 1.35, 95% CI 1.07 to 1.71). Other variables, including LN positive disease, however, were not associated with significant differences in DRFS (Table 5). There was no improvement in DRFS with either ACT or ACRT in patients with LN negative disease (ACT [HR 0.65, 95% CI 0.39 to 1.09]; ACRT [HR 0.80, 95% CI 0.56 to 1.14]); or LN positive disease (ACT [HR 0.73, 95% CI 0.50 to 1.08]; ACRT [HR 0.81, 95% CI 0.56 to 1.17]). When stratifying for margin negative disease, ACRT (HR 0.73, 95% CI 0.55 to 0.97) was associated with improved DRFS, while the use of ACT (HR 0.78, 95% CI 0.57 to 1.09) did not provide any significant benefit compared with surgery alone (Table 4).
Table 6.
Cox Proportional Multivariate Hazards Regression for Distant Recurrence-Free Survival
| Factor | Odds ratio | 95% CI | p Value |
|---|---|---|---|
| Age | 1.00 | 0.99–1.01 | 0.280 |
| Gender (reference, female) | 1.07 | 0.88–1.30 | 0.489 |
| Tumor size | 1.00 | 0.99–1.01 | 0.982 |
| Tumor grade | |||
| 1 (reference) | |||
| 2 | 1.29 | 0.90–1.85 | 0.161 |
| 3 | 1.45 | 0.98–2.15 | 0.060 |
| Preoperative stent | 0.92 | 0.75–1.13 | 0.451 |
| Venous resection | 0.91 | 0.69–1.21 | 0.526 |
| Blood transfusion | 1.14 | 0.91–1.43 | 0.245 |
| Hospital length of stay | 0.99 | 0.98–1.01 | 0.591 |
| Lymph node status | |||
| Negative (reference) | |||
| Positive | 1.14 | 0.93–1.41 | 0.210 |
| No. of lymph nodes resected | 1.01 | 1.00–1.02 | 0.003 |
| Margin status | |||
| Negative (reference) | |||
| Positive | 1.35 | 1.07–1.72 | 0.012 |
| Adjuvant therapy | |||
| None (reference) | |||
| Chemotherapy only | 0.74 | 0.55–0.98 | 0.039 |
| Chemoradiation | 0.80 | 0.62–1.03 | 0.079 |
DISCUSSION
This large multi-institutional study shows that the use of ACT increases both LRFS and DRFS, with a corresponding benefit in OS. Meanwhile, ACRT provides a benefit in LRFS, but not DRFS or OS. These associations are independent of other well-known factors affecting recurrence and survival, including margin status and LN status. Although recurrence after resection for PDAC is common, most studies have focused on the effect of adjuvant therapy on OS, and patterns of recurrence have not generally been reported. Consistent with other reports, most recurrent disease was DR (41%); LR was seen in 22% of patients.3,20 Although several factors, including LN status, margin status, number of LNs removed, and LN ratio, have been associated with OS, their associations with recurrence patterns have not been well described. For example, the presence of LN metastases is one of the most strongly associated prognostic factor for OS.3,19,21–23 Positive resection margin status is another factor that has been associated with decreased OS in some studies,24–28 while others have refuted this finding.3,19,29 In this study, the presence of involved LNs was 57%, and the margin positive (R1) rate was 22%, both consistent with several other reports. Findings from this study showed both positive LN and an R1 margin were independent adverse prognostic factors influencing OS.
In a report from Johns Hopkins, of patients receiving ACRT, LN positive disease was not associated with any site recurrence or LR, while R1 disease was associated with any site recurrence but not LR.20 A recent Japanese study reported that an R1 margin was associated with increased LR but not OS.19 However, this study did not control for receipt of adjuvant therapy. A report from MD Anderson showed that margin status had no association with recurrence or OS, but the majority of patients in that study had received preoperative therapy, which certainly could have affected margin and LN status at the time of resection.3 Findings from the ESPAC-1 randomized controlled trial showed that LN positivity was independently associated with decreased OS, but margin status was only independently associated with OS in the absence of lymph node status and tumor grade.25,29 But that study did not investigate the association of margin status with recurrence patterns. The disparate findings in these studies continue to stress the importance of controlling for these significant confounding factors, as well as different treatment regimens and modalities, in non-randomized studies.
The use of ACT has become the standard of care for patients with resected PDAC, with multiple randomized trials showing improvements in OS compared with surgery alone.6–8 This study also showed a significant OS benefit with the use of ACT when controlling for other confounding factors. The use of ACRT, however, has been more controversial.9,10 In fact, although the ESPAC-1 trial showed a benefit from the use of ACT, ACRT was actually associated with a decreased OS.11 The benefit of ACT in the ESPAC-1 trial was present regardless of margin status, although small numbers prevented the benefit in margin positive patients from reaching statistical significance. This was also the case in our study; small numbers of patients in the margin positive group receiving ACT resulted in large confidence intervals and loss of statistical power.
In the CONKO-001 trial, ACT with gemcitabine was associated with a significant advantage in disease-free survival (DFS) compared with surgery alone.6 This advantage in DFS was present in both margin negative and positive patients as well as LN negative and positive patients. Interestingly, the only OS advantage of ACT in this study was seen in the margin negative patients, similar to the findings in our study.
In the smaller JSAP-2 trial, ACT, also with gemcitabine, showed an improvement in DFS as well, but no difference in OS was seen. Subgroup analysis showed that the beneficial effect of adjuvant gemcitabine on DFS was evident for margin negative and LN negative patients.7 Consistent with this, our results also show a significant OS benefit of ACT in patients with LN negative but not LN positive disease, suggesting that LN positive disease is a harbinger of aggressive disease that is difficult to control with adjuvant therapy.
In our study, ACT was associated with an improvement in LR, DR, and OS. On the other hand, ACRT provided a decrease in LR, but had no significant effect on DR or OS when compared with these parameters from patients who underwent surgery alone. Because most patients die of distant disease rather than local recurrence, it stands to reason that because ACT improves both DR and LR, a benefit in OS would be expected. In the case of ACRT, however, simply reducing LR without affecting DR may not be adequate to provide a benefit in OS. Although most patients receiving ACRT also receive systemic chemotherapy (usually with gemcitabine), the chemotherapy received during the radiation therapy portion of the adjuvant treatment is only radiosensitizing. The lack of a survival benefit with ACRT in our study, and in almost every other study, would suggest that a full 6 months of chemotherapy is perhaps most important in influencing LR and DR, and therefore, OS. These results further raise the question of whether longer periods of systemic chemotherapy, with or without radiation therapy, may improve the patterns of recurrence and OS.
Findings from this study also support the notion that patients with PDAC are a very heterogeneous group, with several confounding factors affecting the course of the disease and the efficacy of treatment regimens. These results clearly show that the relationship between these prognostic factors and LR, DR, and OS is not a simple one. For example, unadjusted OS, as well as LR and DR, did not significantly differ among the 3 treatment groups in this study. Patients in the ACT group, however, had the largest tumors, higher stage, and more LN metastases and R1 resections; the ACRT group had the highest percentage of R1 disease. Only after controlling for these variables did the effect of treatment become clearer.
As with any retrospective study, this study had several limitations. The multi-institutional nature of this cohort helps to overcome limitations such as institutional biases and small patient cohorts and has the advantage of having patients treated at high-volume tertiary referral centers with expertise in multimodality treatment of pancreas cancer. Nevertheless, details of treatment regimens, including the type and duration of adjuvant therapy, were not always available because many patients received their adjuvant treatment at institutions other than the primary facility where the resection was performed. In addition, complete details such as patient comorbidities, complications, and other possible confounders were not always available. Confirmation of LR and DR, as well as palliative treatment for the recurrence, was at the discretion of the treating physician. Although this was a large cohort overall, subset analysis with stratification for LN status and margin status may have led to significant reduction in statistical power in these specific analyses.
CONCLUSIONS
In support of our earlier findings, future studies that investigate the benefits of ACT and ACRT in pancreas cancer should consider longer treatment periods of systemic chemotherapy and should include stratification schemes to account for important confounding factors, including LN status and margin status. As these studies are designed, adequate statistical power for these stratification schemes will be important to draw definitive conclusions. In addition, data on recurrence patterns will be helpful in elucidating mechanisms associated with survival outcomes and ultimately allow us to develop a more tailored and directed approach in the multimodality treatment of pancreas cancer.
Abbreviations and Acronyms
- ACRT
adjuvant chemoradiation
- ACT
adjuvant chemotherapy
- CONKO
Charité Onkologie
- DRFS
distant recurrence-free survival
- ESPAC
European Study Group for Pancreatic Cancer
- HR
hazard ratio
- LN
lymph node
- IQR
interquartile range
- LRFS
local recurrence-free survival
- OS
overall survival
- PDAC
pancreatic ductal adenocarcinoma
Footnotes
Disclosure Information: Nothing to disclose.
Presented at the Southern Surgical Association 127th Annual Meeting, Hot Springs, VA, December 2015.
REFERENCES
- 1.Siegel R, Desantis C, Jemal A. Colorectal cancer statistics, 2014. CA Cancer J Clin 2014;64:104–117. [DOI] [PubMed] [Google Scholar]
- 2.Rahib L, Smith BD, Aizenberg R, et al. Projecting cancer incidence and deaths to 2030: the unexpected burden of thyroid, liver, and pancreas cancers in the United States. Cancer Res 2014;74:2913–2921. [DOI] [PubMed] [Google Scholar]
- 3.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: 52–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Morselli A. The determinant factors of recurrence following resection for ductal pancreatic cancer. JOP 2007;8:1–9. [PubMed] [Google Scholar]
- 5.Winter JM, Tang LH, Klimstra DS, et al. Failure patterns in resected pancreas adenocarcinoma: lack of predicted benefit to SMAD4 expression. Ann Surg 2013;258:331–335. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Oettle H, Neuhaus P, Hochhaus A, et al. Adjuvant chemotherapy with gemcitabine and long-term outcomes among patients with resected pancreatic cancer: the CONKO-001 randomized trial. JAMA 2013;310:1473–1481. [DOI] [PubMed] [Google Scholar]
- 7.Ueno H, Kosuge T, Matsuyama Y, et al. A randomised phase III trial comparing gemcitabine with surgery-only in patients with resected pancreatic cancer: Japanese Study Group of Adjuvant Therapy for Pancreatic Cancer. Br J Cancer 2009; 101:908–915. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Neoptolemos JP, Stocken DD, Tudur Smith C, et al. Adjuvant 5-fluorouracil and folinic acid vs observation for pancreatic cancer: composite data from the ESPAC-1 and −3(v1) trials. Br J Cancer 2009;100:246–250. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Kalser MH, Ellenberg SS. Pancreatic cancer. Adjuvant combined radiation and chemotherapy following curative resection. Arch Surg 1985;120:899–903. [DOI] [PubMed] [Google Scholar]
- 10.Klinkenbijl JH, Jeekel J, Sahmoud T, et al. Adjuvant radio-therapy and 5-fluorouracil after curative resection of cancer of the pancreas and periampullary region: phase III trial of the EORTC gastrointestinal tract cancer cooperative group. Ann Surg 1999;230:776e782; discussion 782–784. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.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; 350:1200–1210. [DOI] [PubMed] [Google Scholar]
- 12.Merchant NB, Rymer J, Koehler EAS, et al. Adjuvant chemoradiation therapy for pancreatic adenocarcinoma: who really benefits? J Am Coll Surg 2009;208:829–838; discussion 838–841. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Corsini MM, Miller RC, Haddock MG, et al. Adjuvant radiotherapy and chemotherapy for pancreatic carcinoma: the Mayo Clinic experience (1975–2005). J Clin Oncol 2008;26: 3511–3516. [DOI] [PubMed] [Google Scholar]
- 14.Hazard L, Tward JD, Szabo A, Shrieve DC. Radiation therapy is associated with improved survival in patients with pancreatic adenocarcinoma: results of a study from the Surveillance, Epidemiology, and End Results (SEER) registry data. Cancer 2007;110:2191–2201. [DOI] [PubMed] [Google Scholar]
- 15.O’Reilly EM. Adjuvant therapy for pancreas adenocarcinoma: where are we going? Expert Rev Anticancer Ther 2011;11: 173–177. [DOI] [PubMed] [Google Scholar]
- 16.Sultana A, Cox T, Ghaneh P, Neoptolemos JP. Adjuvant therapy for pancreatic cancer. Recent Results Cancer Res 2012; 196:65–88. [DOI] [PubMed] [Google Scholar]
- 17.Thomas A, Dajani K, Neoptolemos JP, Ghaneh P. Adjuvant therapy in pancreatic cancer. Dig Dis 2010;28:684–692. [DOI] [PubMed] [Google Scholar]
- 18.Dholakia AS, Kumar R, Raman SP, et al. Mapping patterns of local recurrence after pancreaticoduodenectomy for pancreatic adenocarcinoma: a new approach to adjuvant radiation field design. Int J Radiat Oncol Biol Phys 2013;87: 1007–1015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Sugiura T, Uesaka K, Mihara K, et al. Margin status, recurrence pattern, and prognosis after resection of pancreatic cancer. Surgery 2013;154:1078–1086. [DOI] [PubMed] [Google Scholar]
- 20.Asiyanbola B, Gleisner A, Herman JM, et al. Determining pattern of recurrence following pancreaticoduodenectomy and adjuvant 5-flurouracil-based chemoradiation therapy: effect of number of metastatic lymph nodes and lymph node ratio. J Gastrointest Surg 2008;13:752–759. [DOI] [PubMed] [Google Scholar]
- 21.Schwarz RE, Smith DD. Extent of lymph node retrieval and pancreatic cancer survival: information from a large US population database. Ann Surg Oncol 2006;13:1189–1200. [DOI] [PubMed] [Google Scholar]
- 22.Sohn TA, Yeo CJ, Cameron JL, et al. Resected adenocarcinoma of the pancreas-616 patients: results, outcomes, and prognostic indicators. J Gastrointest Surg 2000;4:567–579. [DOI] [PubMed] [Google Scholar]
- 23.Gebhardt C, Meyer W, Reichel M, Wünsch PH. Prognostic factors in the operative treatment of ductal pancreatic carcinoma. Langenbecks Arch Surg 2000;385:14–20. [DOI] [PubMed] [Google Scholar]
- 24.Verbeke CS, Leitch D, Menon KV, et al. Redefining the R1 resection in pancreatic cancer. Br J Surg 2006;93:1232–1237. [DOI] [PubMed] [Google Scholar]
- 25.Neoptolemos JP, Stocken, Dunn JDA, et al. Influence of resection margins on survival for patients with pancreatic cancer treated by adjuvant chemoradiation and/or chemotherapy in the ESPAC-1 randomized controlled trial. Ann Surg 2001;234:758–768. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Lüttges J, Vogel I, Menke M, et al. The retroperitoneal resection margin and vessel involvement are important factors determining survival after pancreaticoduodenectomy for ductal adenocarcinoma of the head of the pancreas. Virchows Archiv 1998;433:237–242. [DOI] [PubMed] [Google Scholar]
- 27.Winter JM, Cameron JL, Campbell KA, et al. 1423 pancreaticoduodenectomies for pancreatic cancer: A single-institution experience. J Gastrointest Surg 2006;10:1199–1210; discussion 1210–1211. [DOI] [PubMed] [Google Scholar]
- 28.Cleary SP, Gryfe R, Guindi M, et al. Prognostic factors in resected pancreatic adenocarcinoma: analysis of actual 5-year survivors. J Am Coll Surg 2004;198:722–731. [DOI] [PubMed] [Google Scholar]
- 29.Bassi C, Stocken DD, Olah A, et al. Influence of surgical resection and post-operative complications on survival following adjuvant treatment for pancreatic cancer in the ESPAC-1 randomized controlled trial. Dig Surg 2005;22:353–363. [DOI] [PubMed] [Google Scholar]