Neoadjuvant chemoradiation before surgery is an emerging treatment modality for pancreatic ductal adenocarcinoma (PDAC). However, analysis of prognostic factors is limited for patients with PDAC treated with neoadjuvant chemoradiation and pancreaticoduodenectomy (PD).
The study population was comprised of 240 consecutive patients with PDAC who received neoadjuvant chemoradiation and PD and was compared with 60 patients who had no neoadjuvant therapy between 1999 and 2007. Clinicopathologic features were correlated with disease-free survival (DFS) and overall survival (OS).
Among the 240 treated patients, the 1-year and 3-year DFS rates were 52% and 32%, with a median DFS of 15.1 months. The 1-year and 3-year OS rates were 95% and 47%, with a median OS of 33.5 months. By univariate analysis, DFS was associated with age, post-therapy tumor stage (ypT), lymph node status (ypN), number of positive lymph nodes, and American Joint Committee on Cancer (AJCC) stage, whereas OS was associated with intraoperative blood loss, margin status, ypT, ypN, number of positive lymph nodes, and AJCC stage. By multivariate analysis, DFS was independently associated with age, number of positive lymph nodes, and AJCC stage, and OS was independently associated with differentiation, margin status, number of positive lymph nodes, and AJCC stage. In addition, the treated patients had better OS and lower frequency of lymph node metastasis than those who had no neoadjuvant therapy.
Pancreatic cancer is the fourth leading cause of cancer-related death in the United States.1 In 2009, it is estimated that 42,470 men and women were diagnosed with pancreatic cancer and 35,240 died of disease.1 Up to 80% of patients with pancreatic ductal adenocarcinoma (PDAC) present with metastatic disease, with a dismal 5-year survival rate of 5.5%.1 Even in patients with localized disease, the 5-year survival rate is only 22%.1 Surgical resection of localized PDAC is the only potentially curative approach for this deadly disease. However, despite significant improvements in operative techniques and postoperative mortality rates, the overall survival (OS) for patients with PDAC has not changed significantly over the past 4 decades.2 The addition of postoperative adjuvant chemotherapy and/or radiation has been established as the standard of care in light of results of prospective randomized controlled trials showing survival benefit with adjuvant therapy in resected pancreatic cancer patients.3 Recently, several tertiary referral centers worldwide, including our institution, have demonstrated the safety and feasibility of preoperative neoadjuvant chemoradiation for PDAC.4-8 In some studies, survival of patients with resectable or borderline resectable PDAC treated with neoadjuvant therapy has been shown to exceed historical control.4, 8-12 In our institution, neoadjuvant chemoradiation plays an integral role in the treatment of patients with PDAC.
Better survival in patients with PDAC who undergo pancreaticoduodenectomy (PD) has been associated with pathologic primary tumor (pT) stage I disease,13-15 well-differentiated histology,13, 16-18 negative lymph node status,13, 15, 17-21 and negative resection margins.18, 22, 23 However, most of these studies have been based on clinicopathologic data from patients who underwent surgery alone13, 14, 17, 19, 20 or those treated with postoperative chemoradiation.15, 16, 18, 22-25 It remains unclear whether post-therapy pathologic stage predicts disease-free and OS in patients with PDAC treated with neoadjuvant chemoradiation followed by PD. We, therefore, examined the prognostic significance of clinicopathologic characteristics in a large cohort of patients with PDAC who received preoperative chemotherapy and/or radiation and subsequently underwent PD. Our study showed that post-therapy pathologic stage, including lymph node status and the number of positive regional lymph nodes, is an independent prognostic factor in this group of patients.
MATERIALS AND METHODS
Study Population, Patient Characteristics, and Follow-up
The study was approved by the institutional review board of The University of Texas MD Anderson Cancer Center. Two hundred forty consecutive patients with histologically confirmed PDAC who received neoadjuvant chemotherapy and/or radiation therapy and underwent PD (treated group) at The University of Texas MD Anderson Cancer Center between January 1999 and December 2007 were included in this study. This treated group was compared with 60 consecutive patients who underwent PD alone with no neoadjuvant therapy (untreated group) during the same time period. For the purpose of this analysis, patients who underwent distal pancreatectomy for PDAC and those who died from postoperative complications were excluded.
Our treated group consisted of 97 women and 143 men, with a mean age of 62.5 years at the time of surgery (median, 62.9 years; range, 38.5-85.4 years). Neoadjuvant regimens are listed in Table 1. One hundred nineteen patients (50%) received neoadjuvant chemoradiation, 114 patients (47%) received neoadjuvant systemic chemotherapy followed by chemoradiation, and the remaining 7 patients (3%) received neoadjuvant systemic chemotherapy alone. One hundred forty-six of 240 patients (61%) were treated on protocols; 77 (32%) received gemcitabine-based chemoradiation,4 and 69 patients (29%) received systemic chemotherapy with gemcitabine and cisplatin followed by gemcitabine-based chemoradiation.8 All patients underwent restaging evaluation after completion of neoadjuvant therapy, and PD was performed only in patients with resectable disease who did not develop disease progression or metastasis and had no contraindications for major abdominal surgery. The average intraoperative blood loss was 900 mL (range, 100-6000 mL). The patients' clinical follow-up information through December 2009 was extracted from patient medical records and review of the US Social Security Index when necessary. Recurrence status was updated at each follow-up clinic visit.
Table 1. Neoadjuvant Chemotherapy and Radiation Therapy Received by the Patients
No. of Patients
Fluoropyrimidine comprises either 5-fluorouracil or capecitabine (XELODA).
Group 3: Systemic chemotherapy followed by gemcitabine-based chemoradiation
Group 4: Systemic chemotherapy followed by fluoropyrimidine-based chemoradiationa
Group 5: Chemotherapy alone
A standardized system for the pathologic evaluation of PD specimens has been used at our institution since July 1990. All cases were re-reviewed by a gastrointestinal pathologist (Huamin Wang) for tumor type, size, differentiation, extrapancreatic tissue involvement, margin status, and number of positive regional lymph nodes. The pathologic primary tumor classification and stage grouping were determined according to the American Joint Committee on Cancer (AJCC) Staging Manual, seventh edition.26 The final margin status of PD specimens was recorded as R0 (negative surgical margins microscopically), R1 (negative surgical margins grossly but with microscopic presence of tumor cells, including carcinoma in situ/high-grade dysplasia, at any resection margin), or R2 (presence of grossly identifiable tumor at any margin).
Chi-square analysis or Fisher exact tests were used to compare categorical data, and analysis of variance was used to compare continuous variables. Survival curves were constructed using the Kaplan-Meier method, and the log-rank test was used to evaluate the statistical significance of differences. Disease-free survival (DFS) was calculated as the time from the date of surgery to the date of first recurrence after surgery (in patients with recurrence) or to the date of last follow-up (in patients without recurrence). OS was calculated as the time from the date of diagnosis to the date of death or the date of last follow-up (if death did not occur). The prognostic significance of clinical and pathologic characteristics was determined using univariate Cox regression analysis. Cox proportional hazards models were fitted for multivariate analysis. After interactions between the variables were examined, a backward stepwise procedure was used to derive the best-fitting model. In addition, we compared the survival and clinicopathologic parameters of our treated group to 60 patients who underwent PD alone and had no neoadjuvant therapy (untreated group) during the same time period. Statistical analysis was performed using Statistical Package for Social Sciences software (for Windows 12.0, SPSS Inc., Chicago, Ill). A 2-sided significance level of .05 was used for all statistical analyses.
Pathologic Features of the Treated Group
Post-therapy tumor size ranged from 0.1 cm to 8.5 cm, with an average of 2.6 cm. According to the World Health Organization classification standards, 152 cases (63%) were well to moderately differentiated PDAC, and 88 cases (37%) were poorly differentiated PDAC. Dissection of regional lymph nodes from resection specimens yielded, on average, 22 lymph nodes (range, 5-50). Lymph node involvement by metastatic disease was identified in 140 patients, with the number of positive lymph nodes ranging from 1 to 21 nodes. On the basis of the AJCC Staging Manual (seventh edition), 15 patients (6%) had post-therapy pathologic stage I disease, 85 patients (36%) had stage IIA disease, and 140 patients (58%) had stage IIB disease. Three patients with stage IV disease because of positive intraoperative peritoneal washing on cytology were excluded from the study, as the number of cases was too small to be representative. No patient had stage III disease. R0 resection was achieved in 213 patients (89%), 26 patients (11%) had microscopic tumor involvement of 1 or more surgical resection margins (R1), and there were no R2 resections. In patients with R1 resection, 3 cases (1%) had tumor at the common bile duct margin, 8 cases (4%) had tumor (5 cases) or carcinoma in situ/high-grade dysplasia (3 cases) at the pancreatic margin, and 15 cases (6%) had tumor at the retroperitoneal margin.
The median follow-up time in the treated group was 29.8 months (range, 7.6-122.3 months). At the time of last follow-up, 154 patients died of PDAC, 11 patients died of other causes, 15 patients were alive with PDAC, and 60 patients were alive with no clinical or radiographic evidence of PDAC. Among all 240 patients, the DFS rates were 52% at 1 year and 32% at 3 years, with a median DFS of 15.1 months. The OS rates were 95% at 1 year and 47% at 3 years with a median OS of 33.5 months.
The correlations of post-therapy pathologic tumor stage (ypT) and AJCC stage with DFS and OS are shown in Figure 1. Patients with ypT1 or ypT2 had better DFS and OS than those with ypT3 (P = .02 and P = .003, respectively, Fig. 1A, B). Patients with post-therapy stage I disease had a significantly better DFS rate (87% at 1 year and 67% at 3 years) compared with patients with stage IIA (60% at 1 year and 39% at 3 years) and stage IIB disease (43% at 1 year and 23% at 3 years, P = .0004). The median DFS was 81.4 months among patients with stage I disease, 22.0 months among patients with stage IIA disease, and 10.0 months among patients with stage IIB disease (Fig. 1C). OS was also significantly better in patients with post-therapy stage I disease (100% at 1 year and 3 years) than patients with stage IIA disease (98% at 1 year and 50% at 3 years) or patients with stage IIB disease (94% at 1 year and 39% at 3 years) (P < .0001). The median OS was 104.4 months for patients with stage I disease, 39.5 months for patients with stage IIA disease, and 28.3 months for patients with stage IIB disease (Fig. 1D). Both DFS and OS were better in patients with post-therapy stage IIA PDAC than those with stage IIB PDAC (P = .01 and P = .04, respectively, Fig. 1C, D).
Patients with negative regional lymph nodes (ypN0) had a significantly better DFS and OS than those with lymph node metastasis (ypN1)(P = .0003 and P < .0001, respectively, Fig. 2A, B). Because the number of positive lymph nodes has been shown to correlate with prognosis in patients with other gastrointestinal tract malignancies,27-35 we further stratified our patients with regional nodal metastasis into 2 groups: patients with 1 to 3 positive lymph nodes (ypN1a) and those with 4 or more positive lymph nodes (ypN1b). DFS and OS in relation to the number of positive lymph nodes are illustrated in Figure 2C and D. The median DFS times were 27.1 months, 11.5 months, and 9.2 months among patients with ypN0, ypN1a, and ypN1b disease, respectively (P = .0001, Fig. 2C). The median OS times were 53.0 months, 32.8 months, and 24.1 months in patients with ypN0, ypN1a, and ypN1b disease, respectively (P < .0001, Fig. 2D). Our data demonstrated that the number of positive regional lymph nodes correlated significantly with both DFS and OS in patients with PDAC who received neoadjuvant therapy and underwent PD.
Results from univariate Cox regression analysis for DFS and OS are shown in Table 2. DFS was associated with age (P = .002), post-therapy primary tumor stage (ypT) (P = .03), regional lymph node status (ypN) (P < .001), number of positive lymph nodes (P < .001), and pathologic AJCC stage (P = .001). OS was associated with intraoperative blood loss (P = .02), margin status (P = .02), ypT (P = .005), ypN (P < .001), number of positive lymph nodes (P < .001), and pathologic AJCC stage (P < .001). There was no significant correlation with either DFS or OS and sex, type of neoadjuvant therapy administered, or tumor size.
Table 2. Univariate Cox Regression Analysis of Disease-Free and Overall Survival in Relation to Clinicopathologic Features
No. of Patients
HR (95% CI)
HR (95% CI)
Abbreviations: AJCC, American Joint Committee on Cancer; CI, confidence interval; HR, hazard ratio; ref, reference.
Two multivariate survival analyses were performed to determine the prognostic significance of clinicopathologic factors for DFS and OS, and the results are shown in Table 3. In the first model, post-therapy pathologic AJCC stage was an independent predictor for both DFS and OS (P = .002 and P < .001, respectively). In the second model, the number of positive lymph nodes was a prognostic factor for both DFS and OS independent of patient age, intraoperative blood loss, tumor differentiation, ypT, and margin status. Post-therapy primary tumor stage (ypT) was an independent prognostic factor for OS (P = .041), but not an independent prognostic factor for DFS (P = .08). In both models, patient age was an independent prognostic factor for DFS, and tumor differentiation and R0 resection status were independent prognostic factors for OS. Intraoperative blood loss was significant for OS in univariate analysis, but not in the multivariate models.
Table 3. Multivariate Cox Regression Analysis of Disease-Free and Overall Survival in Relation to Clinicopathologic Features
No. of Patients
HR (95% CI)
HR (95% CI)
Abbreviations: AJCC, American Joint Committee on Cancer; CI, confidence interval; HR, hazard ratio; N/A, not available; ref, reference.
To evaluate whether neoadjuvant therapy improved the clinical outcome in patients with PDAC, we compared survival and clinicopathologic features of the treated group to 60 consecutive patients who did not receive any form of neoadjuvant therapy before PD (untreated group) during the same time period at our institution. The median OS in the treated group was 33.5 months (±2.5 months) and was better than the median OS of 26.5 months (±3.4 months) in the untreated group (P = .04, Fig. 3A). The median DFS was also better in the treated group than the untreated group; however, the difference in DFS between these 2 groups was not statistically significant (P = .08, Fig. 3B). In the treated group, metastatic PDAC was identified in 58% (140 of 240) of the patients, which was significantly lower than the untreated group (75% [45 of 60], P = .01). There were no significant differences in tumor size, differentiation, margin status, and other clinical parameters between these 2 groups (data not shown).
In this study, we found that post-therapy pathologic AJCC stage and the number of positive regional lymph nodes are independent prognostic factors for both DFS and OS in a consecutive cohort of 240 patients with PDAC who received neoadjuvant chemoradiation followed by PD treated at a single institution. In addition, our study also found that tumor differentiation, ypT, and margin status are independent prognostic factors for OS. Our study highlights the importance of pathologic evaluation and post-therapy tumor staging in predicting prognosis of patients with PDAC who received neoadjuvant therapy and subsequently underwent PD.
In the treatment of patients with PDAC, neoadjuvant therapy has several potential advantages, including identifying patients who are likely to benefit the most from surgery, providing early treatment of micrometastatic disease, and potentially down-staging the tumor in borderline resectable cases, thereby increasing the rate of complete resection and reducing local recurrence. Several studies, including 2 large trials from our institution, have reported median survival durations of up to 34 months in patients with resectable PDAC treated with neoadjuvant chemoradiation followed by PD.4, 7, 8 Evans et al studied 64 patients who were treated with neoadjuvant gemcitabine-based chemoradiation and subsequent PD, and reported a median OS of 34 months.4 A phase 2 trial reported by Varadhachary et al showed a median OS of 31 months in the 52 patients who received preoperative gemcitabine and cisplatin chemotherapy followed by gemcitabine-based chemoradiation and then underwent PD.8 In the current study of 240 patients, which includes 146 patients (61%) who were treated under the protocols reported by Evans et al and Varadhachary et al,4, 8 the median OS was 33.5 months, with 1-year and 3-year OS rates of 95% and 47%, respectively. No survival difference was seen among patients who received different neoadjuvant regimens.
Our study showed that pT stage, tumor differentiation, and margin status correlated with OS by multivariate analysis, similar to the previous studies of patients with PDAC who underwent surgery with or without adjuvant chemoradiation.13-18, 22, 23 However, we did not find significant correlation of these parameters with DFS. In our study, the R0 resection rate was 89%, which is consistent with the reported R0 resection rates of 72% to 95% in patients with PDAC treated with a variety of neoadjuvant chemoradiation protocols followed by surgery.5, 7, 9, 36 Patients with R0 resection had a significantly better OS than patients with R1 resection (35.5 vs 24.0 months, P = .02). Because the vast majority of our cases (92%, 221 of 240) had ypT3 tumor, with only 19 (8%) cases categorized as either ypT1 (16 cases) or ypT2 (3 cases), the correlation of ypT with survival may not be fully representative for patients with ypT1 or ypT2 tumors. However, our data demonstrated that patients with ypT3 had worse DFS and OS than those with ypT1 and ypT2 tumors by univariate analysis, and had a worse OS by multivariate analysis.
Pathologic AJCC stage and lymph node status are important independent prognostic factors for patient outcome in various types of human malignancies.13, 15, 17-21 Our study demonstrated that post-therapy pathologic AJCC stage is an independent prognostic factor for both DFS and OS in patients with PDAC treated with neoadjuvant therapy followed by PD. Our results are consistent with findings from previous studies demonstrating that post-therapy pathologic AJCC stage predicts outcome for patients who undergo preoperative chemoradiation followed by surgery for carcinoma of the esophagus,37 stomach,38, 39 and rectum.40 In addition, we showed that lymph node status correlated with both DFS and OS in patients with PDAC who received neoadjuvant chemoradiation.
Recent studies suggest that the lymph node ratio (number of positive lymph nodes to the total number of lymph nodes harvested) may be a better predictor of survival than lymph node status alone in patients with PDAC.2, 41, 42 Berger et al evaluated 128 patients who underwent PD, including 48 patients who received neoadjuvant chemoradiation and 80 patients who received postoperative adjuvant chemotherapy.41 By using a cutoff value of 15%, they found that lymph node ratio was a significant prognostic indicator of both DFS and OS by univariate analysis, but failed to detect significance by multivariate analysis. Similar results were reported by Sierzega et al in 96 patients who underwent PD with or without adjuvant chemoradiation using a lymph node ratio of 20% as a cutoff.42 However, the total number of lymph nodes was not found to have prognostic significance for patient survival in both studies.41, 42 Consistent with their findings are results from 2 randomized controlled trials that have examined the role of extended lymphadenectomy in PDAC, and have found no significant advantage in survival among patients with increased lymph node harvest.43, 44 In this study, we grouped our patients into ypN0 (no lymph node metastasis), ypN1a (metastasis in 1-3 lymph nodes), and ypN1b (metastasis in 4 or more lymph nodes) based on the AJCC staging criteria for lymph nodes in other gastrointestinal malignancies.26 We found that the number of positive lymph nodes is an independent prognostic factor for both DFS and OS by univariate and multivariate analyses. The median DFS and OS for patients with ypN1a disease were 11.5 months and 32.8 months, respectively, and were significantly better than the DFS of 9.2 months and OS of 24.1 months for those with ypN1b disease (P = .03 and P = .04, respectively). Our results showed that the number of positive lymph nodes is an important prognostic factor in patients with PDAC who received neoadjuvant therapy.
In this study, we also found that patients with PDAC who received neoadjuvant therapy and subsequently underwent PD had better OS and lower frequency of lymph node metastasis than those who had no neoadjuvant therapy. Retrospectively comparing survival in a cohort that has undergone neoadjuvant therapy followed by selection of patients developing progressive disease before surgery with another cohort that has not undergone such a selection has significant limitations. In addition, we do not have the denominator of total number of patients who received neoadjuvant therapy, but did not undergo PD because of progressive disease. However, among the 240 patients in our treated group, 146 patients (61%) were treated on protocols reported by Evans et al and Varadhachary et al.4, 8 In these 2 trials, PD was performed successfully in 74% and 66% patients who completed neoadjuvant therapy, and the median OS of all unselected patients was 22.7 months and 17.4 months, respectively.4, 8 A comparison group in the literature consisting of all patients who underwent surgery first with anticipation of receiving postoperative adjuvant therapy (including those who did and did not receive intended therapy) is not available. Randomized clinical trials are needed to further examine the role of neoadjuvant therapy in patients with PDAC.
In summary, our study demonstrated that post-therapy pathologic AJCC stage, the number of positive lymph nodes, tumor differentiation, margin status, and ypT were independent prognostic factors for survival in patients with PDAC who received neoadjuvant chemoradiation and subsequently underwent PD. Pathologic evaluation of PD specimen and post-therapy pathologic tumor staging remain the key factors in predicting clinical outcome in this group of patients.
Supported by an institutional research grant at The University of Texas MD Anderson Cancer Center.