Comparison of Overall Survival Between Preoperative Chemotherapy and Chemoradiotherapy for Resectable Pancreatic Adenocarcinoma

View More View Less
  • a Division of Surgical Oncology, Department of Surgery, and Department of Clinical Science, University of Texas Southwestern Medical Center, Dallas, Texas.

Background: Preoperative therapy is being increasingly used in the treatment of resectable pancreatic cancer. Because there are only limited data on the optimal preoperative regimen, we compared overall survival (OS) between preoperative chemotherapy (CT) and preoperative chemoradiotherapy (CRT) in resectable pancreatic adenocarcinoma. Patients and Methods: Patients receiving preoperative therapy and resection for clinical T1–3N0–1M0 adenocarcinoma of the pancreas were identified in the National Cancer Database for 2006 through 2012. We constructed inverse probability of treatment weights to balance baseline group differences, and compared OS between CT and CRT, as well as pathologic and postoperative findings. Results: We identified 1,326 patients (CT: 616; CRT: 710). Differences in OS were not significant between CRT and CT (median survival, 25 vs 26 months; P=.10; weight-adjusted hazard ratio, 0.89; 95% CI, 0.77–1.02). Compared with patients in the CT group, those in the CRT group had lower pathologic T stage (ypT0/T1/T2: 36% vs 21%; P<.01), less lymph node involvement (ypN1: 35% vs 59%; P<.01), and fewer positive resection margins (14% vs 21%; P=.01), but had more postoperative unplanned readmissions (9% vs 6%; P=.01) and increased 90-day mortality (7% vs 4%; P=.03). Those in the CRT group were also less likely to receive postoperative therapy (26% vs 51%; P<.01). Conclusions: Preoperative CT and CRT have similar OS, but CRT is associated with more favorable pathologic features at the cost of higher postoperative morbidity and mortality. Additional trials investigating preoperative therapy are needed for patients with resectable pancreatic cancer.

Approximately 10% to 20% of patients with pancreatic adenocarcinoma present with resectable disease, and treatment often constitutes resection followed by chemotherapy (CT).15 More recently, however, preoperative therapy has been increasing in patients with resectable pancreatic cancer (PaC). Despite the lack of supporting data from randomized clinical trials, this trend reflects both a growing acceptance of preoperative therapy as a management approach for resectable PaC and an increased understanding of the advantages of a preoperative treatment sequencing strategy for this disease. Specifically, patients with PaC are more likely to complete therapy when it is delivered preoperatively.68 Preoperative therapy also allows better selection for surgery by identifying patients with occult metastases who will not benefit from surgical resection, and provides a unique opportunity to evaluate response in the resected surgical specimens. Moreover, in a recent propensity-matched large retrospective population database study, preoperative therapy followed by resection was associated with improved survival compared with upfront resection.9

Preoperative regimens used in the management of resectable PaC have varied markedly. In published nonrandomized studies, preoperative regimens comprised a mix of heterogeneous CT and/or chemoradiotherapy (CRT) regimens.1 Interestingly, patients in these studies often received CRT despite the controversial, and potentially detrimental, role of CRT in the adjuvant setting.2,10 Heterogeneity in preoperative regimens is further compounded by a range of distinct protocols proposed in currently open clinical trials investigating preoperative therapy for resectable PaC.

To guide future trials exploring preoperative treatment of resectable PaC, an evaluation comparing preoperative CT and preoperative CRT is necessary. This study compared overall survival (OS) between preoperative CT and preoperative conventionally fractionated CRT in patients undergoing resection of pancreatic head adenocarcinoma.

Patients and Methods

Database and Study Population

Following approval by the University of Texas Southwestern Institutional Review Board, we identified patients with PaC diagnosed between 2006 and 2012 in the National Cancer Database (NCDB). The NCDB collects information from >1,500 Commission on Cancer centers and captures >70% of incident cancer cases in the United States.11

We abstracted patient characteristics, including age; sex; race/ethnicity; median household income; insurance type; Charlson/Deyo comorbidity score; clinical and pathologic stages according to the 7th edition of the AJCC TNM staging manual; treatment information, including type of procedure, receipt of CT and/or radiotherapy (RT), time to initiating treatment since diagnosis, and sequence of treatment modalities; perioperative data, such as the number of lymph nodes examined and resection margin, postoperative length of stay, 30-day unplanned readmission, 30- and 90-day postoperative mortality; and follow-up outcomes data.

Patients who satisfied the following inclusion criteria were included: aged ≥18 years, diagnosed with clinical stage I or II (T1–3N0–1M0) adenocarcinoma (ICD for Oncology, 3rd edition, morphological codes 814, 821, 825, 831, 844, 845, 847, 850, 856, and 857) involving the head of the pancreas (topographical code C250) who received at least 2 cycles of CT or CRT and underwent curative-intent resection. For patients who received preoperative RT, we included only those undergoing conventionally fractionated radiation dosing. We defined “conventional” as external-beam RT using photon therapy with at least a 6-megavolt energy beam and at least a total radiation dose of 36 Gy (and <66 Gy) administered over at least 15 fractions (and <37 fractions). These criteria encompass commonly used radiation protocols, including the RTOG “standard dose” of 50.4 Gy in 28 fractions used in RTOG 9812, 0020, 0411, and most recently RTOG 1102.1217 We purposefully did not include other radiation modalities and/or lower radiation doses and fractions, because they are not considered standard at present and may overlap with palliative-type RT protocols. We also excluded patients who underwent a palliative resection (denoted under the palliative care variable in NCDB) and those with missing follow-up data.

Statistical Analysis

We defined 2 patient groups: preoperative CT and preoperative CRT. The latter group encompassed patients receiving preoperative CT and preoperative RT. In the CRT group, 67% of patients started RT within 1 or 2 days of receiving their first dose of CT, and overall 95% started RT within 2 weeks of their first CT dose. Although within the CRT group it is not possible to discern patients receiving CRT from those receiving CT followed by CRT, the aforementioned figures indicate those who most likely underwent CRT solely.

We used inverse probability of treatment weight (IPTW) to adjust for observable differences in pretreatment characteristics between patients in the CT and CRT groups.18,19 The probability of treatment, defined as the probability of receiving CT, was estimated using a propensity score. The score was estimated using a multivariable logistic regression model in which we incorporated the following covariates: year of diagnosis; teaching status of the treating hospital; and patient age, sex, race/ethnicity, comorbidity score, income, insurance type, and clinical T and N stages. In this analysis, we included only patients with nonmissing pretreatment covariates. The IPTWs were constructed in a manner to estimate the average treatment effect in the population, rather than in the treated (CT) group. The IPTWs were trimmed to mitigate the effect of influential observations on our findings. The amount of trimming was selected to minimize the mean squared error.20 We evaluated the balance of the pretreatment covariates in the IPTW-adjusted (or weighted) population.

We compared posttreatment covariates in the IPTW-adjusted population between the CT and CRT groups. We used Pearson's χ2 test and the 2-sample t test for categorical and continuous variables, respectively. Both tests were corrected for IPTWs. OS between CT and CRT in the weighted population was also compared. IPTW-adjusted Kaplan-Meier curves were constructed, and the log-rank test and proportional hazards model were used to evaluate the association between OS and type of preoperative therapy. The proportional hazards assumption was plausible according to a plot of the scaled Schoenfeld residuals.

We evaluated for heterogeneity in the effect of preoperative therapy type (CT vs CRT) on OS among levels of pretreatment groups using tests for interaction.21 The treatment effect in the weighted population was estimated within each level of pretreatment groups using the proportional hazards model. Pretreatment groups comprised patient year of diagnosis, age, comorbidity score, and clinical T and N stages.

A sensitivity analysis was conducted to evaluate the influence of missing pretreatment covariates on the treatment effect—covariates that were excluded in the primary analysis. We used multiple imputation by chained equations22,23 to impute missing data for hospital teaching status (n=21; 1.5%), race/ethnicity (n=4; 0.3%), insurance type (n=20; 1.4%), income (n=23; 1.6%), and clinical N stage (n=42; 2.9%). We generated 20 imputed data sets; for each, the propensity score was estimated using multiple logistic regression. Propensity scores from all imputed data sets were then combined according to Rubin's rules.23 The combined propensity score was used to calculate the IPTWs, and OS in the weighted CT and CRT groups was compared using the log-rank test and proportional hazards model.

We conducted our analysis using STATA 14 (see supplemental eAppendix 1, available with this article at JNCCN.org). All statistical tests were 2-sided, and the significance level was .05.

Results

We identified 1,530 patients diagnosed with clinical stage I or II pancreatic head adenocarcinoma who received either CT or CRT and underwent curative-intent pancreatic head resection (Figure 1). A total of 1,326 patients were available for analysis: 710 in the CRT group and 616 in the CT group.

In the unweighted population, CT use increased over time compared with CRT (2006–2008: 12% CT and 21% CRT vs 2011–2012: 56% CT and 45% CRT). Patients in the CT group were more likely to have private or Medicare insurance compared with those in the CRT group (51% and 45% vs 48% and 44%, respectively; P=.01; Table 1) and were also more likely to have higher incomes (above the median: 66% vs 59%; P=.01). Hospital teaching status, patient demographics, and clinical T and N stages did not differ significantly between the groups.

In the weighted population, pathologic findings were more favorable in the CRT group. Patients in the CRT group had a lower pathologic T stage (ypT0/T1/T2: 36% vs 21%; P<.01; Table 2) and had fewer histologically positive lymph nodes (ypN1: 35% vs 59%; P<.01), although the CRT group had a lower lymph node yield compared with the CT group (mean, 13 vs 18 lymph nodes examined; P=.01). Patients receiving CRT also had fewer positive resection margins (14% vs 21%; P=.01). The CRT group, however, had worse postoperative outcomes compared with the CT group, including longer lengths of stay (12 vs 11 days; P=.01), more postoperative

Figure 1.
Figure 1.

CONSORT diagram illustrating patient selection for final cohort.

Abbreviations: CRT, chemoradiotherapy; CT, chemotherapy.

Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 16, 12; 10.6004/jnccn.2018.7068

Table 1.

Pretreatment Characteristics

Table 1.
30-day unplanned readmissions (9% vs 6%; P=.01), and more postoperative 90-day mortality (7% vs 4%; P=.03). More than half of the patients in the CT group received postoperative therapy (57% of which included radiation), whereas only 26% of those in the CRT group received postoperative therapy.

After a median follow-up of 22 months, median OS in the weight-adjusted population was comparable in the CRT and CT groups: 25 and 26 months, respectively (log-rank test: P=.10; Figure 2). OS at 1, 3, and 5 years was 83%, 33%, and 19% for the CRT group and 87%, 37%, and 23% for the CT group. On multivariable analysis, CT was associated with comparable OS to CRT (hazard ratio [HR], 0.89; 95% CI, 0.77–1.02). We noted a similar outcome after imputing missing data (median survival, 25 vs 26 months for CRT vs CT, respectively; P=.06; HR, 0.88; 95% CI, 0.77–1.01). Among patients who received postoperative therapy, the treatment effect remained similar between CRT and CT (median survival, 29 vs 28 months, respectively; P=.97; HR, 1.00; 95% CI, 0.79–1.27).

In the patient subgroup analysis, CT was associated with improved OS among patients aged 55 to 64 years (HR, 0.72; 95% CI, 0.56–0.93; Figure 3) and patients without comorbid conditions (HR, 0.81; 95% CI, 0.68–0.96). No difference was seen in treatment effect related to year of diagnosis, clinical T stage, or clinical N stage. CT had a larger, but non-significant, survival benefit in all year-of-diagnosis categories (P for interaction=.71), clinical T stages 1 and 3 compared with clinical T stage 2 (P=.10), and clinically negative lymph nodes compared with clinically positive nodes (P=.41).

Table 2.

Weight-Adjusted Posttreatment Characteristics

Table 2.

Discussion

In this national cohort of patients with localized pancreatic head adenocarcinoma who underwent preoperative therapy and resection, OS was comparable between preoperative CT and CRT despite more favorable postoperative pathologic findings with CRT. Our results also suggest higher postoperative morbidity and mortality and lower delivery rates of postoperative adjuvant therapy associated with preoperative CRT.

These findings highlight the need for additional randomized clinical trials investigating the role of preoperative therapy and the optimal regimen in resectable PaC. First, we show there is likely clinical equipoise between these preoperative modalities when evaluated in a large matched retrospective cohort. Second, long-term survival with preoperative therapy is comparable to previously reported survival following postoperative CT, the current standard of care for patients with resectable disease.25

Improvement in CT regimens in recent years is yet another compelling reason to invest in trials evaluating preoperative therapy in PaC. For example, multiagent CT regimens, such as FOLFIRINOX24 and gemcitabine/nab-paclitaxel,25,26 which resulted in superior survival outcomes (vs single-agent regimens) in patients with metastatic PaC, have been gradually extrapolated to patients with resectable and borderline resectable disease in the neoadjuvant setting. Currently, CT regimens in emerging preoperative therapy trials are predominantly multiagent regimens. This is clearly highlighted in 2 ongoing multicenter trials evaluating the role of preoperative regimens for PaC: SWOG S1505 (ClinicalTrials.gov identifier: NCT02562716) randomizes patients with

Figure 2.
Figure 2.

Weight-adjusted overall survival for patients who underwent preoperative CRT or CT.

Abbreviations: CRT, chemoradiotherapy; CT, chemotherapy; HR, hazard ratio; IPTW, inverse probability of treatment weight.

Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 16, 12; 10.6004/jnccn.2018.7068

resectable PaC to mFOLFIRINOX or gemcitabine/nab-paclitaxel, and the Alliance for Clinical Trials in Oncology A021501 trial (ClinicalTrials.gov identifier: NCT02839343) compares preoperative mFOLFIRINOX with mFOLFIRINOX + RT for borderline resectable cancer. In our study, accurately evaluating the role of multiagent regimens in the preoperative setting was not possible because CT regimens and treatment duration are not available in NCDB; therefore, it is more than likely that single-agent gemcitabine or 5-fluorouracil were most commonly used, particularly in the period preceding 2011. Nevertheless, exploration of OS between CT and CRT for 2011 and 2012 (the years for which findings from the aforementioned trials in metastatic PaC were reported)24,25 yielded comparable survival profiles (P=.71; Figure 3).

CRT was associated with worse early postoperative outcomes compared with CT. This finding was consistent across all available surrogate variables of morbidity, namely length of stay and 30-day unplanned readmission, as well as 90-day mortality. The underlying reasons for this difference in postoperative outcomes, specifically comparing CT and CRT, have not been explored previously in the literature. Several studies, however, have compared preoperative therapy and pancreatic resection versus upfront resection and have shown no increase in postoperative morbidity and mortality in patients receiving preoperative therapy.2729

Postoperative pathologic findings were more favorable among patients who received CRT. The findings—lower T stage, fewer involved lymph nodes, and fewer positive resection margins—have been shown to predict reduced locoregional recurrence in PaC.30,31 To that end, ongoing advances in RT modalities and refinement of existing protocols are well-poised to leverage these histopathologic benefits, while at the same time offset postoperative morbidity and mortality related to preoperative RT. For example, hypofractionated RT regimens are being more commonly explored in resectable or borderline resectable disease.3234 In our study, we purposefully excluded patients receiving hypofractionated RT regimens because we could not distinguish patients receiving lower radiation amounts (dose or fraction) as part of a hypofractionated definitive courses from patients who required de-escalation or withholding of RT due to toxicity.

Better pathologic response (fewer positive nodes, fewer positive resection margins, and higher rate of downstaging) after CRT did not result, however, in improved OS compared with patients who received CT. Similar findings were noted by Cloyd et al,31 who showed, using a retrospective review of single-institution data, comparable OS between CRT and CT despite lower locoregional recurrence rates after CRT. These findings may reflect the fact that the long-term prognosis of patients with pancreatic adenocarcinoma is driven primarily by impacting systemic dissemination of disease and less so by controlling locoregional disease.3537 In our study, we did not explore the pattern of recurrence after preoperative therapy because the NCDB lacks recurrence data.

A notable finding from our study was the significantly higher proportion of patients in the CT group receiving postoperative adjuvant therapy compared with the CRT group: 51% vs 26%, respectively. One possible explanation is that postoperative CRT was offered to those who underwent preoperative CT (57% of patients who received postoperative therapy had RT); this option is not possible for patients receiving preoperative CRT. Another possible explanation is the increased postoperative morbidity in the CRT group that could have resulted in delay or omission of adjuvant treatment. Merkow et al38 previously reported that the presentation of serious complications after pancreas resection is associated with decreased use of adjuvant therapy (43.6%) compared with patients with no complications (61.8%). A third possible explanation is that the treating physicians may have had a less aggressive inclination to administer postoperative CT to patients in the CRT group, because those patients had, on average, more

Figure 3.
Figure 3.

Forrest plot of the effect of preoperative treatment type, chemotherapy vs chemoradiotherapy, on weight-adjusted HR for death in patient subgroups.

Abbreviation: HR, hazard ratio.

aP value for interaction; the size of each square label is proportional to the precision of the HR estimate.

Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 16, 12; 10.6004/jnccn.2018.7068

favorable pathologic findings: a lower rate of positive margins and positive lymph nodes.

It is important to consider the following limitations when interpreting our study findings. First, the observational nature of the study and the information available in the NCDB did not allow for an intention-to-treat analysis. Our findings were limited to patients who received preoperative therapy and underwent a curative-intent resection. Patients who did not tolerate preoperative therapy as well as those who progressed during therapy or developed distant disease and were denied resection were not represented in the study. Second, selection bias is an unavoidable limitation in this observational study. We used propensity score analysis to minimize selection bias; nonetheless, residual bias cannot be completely excluded. Third, the NCDB does not include information on arterial or venous tumoral involvement or presence of borderline resectable lesions. Therefore, we were limited to using clinical stages I and II (T1–3N0–1M0) as a proxy to resectable disease. Fourth, 44% of patients were missing information on pretreatment CA 19-9 level, which precluded imputing this variable in a meaningful manner. Fifth, although NCDB abstracted the type of resection, there was no information available on vein resection. Last, the CRT group comprised patients who received preoperative CT and RT. Whether those patients underwent CRT or induction CT followed by CRT cannot be accurately verified. Notably, however, 67% of patients started RT within 1 or 2 days of receiving their first CT dose; 95% started RT within 2 weeks of initiating CT. In addition, induction CT followed by CRT was infrequently used in resectable disease during the major part of the study time period.37,39,40 Similarly, NCDB lacks granularity regarding CT agents and the doses and combinations used, and therefore no specific recommendations can be drawn with respect to a specific CT protocol.

Conclusions

Preoperative CT and CRT have similar outcomes for resectable pancreatic head adenocarcinoma. OS is comparable between the 2 modalities, but preoperative CRT is associated with more favorable pathologic features at the cost of a higher rate of postoperative morbidity and mortality. The optimal preoperative therapy should be determined on a patient-by-patient basis, which emphasizes the critical role of multidisciplinary review of all resectable PaC at diagnosis. Advances in our understanding of PaC biology, along with improved systemic therapy and RT, will help personalize this decision.

The authors have disclosed that they have no financial interests, arrangements, affiliations, or commercial interests with the manufacturers of any products discussed in this article or their competitors.

See JNCCN.org for supplemental online content.

References

  • 1.

    Gillen S, Schuster T, Meyer Zum Büschenfelde C et al. . Preoperative/neoadjuvant therapy in pancreatic cancer: a systematic review and meta-analysis of response and resection percentages. PLoS Med 2010;7:e1000267.

    • Search Google Scholar
    • Export Citation
  • 2.

    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;350:12001210.

    • Search Google Scholar
    • Export Citation
  • 3.

    Oettle H, Post S, Neuhaus P et al. . Adjuvant chemotherapy with gemcitabine vs observation in patients undergoing curative-intent resection of pancreatic cancer: a randomized controlled trial. JAMA 2007;297:267277.

    • Search Google Scholar
    • Export Citation
  • 4.

    Neoptolemos JP, Palmer DH, Ghaneh P et al. . Comparison of adjuvant gemcitabine and capecitabine with gemcitabine monotherapy in patients with resected pancreatic cancer (ESPAC-4): a multicentre, open-label, randomised, phase 3 trial. Lancet 2017;389:10111024.

    • Search Google Scholar
    • Export Citation
  • 5.

    Tempero MA, Malafa MP, AL-Hawary M et al. . NCCN Clinical Practice Guidelines in Oncology: Pancreatic Adenocarcinoma. Version 2.2017. Accessed January 22, 2016. To view the most recent version of these guidelines, visit NCCN.org.

    • Search Google Scholar
    • Export Citation
  • 6.

    Tzeng CW, Tran Cao HS, Lee JE et al. . Treatment sequencing for resectable pancreatic cancer: influence of early metastases and surgical complications on multimodality therapy completion and survival. J Gastrointest Surg 2014;18:1624; discussion 24–25.

    • Search Google Scholar
    • Export Citation
  • 7.

    Aloia TA, Aloia TE, Lee JE et al. . Delayed recovery after pancreaticoduodenectomy: a major factor impairing the delivery of adjuvant therapy? J Am Coll Surg 2007;204:347355.

    • Search Google Scholar
    • Export Citation
  • 8.

    Mayo SC, Gilson MM, Herman JM et al. . Management of patients with pancreatic adenocarcinoma: national trends in patient selection, operative management, and use of adjuvant therapy. J Am Coll Surg 2012;214:3345.

    • Search Google Scholar
    • Export Citation
  • 9.

    Mokdad AA, Minter RM, Zhu H et al. . Neoadjuvant therapy followed by resection versus upfront resection for resectable pancreatic cancer: a propensity score matched analysis. J Clin Oncol 2017;35:515522.

    • Search Google Scholar
    • Export Citation
  • 10.

    Klinkenbijl JH, Jeekel J, Sahmoud T et al. . Adjuvant radiotherapy 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;230:776782; discussion 782–784.

    • Search Google Scholar
    • Export Citation
  • 11.

    Boffa DJ, Rosen JE, Mallin K et al. . Using the National Cancer Database for outcomes research: a review. JAMA Oncol 2017;3:17221728.

  • 12.

    Ben-Josef E, Schipper M, Francis IR et al. . A phase I/II trial of intensity modulated radiation (IMRT) dose escalation with concurrent fixed-dose rate gemcitabine (FDR-G) in patients with unresectable pancreatic cancer. Int J Radiat Oncol Biol Phys 2012;84:11661171.

    • Search Google Scholar
    • Export Citation
  • 13.

    Badiyan SN, Olsen JR, Lee AY et al. . Induction chemotherapy followed by concurrent full-dose gemcitabine and intensity-modulated radiation therapy for borderline resectable and locally advanced pancreatic adenocarcinoma. Am J Clin Oncol 2016;39:17.

    • Search Google Scholar
    • Export Citation
  • 14.

    Murphy JD, Adusumilli S, Griffith KA et al. . Full-dose gemcitabine and concurrent radiotherapy for unresectable pancreatic cancer. Int J Radiat Oncol Biol Phys 2007;68:801808.

    • Search Google Scholar
    • Export Citation
  • 15.

    Rich T, Harris J, Abrams R et al. . Phase II study of external irradiation and weekly paclitaxel for nonmetastatic, unresectable pancreatic cancer: RTOG-98-12. Am J Clin Oncol 2004;27:5156.

    • Search Google Scholar
    • Export Citation
  • 16.

    Crane CH, Winter K, Regine WF et al. . Phase II study of bevacizumab with concurrent capecitabine and radiation followed by maintenance gemcitabine and bevacizumab for locally advanced pancreatic cancer: Radiation Therapy Oncology Group RTOG 0411. J Clin Oncol 2009;27:40964102.

    • Search Google Scholar
    • Export Citation
  • 17.

    Rich TA, Winter K, Safran H et al. . Weekly paclitaxel, gemcitabine, and external irradiation followed by randomized farnesyl transferase inhibitor R115777 for locally advanced pancreatic cancer. Onco Targets Ther 2012;5:161170.

    • Search Google Scholar
    • Export Citation
  • 18.

    Rosenbaum PR. Model-based direct adjustment. J Am Stat Assoc 1987;82:387394.

  • 19.

    Guo S, Fraser MW. Propensity Score Analysis: Statistical Methods and Applications. Thousand Oaks, CA: SAGE Publications, Inc.; 2014.

  • 20.

    Potter FJ. A study of procedures to identify and trim extreme sampling weights. In: Proceedings of the Survey Research Methods Section. Washington, DC: American Statistical Association; 1990:225230.

    • Search Google Scholar
    • Export Citation
  • 21.

    Wang R, Lagakos SW, Ware JH et al. . Statistics in medicine—reporting of subgroup analyses in clinical trials. N Engl J Med 2007;357:21892194.

  • 22.

    White IR, Royston P, Wood AM. Multiple imputation using chained equations: Issues and guidance for practice. Stat Med 2011;30:377399.

  • 23.

    Rubin DB. Multiple Imputation for Nonresponse in Surveys. New York, NY: John Wiley & Sons; 2004.

  • 24.

    Conroy T, Desseigne F, Ychou M et al. . FOLFIRINOX versus gemcitabine for metastatic pancreatic cancer. N Engl J Med 2011;364:18171825.

  • 25.

    Von Hoff DD, Ervin T, Arena FP et al. . Increased survival in pancreatic cancer with nab-paclitaxel plus gemcitabine. N Engl J Med 2013;369:16911703.

    • Search Google Scholar
    • Export Citation
  • 26.

    Von Hoff DD, Ramanathan RK, Borad MJ et al. . Gemcitabine plus nab-paclitaxel is an active regimen in patients with advanced pancreatic cancer: a phase I/II trial. J Clin Oncol 2011;29:45484554.

    • Search Google Scholar
    • Export Citation
  • 27.

    Denbo JW, Bruno ML, Cloyd JM et al. . Preoperative chemoradiation for pancreatic adenocarcinoma does not increase 90-day postoperative morbidity or mortality. J Gastrointest Surg 2016;20:19751985.

    • Search Google Scholar
    • Export Citation
  • 28.

    Cooper AB, Parmar AD, Riall TS et al. . Does the use of neoadjuvant therapy for pancreatic adenocarcinoma increase postoperative morbidity and mortality rates? J Gastrointest Surg 2015;19:8086; discussion 86–87.

    • Search Google Scholar
    • Export Citation
  • 29.

    Cho SW, Tzeng CW, Johnston WC et al. . Neoadjuvant radiation therapy and its impact on complications after pancreaticoduodenectomy for pancreatic cancer: analysis of the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP). HPB (Oxford) 2014;16:350356.

    • Search Google Scholar
    • Export Citation
  • 30.

    Groot VP, Rezaee N, Wu W et al. . Patterns, timing, and predictors of recurrence following pancreatectomy for pancreatic ductal adenocarcinoma. Ann Surg 2018;267:936945.

    • Search Google Scholar
    • Export Citation
  • 31.

    Cloyd JM, Crane CH, Koay EJ et al. . Impact of hypofractionated and standard fractionated chemoradiation before pancreatoduodenectomy for pancreatic ductal adenocarcinoma. Cancer 2016;122:26712679.

    • Search Google Scholar
    • Export Citation
  • 32.

    Varadhachary GR, Wolff RA, Crane CH et al. . Preoperative gemcitabine and cisplatin followed by gemcitabine-based chemoradiation for resectable adenocarcinoma of the pancreatic head. J Clin Oncol 2008;26:34873495.

    • Search Google Scholar
    • Export Citation
  • 33.

    Evans DB, Varadhachary GR, Crane CH et al. . Preoperative gemcitabine-based chemoradiation for patients with resectable adenocarcinoma of the pancreatic head. J Clin Oncol 2008;26:34963502.

    • Search Google Scholar
    • Export Citation
  • 34.

    Katz MHG, Shi Q, Ahmad SA et al. . Preoperative modified FOLFIRINOX treatment followed by capecitabine-based chemoradiation for borderline resectable pancreatic cancer: Alliance for Clinical Trials in Oncology trial A021101. JAMA Surg 2016;151:e161137.

    • Search Google Scholar
    • Export Citation
  • 35.

    Iacobuzio-Donahue CA, Fu B, Yachida S et al. . DPC4 gene status of the primary carcinoma correlates with patterns of failure in patients with pancreatic cancer. J Clin Oncol 2009;27:18061813.

    • Search Google Scholar
    • Export Citation
  • 36.

    Gnerlich JL, Luka SR, Deshpande AD et al. . Microscopic margins and patterns of treatment failure in resected pancreatic adenocarcinoma. Arch Surg 2012;147:753760.

    • Search Google Scholar
    • Export Citation
  • 37.

    Tsai S, Evans DB. Therapeutic advances in localized pancreatic cancer. JAMA Surg 2016;151:862868.

  • 38.

    Merkow RP, Bilimoria KY, Tomlinson JS et al. . Postoperative complications reduce adjuvant chemotherapy use in resectable pancreatic cancer. Ann Surg 2014;260:372377.

    • Search Google Scholar
    • Export Citation
  • 39.

    Cloyd JM, Katz MH, Prakash L et al. . Preoperative therapy and pancreatoduodenectomy for pancreatic ductal adenocarcinoma: a 25-year single-institution experience. J Gastrointest Surg 2017;21:164174.

    • Search Google Scholar
    • Export Citation
  • 40.

    Khorana AA, Mangu PB, Berlin J et al. . Potentially curable pancreatic cancer: American Society of Clinical Oncology clinical practice guideline. J Clin Oncol 2016;34:25412556.

    • Search Google Scholar
    • Export Citation

Author contributions: Study concept and methodology: All authors. Analysis and interpretation of results: All authors. Manuscript preparation: Mokdad, Polanco. Critical revision: All authors.

Correspondence: Patricio M. Polanco, MD, Division of Surgical Oncology, Department of Surgery, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, TX 75390-8548. Email: Patricio.Polanco@utsouthwestern.edu
  • View in gallery

    CONSORT diagram illustrating patient selection for final cohort.

    Abbreviations: CRT, chemoradiotherapy; CT, chemotherapy.

  • View in gallery

    Weight-adjusted overall survival for patients who underwent preoperative CRT or CT.

    Abbreviations: CRT, chemoradiotherapy; CT, chemotherapy; HR, hazard ratio; IPTW, inverse probability of treatment weight.

  • View in gallery

    Forrest plot of the effect of preoperative treatment type, chemotherapy vs chemoradiotherapy, on weight-adjusted HR for death in patient subgroups.

    Abbreviation: HR, hazard ratio.

    aP value for interaction; the size of each square label is proportional to the precision of the HR estimate.

  • 1.

    Gillen S, Schuster T, Meyer Zum Büschenfelde C et al. . Preoperative/neoadjuvant therapy in pancreatic cancer: a systematic review and meta-analysis of response and resection percentages. PLoS Med 2010;7:e1000267.

    • Search Google Scholar
    • Export Citation
  • 2.

    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;350:12001210.

    • Search Google Scholar
    • Export Citation
  • 3.

    Oettle H, Post S, Neuhaus P et al. . Adjuvant chemotherapy with gemcitabine vs observation in patients undergoing curative-intent resection of pancreatic cancer: a randomized controlled trial. JAMA 2007;297:267277.

    • Search Google Scholar
    • Export Citation
  • 4.

    Neoptolemos JP, Palmer DH, Ghaneh P et al. . Comparison of adjuvant gemcitabine and capecitabine with gemcitabine monotherapy in patients with resected pancreatic cancer (ESPAC-4): a multicentre, open-label, randomised, phase 3 trial. Lancet 2017;389:10111024.

    • Search Google Scholar
    • Export Citation
  • 5.

    Tempero MA, Malafa MP, AL-Hawary M et al. . NCCN Clinical Practice Guidelines in Oncology: Pancreatic Adenocarcinoma. Version 2.2017. Accessed January 22, 2016. To view the most recent version of these guidelines, visit NCCN.org.

    • Search Google Scholar
    • Export Citation
  • 6.

    Tzeng CW, Tran Cao HS, Lee JE et al. . Treatment sequencing for resectable pancreatic cancer: influence of early metastases and surgical complications on multimodality therapy completion and survival. J Gastrointest Surg 2014;18:1624; discussion 24–25.

    • Search Google Scholar
    • Export Citation
  • 7.

    Aloia TA, Aloia TE, Lee JE et al. . Delayed recovery after pancreaticoduodenectomy: a major factor impairing the delivery of adjuvant therapy? J Am Coll Surg 2007;204:347355.

    • Search Google Scholar
    • Export Citation
  • 8.

    Mayo SC, Gilson MM, Herman JM et al. . Management of patients with pancreatic adenocarcinoma: national trends in patient selection, operative management, and use of adjuvant therapy. J Am Coll Surg 2012;214:3345.

    • Search Google Scholar
    • Export Citation
  • 9.

    Mokdad AA, Minter RM, Zhu H et al. . Neoadjuvant therapy followed by resection versus upfront resection for resectable pancreatic cancer: a propensity score matched analysis. J Clin Oncol 2017;35:515522.

    • Search Google Scholar
    • Export Citation
  • 10.

    Klinkenbijl JH, Jeekel J, Sahmoud T et al. . Adjuvant radiotherapy 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;230:776782; discussion 782–784.

    • Search Google Scholar
    • Export Citation
  • 11.

    Boffa DJ, Rosen JE, Mallin K et al. . Using the National Cancer Database for outcomes research: a review. JAMA Oncol 2017;3:17221728.

  • 12.

    Ben-Josef E, Schipper M, Francis IR et al. . A phase I/II trial of intensity modulated radiation (IMRT) dose escalation with concurrent fixed-dose rate gemcitabine (FDR-G) in patients with unresectable pancreatic cancer. Int J Radiat Oncol Biol Phys 2012;84:11661171.

    • Search Google Scholar
    • Export Citation
  • 13.

    Badiyan SN, Olsen JR, Lee AY et al. . Induction chemotherapy followed by concurrent full-dose gemcitabine and intensity-modulated radiation therapy for borderline resectable and locally advanced pancreatic adenocarcinoma. Am J Clin Oncol 2016;39:17.

    • Search Google Scholar
    • Export Citation
  • 14.

    Murphy JD, Adusumilli S, Griffith KA et al. . Full-dose gemcitabine and concurrent radiotherapy for unresectable pancreatic cancer. Int J Radiat Oncol Biol Phys 2007;68:801808.

    • Search Google Scholar
    • Export Citation
  • 15.

    Rich T, Harris J, Abrams R et al. . Phase II study of external irradiation and weekly paclitaxel for nonmetastatic, unresectable pancreatic cancer: RTOG-98-12. Am J Clin Oncol 2004;27:5156.

    • Search Google Scholar
    • Export Citation
  • 16.

    Crane CH, Winter K, Regine WF et al. . Phase II study of bevacizumab with concurrent capecitabine and radiation followed by maintenance gemcitabine and bevacizumab for locally advanced pancreatic cancer: Radiation Therapy Oncology Group RTOG 0411. J Clin Oncol 2009;27:40964102.

    • Search Google Scholar
    • Export Citation
  • 17.

    Rich TA, Winter K, Safran H et al. . Weekly paclitaxel, gemcitabine, and external irradiation followed by randomized farnesyl transferase inhibitor R115777 for locally advanced pancreatic cancer. Onco Targets Ther 2012;5:161170.

    • Search Google Scholar
    • Export Citation
  • 18.

    Rosenbaum PR. Model-based direct adjustment. J Am Stat Assoc 1987;82:387394.

  • 19.

    Guo S, Fraser MW. Propensity Score Analysis: Statistical Methods and Applications. Thousand Oaks, CA: SAGE Publications, Inc.; 2014.

  • 20.

    Potter FJ. A study of procedures to identify and trim extreme sampling weights. In: Proceedings of the Survey Research Methods Section. Washington, DC: American Statistical Association; 1990:225230.

    • Search Google Scholar
    • Export Citation
  • 21.

    Wang R, Lagakos SW, Ware JH et al. . Statistics in medicine—reporting of subgroup analyses in clinical trials. N Engl J Med 2007;357:21892194.

  • 22.

    White IR, Royston P, Wood AM. Multiple imputation using chained equations: Issues and guidance for practice. Stat Med 2011;30:377399.

  • 23.

    Rubin DB. Multiple Imputation for Nonresponse in Surveys. New York, NY: John Wiley & Sons; 2004.

  • 24.

    Conroy T, Desseigne F, Ychou M et al. . FOLFIRINOX versus gemcitabine for metastatic pancreatic cancer. N Engl J Med 2011;364:18171825.

  • 25.

    Von Hoff DD, Ervin T, Arena FP et al. . Increased survival in pancreatic cancer with nab-paclitaxel plus gemcitabine. N Engl J Med 2013;369:16911703.

    • Search Google Scholar
    • Export Citation
  • 26.

    Von Hoff DD, Ramanathan RK, Borad MJ et al. . Gemcitabine plus nab-paclitaxel is an active regimen in patients with advanced pancreatic cancer: a phase I/II trial. J Clin Oncol 2011;29:45484554.

    • Search Google Scholar
    • Export Citation
  • 27.

    Denbo JW, Bruno ML, Cloyd JM et al. . Preoperative chemoradiation for pancreatic adenocarcinoma does not increase 90-day postoperative morbidity or mortality. J Gastrointest Surg 2016;20:19751985.

    • Search Google Scholar
    • Export Citation
  • 28.

    Cooper AB, Parmar AD, Riall TS et al. . Does the use of neoadjuvant therapy for pancreatic adenocarcinoma increase postoperative morbidity and mortality rates? J Gastrointest Surg 2015;19:8086; discussion 86–87.

    • Search Google Scholar
    • Export Citation
  • 29.

    Cho SW, Tzeng CW, Johnston WC et al. . Neoadjuvant radiation therapy and its impact on complications after pancreaticoduodenectomy for pancreatic cancer: analysis of the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP). HPB (Oxford) 2014;16:350356.

    • Search Google Scholar
    • Export Citation
  • 30.

    Groot VP, Rezaee N, Wu W et al. . Patterns, timing, and predictors of recurrence following pancreatectomy for pancreatic ductal adenocarcinoma. Ann Surg 2018;267:936945.

    • Search Google Scholar
    • Export Citation
  • 31.

    Cloyd JM, Crane CH, Koay EJ et al. . Impact of hypofractionated and standard fractionated chemoradiation before pancreatoduodenectomy for pancreatic ductal adenocarcinoma. Cancer 2016;122:26712679.

    • Search Google Scholar
    • Export Citation
  • 32.

    Varadhachary GR, Wolff RA, Crane CH et al. . Preoperative gemcitabine and cisplatin followed by gemcitabine-based chemoradiation for resectable adenocarcinoma of the pancreatic head. J Clin Oncol 2008;26:34873495.

    • Search Google Scholar
    • Export Citation
  • 33.

    Evans DB, Varadhachary GR, Crane CH et al. . Preoperative gemcitabine-based chemoradiation for patients with resectable adenocarcinoma of the pancreatic head. J Clin Oncol 2008;26:34963502.

    • Search Google Scholar
    • Export Citation
  • 34.

    Katz MHG, Shi Q, Ahmad SA et al. . Preoperative modified FOLFIRINOX treatment followed by capecitabine-based chemoradiation for borderline resectable pancreatic cancer: Alliance for Clinical Trials in Oncology trial A021101. JAMA Surg 2016;151:e161137.

    • Search Google Scholar
    • Export Citation
  • 35.

    Iacobuzio-Donahue CA, Fu B, Yachida S et al. . DPC4 gene status of the primary carcinoma correlates with patterns of failure in patients with pancreatic cancer. J Clin Oncol 2009;27:18061813.

    • Search Google Scholar
    • Export Citation
  • 36.

    Gnerlich JL, Luka SR, Deshpande AD et al. . Microscopic margins and patterns of treatment failure in resected pancreatic adenocarcinoma. Arch Surg 2012;147:753760.

    • Search Google Scholar
    • Export Citation
  • 37.

    Tsai S, Evans DB. Therapeutic advances in localized pancreatic cancer. JAMA Surg 2016;151:862868.

  • 38.

    Merkow RP, Bilimoria KY, Tomlinson JS et al. . Postoperative complications reduce adjuvant chemotherapy use in resectable pancreatic cancer. Ann Surg 2014;260:372377.

    • Search Google Scholar
    • Export Citation
  • 39.

    Cloyd JM, Katz MH, Prakash L et al. . Preoperative therapy and pancreatoduodenectomy for pancreatic ductal adenocarcinoma: a 25-year single-institution experience. J Gastrointest Surg 2017;21:164174.

    • Search Google Scholar
    • Export Citation
  • 40.

    Khorana AA, Mangu PB, Berlin J et al. . Potentially curable pancreatic cancer: American Society of Clinical Oncology clinical practice guideline. J Clin Oncol 2016;34:25412556.

    • Search Google Scholar
    • Export Citation
All Time Past Year Past 30 Days
Abstract Views 33 0 0
Full Text Views 601 103 11
PDF Downloads 301 90 4
EPUB Downloads 0 0 0