eFigure. Study Flowchart
eTable 1. Criteria for defining borderline resectable and locally advanced pancreatic adenocarcinoma
eTable 2. MD Anderson Cancer Center (MDACC) classification of borderline resectable disease
eTable 3. General characteristics of the study population stratified by resectability status (according to NCCN guidelines) at the time of diagnosis (n=680)
eTable 4. Reasons For Not Completing the Planned Primary Chemotherapy Regimen
eTable 5. Surgical, pathological and post-operative characteristics of the patients who completed primary chemotherapy and received resection (n=90)
Customize your JAMA Network experience by selecting one or more topics from the list below.
Identify all potential conflicts of interest that might be relevant to your comment.
Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.
Err on the side of full disclosure.
If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.
Not all submitted comments are published. Please see our commenting policy for details.
Maggino L, Malleo G, Marchegiani G, et al. Outcomes of Primary Chemotherapy for Borderline Resectable and Locally Advanced Pancreatic Ductal Adenocarcinoma. JAMA Surg. 2019;154(10):932–942. doi:10.1001/jamasurg.2019.2277
What are the outcomes of patients with borderline resectable and locally advanced pancreatic ductal adenocarcinoma receiving primary chemotherapy?
In this study including 680 patients with new diagnoses of pancreatic ductal adenocarcinoma, the rates of chemotherapy receipt and completion were 92.9% and 71.6%, respectively. Rate of surgical resection was 24% in patients with borderline resectable disease and 9% in patients with locally advanced disease; the median survival was 12.8 months in the whole cohort and in excess of 35 months in patients who underwent surgical resection.
This article provides real-world outcomes of primary chemotherapy in an unselected cohort of patients with borderline resectable and locally advanced pancreatic ductal adenocarcinoma.
Chemotherapy is the recommended induction strategy in borderline resectable and locally advanced pancreatic ductal adenocarcinoma. However, the associated results on an intention-to-treat basis are poorly understood.
To investigate pragmatically the treatment compliance, conversion to surgery, and survival outcomes of patients with borderline resectable and locally advanced pancreatic ductal adenocarcinoma undergoing primary chemotherapy.
Design, Setting, and Participants
This prospective study took place in a national referral center for pancreatic diseases in Italy. Consecutive patients with borderline resectable and locally advanced pancreatic ductal adenocarcinoma were enrolled at the time of diagnosis (January 2013 through December 2015) and followed up to June 2018.
The chemotherapy regimen, assigned based on multidisciplinary evaluation, was delivered either at a hub center or at spoke centers. By convention, primary chemotherapy was considered completed after 6 months. After restaging, surgical candidates were selected based on radiologic and biochemical response. All surgeries were carried out at the hub center.
Main Outcomes and Measures
Rates of receipt and completion of chemotherapy, rates of conversion to surgery, and disease-specific survival.
Of 680 patients, 267 (39.3%) had borderline resectable and 413 (60.7%) had locally advanced pancreatic ductal adenocarcinoma. Overall, 66 patients (9.7%) were lost to follow-up. The rate of chemotherapy receipt was 92.9% (n = 570). The chemotherapeutic regimens most commonly used included FOLFIRINOX (fluorouracil, leucovorin, oxaliplatin, and irinotecan) (260 [45.6%]) and gemcitabine plus nanoparticle albumin-bound–paclitaxel (123 [21.6%]). Nineteen patients (3.3%) receiving chemotherapy died within 6 months, mainly for disease progression. The treatment completion rate was 71.6% (408 of 570). The overall rate of resection was 15.1% (93 of 614) (borderline resectable, 60 of 249 [24.1%]; locally advanced, 33 of 365 [9%]; resection:exploration ratio, 63.3%). Independent predictors of resection were age, borderline resectable disease, chemotherapy completion, radiologic response, and biochemical response. The median survival for the whole cohort was 12.8 (95% CI, 11.7-13.9) months. Factors independently associated with survival were completion of chemotherapy, receipt of complementary radiation therapy, and resection. In patients who underwent resection, the median survival was 35.4 (95% CI, 27.0-43.7) months for initially borderline resectable and 41.8 (95% CI, 27.5-56.1) months for initially locally advanced disease. No pretreatment and posttreatment factors were associated with survival after pancreatectomy.
Conclusions and Relevance
This pragmatic observational cohort study with an intention-to-treat design provides real-world evidence of outcomes associated with the most current primary chemotherapy regimens used for borderline resectable and locally advanced pancreatic ductal adenocarcinoma.
In the context of localized pancreatic ductal adenocarcinoma (PDAC), chemotherapy is the preferential induction strategy for locally advanced (LA) and borderline resectable (BR) disease.1-3 The recent steady uptake of primary chemotherapy has been fostered by newly introduced combinations, such as FOLFIRINOX (fluorouracil, leucovorin, oxaliplatin, and irinotecan) and gemcitabine plus nanoparticple albumin-bound (nab)–paclitaxel, which demonstrated a high response rate in randomized clinical trials of patients with metastatic disease.4,5 Although the translation of these regimens to earlier stages was empirical, several case series reported encouraging outcomes, with resection rates after primary chemotherapy ranging from 60% to 78%.6-11 However, these data are based on the patient pool selected for surgical exploration and not on the overall number of individuals with new diagnoses. At a collective level, a sizable amount of patients are indeed elderly or present with certain conditions (eg, jaundice, poor hematologic function of poor performance status) that hinder treatment eligibility and enrollment in controlled research protocols.12-14 Furthermore, a limitation of most studies is the inability to determine how many patients receiving primary chemotherapy never qualified for surgery because of treatment discontinuation, disease progression, or death.15-22
This prospective cohort study enrolled consecutive patients with new diagnoses of BR and LA PDAC. The analysis is limited to a recent 3-year period, with the aim to evaluate pragmatically the rate of chemotherapy receipt, the treatment compliance, the rate of surgery, and survival outcomes.
This study was approved by the Verona Ethical Committee (PAD-R, n.1101CESC). Patients provided written informed consent and were prospectively enrolled between January 2013 and December 2015. Inclusion criteria were diagnosis of previously untreated BR or LA PDAC according to the 2013 National Comprehensive Cancer Network (NCCN) guidelines version 1 (eTable 1 in the Supplement).23 Notably, criteria for defining BR and LA PDAC did not change substantially in subsequent guidelines updates. Age and potentially reversible concurrent conditions (including jaundice, inadequate liver and hematologic function, performance status ECOG [Eastern Cooperative Oncology Group] ≥2) did not represent a priori exclusion criteria. Exclusion criteria were distant metastases and anatomically resectable disease (eFigure in the Supplement). The reporting of this study is compliant with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
Demographic and clinical details were collected. Baseline cross-sectional imaging was always reviewed. Additional investigations were requested whenever the available data were of insufficient quality or inconclusive. For BR PDAC, radiologic staging was integrated with the MD Anderson Cancer Center (MDACC) classification (eTable 2 in the Supplement).24 Controversial cases were discussed in a multidisciplinary conference and reconciled. Tissue diagnosis was always obtained before treatment delivery. Chemotherapy was administered either in Verona or at spoke institutions, according to the area of residence or patient preference. In the latter case, patients were assisted with regimen recommendation, although the final decision on the chemotherapy schedule was at the treating oncologic team’s discretion. Information on the ongoing treatment was continuously acquired from spoke centers by dedicated data managers. Protocol-specified treatment modifications were undertaken when significant toxic effects occurred. Common Terminology Criteria for Adverse Events, version 4.0, was used to define and grade toxicity.25
Restaging was normally scheduled on a 3-month basis, multidisciplinary reevaluation of each case was performed at the hub center. After 6 months, based on extrapolation of adjuvant therapy trials, primary chemotherapy was considered completed.26 Tumor response was assessed using the Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1.27 Biochemical response was defined as a cancer antigen 19-9 (CA19-9) decrease more than 50% of the baseline value.13 Patients with jaundice and unmeasurable levels were excluded from CA19-9 analysis. Surgical exploration was indicated when a macroscopically radical resection was considered predictable. Cancer antigen 19-9 response and performance status were also considered in the decision process. Surgical procedures were all carried out at our center. Determinants of intraoperative resectability were absence of distant metastases, reconstructible superior mesenteric vein/portal vein, and no need for superior mesenteric artery resection. Arterial resections included common hepatic artery resection (performed only when a short segment was involved) and Appleby procedure (performed for solid tumor contact with celiac trunk, without aortic involvement). Patients who were considered ineligible surgical candidates underwent oncologic reevaluation for possible treatment prosecution. Follow-up was carried out on a 3-month basis through outpatient visits or telephone up to June 2018.
This study evaluated the rates of receipt and completion of primary chemotherapy, the rates of surgery, and the disease-specific survival (DSS). All results were stratified by age (dichotomized at age 75 years, representing the cut-off for patient inclusion in randomized clinical trials of chemotherapy),4,12 resectability status at diagnosis (BR/LA), MDACC classification, chemotherapy regimen and completion, and, in patients who completed their treatment, RECIST and CA19-9 response. All the analyses were carried out both in patients enrolled at baseline and in the subgroup completing the treatment schedule. For patients who have undergone resection, intraoperative variables, pathological features, and 90-day complications were also registered. Tumors were staged according to the American Joint Committee on Cancer classification, 8th edition.28 Margin status was defined based on the presence of tumor cells within 1 mm from any resection margin.
The a priori sample size calculation (609 patients, 80% power at the 5% significance level) was based on the expected resection rate after chemotherapy in patients with BR and LA PDAC at the time of study design and is described in eMethods in the Supplement.24,29,30
Continuous variables were expressed as medians with interquartile range and compared using the Mann-Whitney test; categorical variables were expressed as frequencies with percentages and compared using the χ2 or Fisher exact test. All tests were 2-tailed. Multivariable analysis was carried out using binary logistic regression (backward stepwise elimination technique, Wald method, P < .05 for entry, P > .10 for removal).
The DSS was calculated from both the date of diagnosis and surgery to the date of last follow-up or death from disease, using the Kaplan-Meier method. Pairwise differences in survival were assessed using the log-rank test. The analysis of DSS predictors was carried out using the proportional hazard Cox regression (backward stepwise elimination technique, Wald method, P < .05 for entry, P > .10 for removal). Statistical significance was determined by a 2-sided P value less than .05. Data were analyzed using SPSS software version 25 (IBM).
Of 680 patients with new diagnoses who were enrolled (eFigure in the Supplement), 267 (39.3%) had BR and 413 (60.7%) had LA PDAC. Baseline characteristics stratified by resectability status are in eTable 3 in the Supplement.
Overall, 66 patients (9.7%) were lost to follow-up, leaving a cohort of 614 patients. Among these, 44 (7.1%) received only supportive care owing to clinical deterioration or because an acute condition present at baseline evaluation did not improve. This was more frequent in elderly patients (15 [3.1%] in individuals aged ≤75 years vs 29 [23.2%] in individuals aged >75 years; P < .001; odds ratio [OR], 9.546; 95% CI, 4.930-18.484), without significant differences based on resectability status (16 individuals with BR [6.4%] vs 28 individuals with LA [7.7%] PDAC; P = .67; OR, 1.210; 95% CI, 0.640-2.287). Within the BR subgroup, there was a significant difference in the frequency of patients receiving only supportive care based on the MDACC classification (2 [1.2%] for class A, 3 [7.7%] for class B, 11 [25.6%] for class C; P = .001; class B: OR, 6.875; 95% CI, 1.108-42.651; class C: OR, 28.359; 95% CI, 5.998-134.079). The most commonly used regimens were FOLFIRINOX/modified FOLFIRINOX (260 [45.6%]; modified FOLFIRINOX was administered without fluorouracil on day 1; referred to as FOLFIRINOX from now on) and gemcitabine plus nab-paclitaxel (123 [21.6%]). Chemotherapy regimens and dosage stratified by age, resectability status, and MDACC classification are shown in Table 1. Throughout the whole disease course, 193 patients (33.9%) received a second-line treatment, and 132 patients (23.2%) underwent additional radiation therapy.
The treatment completion rate was 71.6% (408 of 570); stratification by resectability status, patient age, chemotherapy regimen, and MDACC classification is displayed in Table 2. At multivariable analysis, patients 75 years and younger (OR, 1.863; 95% CI, 1.109-3.129; P = .02) and those receiving platinum-based regimens were more likely to complete the 6-month schedule (FOLFIRINOX: OR, 3.650; 95% CI, 2.094-6.363; P = .001 and gemcitabine and oxaliplatin: OR, 2.589; 95% CI, 1.312-5.111; P = .006). Reasons for treatment discontinuation are summarized in eTable 4 in the Supplement. Among the 162 patients discontinuing chemotherapy, 29 (17.9%) switched to a different treatment protocol, while the remaining 133 (82.1%) never resumed any treatment. Nineteen patients died during chemotherapy (18 for disease progression, 1 for sepsis).
The RECIST response was analyzed in 408 patients completing the planned chemotherapy cycles: 125 patients (30.6%) had a partial response, 128 (31.4%) had a stable disease, and 155 (38.0%) had a disease progression. Cancer antigen 19-9 response occurred in 145 of 318 patients (45.6%) with initially elevated values.
Overall, 147 of 614 patients (23.9%) underwent surgical exploration, with 93 individuals (15.1%) undergoing resection. The resection rates by NCCN clustering were 24.1% (60 of 249) of individuals with BR and 9.0% (33 of 365) of individuals with LA PDAC. The resection/exploration ratios were 63.3% in the overall cohort, 76.9% in patients with BR PDAC, and 47.8% in patients with LA PDAC. The univariable analyses of factors associated with surgical exploration and resection are displayed in Table 3 and Table 4, separately for BR and LA PDAC. The highest resection rate was reached in patients 75 years or younger with BR/MDACC classification A disease who completed primary chemotherapy with FOLFIRINOX (27 of 51 [52.9%]). A multivariable model based on the 570 patients receiving chemotherapy identified age 75 years or younger, (OR, 5.599; 95% CI, 1.681-18.646; P = .005), BR PDAC (OR, 3.275; 95% CI, 2.014-5.326; P < .001), and completion of chemotherapy (OR, 14.323; 95% CI, 4.396-46.692; P < .001) as factors associated with resection. In a subgroup analysis investigating the 408 patients completing treatment, BR PDAC (OR, 7.216; 95% CI, 3.787-13.751; P = .001) and CA19-9 response (OR, 5.010; 95% CI, 2.309-10.873; P = .001) were independently associated with resection, whereas stable and progressive disease was negatively associated with resection rate compared with responders (OR, 0.349; 95% CI, 0.185-0.661; P = .001 for stable disease and OR, 0.005; 95% CI, 0.001-0.037; P = .001 for progressive disease).
Among 54 patients who underwent exploratory but not resection surgery, 37 (68.5%) were deemed ineligible for local resection; unexpected liver metastasis or peritoneal carcinomatosis were found in 10 and 7 patients, respectively. Surgical, pathological, and clinical characteristics of patients who had received resection surgery are displayed in eTable 5 in the Supplement. The R0 resection rate was 57.8%, with no significant differences between patients with BR/LA PDAC and chemotherapy regimens (data not shown).
The median follow-up was 12.5 (interquartile range, 7.5-19.3) months in the overall cohort and 25.1 (interquartile range, 20.0-33.1) months in censored cases. The median DSS of the entire population was 12.8 (95% CI, 11.7-13.9) months. The univariable analysis of factors associated with DSS is shown in Table 5.
A Cox model based on the 570 patients receiving chemotherapy showed treatment completion (hazard ratio [HR], 0.390; 95% CI, 0.319-0.477; P < .001), complementary radiation therapy (HR, 0.700; 95% CI, 0.563-0.872; P = .001), and resection (HR, 0.178; 95% CI, 0.125-0.253; P < .001) to be associated with prolonged survival. For the subgroup of 408 patients completing chemotherapy, radiation therapy (HR, 0.736; 95% CI, 0.577-0.940; P = .01) and resection (HR, 0.185; 95% CI, 0.129-0.267; P < .001) were the only independent prognostic factors.
In patients who had undergone surgical resection, the median DSS was 35.4 (95% CI, 27.0-43.7) months in BR and 41.8 (95% CI, 27.5-56.1) months in initially LA PDAC. No pretreatment and posttreatment factors, including surgical and pathological variables, were associated with survival on multivariable analysis.
Despite accruing evidence demonstrating its effectiveness in controlled settings,12,13 the results of primary chemotherapy for PDAC in real-world conditions remain unclear. This prospective observational study examined the clinical course and the outcomes of patients with LA and BR PDAC observed at a referral center where the PDAC treatment policy having been historically compliant with NCCN guidelines, particularly regarding chemotherapy as opposed to upfront surgery in BR PDAC. The analysis is limited to a recent 3-year period, providing an updated picture of the association of primary chemotherapy with outcomes in the era of new drug combinations.
The results of this study are multifold. First, 7% of patients with new diagnoses received only supportive care owing to clinical deterioration or concurrent acute conditions. This was particularly frequent in individuals older than 75 years (one-fourth of whom did not receive chemotherapy), and it is consistent with a study of patients undergoing palliative biliary drainage for malignant obstruction, demonstrating a 30-day mortality rate of 19.6%, significantly increased among older patients.31 Remarkably, access to chemotherapy was similar for patients with BR and LA PDAC. The most commonly used regimens were FOLFIRINOX (45.6%) and gemcitabine plus nab-paclitaxel (21.6%), confirming the widespread application of these combinations, despite no randomized phase III clinical trial in nonmetastatic PDAC has been completed thus far, to our knowledge.1 The frequency of the applied regimens did not differ between BR and LA PDAC, indicating that the BR/LA distinction did not prompt any specific treatment strategy. As expected, older patients were more likely to receive gemcitabine monotherapy, while younger patients received FOLFIRINOX more frequently.
Second, 28.4% of patients receiving chemotherapy did not complete the treatment schedule, with most of them never resuming any other treatment. Of these, more than 50% were older than 75 years. Interestingly, patients undergoing FOLFIRINOX were more likely to complete the 6-month schedule. The type of chemotherapy was also associated with reasons for treatment discontinuation. While disease progression was more common in gemcitabine-based regimens, toxicity was more likely among patients receiving platinum-based chemotherapy. Nineteen patients died during chemotherapy, mainly for rapid disease progression.
The third key finding was that in an analysis including the baseline cohort, one-fourth of patients qualified for surgical exploration, with 15.1% finally undergoing surgical resection (24.1% in individuals with BR PDAC and 9.0% in individuals with LA PDAC). When focusing on patients completing treatment, the resection rate increased up to 22.1% (34.5% in individuals with BR PDAC and 13.1% in individuals with LA PDAC). These results are somewhat in contrast with previous studies. Although the outcomes of patients receiving primary chemotherapy appear substantially heterogeneous,18,32-37 the rates of surgical conversion seemed more favorable, particularly for patients with BR PDAC. In a study on gemcitabine-based chemoradiation, Takahashi et al16 reported a resection rate of 54%. Similarly, in a 2019 retrospective review of 151 patients by Javed et al,20 63.4% of patients underwent resection. Reni et al13 reported a resection rate of 55.6% in patients with BR and 13.9% in patients with LA PDAC treated with gemcitabine-based combinations at a single center. Finally, a French prospective study on 77 patients with FOLFIRINOX-treated LA PDAC showed a resection rate of 36%,38 and in a 2018 analysis by Gemenetzis et al22 on patients with LA PDAC receiving different chemotherapy regimens, the resection rate was 20%. These discrepancies are likely based on the pragmatic design of this study, prospectively analyzing consecutive patients with new diagnoses regardless of age, performance status, and concurrent conditions. In fact, in individuals with a favorable profile (age <75 years, BR/MDACC classification A disease, completing chemotherapy with FOLFIRINOX) the resection rate was 52.9%, in line with other literature reports.13,14,20 Furthermore, when considering at the denominator the number of patients who underwent surgical exploration, and not the whole cohort of patients with new diagnoses, the resection/exploration ratio was 76.9% for patients with BR PDAC and 47.8% for patients with LA PDAC, again comparable with previous series.10,11
These findings should be interpreted in the context of our institutional and nationwide practice. The indication for primary chemotherapy was always given at the authors’ institution, a national referral center for pancreatic diseases, in compliance with the current treatment paradigm of PDAC. However, most patients received chemotherapy at their local institutions, resulting in a further degree of heterogeneity that might have influenced our results to some extent. Nonetheless, this replicates a realistic scenario in which primary chemotherapy is coordinated through a high-volume center and is consistent with the current oncologic practices because there is limited evidence to recommend specific regimens off study.1,39 Also of relevance is that an array of variables including age, NCCN patient clustering, completion of the planned treatment, RECIST response, and CA19-9 can be used to predict resection after chemotherapy.
When looking at prognostic factors in the whole cohort, treatment completion, and not a specific regimen, was associated with DSS, which on average was slightly in excess of 1 year. Although FOLFIRINOX-treated patients achieved the best results, the value of chemotherapy regimen as a surrogate end point for survival could not be proven. Even the BR/LA distinction at baseline was not an independent prognostic factor. While patients with BR PDAC were more likely to undergo surgery, the 3-fold increase in resection rate was insufficient to affect the collective survival outcome. This is consistent with previous reports, showing that, except for a greater possibility to undergo resection after chemotherapy, BR and LA PDAC do not seem to constitute distinct prognostic groups.13 Remarkably, CA19-9 response was not a significant prognostic factor on multivariable analysis. The prognostic effect of radiation therapy is most likely due to selection bias. Patients whose disease had not progressed after primary chemotherapy had an option for additional radiation therapy, with the intention of possible subsequent surgical resection.22 Resection was the most powerful prognostic factor, being the median DSS of 35.4 months in patients with BR PDAC and 41.8 months in patients with initially LA PDAC. In the subset of patients who underwent resection, no association between pretreatment and posttreatment factors and DSS was identified.
Our analysis has remarkable strengths compared with other case series or pooled analyses of phase I/II trials. Consecutive patients were prospectively enrolled during a short time, and the numeric consistency was very robust, being the largest analysis of a chemotherapy-treated cohort with an intention-to-treat design, to our knowledge.
This study also has limitations. First, enrollment was restricted to patients with BR and LA PDAC, excluding those with resectable disease. The role of primary chemotherapy in this subgroup is particularly controversial, and our institutional policy favors upfront resection in patients without high-risk features. Notably, comparing primary chemotherapy vs upfront surgery in resectable disease was outside the scope of this analysis. Second, after multidisciplinary reassessment, the decision to recommend surgical exploration was individualized based on the degree of radiological and biochemical response, comorbidity profile, and performance status. Because the proportion of patients undergoing surgical resection substantially varies depending on these factors, the reported resection rate reflects the distinctive characteristics of our practice and might not be universally generalizable.
This study shows that 7% of patients with new diagnoses of BR and LA PDAC received only supportive care. In patients receiving chemotherapy, the completion rate was 72%, with significant differences based on age and treatment regimen. One-quarter of patients with BR PDAC and 9% of patients with LA PDAC ultimately received surgical resection. On an intention-to-treat basis, younger age, BR status, and treatment completion were associated with resection. Chemotherapy completion, complementary radiation therapy, and resection were prognostic factors. In the subset of patients who had undergone resection, no pretreatment and posttreatment factor were associated with survival. Although randomized clinical trials are rightfully valued as the primary source of research data that drives changes in practice and directs evidence-based care, the present observational study provides information that can better characterize pragmatically the compliance and outcomes of primary chemotherapy for BR and LA PDAC.
Corresponding Author: Roberto Salvia, MD, PhD (email@example.com), and Giuseppe Malleo MD, PhD (firstname.lastname@example.org), Unit of General and Pancreatic Surgery, Department of Surgery and Oncology, University of Verona Hospital Trust, G.B. Rossi Hospital, P.Le L.A. Scuro 10; 37134 Verona, Italy.
Accepted for Publication: May 17, 2019.
Published Online: July 24, 2019. doi:10.1001/jamasurg.2019.2277
Author Contributions: Drs Malleo and Salvia had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Drs Maggino and Malleo served as co–first authors, each with equal contribution to the manuscript.
Concept and design: Maggino, Malleo, Marchegiani, Landoni, Casetti, Auriemma, Bassi, Salvia.
Acquisition, analysis, or interpretation of data: Maggino, Malleo, Marchegiani, Viviani, Nessi, Ciprani, Esposito, Tuveri, Paiella, Casciani, Sereni, Binco, Bonamini, Secchettin, Merz, Simionato, Zecchetto, D’Onofrio, Melisi, Bassi.
Drafting of the manuscript: Maggino, Malleo, Viviani, Nessi, Ciprani, Paiella, Casciani, Sereni, Binco, Secchettin.
Critical revision of the manuscript for important intellectual content: Marchegiani, Esposito, Landoni, Casetti, Tuveri, Paiella, Bonamini, Auriemma, Merz, Simionato, Zecchetto, D’Onofrio, Melisi, Bassi, Salvia.
Statistical analysis: Maggino, Malleo, Paiella, Binco.
Obtained funding: Bassi.
Administrative, technical, or material support: Viviani, Landoni, Paiella, Sereni, Bonamini, Secchettin, D’Onofrio.
Supervision: Marchegiani, Esposito, Casetti, Tuveri, Auriemma, D’Onofrio, Melisi, Bassi, Salvia.
Conflict of Interest Disclosures: None reported.
Funding/Support: This work was supported by Associazione Italiana per la Ricerca sul Cancro (grants n.12182 and n.17132), Italian Ministry of Health (grant FIMP-CUP_J33G13000210001), and FP7 European Community Grant Cam-Pac (grant n. 602783).
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.