Multimodal treatment of resectable pancreatic ductal adenocarcinoma

After a timing preoperative staging, treatment of resectable pancreatic adenocarcinoma (PDAC) includes surgery and adjuvant therapies, the former representing the initial therapeutic option and the latter aiming to reduce the incidence of both distant metastases (chemotherapy) and locoregional failures (chemoradiotherapy). Herein, we provide a critical overview on the role of multimodal treatment in PDAC and on new opportunities related to current more active poli-chemotherapy regimens, targeted therapies, and the more recent immunotherapy approaches. Moreover, an analysis of pathological markers and clinical features able to help clinicians in the selection of the best therapeutic strategy will be discussed. Lastly, the role of neoadjuvant treatment of initially resectable disease will be considered mostly in patients whose malignancy shows morphological but not clinical or biological criteria of resectability. Depending on the results of these investigational studies, today a multidisciplinary approach can offer the best address therapy for these patients. © 2017 Elsevier B.V. All rights reserved.


Introduction
Pancreatic ductal adenocarcinoma (PDAC) is the fourth leading cause of cancer death in Europe and United States with a 5-year overall survival (OS) rate around 7% (Siegel et al., 2015). Resectable disease comprises only 15-20% of patients at diagnosis and many of them are found to have microscopically positive margins at the time of surgery. Even after curative surgery, most patients will recur with 5-year survival rate around 10%, with locoregional and systemic recurrence rates from 50% to 80% and greater than 70%, respectively (Kamisawa et al., 2016).
Adjuvant treatments include systemic therapy that aims to reduce the incidence of distant metastases and chemoradiotherapy (CRT) potentially able to decrease the risk of locoregional failure. While the role of the former is well established according to the results of randomized clinical trials, the usefulness of the latter remains controversial (Oyasiji and Ma, 2015). Evaluation by a multidisciplinary team of each patient should be mandatory .
Herein, we provide a critical overview on the role of multimodal treatment in PDAC and on new opportunities related to current more active poli-chemotherapy regimens, targeted therapies, and the more recent immunotherapy approaches. Moreover, an analysis of pathological markers and clinical features able to help clinicians in the selection of the best therapeutic strategy will be discussed. Lastly, the role of neoadjuvant treatment of initially resectable disease will be considered mostly in patients whose malignancy shows morphological but not clinical or biological criteria of resectability (Russo et al., 2016).

Morphological and clinical/biological criteria of resectability
After a preoperative staging  surgical resection is the initial therapeutic option of patients with resectable PDAC (Barugola et al., 2013;Serrano et al., 2015). Surgical indication is mainly influenced by the involvement of vessels but the underpinning disease biology as well as the general patients' conditions should also be considered. Vessel involvement is one of the most important features that needs to be carefully assessed in the initial evaluation of resectability. Regarding tumors involving the pancreatic head, there is a general agreement in excluding from immediate surgery patients with PDAC invading the superior mesenteric artery. The main limitations of pancreatectomy with arterial resection are represented by a poor short-and long-term outcome and the absence of prospective studies aimed to clarify a possible survival benefit in highly selected patients (Mollberg et al., 2011). The MD Anderson Cancer Center proposed a clas-sification of PDAC resectability based on vessel involvement in which lesions marginally invading the superior mesenteric artery are defined as borderline resectable pancreatic carcinoma (BRPC) (Katz et al., 2011(Katz et al., , 2013. The same classification identifies the superior mesenteric/portal vein grade of invasion as another key element for distinguishing resectable tumors from borderline and locally advanced forms. The role of superior mesenteric/portal vein resection in patients with PDAC is still a matter of debate. Several meta-analyses reported conflicting results on the possible advantages in terms of survival in patients who underwent pancreatectomy with vein resection (Siriwardana and Siriwardena, 2006;Yu et al., 2014;Giovinazzo et al., 2016). A recent meta-analysis focusing on this issue reviewed 27 studies showing increased postoperative mortality, higher rates of non-radical surgery and shorter survival after pancreatic resection with superior mesenteric/portal vein resection (Giovinazzo et al., 2016).
A distinction has to be made between PDAC involving the head of the pancreas and tumors localized in the pancreatic body/tail due to the high rate of splenic vessels infiltration in the latter group. Several reports analyzed the impact that the invasion of splenic vessels has among patients with distal PDAC (Partelli et al., 2011;Kanda et al., 2010). In particular, the invasion of the splenic artery, contrary to that of the splenic vein, is a crucial prognostic factor of PDAC localized in the body and tail. Vascular involvement is not the only criterion that could preclude immediate surgery. In the absence of clear metastatic disease, the occurrence of distant tumor localization should be suspected when Ca 19.9, the only biomarker with demonstrated clinical usefulness for PDAC, is significantly elevated in the absence of obstructive jaundice (Goonetilleke and Siriwardena, 2007). Ferrone et al. demonstrated a close correlation between the Ca 19.9 value and survival after pancreatic resection (Ferrone et al., 2006). A value of preoperative Ca 19.9 > 200 U/mL in patients with resectable PDAC has been proposed to identify patients with disease that is not suitable for up-front surgery in combination with duration of symptoms and pathological grading (Barugola et al., 2009). The performance status of patients is another important variable to be considered (Boeck et al., 2007). Intriguingly, time interval between preoperative image staging and surgery correlate with the rate of unanticipated metastasis (UM) observed at operation. It was demonstrated that the frequency of UM for 3 shorter time interval groups (0-6, 7-13, and 14-20 days) was 12%. The longest 3 groups (35-41, 42-48, and 49-86 days) showed similar frequencies of UM of 35%, 29%, and 30%, with a linear relationship (Glant et al., 2011). So far, Authors recommend an imaging no longer than 2-3 weeks old by the date of surgery, in all new diagnosed PDAC patients.
In summary, an appropriate surgical indication for PDAC has to take into account several factors that include technical aspects (vascular involvement assessment) as well as clinical/biological features.

Histological features and molecular markers
The most important prognostic factor for resected PDACs is the staging, according to TNM 7th edition (Bosman, 2010;Edge, 2010). Several studies reported that smaller tumors achieved a better survival have a better survival (Phoa et al., 2005;Yamaguchi et al., 1999). Currently, tumor size defines only earlier categories of T (pT1 < 2 cm; pT2 ≥ 2 cm), on the other hand, involvement of peripancreatic soft tissues and extra pancreatic structures identify a more advanced stage of disease (T3 and T4 respectively). A recent prospective study on 757 consecutive resected PDAC patients, has investigated the prognostic value of a revised T-stage protocol: pT1 as ≤2 cm, T2 as 2-4 cm and pT3 as >4 cm; T4 was excluded from the evaluation because it usually refers to BRPC or non-resectable PDAC (Saka et al., 2016). The Authors demonstrated, through univariate and multivariate analyses, a clear association between higher Tstage and shorter survival also in node negative cases and proposed to adopt this novel T-staging system given its linear prognostic value.
Assuming the established prognostic value of staging, several other histopathological factors have been shown to exert a potential prognostic role in resected PDAC, although very limited consensus exists on their evaluation and reporting among major international guidelines .
The importance of negative resection margins (R0, i.e. minimum distance between the neoplasm and the edge of the margin >1 mm) has been largely stressed, although different opinions on what constitute a resection margin and on the definition of microscopic margin involvement still exists (Bockhorn et al., 2014;Jamieson et al., 2013). Importantly, R0-resection is significantly associated with better prognosis, although it is achieved in only 36% of patients . Reliability of the R classification depends on the thoroughness of the clinical and histopathological examinations (Wittekind et al., 2009), so the standardization of reporting may help in defining the prognostic impact of a substantial cut-off for margin (Rau et al., 2012;Verbeka et al., 2006). More recently, a comprehensive archival analysis evaluating all the surgical resection margins (SRM) was performed to determine the effect of R0 vs R1 resection on locoregional recurrence, on survival and on the impact of adjuvant therapy (Osipov et al., 2016). The posterior resection margin resulted the most clinically significant. Furthermore, a resection margin >2 mm was a significant predictor of OS in patients who underwent adjuvant chemotherapy (HR 0.31, p = 0.03) and CRT (HR 0.40,p = 0.19).
In resected PDAC, the presence of lymph nodes metastasis is also an independent prognostic factor and a strong predictor of outcome at multivariate analyses (Schnelldorfer et al., 2008;Shimada et al., 2006;Valsangkar et al., 2013). A minimum of 12-17 lymph nodes is recommended to avoid under-staging (Huebner et al., 2012). Furthermore, for patients with N1 disease, the lymph node ratio (i.e. positive nodes/total nodes examined) appears to be related to prognosis (Opfermann et al., 2014;Pawlik et al., 2007) and it seems to be a stronger prognostic indicator than the number of positive lymph nodes alone (Pawlik et al., 2007). Two recent studies have also demonstrated that metastasis in para-aortic nodes are an indicator of poor prognosis, regardless of regional nodal status (Komo et al., 2016;Paiella et al., 2016). Neoplastic extra-nodal extension is a common event in up to 60% of PDAC. A meta-analysis has demonstrated its statistically significant impact on both OS and disease free survival (DFS), suggesting that extra-nodal involvement should be included in the pathological report (Luchini et al., 2016). Perineural and vascular invasion along with low grade stromal desmoplasia and the presence of a specific subset of macrophage infiltration are also strongly associated with worse prognosis (Garcea et al., 2007;Wang et al., 2016;Hu et al., 2016).
In recent years, several predictors of relapse have been addressed in PDAC at molecular level (Table 1). KRAS is the main driver of PDAC cancerogenesis and, as already established for other cancer types (Silvestris et al., 2014;Izar et al., 2014), distinct codon-mutations may have a different impact on prognosis. Furthermore, KRAS G12V mutation may be related to poor prognosis in T3-resected PDAC . Other frequent genetic alterations, associated with an enhanced malignant behavior, occur in PDAC expressing TP53, CDKN2A and SMAD4 associated to overexpression of Cyclin D1 (Georgiadou et al., 2014;Oshima et al., 2013). However, the real prognostic impact of molecular alterations in PDAC is debated. A tissue microarray study including 127 patients with resected PDAC did not confirm a correlation between SMAD4 loss of expression and recurrence pattern (Winter et al., 2013). Furthermore, another recent analysis failed to demonstrate significant differences in the mutational profiles of KRAS, TP53, SMAD4, and CDKN2A among PDAC groups with different median survivals. This study confirmed that younger age, earlier stage, well/moderate grade of differentiation and negative resection margins are the main prognostic factors defining the very long survivor group of patients (Dal Molin et al., 2015). The lack of a standardized histopathological report and molecular analytical protocols along with the needs of further insight into the PDAC cancerogenesis, may explain these contrasting results. Other molecular markers have been described to play a role as prognostic biomarkers in PDAC. High Rab27A, a RAS related protein, and p53 expression resulted correlated with poor OS; Rab27A expression can be considered an independent prognostic marker and a promising novel therapeutic target for PDAC . Elevated transcript levels of MDM2 P1 promoter and low p53 mRNA expression, significantly correlate with worse prognosis (Grochola et al., 2011).
Single-nucleotide polymorphisms of MAPK8IP1 and SOCS3, involved in the NF-ÄB inflammatory pathway, were examined in 1308 PDAC patients and resulted associated with a 10-6-months survival advantage compared to non-carriers in the resected group (Reid-Lombardo et al., 2013). Polymorphisms in angiogenesis genes such as VEGFR-2, CXCR-2 and PAR-1, may exert a potential impact on PDAC patients' outcome (Uzunoglu et al., 2013). In any case, in spite of the progress achieved in the comprehension of angiogenic gene expression in PDAC, antiangiogenic target therapy remains a promise .
Patients expressing the matricellular glycoprotein secreted protein acidic and rich in cysteine (SPARC) in cancer stromal fibroblasts showed a significantly worse outcome than patients negative for SPARC in the stromal compartment (Infante et al., 2007). Other Authors reported that stromal expression of lumican, a small leucine-rich-repeat proteoglycan, resulted strongly associated with the reduction of metachronous metastases and better OS compared to stromal lumican-negative PDAC patients. Its effect can be most likely related to Akt inhibition .
Recently, the prognostic and predictive value of micro-RNA (miRNA) expression profile has been widely investigated in several cancers. A recent analysis on 538 PDAC samples and 206 non-cancerous controls demonstrated that high expression of miRNA-21 and miRNA-31 and low expression of miRNA-375 are associated with poor OS following resection (Ma et al., 2013). However, the consideration of miRNAs as diagnostic and prognostic biomarkers and as potential targets for therapy would suggest future studies and stronger prospective validations (Brunetti et al., 2015). Other gene expression profile studies indicated a 36-gene signature as an independent prognostic factor of clinical outcome in PDAC (Haider et al., 2014). Chemoresistance to gemcitabine can be restored via p53reactivating molecules (CP-31398 and RITA) in p53 mutants PDAC cell lines (Fiorini et al., 2015) as well as down-regulation of urokinase plasminogen activator (Khanna et al., 2011). Similar sensitization to gemcitabine treatment can be induced by downregulation of cancerous inhibitor of protein phosphatase 2A (CIP2A) expression, which conversely acts as an independent poor prognostic indicator in PDAC (Xu et al., 2016). Moreover, SIRT3 and SIRT7, belonging to the broadly conserved sirtuin gene family, have tumor suppressor properties and SIRT3 can also be considered a potential novel predictive biomarker (McGlynn et al., 2015).
The expression of human equilibrative nucleoside transporter 1 (hENT1) that permits the bidirectional passage into cells of pyrimidine nucleosides such as gemcitabine, capecitabine, and 5FU (Spratlin et al., 2004;Yao et al., 2011) is an important prognostic and predictive biomarker for gemcitabine efficacy in patients with resected PDAC (Greenhalf et al., 2014) but not in metastatic subsets (Poplin et al., 2013). Its use in clinical practice is limited by the lack of large prospective validation studies (Viterbo et al., 2016).
Finally, the advent of high sequencing methods lead to the identification of 4 molecular subtypes in PDAC, each recognized by specific patterns of chromosomal alterations and recurrent mutations in driver genes. These correlate to different histopathological characteristics, inferring differences in the molecular evolution of PDAC subtypes and identifying ulterior chances for therapeutic development (Bailey et al., 2016). In particular, cases characterized by high genomic instability usually exert inactivation of DNA maintenance genes (BRCA1, BRCA2 or PALB2) along with a DNA damage repair deficiency mutational signature. Patients bearing these molecular characteristics seem to benefit from platinumbased therapy with high rate of response (Waddell et al., 2015).
Recently, the presence of circulating tumor cells (CTCs) in peripheral blood of PDAC resected patients has been proposed as a novel poor prognostic factor. In particular, the study performed by Poruk et al. demonstrated the presence of cytokeratin-positive CTCs and vimentin-positive CTCs in the peripheral and portal blood samples in 78% and 67% of 50 PDAC patients, respectively (Poruk et al., 2016). A poorer survival rate was observed only in the group of patients with cytokeratin-positive CTCs (p < 0.01). Furthermore, the presence of cytokeratin-positive CTCs significantly correlated with a better survival rate and was a significant independent prognostic marker at multivariable analyses (p < 0.01). Indeed, detection of both vimentin and cytocheratin positive CTCs correlated with the recurrence rate (p = 0.01).
The results of above studies show that a huge number of biological and molecular markers have been reported over the years as potential prognostic and predictive factors, opening novel scenarios on therapeutical options in PDACs, although their potential use in clinical practice would require larger prospective validation studies. Moreover the advance in molecular analytical technologies has provided an in depth knowledge of genomics alterations that underlie PDAC development and progression. However a fully standardized and widely shared histopathological report along with larger prospective validation studies of robust biomarkers are essential conditions to deeply understand the biology of such an aggressive disease and better address therapy and clinical management in PDAC patients.

Adjuvant chemotherapy
Considering the high risk of recurrence even after radical surgery, the use of adjuvant therapies has largely been studied in resected PDAC. In particular, different randomized studies evaluated the role of postoperative chemotherapy alone for these patients.

Historical chemotherapy regimens
The first available randomized study was conducted about 30 years ago. Sixty-one patients with PDAC (47 patients) or ampullary carcinoma (the remaining 14 patients) were randomized to FAM (5-FU bolus plus doxorubicin plus mitomycin-C every 3 weeks for 6 cycles) or observation alone (Bakkevold et al., 1993). The trial showed an improvement in OS with chemotherapy (mOS: 23 months vs 11 months, p = 0.02; +11% in absolute 2-year survival, from 32% to 43%) even if the curves overlapped after 2 year with no differences in long-term survival and a 5-year survival rate in the chemotherapy group of only 4%. This chemotherapy regimen was associated with consistent toxicity, with 1 toxic death over 31 treated patients, and with severe cardiotoxicity and nephrotoxicity reported in 2 patients. Apart from toxicity, the major limits of this trial were the small number of patients and the inclusion of both pancreatic and ampullary malignancies.
A larger and more complex trial was conducted in Europe in the next decade by the European Study Group for Pancreatic Cancer (ESPAC-1 trial) (Neoptolemos et al., 2001). The trial had a factorial design with two randomizations: the first for chemotherapy vs no chemotherapy and the second for CRT vs no CRT. A total of 285 patients were randomized in the original design of the study while additional 188 patients were randomized either for chemotherapy or observation alone for a total of 473 patients. Chemotherapy was administered with bolus 5-FU at 425 mg/m 2 for 5 consecutive days every 28 days for 6 cycles, combined to folinic acid. Regarding the role of chemotherapy, the ESPAC-1 trial showed an increase in OS with chemotherapy considering all enrolled patients (mOS: 19.7 months vs 14.0 months; p = 0.0005) (Neoptolemos et al., 2001;Neoptolemos et al., 2004). The survival curves began to separate after 6 months and remained distant even after 5 years, with an absolute survival benefit at 2-and 5-years of 10% and 13%, respectively.
Besides 5-FU, also gemcitabine was investigated as adjuvant chemotherapy for resected PDAC. The first study was conducted between 1998 and 2004 in Germany and Austria by the Chariteí Onkologie (CONKO) group (CONKO-001 trial). This study compared weekly gemcitabine (at the dose of 1000 mg/m 2 at days 1, 8, 15 every 28 for 6 cycles) vs observation alone in 368 patients (Oettle et al., 2007). In this trial, gemcitabine prolonged DFS (13.4 months vs 6.7 months, p < 0.001) and OS (22.8 months vs 20.2 months, p = 0.01) with an absolute increase in survival of 10% at 5 years. Furthermore these data were confirmed with a longer follow-up (Oettle et al., 2013). Even if the improvement in terms of OS were modest in this subset of patients (only 2.6 months), the aggressive biological behaviour of this malignancy should be considered.
A second study conducted on 119 Asiatic patients who received gemcitabine for only 3 months compared to observation alone showed an improvement in terms of DSF (HR = 0.60; CI 95% 0.45-0.89; p = 0.01) but not for of OS (HR = 0.77; CI 95% 0.51-1.14; p = 0.19) (Ueno et al., 2009).The low number of enrolled patients influenced these results.
With two effective drugs available, gemcitabine and 5-FU, a comparison between them was performed in 2 randomized trials. The ESPAC-3 trial compared gemcitabine to 5-FU in more than 1000 resected patients. The study initially included a third arm with observation alone that was ended following the results of both ESPAC-1 and CONKO-001 trials (Neoptolemos et al., 2010). The study showed very similar results in terms of quality of life (QoL), DFS, and OS between the two arms with a mOS of about 23 months. The toxicity profile differed between the two arms: haematological toxicity was higher with gemcitabine while gastrointestinal toxicity was higher with 5-FU.
A retrospective analysis from the ESPAC-3 trial revealed that completing all 6 months of planned adjuvant chemotherapy is crucial in deriving benefits while the time of treatment initiation (<12 weeks) is not (Valle et al., 2014). A potential bias related to the timing of treatment could be represented by the inclusion of patients who experienced early death or a progression disease and, therefore did not complete the entire planned chemotherapy. So far, even after the exclusion of patients who died within 8 months after resection, OS and recurrence-free survival of the remaining 889 patients resulted improved.
The RTOG 97-04 trial compared either gemcitabine or 5-FU, before and after CRT (Regine et al., 2011). There was no significant difference in terms of mOS (p = 0.51) with a higher rate of toxicity in the gemcitabine arm (p < 0.001). A benefit correlated with gemcitabine was observed in the multivariate analysis for patients with PDAC localized in the pancreatic head (p = 0.08).
A recent meta-analysis evaluated the role of adjuvant gemcitabine in 2017 resected PDAC patients in four phase III randomized trials (Yu et al., 2015). In particular, two of them compared gemcitabineto5-FU/folinic acid and the other two compared gemcitabine to best supportive care (BSC). This meta-analysis considered an homogenous population of patients among these studies (Q = 4.371; I = 31.37%; p = 0.224) and suggested that adjuvant gemcitabine improved OS compared to both BSC or 5-FU/folinic acid treatments in resected PDAC with an overall HR of 0.88 (p = 0.014).
S-1, an oral fluoropyrimidine, was compared to gemcitabine as adjuvant chemotherapy for Asiatic resected PDAC in two phase III trials (Shimoda et al., 2015;Uesaka et al., 2016). The former (Shimoda et al., 2015) did not demonstrate a significant difference in terms of DSF (14.6 vs 10.5 months; p = 0.293) between patients treated with S1 (29 patients) and gemcitabine (28 patients). On the contrary, the second larger trial (Uesaka et al., 2016) considering 385 patients showed an improvement of 5 year OS between S-1 and gemcitabine (44.1% vs 24.4%). The hypothetical difference in pharmacological metabolism of Asiatic population could influence the applicability of these results in different populations.
Today, single agent chemotherapy both with gemcitabine or 5-FU is considered as standard adjuvant chemotherapy for PDAC (Table 2) in the guidelines of several Western countries (NCCN Guidelines, 2016;AIOM Guidelines, 2015;ASCO guidelines 2016).
Similarly to surgery, the relevance of the center of care in terms of outcome for resectable PDAC patients candidate to adjuvant systemic treatment has been reported. A retrospective analysis (Mandelson and Picozzi, 2016), considering 245 patients, 57% treated at high volume center (HVC) and 43% treated at community oncologist (CC) showed an advantage in terms of mOS and 5-year OS rate for HVC centers compared to CC centers (44 months vs 28 months, p > 0.01; 38.8% vs 24.8%, p > 0.01). These data suggest that, a HVC should be preferred not only for surgery but also for adjuvant systemic treatments.

New combination chemotherapy in the adjuvant setting: may intensification represent a step forward?
Based on the assumption of the early dissemination of pancreatic cancer cells, several ongoing clinical trials are evaluating, in the adjuvant setting, different combination chemotherapy regimens which are efficacious in the advanced setting (Garrido-Laguna and Hidalgo, 2015). PACT-7, an Italian phase II randomized trial, tested the concept of chemotherapy intensification and allocated 102 radically resected stage IB-III PDAC patients to standard gemcitabine or PEFG (cisplatin and epirubicin 40 mg/m 2 , day 1; gemcitabine 600 mg/m 2 , days 1, 8; 5-FU 200 mg/m 2 daily, days 1-28) (Reni et al., 2012). Although producing more haematological toxicities, the combination regimen yielded interesting results in terms of mDFS (15.2 months vs 11.2 months in the gemcitabine arm), 1-year DFS rate (69.4% vs 49%), and mOS (28.9 months vs 24.8 months).
The combination of gemcitabine (1000 mg/m 2 , day 1, 8 and 15) and capecitabine (830 mg/m 2 twice daily on days 1-21 every 28 days) compared to gemcitabine alone has been tested in ESPAC-4, a randomized phase III study (Neoptolemos et al., 2016). This trial considered radically resected PDAC or periampullary cancer patients. Primary endpoint was OS; secondary endpoints were OS rate at 2 and 5 years, toxicity and QoL (NCT00058201, 2016). Early results presented at the last ASCO meeting regarding 730 patients showed that mOS was 28 months and 25.5 months for combination arm and gemcitabine alone, respectively (HR = 0.82; p = 0.032), with a similar toxicity profile. The immature follow-up does not allow to consider these data sufficiently reliable, as evidenced by the large overlap of the confidence intervals of mOSs. In addition, data of 2-year OS rate are not yet available. It is also unclear what the overall patient population of the study is. In fact, a maximum limit of Ca 19.9 was not considered as an inclusion criterion for the inclusion of patients with a CA19.9 baseline value of over 8000. Furthermore, 8% of patients were node positive. Lastly a postoperative restaging was not required thus the inclusion of patients with metastatic disease cannot be excluded. Although promising, these data are not sufficiently mature to consider this combination regimen as a new therapeutic standard.

Adjuvant chemotherapy: ongoing trials
APACT (ABI-007-PANC-003) is a global, open-label, randomized phase III study comparing nab-paclitaxel plus gemcitabine vs gemcitabine alone in over 850 patients with radically resected PDAC (Tempero et al., 2016). This study is based on the results of the phase III MPACT trial, which showed a significant outcome advantage for the combination compared to gemcitabine alone in the advanced setting (Von Hoff et al., 2013). DFS was the primary trial endpoint; secondary endpoints were OS and safety. The trial has recently completed its accrual, with the enrollment of 866 patients.
Following the results of PRODIGE 4/ACCORD 11, that demonstrated the superiority of FOLFIRINOX over gemcitabine alone in the metastatic setting (Conroy et al., 2011), PRODIGE 24/ACCORD 24 was designed as a phase III, multicenter, randomized, openlabel trial to compare modified FOLFIRINOX (without 5-FU bolus) to gemcitabine alone after disease resection. Target sample size was 490 patients with ECOG PS 0-1 and without prior radiotherapy and chemotherapy. The primary study endpoint of this ongoing trial is DFS at 3 years, secondary endpoints are OS and disease specific survival at 3 years.
GIP-2 is an Italian, phase III, multicenter, open-label, randomized phase III study comparing 6 months of FOLFOXIRI (a modified triplet containing irinotecan, oxaliplatin, and 5-FU) vs gemcitabine alone in resected PDAC. This trial is enrolling 310 patients with sur-gically resected stage I-III PDAC, ECOG PS 0-1 and adequate organ function. The primary endpoint is DFS; secondary endpoints are OS, safety and tolerability.
Following the results of PACT-7, the role of postoperative administration of a 4-drug combination is being tested in another randomized Italian study. PACT-15 is a phase II-III multicenter, open-label trial including three arms: (1) adjuvant PEXG (cisplatin, epirubicin, gemcitabine, and capecitabine) perioperative PEXG, and adjuvant gemcitabine.
The study was designed to include 370 previously untreated stage I-II resectable PDAC patients. The primary endpoint is DFS at 1 year; secondary endpoints are radiological, biochemical, and pathological response rates (neoadjuvant arm only), assessment of lymph node status, surgical mortality, morbility and resection rates, and treatment tolerability.
At the ASCO 2016 meeting, the project of a randomized phase II study (Sohal et al., 2016) comparing preoperative mFOLFIRINOX or gemcitabine/nab-paclitaxel combination, followed by surgical resection and 12 weeks of identical postoperative chemotherapy in resectable PDAC, was presented. Primary outcome is 2-year OS, with a total sample size of 112 patients. The results of this study will be very interesting because it is a head to head comparison between the two regimens used in clinical practice in metastatic PDAC.

Adjuvant CRT
CRT is widely used in the adjuvant setting even if the results regarding its effectiveness remain controversial compared to chemotherapy alone.

Adjuvant CRT compared to observation alone
The first randomized trial investigating this topic was performed by Gastrointestinal Tumor Study Group (GITSG) on 43 patients who underwent either CRT or observation alone after R0 radical surgery (Kalser and Ellenberg, 1985). CRT consisted of 40 Gy in 6 weeks with concurrent intermittent bolus of 5-FU at 500 mg/m 2 on days 1-3 and 29-31, followed by further 5-FU for 2 years. An unacceptably low rate of accrual combined with the observation of increasingly large survival differences between the study arms were responsible for a premature termination of the study. A benefit in mOS favored CRT vs observation (21 months vs 10.9 months; p = 0.03), as well as for the 2-year OS (42% vs 15%; p = 0.03). Based on these results, adjuvant CRT became a standard practice in the United States.
EORTC 40801 trial accrued 218 patients operated for pancreatic or ampullary tumor (Klinkenbijl et al., 1999). Among them, 114 presented T3 (or higher) and/or N1 pancreatic malignancy. After surgery, patients were allocated to observation or adjuvant CRT. The therapeutic strategy reproduced the one considered in the GITSG study with the exception of 2 years of therapy with 5-FU after CRT. The main analysis on the whole group (pancreatic and ampullary lesions) failed to detect any significant improvement of OS or DFS. A subgroup analysis, which focused on pancreatic tumors, reported a non-significant trend in favor of CRT. The recurrence patterns were not significantly different between the 2 groups, suggesting the ineffectiveness of CRT. The 11-year follow-up analysis of that study confirmed the initially published short-term findings (Smeenk et al., 2007). Some large retrospective analyses suggested a positive influence of CRT in this subset of patients. Herman et al. compared 271 patients who underwent CRT (50 Gy plus 5-FU) to patients not receiving CRT (Herman et al., 2008). CRT was found to improve mOS (21.2 months vs 14.4 months; p < 0.001). A recent analysis by the US-based Surveillance, Epidemiology and End Results program considering 2532 patients who underwent neither adjuvant radiotherapy nor observation reported a result of median OS and DSF significantly higher in the postoperative radiotherapy group (mOS: 20 months vs 16 months; DFS: 22 months vs 18 months; p < 0.0001) (Sugawara and Kunieda, 2014). The conflicting results of the above studies leave the role of CRT unsolved compared to observation alone in this subset of patients.

Adjuvant CRT compared to chemotherapy alone or observation
The first largest clinical study comparing adjuvant concurrent CRT to chemotherapy alone or observation was the ESPAC-1 trial (Neoptolemos et al., 2004) that ramdomized 289 patients with resected PDAC to receive CRT alone (20 Gy over a two-week period plus 5-FU), 5-FU alone, CRT and sequential 5-FU, or observation. Median OS for adjuvant chemotherapy and concurrent CRT were 20.1 and 15.9 months, respectively (p = 0.009). One-year recurrence rate for concurrent CRT and chemotherapy was 46% vs 55%, associated with both shorter recurrence free survival (10.7 vs 15.2 months, p = 0.04) with respect to a worse toxicity profile for patients treated with CRT. In this trial, adjuvant chemotherapy showed a significant survival benefit compared to adjuvant CRT, even if these results could be influenced by both the suboptimal schedule of CRT adopted and the use of non-conformal 2D treatment planning.
The EORTC-40013-22012/FFCD-9203/GERCOR phase II study randomized 90 PDAC patients within 8 weeks after R0 resection, to receive 4 cycles of gemcitabine (control arm) or gemcitabine for 2 cycles followed by weekly gemcitabine with concurrent radiation (50.4 Gy; CRT arm). Median DFS was 12 and 11 months in the CRT and control arm, respectively, with a mOS of 24 months in both arms. Moreover, local relapse rate was 11% and 24% for CRT and control arm, respectively (Van Laethem et al., 2010).
Koobyet al. analyzed data from 11526 patients from the National Cancer Data Base (NCDB) (Kooby et al., 2013). Of these 1029 (8.9%) underwent chemotherapy, 5292 (45.9%) underwent CRT (excluding patients who received radiotherapy as the only form of adjuvant therapy), and 5205 (45.2%) underwent observation alone. CRT arm showed the best OS (HR: 0.70, CI 95% : 0.61-0.80) in a propensity scored matched comparison with chemotherapy (HR: 1.04, CI 95% : 0.93-1.18) and observation. A similar analysis was performed by Rutteret al (Rutter et al., 2015). These Authors collected data from the NCDB on R0/R1 resected patients; 6165 patients received chemotherapy (38%) or CRT (62%). The median dose of radiotherapy was 50.4 Gy. Subset analyses showed that CRT was associated with improved OS among patients with pT3 or pN1 disease and particularly among patients with R1 resection. It should be observed that CRT was associated with an improvement in OS when it was administered after 1-3 months of systemic chemotherapy. A multicenter retrospective review by Morganti et al. considered 955 patients with R0/R1 resection (Morganti et al., 2014) treated with either CRT or CT alone. CRT was administered with doses >45 Gy (45-60, conventional fractionation), and 5-FU or gemcitabine were adopted as concurrent chemosensitizers. Five-year OS was 41.2% vs 24.8% with and without adjuvant CRT, respectively. Median OS after CRT was 39.9 months vs 27.8 months after chemotherapy alone (p < 0.001). The difference in median survival between patients treated with CRT with or without further adjuvant chemotherapy and patients treated with adjuvant chemotherapy alone was statistically significant (39.5 months vs 27.8 months, p < 0.001). Table 3 summarizes some of the mentioned trials comparing adjuvant CRT to observation alone or to chemotherapy.
Two meta-analyses (the first on 5 randomized trials including 939 patients; the second on 4 randomized trials including 875 patients) did not reveal any significant clinical benefit from the use of CRT, with a small survival benefit favoring the adoption of adjuvant CRT for patients who reported an R1 resection (Stocken et al., 2005;Butturini et al., 2008). Ren et al. conducted another metaanalysis on 15 randomized trials on adjuvant therapy and observed that, compared to observation, chemotherapy improved DFS and OS while CRT did not (Ren et al., 2012). Similarly, Liao et al. performed a meta-analysis on 9 randomized trials (3033 patients) to directly compare adjuvant CRT to chemotherapy (Liao et al., 2013). They did not find an advantage in the use of CRT concerning the survival rate compared to the use of chemotherapy with respect to an increased toxicity profile. These analyses do not provide data on the rate of local recurrence, which could be an important endpoint for these patients in terms of QoL. Furthermore, it should be highlighted that all these meta-analyses included the GITSG, EORTC and ESPAC-1 trials, which administered a suboptimal and obsolete schedule of radiotherapy.
On the basis of published trials and international guidelines (NCCN Guidelines, 2016;AIOM Guidelines, 2015, ASCO guidelines 2016, adjuvant CRT should be proposed to a patient with resected PDAC which presents microscopic positive margins and/or positive node positive disease after 4-6 months of systemic adjuvant chemotherapy. Novel radiotherapic approaches which attempt to shorten courses and intensify radiotherapy with techniques including intensity-modulated radiation therapy and stereotactic body radiation therapy may represent a novel strategy to improve locoregional control and tolerability (Bowenge et al., 2015;Ahn et al., 2016). In this scenario, the currently ongoing American RTOG 0848 and European EORTC 40084-22084 will better define the role of modern 3-dimensional conformal radiation therapy.

Adjuvant target therapy
PDAC shows a peculiar biological complexity leading to the evidence of several, potentially effective molecular targets for specifically designed therapeutic options (Ahn et al., 2016). The tumor transformation process is in fact linked to the acquisition of progressive single mutations inducing genomic instability in several cellular pathways, including KRAS, hedgehog, integrins, transforming growth factor ␤ (TGF-␤), the Wnt/Notch signaling, but also molecular pathways activated by the TGF-␤, the hemophilic cell adhesion and invasion, the regulation of G1/S cell cycle transition, EGFR as well as DNA damage control and apoptosis (Di Marco et al., 2016). As a consequence, adjuvant targeted therapies in this setting of patients may represent a very intriguing field of research (Table 4).

Molecular target agents
The role of cetuximab, a mAb anti-EGFR, in the post resection treatment of PDAC has been investigated in a multicentre phase II clinical trial (ATIP trial) in which 76 resected patients, were enrolled to receive gemcitabine and cetuximab as adjuvant therapy (Fensterer et al., 2013). The primary aim was to determine whether the experimental arm could lead to an increased DFS when compared to the reported results of gemcitabine alone in this setting. Study findings were negative with a mDFS of 10 months and a DFS rate of 27.1% at 18 months. These results were inferior to those reported for gemcitabine alone. There was no pre-planned strati-   fication of the patients by KRAS status, but 69.1% of patients had a mutation in the codon 12 or 13 of the KRAS oncogene. However, no differences in DFS or OS were observed between wild type or mutated KRAS patients. These results were not surprising, due to the lack of activity of cetuximab in the metastatic setting. The efficacy of EGFR-TKI erlotinib as adjuvant therapy in resected PDAC has been evaluated in two major phase III clinical studies: the CONKO-005 and the RTOG-0848 trials. The former study was a phase III randomized trial designed to evaluate the efficacy of erlotinib (100 mg p.o. daily) in combination with gemcitabine after R0 resection (Sinn et al., 2015). Four hundred thirty-six patients were randomized within 8 weeks from resection to gemcitabine-erlotinib vs gemcitabine. Stratification factors included lymph nodes involvement, surgical technique and study centre. Primary endpoint was DFS, secondary endpoint included OS and safety. The study failed to meet its primary endpoint with no differences between the two arms in DFS (11.6 months vs 11.6 months in the gemcitabine-erlotinib and gemcitabine alone arms, respectively) nor in OS (24.6 months vs 26.5 months). The RTOG-0848 was a randomized trial that aimed to evaluate whether erlotinib and/or radiation could improve OS in resected PDAC patients (Li and O'Reilly, 2015). The original study was designed with a first step randomization into two arms: (1) gemcitabine alone for 5 cycles or (2) gemcitabine-erlotinib for 5 cycles. Patients free from tumor recurrence were then randomized to receive one additional cycle of chemotherapy (based on the first randomization -arm 3) or one cycle of chemotherapy followed by 5-FU based CRT (arm 4). This study aimed to provide a better understanding of the role of erlotinib in the adjuvant setting but also attempted to answer the question of the value of combined CRT to systemic chemotherapy in radically resected patients. Based on the disappointing results from LAP 07 phase III trial (Hammel et al., 2013), the study was amended to close enrollment in the erlotinib plus gemcitabine arm in early 2014 and is currently ongoing to determine whether the use of concurrent fluoropyrimidine and radiotherapy is able to improve survival in resected PDAC patients.

Immunotherapy
A further area of increasing interest for research in the adjuvant setting is immunotherapy. Globally, PDAC has always been considered a non-immunogenic malignancy. This is due in part, to its microenvironment that includes inflammatory cells (mainly lymphocytes, plasma cells, macrophages and mast cells) which, rather than reacting against the tumor, seem to promote an inflammatory microenvironment that helps tumor cells to escape from immune surveillance via paracrine cross-talk mechanism (Puzzoni et al., 2016). PDAC is surrounded by a desmoplastic stroma within which neoplastic cells are able to stimulate the proliferation of fibroblasts, specifically pancreatic stellate cells (PSCs). PSCs contribute to the additional fibrosis formation, they promote tumor progression and they seem to facilitate immune evasion, chemoresistance and recurrence of PDAC (Ansari et al., 2015). If on the one hand this complex tumor microenvironment may make an immunotherapeutic approach in the metastatic or advanced disease difficult, on the other hand it could be a winning strategy in the adjuvant setting. Only preliminary data are available at the moment and ongoing trials will hopefully clarify the efficacy of such a treatment approach in the near future.
More than 95% of PDAC are associated with KRAS oncogenic mutations, most as a single point mutation at codon 12 resulting in a constitutive activation of KRAS (Mohammed et al., 2015). This mutation leads to upregulation/dysregulation of the EGFR and induction of proliferation by activation of mitogenic pathway. The amino-acid substitution at position 12 of the RAS protein can be recognized both by helper T-cells and cytotoxic T-cells and can be used as antigen in personalized peptide vaccine corresponding to the KRAS mutations present in the tumor (Wedén et al., 2011). A phase II adjuvant trial considered 176 KRAS mutant, R0-R1 resected, PDAC patients enrolled to receive placebo and gemcitabine or GI 4000 (a whole, heath-killed recombinant S. cerevisiae yeast expressing a mutated RAS protein) and gemcitabine (Muscarella et al., 2012). Primary endpoint was RFS. Only results in 39 patients with R1 resection were presented. The GI 4000 group showed an 11.4 week advantage in mOS, a 16% advantage in 1 year survival (72% vs 56%) and a 4.6 week advantage in mRFS. Authors concluded that GI-4000 plus gemcitabine showed an improvement on survival in R1 subjects with RAS mutant PDAC.
Hyperacute immunotherapy is based on the hypothesis that some antigens presented ad hoc to the immune system (vaccines) are able to trigger a powerful immune response known as hyperacute rejection. This latter characteristically targets xenotransplanted tissue, but in this case it can react against modified cancer cells. This inflammatory reaction/immune response is thought to be responsible for the production of immunity against tumor antigens (Saif, 2011). These vaccines may be made of whole cells, proteins, peptides or DNA encoding tumor antigens. There are at least two vaccines in an advanced phase of evaluation for the adjuvant setting of PDAC patients. The first is the G-VAX −pancreas vaccine that consists of irradiated cells stably transfected to express GM-CSF. This vaccine has been tested in a phase II trial among 60 patients with resected PDAC, in combination with 5-FU based CRT. The vaccine was well tolerated with mDFS and mOS of 17.3 months and 24.8 months, respectively; the 1-year OS was 85% (Lutz et al., 2011). The second is the algenpantucel-L vaccine that consists of 2 irradiated allogenic PDAC cell lines (HAPa-1 and HAPa-2) transfected to express murine alpha-1,3galactosyltransferase (alpha-GT). The murine alpha-GT epitopes are not present in human cells and they can induce a hyperacute rejection of vaccine pancreatic cancer cell allografts. The binding of pre-existing human anti alpha-GT antibodies (which naturally occur against gut flora) to murine alpha-GT epitopes results in a rapid activation of antibody-dependent cell-mediated cytotoxicity toward allograft cells (Saif, 2011;Cid-Arregui and Juarez, 2015). A phase II clinical trial, employing algenpantucel-L, enrolled 70 patients with resected PDAC to receive gemcitabine and 5-FUbased CRT as well as algenpantucel-L (Hardacre et al., 2013). After a median follow up of 21 months, the 12 months DFS was 62% whereas the 12 month OS was 86%. Based on these encouraging results, a phase III clinical trial (IMPRESS) has been recently con-cluded with the aim to assess OS after a regimen of adjuvant therapy (gemcitabine alone or in combination with 5-FU based CRT) with or without algenpantucel-L immunotherapy (NCT01072981, 2016). This trial did not achieve the primary endpoint, since no statistically significant differences have been reported in terms of mOS (30.4 and 27.3 months for the control and algenpantucel-L treated groups, respectively) and long-term survival (3 year OS of 41.4% and 42.1% and 4-year OS of 32.6% and 32.7% for the chemotherapy and chemotherapy plus algenpantucel-L groups, respectively) (NewLink Genetics Corporation, 2016).
At the ASCO 2016 meeting, in a multicenter non randomized phase II study (Yamaue et al., 2016) a peptide vaccine cocktail named OCV-C01 with peptides derived from KIF20A, VEGFR1, and VEGFR2 combined with gemcitabine was tested in 30 resected PDAC patients. At every cycle, patients received subcutaneous injection of 1 mL of OCV-C01 on days, 1, 8, 15, and 22 for 12 cycles, and intravenously 1000 mg/m 2 gemcitabine on days, 1, 8, and 15 for 24 weeks. This safe combination achieved a mDFS of 15.8 months (CI 95% : 11.1-20.6) with a DFS rate at 18 months of 34.6% (CI 95% : 18.3-51.6). Median OS is still not reached but OS rate at 18 month was 69% (CI 95% : 48.8-82.5). Moreover, patients with strong cytotoxic T lymphocytes specific response against KIF20A showed a more statistical significant mDFS. Four R0 resected patients expressing KIF20A displayed no recurrence of PDAC, with a DFS significantly longer than other R0 resected patients not expressing KIF20A expression (p = 0.011).
In conclusion, both molecular target agents and immunotherapy seem to be interesting approaches in the field of multimodal treatment of resectable PDAC, even if more clinical trials are necessary.

Neoadjuvant therapy in resectable PDAC
In the last years, a preoperative approach was studied as a strategy for resectable PDAC patients in order to treat earlier micro-metastatic disease, to increase long-term survival, to avoid morbidity and mortality of surgery in patients with early relapse, to evaluate tumor response, and to improve treatment tolerability . Even if there are many theoretical benefits for this approach, there is still limited evidence on this topic and some potential hurdles should be evaluated, such as a possible missed opportunity for resection, the need of treatment for symptoms of biliary or intestinal occlusion, and the absence of surgical staging.
There are few trials with the use of chemotherapy alone (Table 5). Palmer et al. randomized 50 patients with resectable PDAC to receive gemcitabine alone or in combination with cisplatin (Palmer et al., 2007). A higher resection rate was observed in the combination arm (70%) compared to gemcitabine alone (38%). The mOSs were 15.6 and 9.9 months with a 1-year survival rate of 62% and 42% for the combination arm and the control arm, respectively. The data of this study should be regarded with caution because of the small sample size and the presence of misdiagnosis (pathological examination showed that 12% of patients treated with gemcitabine and 19% with the combination had not received a diagnosis of adenocarcinoma). Another phase II study evaluated 28 patients with resectable disease, who received cisplatin in combination with gemcitabine as neoadjuvant treatment (Heinrich et al., 2008). There was the recording of one partial response (PR) (4%), with 61% of stable disease (SD). Globally, resection rate was 89% with 71% of R0 resections and a mOS of 19.1 months for resected patients. Tajima et al. recently reported the results of the combination of gemcitabine and S-1 in 13 patients (Tajima et al., 2013). They observed 2 PR and 9 SD. Another phase II trial, conducted on 38 patients with resectable disease, evalu- As regards CRT, an MD Anderson retrospective analysis showed that the use of preoperative CRT was associated with a better outcome in resectable patients. Relevant clinical data is represented by the observation of a regional and/or distant progression at preoperative staging in 25% of cases, suggesting a lack of benefit from an upfront surgery in this subgroup of patients (Breslin et al., 2001).
Few Authors have prospectively evaluated the role of CRT in this setting. Varadhachary et al., (2007) conducted a study of chemotherapy with gemcitabine and cisplatin followed by gemcitabine based CRT (Varadhachary et al., 2008). Of the 90 patients enrolled, 79 (88%) completed the integrated treatment. The mOS of the latter was 18.7 months, with a median of 31 months in 52 patients who underwent surgery, compared to 10.5 months in the group of not resected patients (p < 0.001).
Three studies including 189 patients were evaluated in a meta-analysis (Xu et al., 2014) which compared neoadjuvant CRT vs postoperative CRT. The pooled HR of 0.93 (CI 95% : 0.69-1.25; p = 0.62) did not document a benefit from neoadjuvant CRT in this setting. Furthermore, two randomized trials were conducted, aimed at assessing the impact of neoadjuvant CRT compared to surgery upfront. Both studies were interrupted prematurely due to the lack of differences in terms of differences of R0 resection rate (Casadei et al., 2015) or the low recruitment rate (Golcher et al., 2015).
An ongoing phase III trial is comparing preoperarive CRT vs surgery alone (NCT01900327, 2016NCT, 2016eNCT01900327, 2016. Moreover, another ongoing phase III study will be comparing perioperative therapy followed by surgery and adjuvant therapy to surgery followed by adjuvant therapy (NCT01314027, 2016). Later a three arms multicenter Italian Phase II-III study will randomize perioperative chemotherapy (PELF) with the same chemotherapy applied postoperatively and postoperative gemcitabine (NCT01150630, 2016).
Considering the available evidences of neoadjuvant treatment for resectable PDAC, the role of this approach needs to be validated in large, controlled randomized trials compared to surgery followed by standard adjuvant chemotherapy. ASCO guidelines recommend that a total of 6 months of adjuvant therapy (including preoperative regimen) be proposed according to an extrapolation from adjuvant trials (Khorana et al., 2016). One critical clinical issue is the difficult assessment of treatment response, likely owing to the dense stromal component of PDAC which remains radiographycally unchanged even when histological response is reported (Verbeka et al., 2015). In this setting of patients, the use of restaging imaging should primarly be performed to identify local or distant disease progression in combination with the evaluation of performance status and modifications of Ca 19.9 levels.

Conclusions
After a timing preoperative staging focused on both the research of distant metastases and the assessment of local vascular involvement, surgery represents the initial therapeutic option for resectable PDAC patients. According to the available evidences, 6 months of monochemotherapy with gemcitabine or 5-FU represent the standard adjuvant treatment. Similarly to surgery, the relevance of the centre where the patient underwent adjuvant chemotherapy has been reported as well. Today, it is premature to draw conclusion on the potential advantage of combination chemotherapy because they would depend on the results of ESPAC-4 and on the data from randomized phase III trials evaluating FOLFIRINOX and gemcitabine/nab-paclitaxel, which are currently ongoing. Results of studies evaluating the role of CRT are disappointing, since it did not improve survival compared with chemotherapy alone, but it could reduce local recurrence improving the QoL which could be a clinical relevant endpoint for these patients. How to select initially resectable PDAC patients to immediate surgery or preoperative chemotherapy remains a challenge. Probably, this decision should be prioritized by testing more effective systemic regimens and assesing clinico-pathological features and biomarker studies. In fact, fully-standardized and widely-shared histopathological and liquid biopsy tests with larger prospective validation studies are needed for discovering predictive biomarkers, essential in understanding the biology of this malignancy. Depending on the results of the investigational studies evaluating the role of novel modalities of radiotherapy, immunotherapies and other molecular target approaches, a multidisciplinary approach can offer the best address Che significa "address"? therapy for these patients.

Author contributions
NS and OB instructed and designed the original concept, and revised the manuscript critically. MF, MR, MS, AS and SC drafted the manuscript, and revised it critically. EV, VV, AC, FC, IC, and GA searched literatures and revised the manuscript critically. SP, ACG, AGM and VL drafted the manuscript. All authors contributed to review and editing of the manuscript, and approved the final version to be submitted.

Conflict of interest
the Authors of this manuscript declare the absence of conflict of interests.

Fundings
No fundings to be declared.