Radiopathological correlation in borderline resectable pancreatic cancer

UNIVERSITÀ DI PISA

F

M

C

S

S

R

Direttore Prof. Carlo Bartolozzi

Tesi di Specializzazione

RADIOPATHOLOGICAL CORRELATION IN

BORDERLINE RESECTABLE PANCREATIC CANCER

Relatore

Chiar.mo Prof. Carlo Bartolozzi

Candidato

Dr. Andrea Di Giambattista

INDEX

Introduction

3

Materials and Methods

4

Results

8

Discussion

13

Conclusion

18

Figures

19

INTRODUCTION

Despite recent improvements in diagnosis and therapy, ductal adenocarcinoma of the pancreas is among the five most frequent causes of cancer-related death in Europe and the United States, with overall 5-year survival rates of 5–6%. 1, 2

The Cancer Statistics Review (2002–2008) from the US Surveillance Epidemiology and End Results showed that 53% of all patients with pancreatic cancer had concomitant distant metastases at the time of diagnosis; only 15–20% of the remaining patients were eligible for potentially curative resection, and up to 25% were staged as having locally advanced disease and thus were potentially “inoperable” because of involvement of surrounding major vessels. 3

For patients susceptible to surgical treatment, as long as free margins are achieved, surgery is the only treatment which can offer prolonged survival. 4

Whereas venous involvement, as long as venous reconstruction is possible, is a technically complex factor, but not determine unresectability, arterial involvement has long been a contraindication for surgical resection, due to a high morbility and mortality rate and limited oncological benefit.

The term “borderline resectable pancreatic cancer” (BRPC) is now used commonly to describe pancreatic cancer involving the mesentericoportal or arterial axis, in which the multidisciplinary team thinks an R0 resection can likely be achieved.5

So, preoperative staging is a particularly important step in patients with pancreatic cancer. Its purpose is to establish lesions which are considered borderline resectable (BRPC), those which require neoadjuvant treatment and cases where the tumor is inoperable or unresectable.

The purpose of our study is to evaluate the correlation between computed tomography (CT) and pathology in “borderline resectable” pancreatic cancer and discuss its prognostic implications.

MATERIALS AND METHODS

Patient population

A total of 30 patients (12 males, 18 females; mean age 64 years-old – range 52-79) who had undergone surgical resection with associated vascular resection (arterial or arterovenous) for pancreatic ductal adenocarcinoma between February 2003 and January 2014 were examined retrospectively. A neoadjuvant chemotherapy was performed in 20 of 30 patients (16 with FOLFIRINOX treatment protocol, 4 with GEMOX tratment protocol), in particular in all patients included since 2008.

CT examination

All patients underwent preoperative MDCT using a 64-row scanner (Light Speed Plus VCT, GE Medical System, Milwaukee USA) in accordance with standardized evaluation protocol.

Following basal scanning, high-concentration iodinated contrast medium (370-400) was injected intravenously at a flow rate of 4 mL/sec, using a dual-head pump injector. Post-contrastographic study included 3 phases: pancreatic phase (at 35-40 sec), venous phase (at 70 sec) and late phase (at 180 sec). The following scanning parameters were used: section reconstruction 2.5-1.25mm; section interval 1.25-0.625 mm; pitch 6 (High Speed-HS modality) or beam pitch 0.984:1; anode voltage 100-120 kV; 300-350 mA or automatic tube current modulation (smart mA, noise index 21); 0.6/0.8-second rotation speed; 512 matrix size.

Evaluation paramaters

Vascular involvement and evaluation of resectability

In the evaluation of arterial infiltration, we examined the celiac axis (CA) and its main branches (hepatic artery -HA- and splenic artery –SA-) and the superior mesenteric artery (SMA).

The degree of vascular infiltration was defined according to the following grading (Fig.1), published by our group in 2007: 6

• Grade 0: no contact between tumor and artery; • Grade I: focal tumor abutment on artery;

• Grade II: ≤180° encasement (IIa) or >180° encasement (IIb) of the artery, without reduction of its caliber ;

• Grade III: ≤180° encasement (IIIa) or >180° encasement (IIIb) with irregularity and distortion of the artery.

Each artery’s longitudinal extent of contact with the neoplastic tissue was also evaluated using multiplanar reconstructions (MPR) with dedicated software (Advantage Windows 4.4 GE Medical System, Milwaukee USA). The same grading was used to examine the portal (PV), superior mesenteric (SMV), and splenic vein’s involvement.

Each patient was given an overall resectability evaluation according to the criteria of the National Comprehensive Cancer Network (NCCN) version 2.2014. 5

The degree of celiac ganglia involvement was defined according to the following grading (Fig. 2):

• Grade I: celiac ganglia with normal appearance;

• Grade II: enlargement of celiac ganglia with preserved morphology;

• Grade III: enlargement of celiac ganglia with loss of their normal morphology. In the evaluation of the neural plexus, we examined the celiac plexus (and the hepatic and splenic branches) and the superior mesenteric artery neural plexus. Their involvement was graduated according to the following grading (Fig. 3-4):

• Grade 1: slightly thickening of neural fibers;

• Grade 2: a coarse tissue thickening along the course of neural plexus fibers; • Grade 3: gross periarterial fat tissue involvement.

Pathology investigations

Resection specimens are opened and partially sectioned immediately after resection, to aid fixation. The presence of a stent or a named vessel (e.g. portal vein, superior mesenteric vein, or arteries) is noted.

After orientation of the specimen, the common bile duct and pancreatic surgical margins are painted with India ink, the duodenum and common bile duct are opened, the ampulla of Vater is examined. The specimen may then be placed in a large volume of formalin and allowed to fix for 24−48 hours.

The transection margins of the pancreatic neck, common bile duct and duodenum/distal stomach are sampled prior to specimen dissection. Tissue blocks are taken to include the tumour where it approaches or involves anatomical structures relevant to (UICC TNM) T-staging, e.g. duodenum, ampulla, common bile duct, the entire peripancreatic tissue and vessels. Similarly, samples should be taken from the tumour and the adjacent resection margins. It is often difficult to identify accurately the invasive tumour front macroscopically and, therefore, extensive sampling of the tumour and the adjacent margins is performed.

When a segmental resection of the portal/superior mesenteric vein or artery is removed en-bloc with the pancreatoduodenectomy, the vessel with its proximal and distal ends are examined.

The perivisceral adipose tissue including lymph nodes, celiac ganglia and neural plexus fibers are sampled. Finally, samples of the ampulla of Vater, common bile duct and background pancreas are taken.

Statistical analysis

Statistical data analysis was performed using JMP version 7.0 (SAS) statistics software.

Statistical correlation was carried out using one-way ANOVA analysis of variance; χ-squared test was used when appropriate to compare the distribution of individual variables between groups. A survival analysis was performed using Kaplan-Meyer

method, and statistical comparisons of different factors were performed with the bivariate test. The COX model was used in a multivariate analysis. A p value of .05 or less was considered to indicate significance.

A correlation between the presence of arterial infiltration at histology and the CT grade of arterial, perineural plexus and celiac ganglia involvement was evaluated. An evaluation of the correlation between these parameters and survival of patients was also conducted.

RESULTS

MDCT findings

For all patients, CT results showed no suspicion of metastasis or carcinomatosis, and this date was also confirmed during intraoperative evaluation. According to the NCCN criteria, at the CT evaluation 17/30 patients were considered borderline resectable and 13/30 unresectable (including 3 for suspected involvement of the inferior vena cava). A downstaging was observed in 9 of 20 patients after neoadjuvant chemotherapy treatment; in particular, all 9 cases demonstrated a dimensional reduction of the tumor while only 2 cases demonstrated a reduction in the degree of vascular involvement (from IIb to IIa).

Surgery

Pancreaticoduodenectomy (PD) was performed in 2 patients, total pancreasectomy in 22 and distal pancreatectomy in 6; in 13 patients (43%) was associated a multivisceral resection (4 adrenal glands, 9 stomach, 2 kidney, 2 transverse colon, 3 gastric antrum).

In all 30 patiens, 70 arteries were resected: 7 celiac axis, 19 hepatic arteries, 28 splenic arteries and 16 superior mesenteric arteries. In particular, 4 arteries were resected in 1 patient, 3 arteries in 10 and 2 arteries in 17.

The table 1 summarizes the grades assigned to arteries undergoing resection. Twelve arteries were considered grade 0, 11 grade I, 27 grade IIa, 8 grade IIb (3 HA, 5 SMA), 7 grade IIIa (4 HA, 3 SA and 1 AMS) 5 grade IIIb (1HA, 3 SA, 1 SMA). Considering individual patients, 13 cases (45%) presented higher degree of arterial involvement of IIa, in 7 cases IIb, 5 cases IIIa and 5 cases IIIb.

0  

I  

IIa  

IIb  

IIIa  

IIIb  

Celiac  artery  

1  

1  

5  

-­‐  

-­‐  

-­‐  

7  

Hepatic  artery  

-­‐  

2  

9  

3  

4  

1  

19  

Splenic  artery  

11  

7  

5  

-­‐  

2  

3  

28  

Superior  

mesenteric  artery  

-­‐  

1  

8  

5  

1  

1  

16  

12  

11  

27  

8  

7  

5  

In regards to the longitudinal extension of contact with the neoplastic tissue, 22 vessels had a longitudinal contact ≥30mm, the remaining 36 a longitudinal contact <30mm (range 5-80mm, average 27,8mm).

Twenty-four patients (80%) underwent a venous resection. Thirty-three veins were resected (24 mesenteric-portal axes, 6 left renal veins and 3 tangential resection of the inferior vena cava). In these cases, the CT evaluation had assigned a grade I in 3 cases, IIa in 9 cases, IIb in 4, IIIa in 12 and IIIb in 5.

Regarding the evaluation of the involvement of celiac ganglia (Table 2), CT assigned a symmetrical grade 1 to six ganglia, a symmetrical grade 2 to ten ganglia and a symmetrical grade 3 to two ganglia. Asymmetry was observed in 12 patients (9 with grade 1-2 and 3 with grade 2-3).

Table 3 summarizes the scores assigned to the celiac and mesenteric plexus.

                                                                               Right  celiac  ganglia

Left  celiac  

ganglia

1  

2  

3  

1  

6  

4  

-­‐  

2  

5  

10  

1  

3  

-­‐  

2  

2  

                                                                                       AMS  plexus  

Celiac  

plexus  

0  

1  

2  

3  

0  

-­‐  

-­‐  

-­‐  

2  

1  

1  

3  

1  

9  

2  

5  

1  

-­‐  

3  

3  

-­‐  

2  

3  

-­‐  

Pathology

For all cases, diagnosis of ductal adenocarcinoma was confirmed during the histological evaluation. In 30/31 (97%) of cases, a negativity of the resection margins (R0 resection) was obtained; only 1 case demostrated a microscopic residual disease (R1) at the level of pancreatic resection after a distal pancreatectomy.

Vascular infiltration was confirmed in 12/70 (17%) resected arteries, and in 10/30 patients (33%). In particular, no artery classified as grade 0, I or IIa during the CT evaluation was infiltrated (NPV=100% of infiltration). In fact, the thickening of

perivascular fat tissue visible at CT corresponded to perineural infiltration, without involvement of the arterial wall; the infiltration of the adventitia was confirmed in 4/8 cases of arteries classified as grade IIb, 4/7 grade IIIa and in 4/5 grade IIIb (PPV= 50%, 57% and 80% of infiltration respectively).

Circumferential encasement of the artery (grade IIb and IIIb) showed a PPV=61% of infiltration while the reduction of the arterial caliber (grade IIIa and IIIb) a PPV=67%.

Histological arterial infiltration and grade assigned to the artery at the CT evaluation was strongly correlated (p <0.0001) (Fig 5). Histological arterial infiltration was also significantly correlated with average involvement of celiac ganglia (p = 0.007) and with the overall involvement of ganglia and plexus (p = 0.004) (Fig 6).

No correlation was demonstrated between the presence of histologic arterial infiltration and the longitudinal extension of the contact (p = 0.56) and between the group with neoadjuvant chemotherapy and the group treated with upfront surgery (p = 0.17).

Of the 33 resected veins, 0/3 veins classified as grade I, 2/9 IIa, 3/4 IIb, 10/12 IIIa and 5/5 IIIb were infiltrated. Considering venous infiltration, both circumferential encasement (grade IIb and IIIb) and the reduction of the venous caliber (grade IIIa and IIIb) showed a PPV=88% of infiltration. No correlation was demostrated between venous infiltration and the use of neoadjuvant chemotherapy (p = 0.46)

Survival Rates

At the time of last follow- up, 8 patients were alive (follow-up time between 5 and 29 months, median 18.5 months) and 22 were dead (follow-up time between 2 and 36 months, median 10.3 months); 3/30 (10%) patients died for perioperative complications within 3 months after surgery, and therefore excluded from survival statistical analysis.

The 6-month, 9-month and 1-year overall survival rate of the remaining 27 patients was 88%, 76% and 48% respectively; only 4 (16%) patients were alive at 24 months.

16/27 (59%) patients died due to recurrence after surgery (6/10 patients without neoadjuvant chemotherapy), 2 patients due to complications not related to the disease; 6/8 alive patients were free from disease (follow-up time between 4 and 29 months; mean 18.5 months)

In the group of patients with recurrence after surgery (18 cases), 3 had histologic confirmation of arteriovenous infiltration, 4 only of venous infiltration and 3 only of arterial infiltration; the remaining 8 cases, 3 had not performed the venous resection. No correlation was demostrated between survival and neoadjuvant chemotherapy treatment (p = 0.7), nor the type of surgery performed (p = 0.27), nor the number of resected arteries (p = 0.90), nor the type of vascular resection (only arterial or artero-venous) (p = 0.21).

Moreover, no correlation was demostrated between survival and the presence of arterial invasion at histology (p = 0.89) (Fig. 7) nor venous infiltration (p = 0.3). On the contrary, survival was significantly correlated with overall NCCN judgment resectability (p = 0.029) (Fig. 8) and with the presence or absence of recurrence after surgery (p = 0.0031) (Fig. 9).

The months of follow-up were significantly correlated with the perineural involvement, and in particular ganglia involvement (p = 0.0016) (Fig. 10).

Moreover, the presence of recurrence after surgery and death for recurrence were not correlated with the NCNN judgment, with the neoadjuvant chemotherapy, nor with histological arterial or venous infiltration.

DISCUSSION

According to the TNM staging system, pancreatic cancers involving the celiac axis or the superior mesenteric artery are staged as T4 and then clinical T4 pancreatic carcinoma are deemed unresectable because of presumed involvement of these arteries.

The term “borderline resectable pancreatic cancer” (BRPC) was introduced to describe pancreatic cancer in an intermediate stage between straightforwardly resectable and technically unresectable disease.

According to the last National Comprehensive Cancer Network (NCCN) guidelines 5, given absence of distant metastases, tumors with borderline resectability include the following:

• CT findings of venous involvement of the SMV/PV, even including short-segment venous occlusion with proximal and distal sufficient vessel length allowing safe reconstruction;

• Encasement of the gastroduodenal artery up to the HA, with either short-segment encasement or direct abutment of the HA without extension to the CA;

• Tumor abutment of the SMA with no more than 180° of the vessel wall circumference.

Tumors considered to be unresectable demonstrate the following: • Distant metastases

• Greater than 180° SMA encasement or any celiac abutment (for tumors of the head) or SMA or celiac encasement greater than 180 degrees (for tumors of the body-tail)

• Unreconstructible SMV/portal occlusion

• Aortic or inferior vena cava (IVC) invasion or encasement

In recent decades, a small number of groups have been challenge vascular criteria for resectability 7-10 and the exceptionally low rate of “resectability” of pancreatic cancer

establishment of experienced high-volume centers. So, arterial resections are performed more frequently in extended resections of the pancreatic head or body, with the aim of an R0 resection.

However, there is still no consensus on the surgical indication for arterial resection in cases of vascular infiltration recognized at the pre-operative imaging.

In a meta-analysis including 26 studies (published from 1977 to 2010) with 366 patients undergoing arterial resection and 2,243 patients without arterial resection, the survival analysis did not show a benefit compared to patients who underwent only venous resection. However, compared to patients who did not undergo resection, the 1-year survival rate was three times greater for patients with arterial resection.

The authors concluded that arterial resection is only justified in highly selected patients. 11-12

However, the same meta-analysis showed significantly greater perioperative morbidity (median 53.6%) and mortality (median 11.8%) in patients undergoing arterial resection. 11 This conclusion agrees with similar conclusions in previous studies from high-volume centers, although a case-matched controlled study by Bachellier et al reported a similar 3-year survival rate of patients with and without arterial resection. 13-16

In our series, only 3 patients (10%) subject to arterial resection died of perioperative complications. No correlation was found between survial rate and type of surgery (p = 0.27), nor the number of resected arteries (p = 0.90), nor the type of vascular resection (arterial or artero-venous) (p = 0.21), confirming that, in selected patients and in high-volume centers, the arterial resection is not associated with a greater perioperative mortality.

Neoadjuvant chemotherapy has played an important role in patients with arterial infiltration, in order to increase the rate of R0 resection. 17,18

Our study included a total of 30 patients who had undergone surgical resection with associate vascular resection (arterial or arterovenous); according to the NCCN criteria, at the CT evaluation, 17 patients were considered borderline resectable and

13 unresectable. Neoadjuvant chemotherapy was performed in 20 patients. Histology demonstrated a single R1 case due to a microscopic infiltration of the pancreatic resection margin after left pancreatectomy.

According to literature data, 19 a dimensional reduction of the tumor was observed in 9 out of 20 patients; vascular downstaging was observed in only 2 cases, without any change in the overall resectability evaluation.

Also, with limitations due to small sample size and to the different chemotherapy protocols administered to the patients, no significant difference between having carried out neoadjuvant chemotherapy or not, was observed – in regards to both presence of histological arterial infiltration and survival rate.

Therefore, an accurate vascular staging is diriment to select patients for upfront surgery or neoadjuvant chemotherapy.

Multi-detector computed tomography (MDCT) has been widely accepted as the best imaging technique for the diagnosis and staging of pancreatic cancer; most importantly, it is considered the gold standard in assessing vascular infiltration, with a reported diagnostic accuracy ranging from 85 to 93% (80-90% sensitivity and 89-100% specificity), considering both arteries and veins. 20-23

However, the accuracy of CT-based diagnosis of arterial infiltration needs to be evaluated critically. 24

According to the CT-based classification developed at the M.D. Anderson Cancer Center, 25 specific criteria currently used for vascular (arterial and venous) invasion are based mostly on the extent of circumferential vessel involvement. Although a diagnostic sensitivity of up to 97% has been reported for arteries, the specificity ranges from 67%-91%. This is due to inadequate differentiation between fibrous, tumor-mimicking adhesion and true cancerous invasion, particularly in patients subject to neoadjuvant radio-chemotherapy, suggesting that patients could be incorrectly classified as having unresectable disease. 26-29

Moreover, the reported inadequate diagnostic accuracy of the preoperative CT in arterial infiltration staging can be justified by the fact that CT grading systems of

vascular invasion are based on identical semiologic criteria for both veins and arteries, despite their different anatomical structure. In fact, in our study the infiltration was confirmed in 12/70 (17%) resected arteries and in 20/33 (60%) resected veins. On the basis of our CT grading of vascular involvement, both circumferential encasement (grade IIb and IIIb) and reduction of caliber (grade IIIa and IIIb) showed a PPV of infiltration of 88% for veins, while for arteries we observed a PPV of 61% grades IIb and IIIb and a PPV of 67% for grades IIIa and IIIb. To notice that none among the 27 arteries classified as IIa (≤180° encasament) resulted to be infiltrated.

No correlation between histological infiltration and the longitudinal contact between the tumor and the artery was observed, even considering a cut-off ratio of 3cm (p = 0.56).

Therefore, in our study we observed a tendency to overestimate the arterial infiltration at CT, while the problem that more intensely worried the literature was instead the possible underestimation of arterial infiltration. The overestimation of the arterial involvement can be justified on the basis of the characteristic spreading of the ductal carcinoma along perineural plexus fibers, which originate from celiac ganglia and distribute around arterial structures. 30-35 This can further justify the need to use different criteria for grading venous and arterial infiltration in preoperative CT local staging.

The development of modern imaging technology enabled to demonstrate the anatomy of celiac ganglia and the celiac and mesenteric perineural plexus fibers and to identify their pathologic appearance in case of neoplastic invasion. 36-42

Mochizuki et al. 43 made a point-by-point imaging-pathological correlation between surgical specimens of pancreatic head cancer (resected en bloc with the surrounding vessels and connective tissues including plexus) and multiplanar reconstruction of CT images corresponding exactly to the pathological sections. The author identified CT signs suggestive of neural plexus fibers invasion, useful to decide the surgical

strategy, including extended resection with surrounding major vessels. However, in this study, no correlation with possible infiltration of the arterial wall was considered. Zhang et al. 44 instead demonstrated that vascular invasion was often present in the setting of nerve invasion and that the degree of vascular invasion was correlated with that of nerve invasion (P < 0.005)

On the basis of this data, we therefore evaluated perineural plexus and celiac ganglia infiltration assigning each a grade from 1 to 3. We obtained a significant correlation between the presence of histological arterial infiltration and the overall perineural (p = 0.004) and celiac ganglia involvement (p = 0.007). Moreover, the perineural plexus infiltration is well known as one of the factors that predicts a poor prognosis for pancreatic cancer, with an incidence of 35-81% in patients with invasive cancer. 32,39 In our study, no correlation between survival and histological arterial infiltration (p = 0.89) nor venous infiltration (p = 0.3) existed, whereas we demonstrated a significant correlation with the overall resection judgment (p = 0.029). We also demonstrated a correlation between perineural involvement, particularly the involvement of the celiac ganglia, and months of follow-up (p = 0.0016).

On the basis of these results, we evaluated whether the perineural involvement could influence the correlation between survival and NCCN resecability judgment. We demonstrated that all patients with a score <7 of overall perineural involvement were classified as “borderline resectable” with a significant survival curve (p = 0.0047), while the group with a score > 7 there were significant differences between “non resectable” and “borderline resectable” survival curves (p = 0.036) (Fig. 11). The infiltration of the celiac ganglia and perineural plexus played a key role in the preoperative staging, being a prognostic factor to be considered in patients with pancreatic ductal adenocarcinoma and a factor crucial for the interpretation of arterial infiltration imaging.

CONCLUSION

The strong tropism of ductal carcinoma for perineural spreading along arterial structures, justifies the need to use different criteria for grading venous and arterial infiltration in preoperative CT local staging; current CT grading system, originally developed for veins, should therefore be revised because of the many anatomic and structural differences between arteries and veins and the different relationships of these vessels with extrapancreatic nerve plexus. In addition, the involvement of neural plexus and celiac ganglia, being crucial in predicting arterial involvement and prognosis, should be reported in the preoperative CT staging to improve the accuracy in the local spreading description and thus to guide the therapeutic management.

FIGURES

Fig. 1. Grade 0: no contact between lesion and vessel (a). Grade I: focal tumor abutment on vessel (b). Grade IIa: ≤180° tumor encasement without reduction in caliber of the vessel (c). Grade IIb >180° tumor encasement, without reduction in caliber of the vessel (d). Grade IIIa: ≤180° encasement with reduction of the lumen

Fig. 2. Celiac ganglia involvement: Grade I: celiac ganglia with normal appearance (a). Grade II: enlargement of celiac ganglia with preserved morphology (b). Grade III: enlargement of celiac ganglia with loss of their normal morphology (c).

Fig. 3. Celiac plexus involvement: Grade 1: slightly thickening of neural fibers (a). Grade 2: coarse tissue thickening along the course of neural plexus fibers (b). Grade 3: gross periarterial fat tissue involvement (c).

Fig. 4. AMS plexus involvemen: Grade 1: slightly thickening of neural fibers (a). Grade 2: coarse tissue thickening along the course of neural plexus fibers (b). Grade 3: gross periarterial fat tissue involvement (c).

Fig. 5. Correlation between histological arterial infiltration and grade assigned to the artery at the CT (p <0.0001) .

Fig. 6. Correlation beetween histological arterial infiltration and total involvement of ganglia and plexus (p = 0.004) and average involvement of celiac ganglia (p = 0.007).

Fig. 7. Comparison of survival in patients with or without histological arterial infiltration (p = 0.89)

Fig. 8. Comparison of survival between patients classified as borderline resectable and patients classified as non resectable (p = 0.029).

Fig. 9. Comparison of survival between patients with or without recurrence after surgery (p = 0.0031).

Fig. 10. Correlation between months of follow-up and ganglia involvement (p = 0.0016)

Fig. 11. Survival curve in group with a score <7 of overall perineural involvement (a) (p = 0.0047). The survival curve in group with a score >7 of overall perineural involvement (b) demostrate a significant difference between “non resectable” and “borderline resectable”patients (p = 0.036).                                        

REFERENCES

1. Malvezzi M, Bertuccio P, Levi F, La Vecchia C, Negri E. European cancer mortality predictions for the year 2012. Ann Oncol 2012; 23:1044-52.

2. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2012. CA Cancer J Clin 2012; 62:10-29

3. National Cancer Insitute. SEER Cancer Statistics Review, 1975-2009 (Vintage 2009 Populations). Available from:

http://seer.cancer.gov/csr/1975_2009_pops09/

4. Gaedcke J, Gunawan B, Grade M, Szoke R, Liersch T, Becker H, et al. The mesopancreas is the primary site for R1 resection in pancreatic head cancer: relevance for clinical trials. Langenbecks Arch Surg. 2010;395:451–8

5. Tempero MA, Malafa MP. Pancreatic adenocarcinoma, version 2.2014: featured updates to the NCCN guidelines. 2014. J NatlComprCancNetw. 12(8):1083-93),

6. Mazzeo S, Cappelli C, Caramella D, Del Chiaro M et al. Evaluation of vascular infiltration in resected patients for pancreatic cancer: comparison among multidetector CT, intraoperative findings and histopathology. Abdominal Imaging 2007; 32(6):737-42

7. Amano H, Miura F, Toyota N, Wada K, Katoh K, Hayano K, et al. Is pancreatectomy with arterial reconstruction a safe and useful procedure for locally advanced pancreatic cancer? J HepatobiliaryPancreat Surg. 2009; 16:850–7.

8. Ravikumar R, Holroyd D, Fusai G. Is there a role for arterial reconstruction in surgery for pancreatic cancer? World J Gastrointest Surg. 2013; 5:27–9. 9. Takahashi S, Kinoshita T, Konishi M, Gotohda N, Kato Y, Kinoshita T, et al.

Borderline resectable pancreatic cancer: rationale for multidisciplinary treatment. J HepatobiliaryPancreat Sci. 2011; 18:567–74.

10. Callery MP, Chang KJ, Fishman EK, Talamonti MS, William Traverso L, Linehan DC. Pretreatment assessment of resectable and borderline resectable pancreatic cancer: Expert consensus statement. Ann SurgOncol. 2009; 16:1727–33

11. Mollberg N, Rahbari NN. Arterial resection during pancreatectomy for malignant disease: a systematic review and meta-analysis. AnnSurg 2011; 254: 882-93

12. Bockhorn M, Uzunoglu FG. Borderline resectable pancreatic cancer: A consensus statement by the International Study Group of Panceratic Surgery. Surgery 2014; 155:977-88].

13. Ouaissi M, Hubert C, Verhelst R, Astarci P, Sempoux C, Jouret-Mourin A, et al. Vascular reconstruction during pancreatoduodenectomy for ductal adenocarcinoma of the pancreas improves resectability but does not achieve cure. World J Surg 2010; 34:2648-61.

14. Bockhorn M, Burdelski C, Bogoevski D, Sgourakis G, Yekebas EF, Izbicki JR. Arterial en bloc resection for pancreatic carcinoma. Br J Surg 2011; 98:86-92.

15. Hartwig W, Hackert T, Hinz U, Hassenpflug M, Strobel O, Buchler MW, et al. Multivisceral resection for pancreatic malignancies: risk-analysis and long-term outcome. Ann Surg 2009; 250:81-7.

16. Bachellier P, Rosso E, Lucescu I, Oussoultzoglou E, Tracey J, Pessaux P, et al. Is the need for an arterial resection a contraindication to pancreatic resection for locally advanced pancreatic adenocarcinoma? A case-matched controlled study. J SurgOncol 2011; 103:75-84.

17. Papalezova KT, Tyler DS, Blazer DG 3rd, Clary BM, Czito BG, Hurwitz HI, et al. Does preoperative therapy optimize outcomes in patients with resectable pancreatic cancer? J SurgOncol 2012; 106:111-8)

18. Gillen S, Schuster T, Meyer Zum BC, Friess H, Kleeff J. Pre- operative/neoadjuvant therapy in pancreatic cancer: a sys- tematic review and meta-analysis of response and resection percentages. PLoSMed 2010; 7:e1000267

19. Katz MH, Fleming JB, Bhosale P, Varadhachary G, Lee JE, Wolff R, et al. Response of borderline resectable pancreatic cancer to neoadjuvant therapy is not reflected by radiographic indicators. Cancer 2012; 118:5749-56

20. Tamm EP, Bhosale PR, Imaging of pancreatic ductal adenocarcinoma: State of the art. World J Radiol 2013; 5(3): 98-105.

21. Brugel M, Rummeny EJ. Vascular invasion in pancreatic cancer: value of multislice helical CT. Abdominal Imaging 2004; 29: 239-245.

22. Phoa SS, Tilleman EH. Value of CT criteria in predicting survival in patients with potentially resectable pancreatic head carcinoma. J SurgOncol 2005; 91:33-40.

23. Al-Hawary MM, Francis IR. Pancreatic ductal adenocarcinoma radiology reporting template: consensus statment of the Society of Abdominal Radiology and the American Pancreatic Association. Radiology 2014; 270(1): 248-60

24. Buchs NC, Chilcott M, Poletti PA, Buhler LH, Morel P. Vascular invasion in pancreatic cancer: imaging modalities, preoperative diagnosis and surgical management. World J Gastroenterol. 2010; 16:818–31.

25. Varadhachary GR, Tamm EP, Abbruzzese JL, Xiong HQ, Crane CH, Wang H, et al. Borderline resectable pancreatic cancer: definitions, management, and role of preoperative therapy. Ann SurgOncol 2006; 13:1035-46.

26. Wong JC, Lu DS. Staging of pancreatic adenocarcinoma by imaging studies. ClinGastroenterolHepatol 2008; 6:1301-8.

27. Yekebas EF, Bogoevski D, Cataldegirmen G, Kunze C, Marx A, Vashist YK, et al. En bloc vascular resection for locally advanced pancreatic malignancies infiltrating major blood vessels: perioperative outcome and long-term survival in 136 patients. Ann Surg 2008; 247:300-9.

28. Nakayama Y, Yamashita Y, Kadota M, Takahashi M, Kane- mitsu K, Hiraoka T, et al. Vascular encasement by pancreatic cancer: correlation of CT findings with surgical and patho- logic results. J Comput Assist Tomogr 2001; 25:337-42.

29. Li H, Zeng MS, Zhou KR, Jin DY, Lou WH. Pancreatic adenocarcinoma: the different CT criteria for peripancre- atic major arterial and venous invasion. J Comput Assist Tomogr 2005; 29:170-5

30. Noto M, Miwa K, Kitagawa H, Kayahara M, Takamura H, Shimizu K, OhtaT. Pancreas head carcinoma: frequency of invasion to soft tissue adherent to the superior mesenteric artery. Am J SurgPathol2005; 29: 1056-1061

31. Nano M, Lanfranco G, Ferronato M, Dal Corso H, Solej M. Contribution to the study of innervation of the pancreas with a view to its relevance to neoplasm surgery. ChirItal 2001; 53: 587-594

32. Mitsunaga S, Hasebe T, Kinoshita T, Konishi M, Takahashi S, Gotohda N, Nakagohri T, Ochiai A. Detail histologic analysis of nerve plexus invasion in invasive ductal carcinoma of the pancreas and its prognostic impact. Am J SurgPathol 2007; 31: 1636-1644

33. Tian H, Mori H, Matsumoto S, Yamada Y ,Kiyosue H, Ohta M, Kitano S. Extrapancreatic neural plexus invasion by carcinomas of the pancreatic head region: evaluation using thin-section helical CT. RadiatMed 2007; 25: 141-147

34. Takahashi T, Ishikura H, Kato H, Tanabe T, YoshikiT. Intra-pancreatic, extra-tumoral perineural invasion (nex). An indicator for the presence of retroperitoneal neural plexus invasion by pancreas carcinoma. Acta PatholJpn 1992; 42: 99-103

35. Yi SQ, Miwa K, Ohta T, Kayahara M, Kitagawa H, Tanaka A, Shimokawa T, Akita K, Tanaka S. Innervation of the pancreas from the perspective of perineural invasion of pancreatic cancer. Pancreas 2003; 27: 225-229

36. Hou-Dong Zuo, Xiao-Ming Zhang. CT and MR imaging patterns for pancreatic carcinoma invading the extrapancreatic neural plexus (PartI): Anatomy, imaging of the extrapancreatic nerve. World J Radiol. 2012. 28; 4(2): 36-43.].

37. Hou-Dong Zuo, WeiTang, Xiao-Ming Zhang, Qiong-Hui Zhao, Bo Xiao. CT and MR imaging patterns for pancreatic carcinoma invading the extrapancreatic neuralplexus (Part II): Imaging of pancreatic carcinoma nerve invasion. World J Radiol 2012. 28; 4(1): 13-20

38. Fukuda T, Iwanaga S, Sakamoto I, Aso N, Nagaoki K, Hayashi K, Yamaguchi H, Okudaira S, Tomioka T, OkimotoT. CT of neural plexus invasion in common bile duct carcinoma. J Comput Assist Tomogr 1998; 22: 351-356

39. Nakao A, Harada A, Nonami T, Kaneko T, TakagiH. Clinical significance of carcinoma invasion of the extrapancreatic nerve plexus in pancreatic cancer. Pancreas 1996; 12: 357-361

40. Tezel E, Kaneko T, Sugimoto H, Takeda S, Inoue S, Nagasaka T, Nakao A. Clinical significance of intraportal endovascular rultrasonography for the diagnosis of extrapancreatic nerve plexus invasion by pancreatic carcinoma. Pancreatology 2004; 4: 76-81

41. Zhang XM, Mitchell DG, Byun JH, Verma SK, Bergin D, Witkiewicz A. MR imaging for predicting the recurrence of pancreatic carcinoma afte rsurgical resection. Eur J Radiol 2010; 73: 572-578

42. Makino I, Kitagawa H, Ohta T, Nakagawara H, Tajima H, Ohnishi I, Takamura H, Tani T, Kayahara M. Nerve plexus invasion in pancreatic cancer: spread patterns on histopatho- logic and embryological analyses. Pancreas 2008; 37: 358-365

43. Mochizuki K, GabataT. MDCT findings of extrapancreatic nerve plexus invasion by pancreas head carcinoma: correlation with en bloc pathological specimens and diagnostic accuracy. EurRadiol 2010; 20:1757–1767

44. Zhang XM, Mitchell DG, Witkiewicz A, Verma S, Bergin D. Extrapancreatic neural plexus invasion by pancreatic carcinoma: characteristics on magnetic resonance imaging. AbdomImaging 2009; 34: 634-641.