Pancreatic Cancer (Staging)

Imaging is of paramount importance in both the diagnosis and staging (and thus management) of pancreatic ductal adenocarcinoma (PDAC). Only about 20% of patients with PDAC have resectable tumours at presentation1-3 and it is therefore an important role of imaging to identify those with potentially curable disease (surgery +/- neoadjuvant therapy), and to prevent unnecessary surgery in those with non-resectable lesions (surgical morbidity decreases the net survival benefit) or metastases.

Staging of PDAC

• The 8^th Edition of the American Joint Committee on Cancer (AJCC) employs the TNM methodology (T-tumour, N-nodal involvement, M- non-nodal /distant metastases) 4, 5. This latest iteration of the staging system is based on pathology and has, as its focus, the improved assessment of prognosis rather than management guidance. It emphasises the importance of the size of the tumour as a surrogate for tumour biology aiming to correlate with overall survival. Therefore, the T stage is now size-based apart from the T4 tumours (see Table 1, below).

• The 8^th edition of the AJCC pathology staging system 6, 7 (see Table 1, below) removes resectability from the T4 category, since there is inconsistency among institutions, and the criteria for resectability change as surgical techniques advance. With regard to the identification of non-resectability on imaging, this has been subject to significant interobserver variation in the past.

• N1 is defined as one to three metastatic lymph nodes, and N2 as four or more metastatic lymph nodes. A number-based categorization of metastatic lymph nodes results in better prognostic stratification.

• Studies designed to validate the AJCC 8th edition 8-11 have shown increased consistency of T staging (compared to the 7^thedition) but the revised pathology T stage remains poorly associated with survival, whereas the revised N-staging is highly prognostic .

• In view of the emphasis of the 8^th Edition of the AJCC system on tumour size at pathology, it is important that staging by imaging enables tumour measurement to correlate with the size of the resected specimen. It has previously been shown that CT and MRI underestimate tumour size. However a recent study 12 comparing parenchymal and portal venous contrast-enhanced  CT phases and multiple MRI sequences , as well as interobserver variation, indicated that the parenchymal phase of contrast-enhanced CT was the most accurate of the CT phases in measuring tumour size, and that interobserver agreement was good for all CT and MR images. MRI significantly underestimated tumour size. However, it should be noted that this study only included 29 patients.

• For guidance to management, it is important to assess resectability on imaging in order to identify:

• those with resectable, potentially curable disease

• to prevent unnecessary surgery in those with non-resectable lesions, and/or distant metastases

• to identify those patients with “borderline resectable” tumours who may benefit from neoadjuvant therapy by increasing the likelihood of R0 resections at surgery 7.

• The most widely used guidelines for assessing resectabilty PDAC Resectability of PDAC at imaging are the National Comprehensive Cancer Network (NCCN) guidelines 6.

Table 1: 8^th Edition AJCC Pathology staging

(from Kulkarni et al (2020)  7

Abbreviations:

CA =Coeliac axis/artery; CHA = Common Hepatic artery; GDA = Gastroduodenal Artery

HA = Hepatic artery; IVC = Inferior Vena Cava; PV = Portal vein; SMA = Superior Mesenteric Artery; SMV = Superior Mesenteric Vein; Ao = Aorta

• Abbreviations:

CA =Coeliac axis/artery; CHA = Common Hepatic artery; GDA = Gastroduodenal Artery; HA = Hepatic artery; IVC = Inferior Vena Cava; PV = Portal vein; SMA = Superior Mesenteric Artery; SMV = Superior Mesenteric Vein; Ao = Aorta

• Definitions: Abutment is < 180 degree contact; Encasement is  >180 degree contact

PDAC Resectability

• Management of PDAC depends on assessment of primary tumour resectability and the presence of metastatic disease. Only 10-20% of patients with PDAC have surgically resectable disease at presentation 13, 14

• Of those patients deemed resectable prior to surgery, 14-30% are found to be unresectable at the time of surgery 15due to  locally advanced disease and/or distant metastases.

• The rate of R0 (resection margin-negative) resection rates in patients presenting with PDAC remains low.

• However, increasingly, surgeons in high-volume specialised centres are performing more radical resections, with vascular reconstructions where peri-pancreatic vessels are involved.

• A study published in 2017 based on the Netherlands Cancer Registry 2 found that, between 2009 and 2013, exploration and resection rates increased, but one-third of patients who had surgical exploration for pancreatic cancer did not undergo resection. Non-resectional surgery doubled the 30-day mortality rate compared with that in patients undergoing tumour resection.

• The most widely used guidelines for assessing resectabilty of PDAC at imaging are the National Comprehensive Cancer Network (NCCN) guidelines 6.

• However, there are several other published guidelines from international authorities related to resectability, which combine published evidence with expert consensus.

These include:

• Alliance for Clinical Trials in Oncology (ACTO) 16,

• American Hepato-Pancreato-Biliary Association (AHPBA), Society for Surgery of the Alimentary Tract (SSAT), Society for Surgical Oncology (SSO) 17

• International Association of Pancreatology 18

• International Study Group of Pancreatic Surgery (ISGPS)19

• Japanese Society of Hepato -Biliary-Pancreatic Surgery 20

• Of these the NCCN guidelines (2023) are the most recent.6

• Japan Pancreas Society 21

For guidance to management, it is important to assess resectability on imaging:

To identify those patients with: 

1. RESECTABLE TUMOURS, potentially curable disease and thus suitable for upfront surgical resection. 

• There is general agreement among the various guidelines that a resectable tumour has:

• no contact with the coeliac axis (CA), Superior Mesenteric Artery (SMA) or Common Hepatic Artery (CHA).

• With regard to venous structures, there is no contact with Superior Mesenteric Vein (SMV) or Portal Vein (PV), but there may be abutment (contact < 180 degrees) without contour abnormality of the veins.

• There must be absence of distant metastases.

• Categorisation of tumours into the resectable category may depend on tumour location within the pancreas; those cancers involving the body or tail of the pancreas that involve the splenic artery or vein are deemed resectable as  the patient may undergo splenectomy at the time of surgery.

• However, many patients with resectable PDAC who undergo upfront surgery have an incomplete (R1) resection, and /or have a node-positive resection 22. Patients must be selected for surgery based on the likelihood of achieving complete curative resection with negative resection margins (R0) 23.

• Therefore, there is a move towards recommending treatment of at least some patients with resectable tumours with multi-agent neoadjuvant therapy (NAT) prior to surgery 24. Further evidence is required 25. A recent meta-analysis 26 reported that NAT in patients with resectable PDAC  increases the R0 resection rate and decreases positive node resection rate, but without significant effect on overall survival time.   It remains to be fully determined which patients with resectable tumours based on imaging +/- biological and or other markers, are likely to benefit from NAT. 27, 28. However there is evidence that features of resectable PDAC that may indicate benefit of NAT include tumour location in the pancreatic head, CA 19-9 >100units/ml, tumour size >20mm. -these are independent predictors of poor prognosis and together may provide a prognostic scoring system and thus a selection process for NAT 29 - and extra-pancreatic peri-neural invasion (EPNI)30. EPNI is thought to be a significant factor in positive surgical resection margins and tumour recurrence. A recent limited study 30 reported high sensitivity of CT for detection of EPNI, but only moderate specificity (since fibrosis and inflammation are indistinguishable). Nevertheless, the authors suggest that detection of EPNI in patients with otherwise resectable tumours may be an indication for neoadjuvant therapy.

The NCCN guidelines 6 and the European Society for Medical Oncology (ESMO)31 recommend that only patients with a high probability of R0 resection should have upfront surgery and that where there is doubt or where a R1 resection is of high probability, patients should receive NAT. 

2. NON-RESECTABLE PDAC in order to prevent unnecessary surgery. This is usually based on the finding of locally advanced disease due to vascular involvement and/or distant metastases. These patients may be palliated or, in the absence of detectable metastases, sometimes given induction chemo(radio)therapy.

The criteria for locally advanced disease  listed below are derived from the latest (2023) version of the National Comprehensive Cancer Network (NCCN) guidelines include 6:

• Encasement (>180 degrees) of SMA, CA or first branch jejunal artery branches of the SMA or involvement of the aorta.

• Venous involvement that renders the tumour unresectable includes involvement (or thrombus) of the SMV or PV that is not amenable to venous reconstruction, or involvement of proximal jejunal veins that drain into the SMV

3. BORDEELINE RESECTABLE PDAC. These tumours are those involving nearby structures and are neither clearly resectable nor clearly unresectable, with a high chance of RI resection margins. Such patients may benefit from neoadjuvant therapy in an attempt to downsize tumour staging prior to surgery and thus improve surgical outcomes by increasing the likelihood of R0 resections 7.

• The current standard of care for borderline resectable PDAC is neo-adjuvant therapy followed by surgical resection. A meta-analysis of large studies has shown that neoadjuvant chemo(radio)therapy can improve survival in these patients as well as some patients with resectable tumours 28

• The various published guidelines differ to some degree in the criteria used to define “borderline resectable” PDAC 7 18, 32

The NCCN guidelines , in the 2023  version 6, include the following criteria:

• SMV/PV encasement, abutment or contour distortion /narrowing or thrombosis of the vein but amenable to venous reconstruction.

• Abutment of / contact with the IVC.

• For tumours involving the head or uncinate process of the pancreas:

•  Contact of solid tumour with the CHA without extension to the CA or HA bifurcation allowing for resection and arterial reconstruction

• Abutment of the SMA

• Tumour contact with variant arterial anatomy.

• For tumours of the body and tail of the pancreas:

• Abutment of the CA (<180 degrees).

• The most common criterion for defining borderline resectability is <180 tumour interface to SMA 3(3)

Whereas there is largely agreement on what constitutes resectable and non-resectable tumours , there may be disagreement among institutional MDTs ( as shown by one European study of a small number of patients 33 ). In addition, there is some disagreement among the various authorities on the definition of “borderline resectable”, which may also vary over time, in part due to advances in surgical techniques. Since the question of borderline resectability is concerned with the relationship of the tumour to local blood vessels, it has been suggested that, rather than attempt to categorise cases into “borderline resectable”, radiologists should report in detail the relationship with the relevant vessels.

It is important to note that the definitions regarding resectability are likely to change in the coming years as systemic treatment improves also that some large refer centres are now routinely undertaking selected arterial resections en bloc and resecting solitary liver metastases.

Notes on assessing resectability:

• Some consensus guidelines differ with regard to any contact of solid tumour with the CA. Some categorise this as unresectable 32, particularly for tumours of the body/tail of pancreas while others regard this as borderline resectable 6

• It is important that all decisions regarding resectability are made by discussion at multidisciplinary meetings and informed by results of all relevant staging and clinical data.6, 19

• It is generally regarded that pancreatic surgery should be carried out at high-volume centres.

• Expert consensus and relevant studies have shown that the use of structured reporting of CT scans in PDAC provides superior evaluation of PDAC and facilitates surgical planning 34, 35 36. Templates may be locally devised by multidisciplinary consensus or are available on-line 6.

CT in assessing resectability – Computed Tomography

Multidetector CT is the preferred staging investigation of first choice 7, 34.

Staging is based on the determination of tumour size, location within the pancreas, local extent including involvement of surrounding vessels and other structures, and the presence of nodal and distant metastatic lesions and requires a dual phase examination at minimum – pancreatic (parenchymal) post-contrast phase and portal venous phase.34. The pancreatic phase is optimal for demonstrating tumour size 12 and will show peri-pancreatic arteries. Portal venous phase will demonstrate the porto-mesenteric venous system and hepatic metastases.

CT Protocol

• CT should be performed, if possible, before any planned placement of a biliary stent, due to the possible artifact resulting from the stent.

• The CT protocol should be undertaken on a helical scanner using thin section thickness (preferably submillimetre) at intervals equivalent to the section thickness.

•  Oral contrast should preferably be a neutral agent.

•  Intravenous contrast medium should be of high concentration (>300mg I/ml) injected at 3-5ml/sec.

• The pancreatic parenchymal phase is obtained at 40-45 seconds and the portal venous phase at 65-70 seconds following the start of injection.

• Image reconstruction in the axial plane should be at 2-5mm thickness; multiplanar reformats (coronal +/- sagittal) at 2-3mm thickness. Maximal Intensity Projections or 3D volumetric sections for vascular evaluation. 6, 34.

• A triple phase or quadruple phase scan should be considered in light of evidence that this gives a greater sensitivity for liver metastases 37.

• The Expert Panel constituted by the AJR recommends use of Dual-energy CT, if available, with a dedicated biphasic CT to increase tumour conspicuity and improve staging 38.

CT and MRI are approximately equivalent for local staging of PDAC. However, a recent meta-analysis comparing CT and MRI for the detection of liver metastases in PDAC found that MRI was significantly more sensitive than CT, although, despite the Society of Abdominal Radiology recommending a dual phase protocol (see above), there is some evidence that a triple or quadruple phase CT scan improves sensitivity for liver metastases 37.

Chest CT

In order to exclude thoracic metastases, the NCCN guidelines 6 state that CT scan coverage “can be extended to cover the chest and pelvis for complete staging as per institutional preferences”. However, the utility of chest CT as a routine part of staging PDAC has been questioned. 39. The prevalence of indeterminate pulmonary nodules in patients with PDAC is reported to range from 18 to 71% and the preoperative presence of such nodules does not demonstrate an association with overall survival after surgery 39, 40

Structured reporting

Structured reporting of CT scans provides superior evaluation of PDAC, facilitates surgical planning 34-36 and improves inter-reader agreement 41. 12 key features have been included in a reporting template. These include: tumour location, probable diagnosis, TNM staging, superior mesenteric artery and vein involvement, presence of vessels with thrombosis, presence of replaced right hepatic artery, additional aberrant arterial anatomy, presence of atherosclerotic plaques, and distance of tumour to the superior mesenteric vein.

Prediction of resection margin status by CT

• 70-85% of patients determined to have resectable tumours on CT are able to undergo resection 6.

• Multidetector CT performance for detecting vascular invasion demonstrates high specificity (82% to 100% ) but with less sensitivity (70-96%)23. A 2014 meta-analysis showed a pooled specificity for vascular invasion of 92% and sensitivity of  63% 42. False-positive vascular invasion at CT is rare in patients who have not received neoadjuvant therapy.

• Park et al 43 in a series of 75 patients with either resectable or borderline resectable PDAC of the pancreatic head found CT to have high specificity, but poor sensitivity, in predicting pathological CRM involvement. This study, however, included a relatively small number of patients and excluded those patients with tumours deemed unresectable on CT.

• Bae et al 44 found tumour size, peri-tumoural fat stranding, suspicious distant metastasis and portal vein involvement were independent predictors of a positive resection margin at surgery. These authors also found that Common Hepatic Artery involvement, N1 stage and R1/2 resection margin were predictors of poor overall survival. It should be noted, however, that peri-tumoural infiltration/peri-neural infiltration and/or lymphatic invasion is an unreliable sign due to the difficulty of distinguishing tumour infiltration from peri-pancreatic inflammation on imaging.

• Extra-pancreatic peri-neural invasion (EPNI) is thought to be a significant factor in positive surgical resection margins and tumour recurrence. A recent limited study 30 reported high sensitivity of CT for detection of EPNI, but only moderate specificity (since fibrosis and inflammation are indistinguishable). Nevertheless, the authors suggest that detection of EPNI in patients with otherwise resectable tumours may be an indication for neoadjuvant therapy.

• There is evidence that features of resectable PDAC that may indicate benefit of NAT include location in pancreatic head, CA 19-9 >100units/ml, tumour size >20mm -these are independent predictors of poor prognosis and together may provide a prognostic scoring system and thus a selection process for neo-adjuvant therapy.

• Hong et al 45 reported that tumour size >4cm and abutment of porto-mesenteric veins were independent predictors of resection margin positivity.  Their results support the use of NCCN criteria in predicting margin status.

• A 2017 study 46 reported that patients with tumour size > 30mm (except when the tumour is located at the pancreatico-duodenal interface) or size > 20mm and located in the pancreatic neck or uncinate process were at high risk of a R1 resection.

Using a pancreas CT protocol, positive predictive values in predicting surgical resectability is above 80% 47=49 and for R0 resectability is reported to be about 73% 45, 46.

Interobserver agreement

A 2018 study showed a high degree of interobserver agreement in categorising patients by CT using the NCCN criteria 14. Another study 50, however, in which the radiologists used a modification of the borderline resectability criteria of the NCCN guidelines, found only moderate agreement for categorisation into resectable/nonresectable/borderline resectable with the greatest diversity of opinion for the borderline resectable category. Interobserver agreement increased with radiologists’ experience.

In addition, a study suggests considerable discrepancies between “real world” radiology reporting and reporting by subspecialist radiologists 51.

Distant metastases

The sensitivity of CT for detecting small liver metastases and peritoneal deposits is limited. Hence, the suggestion that in patients with tumours deemed resectable on CT, who are considered high-risk for metastatic disease (markedly elevated CA 19-9, large primary tumours, large regional lymph nodes, and patients who are very symptomatic) and patients with borderline resectability, should undergo additional imaging with other modalities (e.g MRI, PET/CT, diagnostic laparoscopy).

Innovations

Innovations in imaging PDAC that have not yet entered the mainstream have been summarised in several recent reviews 52-56.

These require further study and include: 

CT Radiomics

Radiomics 57 involves extracting quantitative data from the images and assessing them to identify diagnostic or prognostic features such as tumour grade, resectability, tumour response to neoadjuvant therapy, and survival. Combined with Artificial Intelligence and Machine Learning techniques, radiomics has great potential in characterising pancreatic tumours, assessing tumour burden, tumour aggressiveness 58, vascular involvement and resectability 59, 60 and improving survival prediction 58, 61 but requires further validating studies. 

A retrospective study combining deep-learning radiomics models of dual energy CT with key clinical features found excellent correlation with the presence of lymph node metastasis 62.

Dual energy CT

Dual energy CT (DECT) offers the generation of so called virtual monoenergetic images (VMI) which can be used to increase CT attenuation of iodinated structures, to improve CT attenuation stability or to reduce metal artefacts. They can be reconstructed at different energies in kiloelectron Volt (keV) ranging from 40 to 200 keV. VMI at low-keV improve the CT attenuation and therefore the contrast of iodinated structures, potentially increasing the contrast-to-noise ratio (CNR). They are therefore often used in vascular and contrast-enhanced parenchymal CT imaging.

This allows the calculation of iodine concentrations in tissues indicating the uptake of iodine in regions of interest. Dual-layer spectral technology improves virtual monoenergetic image (VMI) quality in patients with PDAC, resulting in better or equivalent tumour conspicuity and peripancreatic vascular assessment compared to convention pancreatic parenchymal phase CT images 63.

Some studies have shown DECT to offer greater accuracy in measuring tumour size, and in increasing specificity in determining vascular involvement and metastatic disease 64, 65. 

A recent retrospective study has suggested that dark-blood dual energy CT images may aid decisions regarding neoadjuvant therapy and surgical planning for PDAC 66

A study published in 2020 67, using dual energy CT suggested that Increased quantitative and qualitative border conspicuity (the tumour-pancreas interface) is associated with shorter survival in patients with PDAC, and that this is best assessed qualitatively using iodine material density and lower-energy simulated monoenergetic images at pancreatic protocol dual-energy CT.

Quantitative CT Perfusion imaging

A recent systematic review (68) suggests that quantitative CT perfusion may serve as a tool for tumour grading and assessment of treatment response.

A further small study published in 2022 showed that lower and more heterogeneous perfusion parameters correlated with an unfavourable response to therapy 69. Further study is required.

Diffusion-weighted MRI (DWI)

A multiparametric approach, using Dual Energy CT, radiometrics texture analysis and DWI allowed for accurate discrimination of pancreatic cancer and revealed great potential to provide independent prognostic information on all‐cause mortality 70.

• MRI is of equal accuracy to CT for assessing vascular involvement by PDAC 71 and for categorisation of local resectability/non-resectability/borderline resectability 72.

• However, for the detection of liver metastases, MRI is more sensitive than CT but of equal specificity 37, 73. A multicentre prospective study reported that preoperative MRI improves the detection of liver metastases in patients with potentially resectable pancreatic ductal carcinoma and may change management – for example, the rate of unnecessary laparotomy and pancreatectomy - for 10-14.4% of patients 74, 75. 

• MRI is not as widely available as CT, has less spatial resolution and slower acquisition time and has been shown to underestimate tumour size 12.

• Therefore most institutions and international guidelines (7) recommend CT as the preferred initial examination for staging , reserving MRI as an adjunct for problem solving (for example, for indeterminate liver lesions seen on CT) or if CT is equivocal, or for patients who have an iodine allergy. However some institutions do perform MRI routinely as their investigation of choice 76. 

• Given the superiority of MRI in detecting liver metastases, there is a reasonable argument for performing MRI in patients with potentially resectable primary tumours on CT 38, 74, 77, 28, with inclusion of Diffusion Weighted imaging.  In addition, there is some evidence that MRI may be superior to CT for imaging of small masses 76, 79(76, 79).

Diffusion-weighted MRI (DWI)

DWI has been shown in three small studies to be useful in assessing treatment response in patients undergoing neoadjuvant treatment 80-82 and, in patients with borderline resectable tumours, may be a predictor of R0 resection margins at subsequent surgery 82. Acquiring a baseline MRI prior to treatment may therefore be desirable.

A multiparametric approach, using Dual Energy CT, radiometrics texture analysis and DWI allowed for accurate discrimination of pancreatic cancer and revealed great potential to provide independent prognostic information on all‐cause mortality in one study 70.

• Patients require sedation for EUS and complications, albeit rare, do occur.

• EUS has been shown to be superior to CT and MRI in the detection of small pancreatic tumours (83), but EUS (with EUS-guided FNA where required) is more commonly used in the diagnosis of pancreatic masses than in staging.

• The primary role of EUS in pancreatic masses is to enable targeted biopsies. 

• Its use for staging is complementary to CT for problem solving with regard to T-staging and local invasion of vessels or local nodal involvement. It is not recommended as a routine staging tool, in part because of operator dependability but also because of the high incidence of variant vascular anatomy.

• EUS is unable to adequately stage PDAC for distant metastases.

• Meta-analyses 84-86 concluded that EUS is reliable in local staging of PDAC and assessment of vascular invasion, although better for venous invasion than arterial involvement. Sensitivity and specificity for vascular invasion were 0.85-0.87 and 0.91-0.92 respectively84, 85 . Nodal staging was less accurate. Two of these meta-analyses 84, 86 assessed the potential benefit of adding EUS to pancreatic cancer staging. Their results support  EUS being useful in some cases: (a) in small (<2 cm) and ill-defined and/or iso-attenuating tumours on CT; (b) in determining tumour involvement of the portal vein and portal venous /SMV confluence; (c) in detection of malignant loco-regional lymph nodes; (e) in detecting minimal amounts of ascites, (a possible accompaniment of peritoneal involvement with malignancy and thus an indication for a diagnostic staging laparoscopy).

• One of these meta-analyses 86 which evaluated the potential benefit of EUS for identifying borderline/unresectable disease after initial pancreatic protocol CT, found that EUS results might change patient management from upfront surgery to chemotherapy in 14% of cases.However, these meta-analyses and a further Cochrane Systematic review 87 suffer from a heterogeneity of included studies and a paucity of recent data. Many of the included studies were from the 1990s and early 2000s, so that comparisons with CT were with old CT technology, thus raising the question of whether similar findings would result from more up-to-date technology. Similarly, the EUS methodology was poorly defined and would have used old technology, without adjunct techniques such as contrast-enhancement.

• The Cochrane review 87concluded that, with a paucity of data, there is no evidence to support the routine use of EUS in patients with pancreatic cancer found to be resectable on CT scan.

• However, in a recent multinational retrospective series of patients with resectable PDAC 88, EUS was found to be superior in sensitivity, PPV and NPV to CT for identification of peri-pancreatic lymph node, vascular and adjacent organ involvement. However, this study was flawed by the the exclusion of patients who were found to have non-resectable or metastatic disease during operation. In addition, the CT methodology was not detailed, nor the details of EUS (linear or convex-array).

• A further recent retrospective study 89 of EUS in detecting vascular invasion resulted in excellent accuracy for venous and arterial invasion (91% and 93%, respectively). However, the EUS operators were not blinded to previous CT and MRI results.

• In summary, satisfactory high quality series comparing state-of-the-art CT and EUS technologies (including contrast-enhanced EUS) are required.

• The current recommendation of the NCCN guidelines 6 is that:

“EUS is not recommended as a routine staging tool. In select cases, EUS may be complementary to CT for staging.” Its use for staging may be for problem solving with regard to T-staging and local invasion of vessels or local nodal involvement. It is not recommended as a routine staging tool, in part because of operator dependability but also because of the high incidence of variant vascular anatomy.

• Note that biopsy is not mandated preoperatively. Although a pathologic diagnosis is not required before surgery, it is necessary before administration of neoadjuvant therapy and for patients staged with locally advanced pancreatic cancer or metastatic disease. 

• In addition, prior to a planned pancreaticoduodenectomy, most centres will mandate tissue acquisition as there is a 5-10 % rate of benign disease at surgery, in part due to the prevalence of IgG4 auto-immune pancreatitis masquerading as malignancy 90.

• The preferred method of tissue acquisition is EUS-guided fine needle biopsy because of better diagnostic yield, safety, ability to perform molecular profiling (see below) and potentially lower risk of peritoneal seeding when compared with the percutaneous (CT-guided) approach 6.

• In a small proportion of patients with advanced pancreatic cancer, targeted immunotherapy is now utilised 91; this requires adequate tissue for molecular profiling – and thus EUS-guided fine needle (EUS-FNB) rather than merely fine needle aspiration (EUS-FNA) for cytology 92.

• The reported incidence of needle track seeding following trans-gastric EUS-guided tissue acquisition of resectable tumours of the body and tail of the pancreas is 0.86% 93.

New technical developments such as elastography and the use of ultrasound contrast agents have recently been applied to EUS, improving its diagnostic capability. Contrast-enhanced EUS in recent studies has been found to not significantly increase sensitivity of EUS-guided biopsy 94, 95. However, the effect of contrast-enhanced EUS on staging PDAC does not appear to have been reported yet.

• Morphological imaging (CT, MRI, EUS) is superior to FDG PET/CT for local staging, particularly for determining vascular invasion.

• Nuclear medicine imaging for PDAC is mostly performed with 18F fluoro-2-deoxy-D-glucose -Positron Emission Tomography (FDG-PET). In modern times images are fused with CT (PET/CT) – which is 96-97% sensitive in demonstrating malignant pancreatic lesions – or, more recently with MRI (PET/MRI) 96, 97.

• The role of PET/MRI in staging pancreatic cancer awaits elucidation by further studies.

• Traditionally in PET/CT the PET images have been fused with non-contrast CT images, but there is some evidence that contrast-enhanced CT (sometimes referred to as “1-stop-shop” staging) provides more precise assessment of distant metastases and supraclavicular nodes as well as peritoneal metastases 98, 99. It is possible to employ cross-modality image fusion of PET/CT and contrast-enhanced CT 100. However, contrast-enhanced PET/CT is not widely used for practical reasons; usually the patient has undergone pancreatic protocol CT prior to referral for PET/CT. Also, the protocol for CT should include a multiphase examination32.

• FDG PET/CT must be performed in patients with a plasma glucose level of less than 11mmol/litre – an important point given that many patients with PDAC are hyperglycaemic 101./p>

• The detection of peri-pancreatic malignant nodal disease may be problematic due to obscuration by activity in the primary tumour. Three levels of nodal disease are described in PDAC. Level 1 nodes are peripancreatic; level 2 nodes are along the hepatic hilum or main arteries (SMA, coeliac axis, splenic artery); level 3 nodes are para-aortic. Prognosis worsens with increasing level of nodal involvement 101. Accuracy of FDG-PET/CT in diagnosis of nodal disease has been shown to be only poor to moderate, partly due to the presence of reactive peri-pancreatic nodes, but is best for Level 3 nodal disease.101, 102.

• Specificity of FDG PET is also reduced by false positive findings from, for example, focal pancreatitis. False negatives also occur due to hyperglycaemia and the presence of small tumours  96.

• PET or PET/CT can prevent futile resection of PDAC when distant metastases (liver, lung, peritoneum, bone) are present. Several studies have demonstrated the value of FDG PET/CT in detecting distant metastases not shown by CT 103, 104. PET/CT has been shown to change PDAC management in 11-27 % 96, 105-107 of patients thought to have resectable or borderline resectable disease prior to the PET/CT. Ghaneh et al (108) reported that PET/CT correctly changed staging in 10% of their patients with PDAC and influenced management in 45%. In particular, it stopped resection in 20% of those patients due to have surgery.

• Sensitivity (i.e. determining whether patients have liver metastases ) for hepatic metastases >1cm has been shown to be in the order of 90%  109 , but considerably lower for metastases <1cm 109, 110, although earlier studies ,for example 109, employed stand-alone PET without fusion of CT images. A study comparing PET, PET/CT and contrast-enhanced PET/CT ( a “one-stop-shop” approach to staging) demonstrated moderately good sensitivity for liver metastases, increasing from 46% for non-enhanced PET/CT to 82% for contrast-enhanced PET/CT 99. In addition, enhanced PET/CT was more accurate than unenhanced PET/CT in assessing overall resectability. However, this study did not compare enhanced PET/CT with PET/CT and enhanced CT performed as separate procedures.

• Several studies have shown a high Standard Uptake Value (SUV) of the primary tumour to correlate with tumour aggressiveness and poorer prognosis 103, 111-115; studies include a meta-analysis 111.

• A Health Technology Assessment undertaken for the National Institute for Health Research (UK) , which involved a multicentre prospective study published in 2018, found that PET/CT significantly influenced the staging and management of patients and was seen to be most cost-effective for the subgroup of patients with suspected pancreatic cancer who were thought to be resectable on CT 108. The conclusion from this study was that PET/CT adds significant benefit to patients in terms of diagnosis, staging and management of pancreatic cancer. The most cost-effective use of PET/CT was in the subgroup of patients who were suspected of having pancreatic cancer on MDCT and who were planned for surgery.

• The UK NICE (National Institute for Clinical Excellence) Guidelines of 2018 116 recommend FDG PET/CT as the next procedure for patients in whom a pancreas protocol CT that includes  CT of the chest , abdomen and pelvis shows localised disease and who are prospective candidates for surgery, radiotherapy or systemic therapy.

• However, currently accepted practice 6 appears to be to reserve PET/CT for selected patients – in particular patients at “high risk” for metastatic disease. These may include patients with borderline resectable disease, markedly elevated CA 19-9, large primary tumours, large regional lymph nodes, and patients who are very symptomatic.

PET/MRI

Some recent studies have investigated the fusion of PET and MRI instead of CT 117, 118. Most of the current literature reports studies in which retrospective fusion of MRI images with PET has been utilised, but simultaneous PET/MRI is being introduced. A study published in 2017 reported that FDG PET/MR imaging showed a diagnostic performance similar to that of PET/CT plus contrast-enhanced multidetector CT in the preoperative evaluation of the resectability and staging of pancreatic tumours 119. A small recent study showed accuracy of TNM staging of pancreatic cancer by F-FDG PET/MR was 73.3% 120.Further studies are awaited.

Novel PET tracers

New tracers for use in PET/CT or PET/MRI are under investigation, which aim to counter the lack of specificity of the [18F]-2-Fluoro-Deoxy-Glucose (FDG) tracer in common use. Cancer-associated fibroblasts show a high expression of fibroblast activation protein (FAP), while expression levels in normal human tissues are generally very low. This has led to fibroblast activation protein inhibitor (FAPI) labelled to fluorine-18 or gallium-68 being used for PET imaging. Early studies have shown promise suggesting FAPI PET superiority over FDG-PET/CT/MRI for the detection of primary tumour, lymph node metastases and distant metastases including peritoneal deposits 121-123

Also showing promise for detection of metastatic disease in a preliminary study is (4S)-4-(3-(18)F-Fluoropropyl)-L-glutamate (FSPG) positron emission tomography 124.

• Laparoscopy can detect peritoneal, capsular, or serosal metastases or small deposits on the surface of the liver that may be missed by CT (occult metastases). 

• In addition, peritoneal washings can be obtained for cytology examination. Positive cytology constitutes M1 disease.

• Laparoscopic ultrasound (LUS) may be combined with diagnostic laparoscopy to examine the liver and vascular involvement 6, 125, 126.

• In some institutions DL is used routinely to rule out metastases in patients with PDAC deemed resectable and without metastatic disease on radiological imaging, as well as patients with borderline resectable disease in whom neoadjuvant treatment is planned. 

• Whereas routine use of laparoscopy is controversial, other institutions use the technique selectively (as per NCCN guidelines 6) to rule out metastatic disease prior to surgery or neoadjuvant treatment in borderline resectable patients , as well as in selected patients with resectable disease in whom metastases are not detected on diagnostic imaging and who are deemed at high risk of disseminated disease (see below).

• A Cochrane review published in 2016 127 of the utility of DL in pancreatic and peri-ampullary cancer found the quality of evidence of included studies to be low with considerable heterogeneity among included studies. In addition, it is unclear what proportion of the included studies employed peritoneal washings and cytology and laparoscopic ultrasound as part of the laparoscopic methodology. However, the review found that adding staging laparoscopy to CT scan decreases the probability of unresectable disease at surgery from 41% to 20% ie that 21 patients in every 100 will avoid laparotomy compared to those undergoing CT alone. Using DL could halve the rate of unnecessary surgery for potentially surgically curable pancreatic cancer. This review included studies reporting results of laparoscopic staging of both pancreatic cancers and peri-ampullary cancers. In the subgroup of patients with pancreatic cancer the post-test probability of unresectable disease after negative diagnostic laparoscopy was 0.18.

• A further systematic review 128 also reported heterogeneity of included studies but reported similar findings, in that 20% of CT-resectable patients had a positive laparoscopy i.e. were found to be unresectable due to metastatic disease. There was a 5% rate of false negative laparoscopies – that is patients with a negative laparoscopy but found to be non-resectable at subsequent laparotomy.  In most of these patients this was due to vascular involvement with tumour. In this systematic review, studies that included laparoscopic ultrasound as an adjunct to laparoscopy were excluded.

• Limitations of both of these systematic reviews 127, 128 is that: both reviews reported heterogeneity of the included studies; a variable proportion of patients included in these two reviews had undergone other imaging (MR or PET or PET/CT or endoscopic ultrasound) apart from CT, prior to staging laparoscopy; many of the studies analysed would have used CT scanners and protocols which would not now be regarded as state of the art, thus raising the possibility that with modern CT techniques the number of falsely diagnosed CT-resectable tumours would be less – that is, the yield of laparoscopy subsequent to a CT scan showing resectable disease would decrease 129. Indeed a 2023 publication reported that staging laparoscopy could avoid an unnecessary laparotomy in 10% of the patients with a resectable or borderline resectable PDAC based on multiphase CT scan imaging 130. However, a retrospective series reported >1000 patients examined between 2017 and 2021 (and therefore, presumably having high quality CT scanning) with radiographically resectable PDAC. Despite (a) the more modern CT techniques in these patients and (b) many of the patients undergoing MRI and/or PET (CT or MRI), in those who had not received prior chemotherapy, there were 23% had who positive laparoscopy. This rose to 42% in younger patients with large primary tumours and elevated CA 19-9, but was only 1.6% in those with no high-risk factors (see below)131. However, it should be noted that the majority of patients in this series had peritoneal washings taken at the time of laparoscopy for cytology and examination for CA 19-9 and CEA.

• Peritoneal washings with cytology examination yields more positive evidence of disseminated disease than DL alone 131, 132. A recent relatively small retrospective study 133 reporting patients examined between 2014 and 2019, found that laparoscopy, including peritoneal washings  resulted in upstaging of 15% of patients with resectable disease (on radiological imaging) and 9% with borderline resectable. 7% of their patients had positive peritoneal cytology without macroscopic evidence of metastatic disease. Only 3% of patients with a negative diagnostic laparoscopy had an abandoned resection due to liver metastases. There was a 12.5% absolute risk reduction of a subsequent non-therapeutic laparotomy.

• It is to be noted that laparoscopy (without laparoscopic ultrasound) cannot detect vascular invasion, lymph node involvement and deep hepatic metastases, and hence must be used in addition to other imaging modalities.

• Several studies have included the use of laparoscopic ultrasound (LUS) as an adjunct to DL. A 2016 meta-analysis 126 reported that DL with laparoscopic ultrasound prevented noncurative laparotomies in 33%. LUS facilitates the identification of vascular involvement and liver metastases not visible on the liver surface. In a recent exploratory study, Oba et al 125 used contrast-enhanced laparoscopic ultrasound (as well as fluorescence imaging).

Selective use of staging laparoscopy in “High-risk” patients

Identifying which patients are at “high risk” of having metastatic disease despite having apparent upfront resectable or borderline resectable primary tumours (and thus benefiting from staging laparoscopy, when this is utilised selectively) has varied with various publications.

Factors that constitute high risk may include:

• location of PDAC in the body or tail of the pancreas 131, 134-136.

• elevated CA 19-9 131, 134, 135, 137-139 .

• large primary tumour  131, 135-137, 139,

• large regional lymph nodes 

• younger patients (<60 years old) 131.

• Indeterminate lesions on previous imaging 131, 139

• Highly symptomatic patients (including weight loss and jaundice). However, the evidence for correlation of symptoms with unresectability is conflicting.

However, one recent publication found no correlation between a positive diagnostic laparoscopy and Ca 19-9 levels or tumour location or size.133

NCCN guidelines 6 suggest the use of staging laparoscopy in patients with borderline resectable PDAC prior to neo-adjuvant therapy and in those with resectable disease at increased risk of disseminated disease (i.e. high-risk patients as described above).

The sensitivity of CT for detecting small liver metastases and peritoneal deposits is limited. Hence, the suggestion that in patients with tumours deemed resectable on CT, who are considered high-risk for metastatic disease, and patients with borderline resectability, should undergo imaging with other modalities (eg MRI, PET/CT, diagnostic laparoscopy).

The following patients may be selected to undergo further imaging after CT with MRI +/-EUS +/-PET/CT due to a higher risk of metastatic disease. Patients with:

• elevated CA 19-9

• large primary tumours 

• large regional lymph nodes at CT

• marked symptoms.

The following criteria have been used for selecting patients to undergo diagnostic /staging laparoscopy:

• location of PDAC in the body or tail of the pancreas 131, 134-136,

• elevated CA 19-9 131, 134, 135, 137-139, 

• large primary tumour 131, 135-137, 139,

• large regional lymph nodes 

• younger patients (<60 years old) 131.

• indeterminate lesions on previous imaging 131, 139

• highly symptomatic patients (including weight loss and jaundice). However, the evidence for correlation of symptoms with unresectability is conflicting.

Li et al 140 have constructed a risk model and nomogram to predict distant metastasis in patients with PDAC . Risk factors include age (< 60), primary site (body and tail of pancreas), histological grade and lymph node status.

• In addition to patients with borderline resectable PDAC, there is some evidence that selected patients with tumours deemed to be resectable on imaging may benefit from NAT 24, although further evidence is required 25.

• A recent meta-analysis 26 reported that NAT in patients with resectable PDAC  increases the R0 resection rate and decreases positive node resection rate, but without significant effect on overall survival time.   It remains to be determined which patients with resectable tumours based on imaging +/- biological and or other markers, are likely to benefit from NAT. 27, 28.

• However there is evidence that features of resectable PDAC that may indicate benefit of NAT include location in pancreatic head, CA 19-9 >100units/ml, tumour size >20mm -these are independent predictors of poor prognosis and together may provide a prognostic scoring system and thus a selection process for NAT 29 - and extra-pancreatic peri-neural invasion 30.

• Extra-pancreatic peri-neural invasion (EPNI) is thought to be a significant factor in positive surgical resection margins and tumour recurrence. A recent limited study 30 reported high sensitivity of CT for detection of EPNI, but only moderate specificity (since fibrosis and inflammation are indistinguishable). Nevertheless, the authors suggest that detection of EPNI in patients with otherwise resectable tumours may be an indication for neoadjuvant therapy.

The NCCN guidelines and the European Society for Medical Oncology (ESMO) recommend that only patients with a high probability of R0 resection should have upfront surgery and that where there is doubt or where a R1 resection is of high probability, patients should receive NAT. 

A review, including work-in-progress has been provided recently by Zhang et al. 141

Recent publications by an American Journal of Roentgenology Expert Panel 38 and by representatives of the Society of Abdominal Radiology 22 state that current morphological imaging techniques cannot reliably make the distinction between treatment response and stable or worsening disease.

Multidetector computed tomography has limited roles in the assessment of tumour resectability after NAT for PDAC because of the similar appearance of tissue fibrosis/ inflammation induced by NAT and tumour infiltration 142. This can lead to an overestimation of non-resectability on the basis of apparent vascular involvement with tumour. Standard criteria for predicting vascular invasion, based on the amount of tumor-vessel contact, are not valid after neoadjuvant therapy 23.

Various authors have attempted to improve diagnostic CT performance in the post-neoadjuvant setting by modifying the conventional imaging criteria for upfront surgery in various ways. A meta-analysis of these criteria 143  indicated that sensitivity for assessing resectability post- neoadjuvant therapy was increased but specificity decreased. Further modifications need to be investigated.

Signs of regression on CT include: 22, 38.

• Decrease in size

  A recent study has reported that imaging-based tumour downsizing and not downstaging can guide the selection of patients with a realistic chance of R0-resection in locally advanced PDAC after multi-agent induction chemotherapy 144.

  CT often underestimates tumour size following NAT, compared to the resected specimen. Dual-energy CT has been recommended to increase reproducibility and accuracy of measurement.38

• Decrease in tumour-vessel contiguity

  Complete regression is rare, but in terms of vascular involvement the objective sign of treatment response is decrease in tumour-vessel contiguity 38. CT assessment of venous involvement is more reliable than arterial involvement. The latter remains difficult after NAT 38.

• Other signs of favourable tumour response include:  replacement of solid tumour contact of vessel by a perivascular halo; increased tumour attenuation 22.

With regard to predicting nodal positivity at surgery following NAT, dual energy (CT) based minimum normalized iodine value of all nodes in the surgical field on preoperative studies has modest utility in differentiating N0 from N1/2, and generally outperformed conventional features for identifying nodal metastases 145.

CT perfusion (which uses blood perfusion as a biomarker) may predict the response to neoadjuvant therapy 146.

CT radiomics.CT texture analysis has been explored with promise in predicting tumour response and disease free survival after NAT 147. Further large-scale studies are required to determine its real potential 141.

Diffusion-weighted MRI. Post-treatment whole-tumour ADC value may be a predictor of R0 resectability in patients with borderline resectable PDAC after chemotherapy 22, 82.

PET/CT148 can monitor tumour response to therapy by measuring reduction in tumour SUV (max).A 25-50% decrease in maximum SUV in response to treatment , compared to baseline metabolic uptake , suggests a favourable response, with a greater decrease in SUV max correlating with longer progression-free survival and overall survival 22

In light of increased perioperative complications the role of routine pre-operative biliary drainage (PBD) is being reassessed in recent years.

• Current indications for PBD include cholangitis, delayed surgery due to logistics, and relief of jaundice in patients planned to receive neoadjuvant therapy. 149

• Biliary drainage may be achieved by ERCP-placement of a stent or by Endoscopic Ultrasound-guided biliary drainage (EUS-BD). If these options are unavailable or unsuccessful, percutaneous transhepatic biliary drainage (PTBD) is a fall-back option.

• Meta-analyses have found that EUS-BD and ERCP have similarly high rates of technical success and clinical success when used as the primary option for biliary obstruction 150.

• EUS-BD was non-inferior to ERCP as a primary option for malignant biliary obstruction the study also found lower rates of overall adverse events for EUS-BD relative to ERCP (6.3% vs. 19.7%, respectively), including post-procedure pancreatitis (0 vs. 14.8%) and reintervention (15.6% vs. 42.6%), as well as a higher rate of stent patency (85.1% vs. 48.9%) with EUS-BD 151