Definition andManagement of Borderl ine

Resectable Pancreatic Cancer

Jason W. Denbo, MD, Jason B. Fleming, MD*

KEYWORDS

� Borderline resectable pancreatic cancer � Definitions � Management� Pancreatic ductal adenocarcinoma

KEY POINTS

� Pretreatment evaluation of patients can aid in accurate communication and treatmentplanning for patients with borderline resectable pancreatic cancer; many patients havetechnically resectable tumors but are not adequate candidates for surgery.

� Imaging using contrast-enhanced computed tomography is necessary to stage the pa-tient and evaluate extrapancreatic extent of the primary tumor.

� Even though small variations in the definition of borderline resectable tumors exist, thesurgeon remains responsible for the complete and safe resection of the primary tumor.

� Safe surgical resection requires detailed preoperative knowledge of pertinent vascularanatomy, careful retroperitoneal dissection, and vascular isolation before vascular resec-tion if necessary.

INTRODUCTION

Recent genetic studies have shown that initial clinical evaluation of pancreatic ductaladenocarcinoma (PDAC) occurs very late in the natural history of the disease,1

because only 10% to 20% of patients present with surgically resectable disease.2–5

Nevertheless, the possibility of a potentially curative surgical resection provides apowerful impetus toward surgical resection of all possible patients. The impetusmust be governed by the discipline and insight necessary to perform a safe and onco-logically effective pancreatic resection, which is just one part of a multidisciplinaryeffort engineered to achieve long-term patient survival.Over the last several decades, particularly at high-volume centers, the postopera-

tive mortality following pancreaticoduodenectomy (PD) has decreased from 30% to1%.6 However, despite these advances in surgical safety, patients with PDAC have

Disclosure: The authors have nothing to disclose.Department of Surgical Oncology, The University of Texas MD Anderson Cancer Center, 1400Pressler Street, Unit 1484, Houston, TX 77030, USA* Corresponding author.E-mail address: jbflemin@mdanderson.org

Surg Clin N Am 96 (2016) 1337–1350http://dx.doi.org/10.1016/j.suc.2016.07.008 surgical.theclinics.com0039-6109/16/ª 2016 Elsevier Inc. All rights reserved.

Denbo & Fleming1338

not benefited from improved long-term survival when surgery is used as initial ther-apy.3,6–8 Furthermore, despite multiple trials showing a survival benefit with adjuvanttherapy,7,9–11 up to 47% of patients treated with up-front surgical resection fail toreceive any adjuvant therapy,12 usually because of delayed postoperative recoveryor early disease recurrence.13,14 These pitfalls of surgery as primary therapy are ampli-fied in borderline resectable (BR) patients for whom careful staging, meticulous patientevaluation, and preoperative therapy are necessary to identify the subset of patientsmost likely to benefit from the aggressive surgical procedures necessary for completesurgical resection.

IDENTIFICATION OF BORDERLINE RESECTABLE PATIENTS

Any patients with PDAC evaluated by a pancreatic surgeon may require a high-risksurgical procedure that offers the only chance of cure. A new diagnosis of PDAC isoften made in patients with multiple underlying medical conditions of variable signifi-cance with respect to the risks of pancreatectomy. This high-stakes clinical scenariomandates that the surgeon uses an organized approach to ensure a thorough and effi-cient initial evaluation. The anatomic relationship of the tumor and critical vessels asdetermined by a pancreas protocol computed tomography (CT) scan is of crucial sig-nificance, but other nonanatomic factors must also be evaluated, such as suspicionfor extrapancreatic disease, comorbidities, and functional status. Using this approach,the whole patient and not just tumors are classified as potentially resectable or border-line candidates for surgical resection of the primary tumor.Our center has developed a systematic approach in which all patients with localized

PDAC receive a physical examination, a review of laboratory studies, and radiographicimaging as part of a comprehensive evaluation in a surgical clinic. These data are thencollated using a system denoted by the acronym ABC in which A refers to tumoranatomic considerations for surgery, B to cancer biology or stage; and C to patientcondition or performance status and fitness for surgery (Fig. 1). In the course of treat-ment planning and communication across our multidisciplinary care team, patients areclassified as clinically resectable (CR) or BR using the common nomenclature BR-A,BR-B, or BR-C.15,16 BR-A patients have no major comorbidities, no clinical findingsthat are suspicious for extrapancreatic disease, and meet anatomic imaging criteriafor a BR tumor, as outlined later. BR-B patients have no major comorbidities oranatomic imaging criteria for a borderline resectable tumor, and have clinical findingssuspicious for extrapancreatic disease: (1) indeterminate liver lesions, (2) serum car-bohydrate antigen 19-9 (CA19-9) level greater than or equal to 1000 U/mL in thesetting of a normal bilirubin level, or (3) biopsy-proven involvement of regional lymphnodes. BR-C patients are advanced in age (�80 years old) or possess severe comor-bidities requiring extensive evaluation or optimization, or depressed performance sta-tus (Eastern Cooperative Oncology Group [ECOG] �2). They may or may not haveclinical findings suspicious for extrapancreatic disease.In practice, the fitness of each patient for pancreatic surgery is evaluated first

(see Fig. 1). Patients who are too frail for surgery secondary to uncorrectable comor-bidities do not need to undergo extensive evaluation for surgical resection or consid-eration for preoperative therapy because surgical resection will not be the end result.These patients can therefore be efficiently triaged for palliative therapy or supportivecare. If the patient is not currently fit for surgery, but has a potentially reversible con-dition, then medical optimization or prehabilitation during preoperative therapy is thegoal. These patients are referred to as BR-C and are generally older (median age,75 years) with a higher ECOG status (44% ECOG 2) usually secondary to

Fig. 1. Initial evaluation and categorization of patients with pancreatic adenocarcinoma. CA19-9, carbohydrate antigen 19-9; Chemo, chemotherapy;XRT, X-ray therapy.

Borderlin

eResecta

ble

Pancre

atic

Cancer

1339

Denbo & Fleming1340

cardiopulmonary disease (63%).15 If the patient is fit for surgery, biological staging isnext. Evidence of distant metastases on radiographic imaging is a contraindication toresection, but in many cases there are suspicious radiographic findings, but not diag-nostic for distant metastatic disease. These patients are termed BR-B and receivechemotherapy followed by restaging. Similarly, patients with a high CA19-9 level(1,000 U/mL) even with negative imaging are considered BR-B and receive chemo-therapy followed by restaging. In addition, local anatomic factors related to the pri-mary tumor are considered in patients who are without metastases and are fit forsurgery, which necessitates careful review of radiographic image using standardanatomic criteria designed to categorize tumors as resectable, borderline, or locallyadvanced. Patients who are fit for surgery with no evidence or suspicion of metastasesand borderline tumors are considered BR-A, and usually receive chemotherapy plusor minus chemoradiation and restaging before proceeding with surgical resection.Clinical application of this approach to the initial evaluation identifies that at least

50% of patients have borderline clinical features. When treated preoperatively, only37% of BR-C patients can be expected to undergo resection, whereas the othersfail to reach surgical resection because of poor performance status (31%) or intervalidentification of metastatic disease (26%). BR-C patients who undergo resectionexperienced a median survival of 38.6 months versus 13.3 months (P 5 .02) for thosenot receiving surgery. Roughly 46% of BR-B patients receive surgical resection afterpreoperative therapy and an equal portion (46%) have distant metastases precludingsurgery. The loss of performance status is uncommon (4.9%) in BR-B patients. Resec-tion conferred a 33-month median survival versus 11.8 months in those patients un-able to proceed to resection (P<.001). Of note, local progression was rarelyobserved during preoperative therapy in either BR-B or BR-C patients (2.6%)15,16

(Fig. 2). Management of patients with BR type A is considered later.

IMAGING

All patients with apparent localized disease are evaluated with a contrast-enhancedCT scan, which provides essential information about the presence of regional ordistant metastases and the site and local extent of the primary tumor. This evaluationallows the surgeon to determine whether the patient has a resectable tumor and thelikelihood of achieving a margin-negative resection. Multidetector row CT is the

Fig. 2. Typical distribution of outcomes for patients with type BR-B and type BR-C treatedwith preoperative therapy. Identification of distant metastases or poor performance status(PS) is most common in patients with BR-B or BR-C, respectively.

Borderline Resectable Pancreatic Cancer 1341

most widely used staging modality for pancreas cancer and a workhorse for new pa-tient evaluation. When performed and interpreted correctly, it provides valuable stag-ing for both distant and regional metastases as well as local extrapancreatic extensionof the primary tumor to adjacent critical vascular structures.17 The National Compre-hensive Cancer Network (NCCN) recommends that all patients with suspicion forPDAC have a dedicated pancreas protocol CT scan as part of the initial evaluation(Version 2.2015). At MD Anderson Cancer Center (MDACC), a pancreas protocol CTscan uses water as a negative oral contrast agent and starts with precontrast imagingfrom the dome of the liver extending caudally to include the entire liver reconstructedto 2.5-mm slice thickness. Next, 125 mL of iodinated contrast is administered intrave-nously at a rate of 3 to 5 mL/s. The pancreas phase/arterial phase uses bolus trackingand images are obtained 10 seconds after a Hounsfield unit value of 100 is reached inthe aorta at the level at the celiac axis from the dome of the liver to the iliac crests. Im-ages for the portal venous phase are obtained at a 20-second delay from the pancreasphase. Hepatic metastases are usually best visualized on the portal venous phase.Delayed images are obtained 15 seconds after the portal venous phase. The imagesare reconstructed to 2.5-mm slice thickness for imaging review and at 0.625-mm or1.25-mm slice thickness to create coronal and sagittal reformatted images.18

The primary pancreatic tumor is best seen on the pancreas phase of the CT scanand is usually a hypodense mass, because the surrounding normal pancreatic paren-chyma enhances. The pancreas phase/arterial phase illuminates the branches of theceliac axis and superior mesenteric artery (SMA), enabling clinicians to identify impor-tant arterial anatomy variants and discern whether the tumor has any arterial involve-ment. As many as 40% to 45% of patients have variants of normal hepatic arterialanatomy, which are of vital importance to appreciate on preoperative imagingbecause these variants can affect operative planning.19 A replaced or accessory righthepatic artery is present in up to 15% of patients, and most commonly arises from theSMA and courses posterior to the pancreas and posterolateral to the bile duct. Anadditional 2.5% of patients have a replaced common hepatic artery (CHA) that arisesfrom the SMA and follows a similar path to a replaced or accessory right hepaticartery. The superior mesenteric vein (SMV) and portal vein (PV) are best evaluatedon the portal venous phase. The initial staging CT scan has 94% sensitivity and84% specificity for determining vascular involvement, and the surgeon should care-fully note the tumor-vein interface, vein contour, and/or deformity; there are multipleclassification schemes that predict venous involvement based on imaging character-istics and these should be used for operative planning.20–23 In addition, the surgeonshould identify the location and relationship of the gastroepiploic vein, colic veins, infe-rior mesenteric vein (IMV), and jejunal/ileal branches of the SMV because these havevariable courses and the drainage pattern directly affects the surgical options forreconstruction of the SMV-PV confluence, which can be expected in more than40% of cases. Terminology that describes vascular involvement has become stan-dardized and is reviewed in detail later. Vascular involvement of less than or equalto 180� of the circumference of the vessel is termed abutment. Vascular involvementgreater than 180� of the circumference of the vessel is termed encasement. Properlystaging patients and determining potential vascular involvement is a cornerstone oftreatment planning and its importance cannot be overstated.24

DEFINITIONS OF RESECTABILITY

As patient assessment, imaging, and multidisciplinary treatment techniques for pa-tients with localized PDAC were refined in the late 1990s, it became evident that

Denbo & Fleming1342

this patient group included a spectrum of primary tumor types from removable to unre-sectable. To allow common classification, the multidisciplinary team at MDACC devel-oped imaging criteria that are still in use that define CR tumors by (1) absence ofextrapancreatic disease; (2) clear tissue plane between the tumor and the celiacaxis, hepatic artery, and SMA; (3) a patent SMV-PV confluence, abutment, or encase-ment is allowed as long as the vessel is patent.4,16,25 Locally advanced (LA) tumors aredefined by (1) encasement of the celiac axis; (2) encasement of the hepatic artery withno options for vascular reconstruction; (3) encasement of the SMA greater than 180�;(4) occlusion of the SMV-PV confluence with no options for vascular reconstruc-tion.4,16,25 Patients who were classified as CR based on these imaging criteria werelikely candidates for an R0 resection, whereas patients with LA tumors were unlikelyto respond to chemotherapy and/or chemoradiation to a degree that would allow sur-gical resection; however, improved response to systemic therapy is occurring morefrequently and has allowed patients with advanced tumors to undergo resection.At present, NCCN defines resectable PDAC as a tumor with no contact of celiac

axis, SMA, or CHA, and no contact with the SMV-PV or less than or equal to 180� con-tact without vein contour irregularity (Version 2.2015). LA PDAC of the head/uncinateprocess is a tumor that contacts the SMA greater than 180�, the celiac axis greaterthan 180�, the first jejunal SMA branch, the most proximal draining jejunal vein, orhas unreconstructable SMV-PV involvement or occlusion. Tumors of the body/tailare LA when there is contact of greater than 180� with the SMA or celiac axis,contact with the aorta, or unreconstructable SMV-PV involvement or occlusion(Version 2.2015).In 2001, Mehta and colleagues26 described a group of marginally resectable pa-

tients who were treated with chemoradiation preoperatively in order to downstagethe tumor and increase the likelihood of a margin-negative resection. Marginallyresectable was defined as a tumor in which the perivascular fat plane was absentover 180� of the SMA, SMV, or PV and persisted for a length greater than 1 cm.26

The NCCN adopted the term BR in 2006 to describe patients with localized tumorswho blurred the distinction between CR and LA tumors. These patients were thoughtto be at higher risk of a margin-positive resection if up-front surgery was used; thus theNCCN suggested the use of preoperative therapy.Over the last several years, several groups developed specific radiographic features

to define BR PDAC. At MDACC, the imaging criteria used to define a BR tumor are (1)abutment of the celiac axis; (2) abutment of the hepatic artery or short-segmentencasement; (3) abutment of the SMA less than or equal to 180�; (4) short-segmentocclusion of the SMV-PV confluence amenable to resection and reconstruction.4,16,25

The American Hepato-Pancreato-Biliary Association (AHPBA), the Society of SurgicalOncology (SSO), and the Society for Surgery of the Alimentary Tract (SSAT) define BRPDAC as tumors with no abutment or encasement of the celiac axis, short-segmentabutment or encasement of the CHA amenable to reconstruction, abutment lessthan 180� of the SMA, or abutment with or without impingement or narrowing of theSMV-PV or encasement with or without occlusion with suitable vein proximal anddistal to allow resection and reconstruction.27,28 The difference between the MDACCand AHPBA/SSO/SSAT definitions depends on the extent of SMV-PV involvementthat differentiates BR from resectable tumors. The NCCN definition for BR PDAChas changed multiple times over the years, but currently includes tumors of thehead/uncinate process that contact the CHA without extension to the celiac axis orthe hepatic artery bifurcation, contact less than or equal to 180� of the SMA, contactgreater than 180� of the SMV or PV, contact less than or equal to 180� with a contourirregularity or thrombosis of the SMV-PV with suitable vessel proximal and distal that

Borderline Resectable Pancreatic Cancer 1343

allows venous resection and reconstruction, or contact with inferior vena cava.Tumors of the body/tail are classified as BR when there is contact of less than or equalto 180� with the celiac axis or contact of greater than 180� with the celiac axiswithout involvement of the aorta and an intact and uninvolved gastroduodenal artery(Version 2.2015) (Table 1).A current multi-institutional treatment trial investigating preoperative FOLFIRINOX

(folinic acid, fluorouracil, irinotecan, oxaliplatin) and chemoradiation defines BRPDAC as radiographically localized tumors with 1 or more of the following: (1) an inter-face between the tumor and SMV-PV of greater than or equal to 180�, (2) short-segment occlusion of the SMV-PV with normal vein above and below that is amenableto resection and reconstruction, (3) short-segment interface between the tumor andhepatic artery with normal artery proximal and distal that is amenable to resectionand reconstruction, (4) interface between the celiac axis or the SMA less than 180�.29

Although no consensus definition for BR has been reached, common themes can beappreciated. All BR criteria include statements regarding the ability or inability of thesurgeon to reconstruct the SMV-PV or the hepatic artery involved with tumor. Thesestatements imply that anatomic resectability depends heavily on the judgment andexperience of the surgeon. The importance of this expertise cannot be overempha-sized. In contrast, another common theme of borderline criteria is the exclusion of tu-mors involving greater than 180� of the SMA: a practice largely derived from theconcept that tumor involvement of the nerves and periadventitial tissue reflects anaggressive tumor biology that likely cannot be overcome through surgical techniquealone.

MANAGEMENT

Despite differences in definitions of BR PDAC, there is agreement that these patientsare at a higher risk for margin-positive resection, and that preoperative therapy is pru-dent. Since initially described by Rich and Evans30 in 1995, the potential benefits ofpreoperative therapy have been itemized and include (1) early treatment of

Table 1Definitions of BR PDAC

MDACC AHPBA/SSO/SSAT NCCN

CA Abutment No abutment orencasement

Contact �180�

CHA Abutment of short-segment encasement

Short-segment abutmentor encasementamenable toreconstruction

Contact withoutextension to celiac axisor hepatic arterybifurcation amenableto resection andreconstruction

SMA Abutment <180� Abutment <180� Contact �180�

SMV/PV Short-segment occlusionamenable to resectionand reconstruction

Abutment >180� orocclusion amenable toresection andreconstruction

Contact of >180�, contactof �180� withirregularity of vein orthrombosis amenableto resection andreconstruction

Abbreviations: CA, celiac axis; CH, common hepatic artery; SMA, superior mesenteric artery;SMV/PV, superior mesenteric vein/portal vein.

Denbo & Fleming1344

micrometastatic disease; (2) higher proportion of patients receiving multimodal ther-apy; (3) select patients with localized disease and more favorable tumor biology,who are most likely to benefit from surgical resection; (4) increased likelihood of anR0 resection; and (5) smaller radiation fields with well-oxygenated tissue. At MDACC,all patients with BR PDAC receive chemotherapy, chemoradiation, or both before sur-gical resection. Chemotherapy regimens have continued to evolve over the years;currently, most patients receive either gemcitabine with nab-paclitaxel or FOLFIRI-NOX. External beam radiation therapy is used and consists of 50.5 Gy delivered in28 fractions or 30 Gy in 10 fractions with a concomitant radiosensitizing dose of 5-fluo-rouracil, gemcitabine, or capecitabine. The most common treatment sequence for BRPDAC is 2 to 4 months of chemotherapy, followed by chemoradiation, and a 6-weektreatment break before surgical resection. Patients are typically restaged every2 months. Patients only undergo surgical resection if the operating surgeon and multi-disciplinary treatment group reach consensus that pancreatectomy will safely achievean R0 resection and provide a reasonable chance for cure.Restaging should include a pancreas protocol CT scan and measurement of

CA19-9. Katz and colleagues31 evaluated the radiographic response, using responseevaluation criteria in solid tumors (RECIST) criteria, of 129 patients with BR PDAC aftercompletion of preoperative therapy (Fig. 3). The preoperative therapy consisted ofgemcitabine-based chemotherapy followed by chemoradiation (30 Gy or 50.4 Gy)or chemoradiation alone. One-hundred and twenty-two patients completed therapyand were restaged, 84 (69%) had stable disease, 15 (12%) had a partial response,23 (19%) had progressive disease (development of metastases, n5 21; primary tumorgrowth, n5 2), and no patient had a complete response. All patients were classified bythe MDACC and AHPBA/SSO/SSAT definitions, and only 1 patient was downstaged,whereas approximately 80% remained at the same stage and 20% were upstaged.31

Fig. 3. Outcomes after preoperative therapy for 129 patients with borderline tumor criteriaclassified by AHPBA/SSO/SSAT and MDACC criteria. Regardless of criteria, local downstagingor progression is uncommon. After initial staging, 5 patients did not return for restaging forunclear reasons, possibly a decline in PS. BLR, borderline resectable; Cx, chemotherapy;CXRT, chemoradiation; LA, locally advanced; LP, local progression; PR, potentially resectable;PS, performance status.

Borderline Resectable Pancreatic Cancer 1345

The main purpose of a restaging CT scan is to rule out disease progression, notnecessarily to look for downstaging. Donahue and colleagues32 reported a series ofpatients with BR and LA pancreatobiliary malignancies who were treated with preop-erative chemotherapy and were restaged with CT/MRI imaging, which only had a 71%sensitivity and 58% specificity for vascular involvement after completion of preopera-tive therapy.32 Ferrone and colleagues33 reported a series of patients with BR and LAPDAC treated with preoperative FOLFIRINOX with or without radiation therapy, and30%were deemed resectable on posttreatment imaging. Most patients were still clas-sified as LA (48%) and BR (22%), because there were no clear fat planes around thecritical vascular structures. Nonetheless, an R0 resection was achieved in 92% of thepatients.33 Current cross-sectional imaging does not differentiate viable tumor fromfibrosis.Another important data point to consider during a restaging visit is the CA19-9 level.

Tzeng and colleagues34 compared pretreatment and posttreatment CA19-9 levels inpatients with BR PDAC. All patients had a pretreatment CA19-9 level greater than orequal to 40 U/mL and a total bilirubin level less than or equal to 2 mg/dL. A decline inCA19-9 level was seen in 116 (82%) patients and 47 (33%) had normalization ofCA19-9. Posttreatment CA19-9 level was a predictor of failure to undergo pancreatec-tomy. Normalization of CA19-9 level was associated with improvedmedian overall sur-vival in resected (38 vs 26 months, P<.02) and unresected patients (15 vs 11 months,P 5 .02).34 After completion of preoperative chemotherapy/chemoradiation, patientswithout evidence of radiographic disease progression and a decrease in CA19-9 levelshould undergo attempted surgical resection, if medically fit for an operation.The application of these management approaches is described in a recent report in

which 160 patients with BR PDAC (BR-A 84, BR-B 44, BR-C 32) were followed pro-spectively. One-hundred and twenty-five (78%) patients completed induction therapyand a restaging evaluation. Forty-three patients were determined to not be surgicalcandidates: poor performance status (n 5 10), distant disease progression (n 5 16),and unresectable local-regional disease (n 5 17). Seventy-nine patients were takento the operating room, 9 were found to have radiographically occult distant metasta-ses, 4 had LA disease, and the other 66 underwent a grossly complete resection of theprimary tumor; 53% of the patients underwent restaging. Most underwent a PD (86%),27% required an SMV-PV resection, and an additional 3% required a hepatic arteryresection. An R0 resection was achieved in 94% of the patients, 4 patients had micro-scopically positive margins (2 SMA, 1 pancreatic duct, 1 bile duct). Twenty-six (39%)patients had nodal metastases. A partial or complete pathologic response (<50%remaining viable tumor cells) was seen in 56% of patients, and 4 patients (6%) hada complete pathologic response. Considering the cohort of 160 patients, 41% of pa-tients underwent resection; the resection rates for BR-A, BR-B, and BR-C were 38%,50%, and 38%, respectively. Themedian overall survival for the cohort was 18months,with a 5-year survival of 18%. For the 66 patients who completed all therapy, themedian survival was 40 months with a 5-year survival of 36%.16 Together, these pro-spective data provide support for planned and ongoing single-institution and multi-institutional prospective studies evaluating this multidisciplinary approach for patientswith BR PDAC.

SURGICAL TECHNIQUE

Once the decision to perform surgical resection is reached, the surgeon must adhereto certain principles necessary for safe and oncologically effective pancreatectomy,so that the patient has a chance for cure. There are several variables associated

Denbo & Fleming1346

with improved patient survival after up-front surgical resection of localized PDAC.These variables include smaller tumor diameter, histologic grade (well differentiatedor moderate differentiation), absence of lymph node metastases, negative surgicalmargins, and receipt of adjuvant therapy.3,6,7 Although many of these variables areoutside a surgeon’s control, pathologically negative margins and a patient’s abilityto receive therapy after surgery are heavily influenced by surgical technique and post-operative morbidities. However, in large series, up to 42% of patients underwent amargin-positive PD.2,6 A recent rapid autopsy series of patients with localized PDACshowed that 80% of patients treated with up-front surgical resection died with localrecurrence, and 15% died of local recurrence alone.35 Of note, 5 of 6 patients who un-derwent a positive-margin PD recurred locally.35 A positive margin pancreatectomyshould be viewed as a failure of patient selection, treatment strategy, and/or surgicaltechnique. Routine use of planned segmental resection of the SMV-PV during PD wasinitially described in 1996 as a safe method of achieving negative margins when theprimary tumor and SMV-PV interface were obliterated. Importantly, no difference in tu-mor biology or cancer-related outcome were observed suggesting that loss of thedissection plane between tumor and SMV-PV is an anatomic consequence morethan an indicator of aggressive tumor biology.36

Performing a PD for a BR-A tumor requires a high level of technical expertise. Theprinciples of oncologic and vascular surgery must simultaneously be applied in orderto achieve the following surgical objectives:

1. Maintenance of hepatic artery blood flow2. Exposure and meticulous dissection of the SMA3. Preservation of portal venous flow from the stomach, spleen, and intestines, while

minimizing the risk of sinistral portal hypertension37

Tumor encasement of the SMV-PV confluence is challenging, because it limits theexposure of the proximal SMA. During a procedure without vascular involvement, afterthe SMV-PV is freed from the pancreas and uncinate process, the venous confluence isretracted to thepatient’s left, so theproximal SMAcanbeexposed; thismaneuver is notpossible when there is encasement of the SMV-PV. When the SMV-PV is encased, di-vision of the splenic vein allows retraction of the SMV-PV to the patient’s right. Next, theperiadvential plane on the SMA is accessed sharply caudal to the tumor and followedcephalad to the proximal portion of the SMA. After division of the remainder of the retro-peritoneal tissues, the tumor and specimen are attached by only the SMV-PV conflu-ence. After obtaining proximal and distal vascular control and heparinizing thepatient, the vein is resected at a level to ensure negative margins but also being cogni-zant to preserve as much vein as possible. Portal venous flow can be restored with aprimary venous anastomosis, vein patch, or internal jugular interposition, dependingon thedefect after venous resection.Resectionof infiltrating tumors commonly requiresligation of the 3main veins that drain the stomach: the coronary, right gastric, and rightgastroepiploic veins. Venous outflow from the stomach can be comprised, especiallyafter splenic vein ligation. If the IMVdrainsdirectly into theSMV, thenweperformasple-norenal shunt to prevent sinistral hypertension. If the IMV drains into the splenic vein,then we do not perform a shunt, and rely on retrograde flow through the IMV.38 Arterialresection canbenecessary for tumors in the headof thepancreas that extend cephaladalong the gastroduodenal artery to involve the common or proper hepatic arteries. Inaddition, an aberrant right hepatic artery or CHA arising from the SMA can easily beinvolvedbya tumor in thehead. The sameprincipals discussedearlier for venous resec-tion apply to arterial resection. After arterial resection, hepatic artery flow is restoredwith a saphenous vein interposition graft or polytetrafluoroethylene graft.37,39

Fig. 4. The microscopic extent of pancreatic adenocarcinoma in the pancreatic head withrespect to the SMV and SMA. The effect of radiation on tumor periphery is noted. Thedark line represents the line of retroperitoneal surgical resection necessary to achieve anegative margin. (Adapted from a drawing courtesy of RA Wolff.)

Borderline Resectable Pancreatic Cancer 1347

The SMAmargin is themargin that is most frequently positive after a PD.40,41 As pre-viously mentioned, the Johns Hopkins rapid autopsy series revealed that 80% of pa-tients had evidence of local recurrence, and 15% had isolated local recurrence.35

This finding highlights the importance of adherence to the surgical principals outlinedearlier. Katz and colleagues42 evaluated recurrence patterns in 194 patients with CRand BRPDACwith respect to the use of preoperative chemoradiation and SMAmarginstatus. Fifteen (8%) patients had an R1 resection; the SMA margin was the mostcommonly involved margin. An additional 40 (22%) patients had a negative margin,but the margin distance was less than or equal to 1 mm. Patients who received preop-erative chemoradiation had longer SMA margin distances (P 5 .01). Patients who un-derwent up-front surgical resection and had an SMA margin distance less than orequal to 1 mm had the highest overall (82%), locoregional (36%), and distant (50%)recurrence rates. CT scan overestimated the distance between the cancer and theSMA in 73% of cases. Liu and colleagues43 found that the SMA margin distance is anindependent predictor of both disease-free and overall survival in patients with PDACtreated with preoperative therapy. Adherence to meticulous surgical technique withperiadvential SMA dissection and preoperative chemoradiation should be used tomaximize the SMA margin (Fig. 4).

SUMMARY

Patients with BR PDAC represent a heterogeneous group of patients at high risk ofregional and distant recurrence after therapy. The first step is to evaluate, accuratelyidentify, and stage, so that a treatment plan canbederived accordingly for eachpatient.Amultidisciplinary approach using preoperative chemotherapy and/or chemoradiationis necessary to select patients who are most likely to benefit from pancreatectomy. Inthe patients receiving surgery, a safe and oncologically sound resection of the primarytumor is achieved throughmeticulous retroperitoneal dissection and vascular resectionif necessary. Postoperative recovery that allows additional systemic therapy providespatients with the greatest opportunity for long-term survival.

Denbo & Fleming1348

REFERENCES

1. Yachida S, Jones S, Bozic I, et al. Distant metastasis occurs late during the ge-netic evolution of pancreatic cancer. Nature 2010;467(7319):1114–7.

2. Cameron JL, Riall TS, Coleman J, et al. One thousand consecutive pancreatico-duodenectomies. Ann Surg 2006;244(1):10–5.

3. Lim JE, Chien MW, Earle CC. Prognostic factors following curative resection forpancreatic adenocarcinoma: a population-based, linked database analysis of396 patients. Ann Surg 2003;237(1):74–85.

4. Varadhachary GR, Tamm EP, Crane C, et al. Borderline resectable pancreaticcancer. Curr Treat Options Gastroenterol 2005;8(5):377–84.

5. Katz MH, Wang H, Fleming JB, et al. Long-term survival after multidisciplinarymanagement of resected pancreatic adenocarcinoma. Ann Surg Oncol 2009;16(4):836–47.

6. Winter JM, Cameron JL, Campbell KA, et al. 1423 pancreaticoduodenectomiesfor pancreatic cancer: a single-institution experience. J Gastrointest Surg 2006;10(9):1199–210 [discussion: 210–1].

7. Oettle H, Post S, Neuhaus P, et al. Adjuvant chemotherapy with gemcitabine vsobservation in patients undergoing curative-intent resection of pancreatic can-cer: a randomized controlled trial. JAMA 2007;297(3):267–77.

8. Klinkenbijl JH, Jeekel J, Sahmoud T, et al. Adjuvant radiotherapy and 5-fluoro-uracil after curative resection of cancer of the pancreas and periampullary region:phase III trial of the EORTC gastrointestinal tract cancer cooperative group. AnnSurg 1999;230(6):776–82 [discussion: 82–4].

9. Neoptolemos JP, Dunn JA, Stocken DD, et al. Adjuvant chemoradiotherapy andchemotherapy in resectable pancreatic cancer: a randomised controlled trial.Lancet 2001;358(9293):1576–85.

10. Oettle H, Neuhaus P, Hochhaus A, et al. Adjuvant chemotherapy with gemcita-bine and long-term outcomes among patients with resected pancreatic cancer:the CONKO-001 randomized trial. JAMA 2013;310(14):1473–81.

11. Ueno H, Kosuge T, Matsuyama Y, et al. A randomised phase III trial comparinggemcitabine with surgery-only in patients with resected pancreatic cancer: Jap-anese Study Group of Adjuvant Therapy for Pancreatic Cancer. Br J Cancer2009;101(6):908–15.

12. Hsu CC, Herman JM, Corsini MM, et al. Adjuvant chemoradiation for pancreaticadenocarcinoma: the Johns Hopkins Hospital-Mayo Clinic collaborative study.Ann Surg Oncol 2010;17(4):981–90.

13. Tzeng CW, Tran Cao HS, Lee JE, et al. Treatment sequencing for resectablepancreatic cancer: influence of early metastases and surgical complications onmultimodality therapy completion and survival. J Gastrointest Surg 2014;18(1):16–24 [discussion: 5].

14. Labori KJ, Katz MH, Tzeng CW, et al. Impact of early disease progression andsurgical complications on adjuvant chemotherapy completion rates and survivalin patients undergoing the surgery first approach for resectable pancreaticductal adenocarcinoma - a population-based cohort study. Acta Oncol 2016;55(3):265–77.

15. Tzeng CW, Fleming JB, Lee JE, et al. Defined clinical classifications are associ-ated with outcome of patients with anatomically resectable pancreatic adenocar-cinoma treated with neoadjuvant therapy. Ann Surg Oncol 2012;19(6):2045–53.

Borderline Resectable Pancreatic Cancer 1349

16. Katz MH, Pisters PW, Evans DB, et al. Borderline resectable pancreatic cancer:the importance of this emerging stage of disease. J Am Coll Surg 2008;206(5):833–46 [discussion: 46–8].

17. Fuhrman GM, Charnsangavej C, Abbruzzese JL, et al. Thin-section contrast-enhanced computed tomography accurately predicts the resectability of malig-nant pancreatic neoplasms. Am J Surg 1994;167(1):104–11 [discussion: 11–3].

18. Balachandran A, Bhosale PR, Charnsangavej C, et al. Imaging of pancreatic neo-plasms. Surg Oncol Clin North Am 2014;23(4):751–88.

19. Delbeke D, Pinson CW. Pancreatic tumors: role of imaging in the diagnosis, stag-ing, and treatment. J Hepatobiliary Pancreat Surg 2004;11(1):4–10.

20. Karmazanovsky G, Fedorov V, Kubyshkin V, et al. Pancreatic head cancer: accu-racy of CT in determination of resectability. Abdom Imaging 2005;30(4):488–500.

21. Ishikawa O, Ohigashi H, Imaoka S, et al. Preoperative indications for extendedpancreatectomy for locally advanced pancreas cancer involving the portalvein. Ann Surg 1992;215(3):231–6.

22. Loyer EM, David CL, Dubrow RA, et al. Vascular involvement in pancreaticadenocarcinoma: reassessment by thin-section CT. Abdom Imaging 1996;21(3):202–6.

23. Tran Cao HS, Balachandran A, Wang H, et al. Radiographic tumor-vein interfaceas a predictor of intraoperative, pathologic, and oncologic outcomes in resect-able and borderline resectable pancreatic cancer. J Gastrointest Surg 2014;18(2):269–78 [discussion: 78].

24. Al-Hawary MM, Francis IR, Chari ST, et al. Pancreatic ductal adenocarcinomaradiology reporting template: consensus statement of the Society of AbdominalRadiology and the American Pancreatic Association. Gastroenterology 2014;146(1):291–304 e1.

25. Varadhachary GR, Tamm EP, Abbruzzese JL, et al. Borderline resectable pancre-atic cancer: definitions, management, and role of preoperative therapy. Ann SurgOncol 2006;13(8):1035–46.

26. Mehta VK, Fisher G, Ford JA, et al. Preoperative chemoradiation for marginallyresectable adenocarcinoma of the pancreas. J Gastrointest Surg 2001;5(1):27–35.

27. Callery MP, Chang KJ, Fishman EK, et al. Pretreatment assessment of resectableand borderline resectable pancreatic cancer: expert consensus statement. AnnSurg Oncol 2009;16(7):1727–33.

28. Landry J, Catalano PJ, Staley C, et al. Randomized phase II study of gemcitabineplus radiotherapy versus gemcitabine, 5-fluorouracil, and cisplatin followed byradiotherapy and 5-fluorouracil for patients with locally advanced, potentiallyresectable pancreatic adenocarcinoma. J Surg Oncol 2010;101(7):587–92.

29. Katz MH, Marsh R, Herman JM, et al. Borderline resectable pancreatic cancer:need for standardization and methods for optimal clinical trial design. AnnSurg Oncol 2013;20(8):2787–95.

30. Rich TA, Evans DB. Preoperative combined modality therapy for pancreatic can-cer. World J Surg 1995;19(2):264–9.

31. Katz MH, Fleming JB, Bhosale P, et al. Response of borderline resectablepancreatic cancer to neoadjuvant therapy is not reflected by radiographic indica-tors. Cancer 2012;118(23):5749–56.

32. Donahue TR, Isacoff WH, Hines OJ, et al. Downstaging chemotherapy and alter-ation in the classic computed tomography/magnetic resonance imaging signs ofvascular involvement in patients with pancreaticobiliary malignant tumors: influ-ence on patient selection for surgery. Arch Surg 2011;146(7):836–43.

Denbo & Fleming1350

33. Ferrone CR, Marchegiani G, Hong TS, et al. Radiological and surgical implica-tions of neoadjuvant treatment with FOLFIRINOX for locally advanced andborderline resectable pancreatic cancer. Ann Surg 2015;261(1):12–7.

34. Tzeng CW, Balachandran A, Ahmad M, et al. Serum carbohydrate antigen 19-9represents a marker of response to neoadjuvant therapy in patients with border-line resectable pancreatic cancer. HPB (Oxford) 2014;16(5):430–8.

35. Iacobuzio-Donahue CA, Fu B, Yachida S, et al. DPC4 gene status of the primarycarcinoma correlates with patterns of failure in patients with pancreatic cancer.J Clin Oncol 2009;27(11):1806–13.

36. Fuhrman GM, Leach SD, Staley CA, et al. Rationale for en bloc vein resection inthe treatment of pancreatic adenocarcinoma adherent to the superiormesenteric-portal vein confluence. Pancreatic Tumor Study Group. Ann Surg1996;223(2):154–62.

37. Katz MH, Lee JE, Pisters PW, et al. Retroperitoneal dissection in patients withborderline resectable pancreatic cancer: operative principles and techniques.J Am Coll Surg 2012;215(2):e11–8.

38. Katz MH, Fleming JB, Pisters PW, et al. Anatomy of the superior mesenteric veinwith special reference to the surgical management of first-order branch involve-ment at pancreaticoduodenectomy. Ann Surg 2008;248(6):1098–102.

39. Tseng JF, Raut CP, Lee JE, et al. Pancreaticoduodenectomy with vascular resec-tion: margin status and survival duration. J Gastrointest Surg 2004;8(8):935–49[discussion: 49–50].

40. Raut CP, Tseng JF, Sun CC, et al. Impact of resection status on pattern of failureand survival after pancreaticoduodenectomy for pancreatic adenocarcinoma.Ann Surg 2007;246(1):52–60.

41. Katz MH, Merchant NB, Brower S, et al. Standardization of surgical and patho-logic variables is needed in multicenter trials of adjuvant therapy for pancreaticcancer: results from the ACOSOG Z5031 trial. Ann Surg Oncol 2011;18(2):337–44.

42. Katz MH, Wang H, Balachandran A, et al. Effect of neoadjuvant chemoradiationand surgical technique on recurrence of localized pancreatic cancer.J Gastrointest Surg 2012;16(1):68–78 [discussion: 78–9].

43. Liu L, Katz MH, Lee SM, et al. Superior mesenteric artery margin of posttherapypancreaticoduodenectomy and prognosis in patients with pancreatic ductaladenocarcinoma. Am J Surg Pathol 2015;39(10):1395–403.