Multidisciplinary standards of care and recent progress in pancreatic ductal adenocarcinoma

Aaron J Grossberg; Linda C Chu; Christopher R Deig; Eliot K Fishman; William L Hwang; Anirban Maitra; Daniel L Marks; Arnav Mehta; Nima Nabavizadeh; Diane M Simeone; Colin D Weekes; Charles R Thomas, Jr

doi:10.3322/caac.21626

. Author manuscript; available in PMC: 2021 Sep 1.

Published in final edited form as: CA Cancer J Clin. 2020 Jul 19;70(5):375–403. doi: 10.3322/caac.21626

Multidisciplinary standards of care and recent progress in pancreatic ductal adenocarcinoma

Aaron J Grossberg ^1,^2,³, Linda C Chu ⁴, Christopher R Deig ¹, Eliot K Fishman ⁴, William L Hwang ^5,⁶, Anirban Maitra ⁷, Daniel L Marks ^2,⁸, Arnav Mehta ^6,⁹, Nima Nabavizadeh ¹, Diane M Simeone ¹⁰, Colin D Weekes ¹¹, Charles R Thomas Jr ¹

PMCID: PMC7722002 NIHMSID: NIHMS1615587 PMID: 32683683

Abstract

Despite tremendous gains in the molecular understanding of exocrine pancreatic cancer, the prognosis for this disease remains very poor, largely owing to delayed disease detection and limited effectiveness of systemic therapies. Incidence rates and mortality rates for pancreatic cancer have both increased during the past decade, in stark contrast to most other solid tumor types. Recent improvements in multimodality care have substantially improved overall survival, local control, and metastasis-free survival for patients with localized tumors amenable to surgical resection. The widening gap in prognosis between patients with resectable and unresectable or metastatic disease reinforces the importance of detecting pancreatic cancer sooner to improve outcomes. Furthermore, the developing use of therapies that target tumor-specific molecular vulnerabilities may offer improved disease control for patients with advanced disease. Finally, the substantial morbidity associated with pancreatic cancer, including wasting, fatigue, and pain, remains an underaddressed component of this disease, which powerfully impacts quality of life and limits tolerance to aggressive therapies. In this paper we review the current multidisciplinary standards of care in pancreatic cancer with a focus on emerging concepts in pancreatic cancer detection, precision therapy, and survivorship.

Background and Epidemiology

Pancreatic cancer has the poorest prognosis of any common solid malignancy, with a 5-year overall survival of approximately 10%.¹ Although this represents a modest improvement in survival, the absolute number of individuals who succumb to this disease continues to rise. In 2020, it is estimated that 57,600 people will be diagnosed with and 47,050 deaths will be attributed to pancreatic cancer in the United States, recently eclipsing breast cancer as the 3^rd leading cause of overall cancer death.² The median age at diagnosis of pancreatic cancer is 70 years.^{1, 3} Incidence rates during 2013–2017 were higher among males than females (14.9 and 11.6 cases annually per 100,000 persons, respectively), as were mortality rates (12.7 and 9.6 deaths annually per 100,000 persons, respectively).^{1, 3} Incidence and mortality rates during this period were highest for blacks (15.3 cases and 13.3 deaths annually per 100,000 persons, respectively), followed by non-Hispanic whites (13.1 cases and 10.9 deaths annually per 100,000 persons, respectively), with lower rates among Hispanics, and especially among Asians/Pacific Islanders and American Indians/Alaska Natives.¹ Lost earnings from person-years of life lost from pancreatic cancer in 2015 are estimated to be over $6 billion.⁴ Incidence and mortality rates both increased by an average of 0.3% per year during the past decade.¹ Underlying these trends is a combination of an aging population, longer expected lifespan, and the public health pandemics of obesity and diabetes.

Approximately 95% of pancreatic cancers are exocrine cell tumors, most commonly ductal adenocarcinomas (PDAC). Endocrine pancreas cancers are generally more indolent tumors with a more favorable prognosis, reviewed elsewhere.⁵ There are four fundamental challenges that underlie the high mortality of PDAC. First, the pancreas is situated deep within the upper abdomen, seated behind the stomach and between the aorta and its major upper abdominal branches. Not only does this shield growing tumors from detection, but, because the cancer often grows around and encases these vessels, only 15–20% can undergo surgical resection, the foundation of curative treatment.⁶ Second, PDAC exhibits an aggressive biology characterized by early metastasis. Greater than 50% of patients have distant metastatic disease on presentation, and the majority of patients who undergo resection will develop metastases within 4 years of surgery, suggesting the de facto presence of micrometastases in patients with apparently localized tumors.^{1, 3, 7, 8} Third, the physiologic effects of PDAC can dramatically weaken patients, limiting their ability to withstand aggressive treatment. The wasting syndrome of cachexia is present in up to 80% of PDAC patients at diagnosis,⁹ and this may be further complicated by exocrine and endocrine pancreatic dysfunction.^{10, 11} Cachectic patients exhibit poor treatment tolerance, as evidenced by decreased survival following pancreatectomy^{12, 13} or chemotherapy.^{14, 15} Finally, PDAC exhibits resistance to many antineoplastic therapies, with rapid progression and low rates of pathologic complete response even with the most effective systemic agents and radiotherapy.¹⁶ Indeed, fewer than 3% of patients who present with metastatic disease are alive after 5 years, whereas this number jumps to over 70% in patients with localized, stage IA disease.^{17, 18} The impact of these challenges is reflected in the prognostic factors predicting poorer survival: advanced T-category, the presence of nodal metastasis or distant metastasis, the presence of gross or microscopic residual disease following resection, high histologic grade, invasion of major blood vessels, and poor performance status.^19–21

Presentation

Although most patients are symptomatic at presentation, symptoms of PDAC are often non-specific, leading to a median delay between presentation and diagnosis of over 2 months.²² The most commonly reported symptoms are fatigue (86%), weight loss (85%), anorexia (83%), jaundice (56%), nausea (51%), abdominal pain (79%), diarrhea (44%), pruritis (32%), and steatorrhea (25%).²³ Clinical signs of PDAC, including jaundice (55%), hepatomegaly (29%), cachexia (13%), epigastric mass (9%), or ascites (5%), are much less common.²³ This can make it difficult for primary care and front-line physicians to know when it is appropriate to escalate a workup, as there is no specified diagnostic algorithm for PDAC. The development of any of these symptoms in the context of newly diagnosed diabetes, a family history of PDAC, or a history of recurrent or chronic pancreatitis should alert the managing physician to strongly consider PDAC in the differential.^{24, 25} Recent evidence indicates that PDAC-associated metabolic disruptions, leading to skeletal muscle and adipose wasting, provide early evidence of PDAC growth.^{10, 26, 27}

Risk Factors and Early Detection

Despite the poor prognosis of PDAC overall, data from the National Cancer Institute’s Surveillance, Epidemiology and End Results (SEER) program underscore the impact of earlier detection on outcomes.³ For example, analysis of SEER data between 2010–2016 shows that while only 2.9% of PDAC patients with distant metastases survived 5-years or beyond, 39.4% of patients with localized disease survived that duration. Patients with PDAC smaller than 1 cm who were diagnosed by endoscopic ultrasound (EUS) had even higher 5-year survival rates (reaching ~70% for patients with no obstructive symptoms or detectable “mass” on CT)¹⁸, while a more recent study from the United States of individuals at high risk for PDAC undergoing longitudinal surveillance has demonstrated that resecting advanced preneoplastic lesions is essentially curative.²⁸ These findings reiterate the importance of earlier detection of PDAC, possibly even at the stage of carcinoma-in-situ (a.k.a. pancreatic intraepithelial neoplasia [PanIN]-3).

Nonetheless, the United States Screening and Preventive Task Force (USPSTF), an expert body charged with making recommendations for screening and other clinical preventive services for the nation, has recently reaffirmed its longstanding recommendation of discouraging screening for PDAC in the general population, concluding that, “there is moderate or high certainty that the service has no net benefit or that the harms outweigh the benefits” and thus service a failing “D” grade.²⁹ On the one hand, the USPSTF guidelines sound counterintuitive in light of the aforementioned impact of earlier detection of PDAC on stage specific survival. On the other hand, there is compelling rationale against screening the general population for PDAC. At an incidence rate of ~13 cases per 100,000 adults,³ PDAC is still relatively uncommon. In contrast, the incidence rate of two cancers for which general population screening is recommended by the USPSTF - breast cancer and colorectal cancer – are ~69 cases and ~38 cases per 100,000 adults, respectively.^30–33 This implies that even a “perfect” PDAC biomarker with a sensitivity of 100% (i.e. not a single cancer being missed) and a specificity of 99% (one false positive of 100 abnormal tests), would only have a positive predictive value close to 1%^{34, 35}, leading to a large number of individuals undergoing unnecessary imaging tests or potentially harmful procedures and adding greatly to healthcare costs and patient morbidity. In order to circumvent this pitfall, the USPSTF has excluded defined cohorts that are at higher than average risk for PDAC from its screening recommendation. In the following paragraphs, we will discuss some of these high-risk cohorts, and other emerging paradigms in early detection.

Approximately 10% of PDAC patients harbor a pathogenic germline mutation in a cancer predisposing gene, of which BRCA2 and ATM are the two most common candidates, followed by BRCA1, PALB2, CDKN2A/p16, LKB1/STK11, and the mismatch repair genes (hMLH1, hMSH2 and hPMS6), and other rarer variants (Table 1).^{36, 37} Of note, only half of the patients with a deleterious germline mutation report an overt family history of PDAC, in light of which the American Society of Clinical Oncology (ASCO) and the National Comprehensive Cancer Network (NCCN) recently updated their guidelines to recommend universal germline mutation testing for all patients diagnosed with PDAC (instead of only those with a suspicious family history).^{38, 39} This has the added benefit of identifying patients with BRCA1/2 mutations who might benefit from PARP inhibitor therapy (see later).⁴⁰ The first-degree relatives of positive index cases can then be approached for testing done of the pathogenic mutation of interest (a process known as “cascade testing”), and the ready availability of relatively inexpensive blood and saliva-based multi-gene panels has greatly simplified the process.⁴¹ Asymptomatic germline mutation carriers represent a rich pool of high-risk individuals for cancer interception, and in fact, this subset has been explicitly excluded from the USPSTF recommendations against screening in the general population at average risk.²⁹ There is currently no broad consensus on how to conduct longitudinal surveillance on germline mutation carriers, including the optimal imaging modalities to use in this population, although a recent international consortium has suggested some overall guidelines that can be pursued within an academic research setting.¹⁷ Nonetheless, retrospective data from two of the largest familial PDAC registries have shown the impact of longitudinal surveillance on outcomes, with the majority (75–90%) of incident cancer cases being diagnosed at a resectable stage, which in turn, translated into disease specific survival exceeding 3 years.^{17, 42}

Table 1.

Pancreatic cancer susceptibility genes and estimated risk

Syndrome	Gene	Relative Risk
Familial atypical multiple mole melanoma (FAMMM)	CDKN2A	13 to 39-fold^{37, 269, 270}
Familial breast and ovarian	BRCA1 and 2	2-fold and 3–9 fold^{37, 271}
Fanconi anemia, Breast CA	PALB2	Unknown²⁷²
Familial adenomatous polyposis (FAP)	APC	5-fold²⁷³
Lynch syndrome	MLH1, MSH6, MSH2, PMS2, EPCAM	9 to 11-fold^{37, 274}
Peutz-Jeghers syndrome	STK11/LKB1	132-fold²⁷⁵
Hereditary pancreatitis	PRSS1	53-fold²⁷⁶
Li-Fraumeni syndrome	p53	7-fold³⁷
Ataxia-telangiectasia	ATM	~3-fold³⁷

Open in a new tab

A separate subgroup of patients, with pancreatic cysts, may benefit from early detection efforts as well. Pancreatic cysts are categorized as inflammatory (including pancreatic pseudocysts) and non-inflammatory (including mucinous and non-mucinous lesions).⁴³ Mucinous cysts of the pancreas are comprised of two distinct entities – intraductal papillary mucinous neoplasms (IPMNs) and mucinous cystic neoplasms (MCNs), and both are considered bona fide precursor lesions of PDAC.⁴⁴ In contrast to microscopic PanINs, which are the most common (~90%) precursor subtype associated with an invasive adenocarcinoma⁴⁵, the macroscopic (cystic) precursors can be readily imaged using CT or MRI scans, and are thus amenable to longitudinal surveillance for progression to cancer. Notably, retrospective studies on imaging data have shown that as many as 2–3% of the general population might harbor asymptomatic pancreatic cysts, and this number rises to over 10-fold higher in elderly individuals.⁴⁶ This “man-made” epidemic in pancreatic cysts can be attributed to the tens of millions of abdominal scans that are conducted each year in the United States for unrelated causes. Given that no more than 5–10% of PDAC annually arise in the backdrop of a cystic lesion, the vast majority of these asymptomatic pancreatic cysts are essentially benign and can be followed with conservative surveillance. The importance of early detection in cyst patients is underscored by the data that patients with non-invasive cystic lesions are usually cured upon surgical resection, while those with an invasive component can see their 5-year survival drop by 50% or greater.⁴⁴ Therefore, identifying the minor subset of mucinous cysts that have either progressed to high-grade dysplasia or PDAC, or harbor an intrinsic biological potential for progression during the patient’s lifetime, is of paramount importance. However, based on SEER data, a cyst detected incidentally on MRI has a 17 in 100,000 chance of being a ductal cancer, indicating a very elevated number needed to surveil to prevent one premature death.⁴⁷ Despite this, several international societies have published largely overlapping recommendations on cyst surveillance. The evidence driving these recommendations is derived from case series and retrospective reports, and graded as “very low quality”; thus it is unclear that the benefits of imaging surveillance outweigh potential harm.⁴⁷ A compendium of clinical and imaging-based criteria have been suggested by various expert bodies (reviewed in⁴⁸) that can support the clinician in their management decision making, but all have various shades of imperfections, resulting in some cases of both over-treatment and missed cancer diagnoses. Recently, molecular testing of endoscopically aspirated pancreatic cyst fluid for molecular biomarkers of mucinous cysts, as well as progression to cancer, have been implemented in the clinical domain, and have resulted in improved performance over the clinical/imaging criteria alone.^{49, 50} The application of machine learning algorithms towards designing an integrated approach to cyst classification will result in further accuracy in predicting underlying biology and improved management.⁵¹

A third high risk subset for PDAC are patients with chronic pancreatitis, most commonly secondary to chronic alcohol dependence, smoking, hypertriglyceridemia, diabetes, or renal failure.⁵² Approximately 5% of patients with chronic pancreatitis of 20 years duration will progress to PDAC, and concomitant smoking enhances the risk of neoplastic progression.⁵³ Patients with sporadic chronic pancreatitis are currently not recommended to undergo PDAC surveillance under the USPSTF screening guidelines.²⁹ However, the USPSTF recommendation against screening does not apply to a rare subset of patients with so-called “hereditary pancreatitis” secondary to germline mutations in the PRSS1 gene, which encodes for cationic trypsinogen.⁵⁴ The mutation renders trypsin resistant to inactivation, resulting in recurrent episodes of acute pancreatitis beginning in childhood. These patients have a ~50-fold higher lifetime risk of PDAC⁵⁵, redemonstrating the intimate link between inflammation and cancer.

While the aforementioned risk factors cumulatively impact ~15–20% of diagnosed PDAC patients annually, the majority still fall under the category of what would be considered as “sporadic” cancer (Table 2). How do we enable early detection in individuals with no apparent clinical risk factor like cysts or family history? One can certainly implement public health approaches like smoking avoidance and cessation, and maintaining a healthy body mass index (BMI), since both of these modestly (~1.5–2-fold) increase the lifetime risk of PDAC.⁵⁶ The use of genome wide association studies (GWAS), where thousands of polymorphisms across the genome are compared in cases versus controls, has enabled the identification of multiple susceptibility alleles in PDAC, such as alleles within the ABO blood group genes and the gene encoding telomerase reverse transcriptase (TERT).^{57, 58} In contrast to deleterious germline mutations described above (BRCA1/2, ATM, etc.), these variant alleles individually only have a very modest impact on lifetime risk, but their cumulative risk could become appreciable. The incorporation of predisposing allelic information into a so-called “polygenic risk score” (PRS)⁵⁹ could then identify individuals at the highest quartile of lifetime risk who could be enrolled into longitudinal surveillance programs. In the past decade, the emergence of new onset hyperglycemia or frank diabetes has been identified as the presenting symptom of an otherwise asymptomatic PDAC in up to half of all patients.⁶⁰ The deregulation in glucose homeostasis is a paraneoplastic syndrome caused by the underlying PDAC, which can start to appear as early as 36 months prior to clinical diagnosis and is accompanied by changes in subcutaneous adipose tissue.²⁶ Circulating factors, including antigen and microRNA panels, show promise in discriminating patients with early PDAC from controls or individuals with benign pancreatic conditions, offering the promise of liquid biopsy approaches toward PDAC early detection.^61–64 Additionally, recent data establishes that alterations in the gut and pancreatic microbiome directly promote pancreatic oncogenesis and influence survival, indicating that changes in fecal microbial composition may also help identify disease development.^{65, 66} In a “cancer interception paradigm of the future”, one can envision a pipeline wherein asymptomatic individuals are identified at higher than average lifetime risk for PDAC based on a combination of PRS, family history, smoking history, and BMI, and enrolled into a risk reduction program for PDAC (which includes guidance on lifestyle modification) and a surveillance program for individuals for whom benefit is anticipated to exceed harm. The onset of hyperglycemia or frank diabetes, as well as a catalog of longitudinally structured data readily available from the EMR would then “trigger” additional work up, such as imaging studies, in these high-risk individuals, eventually leading to earlier diagnosis of PDAC at a resectable stage and improved long term survival.

Table 2.

Risk factors for the development of pancreatic cancer

Factor	Relative Risk
Tobacco Smoking	1.7 to 2.6-fold^277–279
Obesity	1.1 to 1.5-fold^{280, 281}
Diabetes	1.5 to 2-fold¹¹
Family History	1.7 to 2.3-fold^282–284
Chronic pancreatitis	13.3-fold²⁸⁵

Open in a new tab

Smoking appears to increase risk of PDAC in women more than men.^{286, 287}

Diagnosis and Imaging

Diagnostic Workup

The diagnosis of PDAC cannot be made based on symptoms and signs alone. Patients presenting with jaundice or epigastric pain should be evaluated with complete blood count, blood chemistry panel, and liver function tests, including serum aminotransferases, alkaline phosphatase, and bilirubin. These values can help assess the extent of cholestasis (bilirubin), liver metastasis (alkaline phosphatase), hepatitis (aminotransferases), and nutritional status (albumin, prealbumin). Those with epigastric pain should also have serum lipase measured to evaluate for acute pancreatitis. The tumor marker sialylated Lewis^a blood group antigen CA19–9 is frequently used in the workup for PDAC. In symptomatic patients, the sensitivity and specificity of CA19–9 range from 70–90%, but the positive predictive value of elevated CA19–9 in asymptomatic patients was only 0.9%, making it inadequate as a diagnostic in this population.^{35, 67, 68} Its limited utility is based on elevations in benign pancreaticobiliary diseases, cancers other than PDAC, and the fact that 5–10% of the population do not express Lewis antigens.^{69, 70} Emerging data suggest that the combination of serum CA19–9 with additional biomarkers, such as MUC5AC or thrombospondin-2 improves the specificity of serum testing, offering potential for a future blood-based diagnostic approach.^{63, 69} Serum CA19–9 levels are closely related to tumor size, and the degree of elevation in CA19–9 is associated with prognosis.⁶⁸ In a study of patients with apparently localized disease, values above 130 units/mL predicted occult unresectable disease and were prognostic for survival among >1,500 patients with resectable cancers.^{71, 72} Although patients with apparently localized PDAC and high levels of CA19–9 are commonly recommended for staging laparoscopy and neoadjuvant therapy, ASCO guidelines do not specify a cutoff value of CA19–9 to be used in this manner.⁷³ Because elevations in serum CA19–9 can be induced by either tumor production or cholestasis, CA19–9 should be remeasured after stent placement in patients with biliary obstruction to estimate “true” tumor burden, accounting for its 4–8 day half-life.⁷⁴ Serial monitoring of CA19–9 is commonly used to track response to therapy in patients who present with elevated CA19–9. A failure in CA19–9 normalization after surgery is associated with poor survival and thought to represent occult metastatic disease.^75–77 Similarly, declining CA19–9 during systemic therapy correlates with improved patient survival, although it is unclear what magnitude of decline is most prognostic.^{78, 79} Rises in CA19–9 after a nadir can represent treatment failure and often precede imaging evidence of recurrent or progressive cancer.⁸⁰ Serum CA19–9 changes are not considered to be a substitute for imaging evidence of treatment response or recurrence.⁸¹ In some tumors additional cancer-specific biomarkers, such as carcinoembryonic antigen or CA-125, are elevated and can also be used to track response to therapy and recurrence. Because these markers are elevated in only a subset of patients with PDAC, their utility in diagnosis is limited.⁸²

Imaging Techniques

Computed tomography (CT) is the first-line imaging modality for initial evaluation of suspected PDAC and is preferred over magnetic resonance imaging (MRI) due to its lower cost and widespread availability.⁸³ Both CT and MRI have comparable sensitivity in detection of PDAC, ranging from 76% to 96% for CT and from 83% to 94% for MRI.⁸⁴ MRI is usually reserved as a second-line modality in patients with contraindications to CT (e.g. severe iodinated contrast allergy or renal insufficiency). MRI is also used as a problem-solving tool in cases with equivocal CT features and for characterization of indeterminate liver lesions.⁸³ Position emission tomography/computed tomography (PET/CT) has been shown to detect extrapancreatic metastatic disease that was not detected based on traditional staging examination.⁸⁵ Although not recommended as part of routine staging, PET/CT may be considered in patients at high-risk of extrapancreatic metastases.⁸³ The primary role of endoscopic ultrasound (EUS) is to guide needle biopsies to confirm the diagnosis of PDAC. In select cases, EUS may be helpful in detecting a small pancreatic mass that may be difficult to visualize on CT or MRI, and is therefore the preferred imaging modality in some early detection surveillance programs.^{28, 86, 87}

Reported accuracy in determining tumor resectability ranges from 73% to 87% for CT and from 70% to 79% for MRI.⁸⁴ CT offers superior spatial resolution and is less susceptible to respiratory motion artifacts than MRI, which is essential in demonstrating the critical relationship between the tumor and adjacent vasculature. The accuracy of PDAC detection and staging critically depends on the appropriate imaging protocol, post-processing technique, and experience of radiologists.

Pancreatic cancer CT protocol, endorsed by both Society of Abdominal Radiology and American Pancreatic Association, states that CT examination should be performed with intravenous contrast (>300 mg I/mL) at an injection rate of 3–5 mL/s with scans obtained at pancreatic parenchymal phase (40–50 s) and portal venous phase (65–70 s). A neutral or low-Hounsfield unit oral agent should be administered. The dataset should be obtained with submillimeter slice thickness, reconstructed into 0.75 to 3 mm axial slices, with multiplanar reconstruction, and 3D reconstruction to allow for full assessment of vascular involvement.⁸⁸ Cinematic rendering is a recently described 3D rendering technique that can provide photorealistic detail, and has the potential to improve visualization of tumor-vessel relationships (Figures 1 & 2).⁸⁹

Figure 1. — 44-year-old woman with resectable PDAC arising from pancreatic body. A. Axial IV contrast-enhanced CT image shows a hypoenhancing pancreatic body mass (arrow) associated with pancreatic duct dilatation (white arrowhead). Tumor abuts splenic artery and splenic vein (black arrowhead). B. Coronal IV contrast-enhanced CT image with cinematic rendering improves conspicuity of pancreatic tumor (arrow) relatively to background pancreas and improves visualization of minimal tumor abutment of splenic artery (arrowhead).

Figure 2. — 50-year-old man with locally advanced PDAC arising from head of pancreas. A, B. Axial IV contrast-enhanced CT images show (A) tumor (arrow) with SMA encasement (arrowhead), and (B) tumor (arrow) with encasement and near occlusion of portal vein (arrowhead). C. Coronal IV contrast-enhanced CT image with cinematic rendering improves visualization of tumor encasement (arrow) of SMA (white arrowhead) and portal vein (black arrowhead).

Staging Systems

The American Joint Committee on Cancer (AJCC) stages PDAC based on a TNM staging system. The revised 8^th edition of the AJCC manual addressed some of the criticisms of earlier versions with changes to T and N categories.⁹⁰ T categories are mostly based on tumor size. T4 is defined as tumor with arterial involvement regardless of size. N categories are further classified in the 8^th edition based on absence of lymph node involvement (N0) and number of regional lymph nodes involved (1–3 for N1, ≥4 for N2), rather than only absence (N0) or presence (N1) of lymph nodes.⁹⁰ The primary goal of AJCC system is to provide prognostic information, rather than guiding management. A number of organizations have issued management guidelines.^{83, 91–93} NCCN guidelines classify the resectability of localized PDAC based on preoperative imaging findings into resectable, borderline resectable, and locally advanced disease, and are summarized in Table 3. Arterial abutment (<180°) is considered borderline resectable, whereas arterial encasement (≥180°) is usually considered locally advanced (exception noted below) (Figure 2). Venous abutment, encasement, or thrombosis are considered borderline resectable, as long as the venous segment is reconstructable. Unreconstructable venous involvement is considered locally advanced.⁸³ NCCN guidelines share many common features with other guidelines (Table 4) with notable exception of celiac artery encasement (>180°). A subset of patients with pancreatic body or tail tumor with celiac artery encasement, but without involvement of the aorta or gastroduodenal artery, can be considered borderline resectable with a modified Appleby procedure, a distal pancreatectomy with en bloc resection of the celiac axis (Figure 3). However, in other guidelines, the presence of celiac artery encasement is considered unresectable.

Table 3.

NCCN Criteria in Defining Resectability of PDAC at Diagnosis83

Resectability Status	Arterial Involvement			Venous Involvement
Resectability Status	Celiac Artery	SMA	Common Hepatic Artery	Portal Vein/SMV
Resectable	None	None	None	None ≤180° contact without contour irregularity
Borderline Resectable	≤180° >180°, without involvement of aorta or GDA (body/tail)	≤180°	Solid tumor contact without extension into CA or hepatic artery bifurcation	>180° ≤180° with contour irregularity or thrombosis, with reconstructible PV/SMV Solid tumor contact with IVC
Locally Advanced	>180° (head/uncinate) Solid tumor contact with CA and aorta	>180°		Unreconstructible PV/SMV due to tumor involvement or occlusion

Open in a new tab

Table 4.

Comparison of Resectability Criteria among Organizations

Vessel Involvement	NCCN 2019⁸³	MD Anderson⁹¹	ACTO⁹²	AHPBA/SSAT/SSO⁹³
CA Abutment (≤180°)	Borderline	Borderline	Borderline	Unresectable
CA Encasement (>180°)	Borderline (body/tail) Locally Advanced (head/uncinate)	Unresectable	Unresectable	Unresectable
SMA Abutment (≤180°)	Borderline	Borderline	Borderline	Borderline
SMA Encasement (>180°)	Locally Advanced	Unresectable	Unresectable	Unresectable
CHA Abutment or Encasement	Borderline	Borderline	Borderline	Borderline
PV/SMV Encasement (>180°) or Abutment (≤180°) with Contour Abnormality	Borderline	Borderline	Borderline	Borderline

Open in a new tab

ACT = Alliance for Clinical Trials in Oncology; AHPBA = American Hepato-Pancreato-Biliary Association; SSAT = Society for Surgery of the Alimentary Tract; SSO = Society for Surgical Oncology

Figure 3. — 70-year-old man with borderline resectable PDAC arising from pancreatic body. A. Axial IV contrast-enhanced CT image in maximum intensity projection shows tumor (arrow) encasement of celiac artery (arrowhead), common hepatic artery, and splenic artery. B. Sagittal IV contrast-enhanced CT image shows tumor encasement of the celiac artery (arrow) and sparing of the SMA (arrowhead), which allows for modified Appleby procedure.

Principles of Multidisciplinary Treatment

Metastatic Disease Therapy: An Evolving Treatment Paradigm

Metastatic disease represents the most common clinical presentation of PDAC.² Historically, gemcitabine was the standard of care for the first-line treatment of metastatic disease based on a 5.65 month improvement in median overall survival (OS) in comparison to fluorouracil (5FU).⁹⁴ Major studies establishing standard of care first-line therapy for PDAC are listed in Table 5. In 2011, the PRODIGE 4/ACCORD 11 phase III trial demonstrated an improved median OS and median progression-free survival (PFS) in patients treated with FOLFIRINOX (fluorouracil, leucovorin, irinotecan and oxaliplatin) in comparison to those receiving gemcitabine as first-line therapy for metastatic PDAC (median OS: 11.1 v 6.8 mo, hazard ratio (HR) 0.57, p<0.001 and median PFS: 6.4 v 3.3 mo, HR 0.47, p<0.001).⁹⁵ FOLFIRINOX now embodies standard first-line therapy for fit patients. The combination of gemcitabine plus nab-paclitaxel represents another first-line therapy for the disease. The phase III MPACT trial demonstrated an improvement of 1.8 months in both median OS (HR 0.72, p<0.001) and median PFS (HR 0.69, p<0.001), respectively for patients with metastatic PDAC treated with gemcitabine plus nab-paclitaxel compared to gemcitabine.⁹⁶ Collectively, these studies demonstrate that patients with metastatic PDAC both tolerate and benefit from aggressive combination chemotherapy. In the absence of a head-to-head comparison of these treatments, many providers select therapy based on patient performance status and ability to adhere to the treatment schedule, with FOLFIRINOX considered the more challenging, but perhaps more effective regimen.

Table 5.

Trials evaluating chemotherapy for metastatic pancreatic cancer.

Trial	N	Chemotherapy	MS
Cullinan et al. (1985)²⁸⁸	144	5FU vs 5FU+dox vs 5FU+dox+mitomycin	5.5 vs 5.5 vs 4.5 mo. NS
Burris et al. (1997)⁹⁴	126	5FU vs Gem	4.4 vs 5.6 mo. P=.0025
Tempero et al. (2003)²⁸⁹	92	Gem vs Gem (fixed rate)	5 vs 8 mo. P=.013
Heinemann et al. (2006)²⁹⁰	195	Gem ± cisplatin	6.0 vs 7.5 mo. P=.015
NCIC CTG PA.3 (2007)²⁹¹	569	Gem ± erlotinib	5.9 vs 6.2 mo. P=.038
Cunningham et al. (2009)²⁹²	533	Gem ± capecitabine	6.2 vs 7.1 mo. P=.08
CALGB 80303²⁹³	602	Gem ± bevacizumab	5.9 vs 5.8 mo. P=.95
SWOG S0205 (2010)²⁹⁴	745	Gem ± cetuximab	5.9 vs 6.3 mo. P=.23
PRODIGE 4/ACCORD 11 (2011)⁹⁵	342	Gem vs FOLFIRINOX	6.8 vs 11.1 mo. P<.0001
MPACT (2013)⁹⁶	861	Gem ± nab-paclitaxel	6.7 vs 8.5 mo. P<.001

Open in a new tab

NS = not significant; dox = doxorubicin; Gem = gemcitabine.

Second-line therapy primarily consists of doublet therapy utilizing the alternative pyrimidine backbone to what was used in the first-line setting. Patients receiving FOLFIRINOX for first-line therapy are commonly transitioned to gemcitabine plus nab-paclitaxel as second line therapy based upon the finding of the MPACT study.^{97, 98} In contrast, patient receiving gemcitabine plus nab-paclitaxel in the first-line setting generally are treated with FOLFIRINOX, FOLFOX (5FU and oxaliplatin), or 5FU plus NALIRI (nanoliposomal irinotecan).^{95, 98–100} These additional options are important for patients with impaired performance status or those who exhibit dose-limiting toxicities precluding FOLFIRINOX treatment. FOLFOX has been evaluated in several trials demonstrating a median PFS of ~ 3.5 months.^{98, 99} However, the randomized phase III PANCREOX study, which compared second-line FOLFOX against infusional 5FU/leucovorin found that the FOLFOX was associated with similar PFS, but worse OS and toxicity than the comparison arm, calling into question the advantage of adding oxaliplatin in this setting.⁹⁹ The randomized phase III NAPOLI-1 trial demonstrated an improved OS of 6.1 months associated with NALIRI treatment in comparison to 4.2 months for 5FU (HR for death 0.67, p = 0.012). The use of 5FU as the control versus FOLFOX is a common criticism of this study, but perhaps unwarranted, given the PANCREOX results.¹⁰⁰

The addition of new combinations of cytotoxic chemotherapies is being evaluated as an optimization strategy. The results of a phase IB/II study of gemcitabine, nab-paclitaxel and cisplatin as first-line therapy demonstrated the safety of the regimen, whilst providing early evidence of potential clinical benefit, with a remarkable PFS of 10.1 months.¹⁰¹ Although these results suggest increased efficacy with the addition of platinum, these data should be interpreted with caution in the absence of phase III data.

Maintenance Therapy

The improved outcomes associated with multimodality chemotherapy has led to prolonged exposure of patients to chemotherapy-related toxicities. The PANOPTIMOX-PRODIGE 35 phase II study was designed to evaluate the role of employing a strategy of abbreviated FOLFIRINOX plus maintenance FU or sequential gemcitabine and FOLFIRI3 versus continuous FOLFIRINOX until disease progression to mitigate oxaliplatin induced neuropathy.¹⁰² Treatment with alternating gemcitabine and FOLFIRI3 demonstrated inferior clinical outcomes compared to the FOLFIRINOX regimens. Patients treated with FOLFIRINOX were randomly assigned to continuous treatment for 6 months or 4 months followed by 5FU maintenance with reintroduction of FOLFIRINOX at disease progression. The primary endpoint of the study was 6-month PFS. Treatment with the FOLFIRINOX regimens resulted in equivalent 6-month PFS and OS. Patients treated with maintenance 5FU therapy experienced increased grade 3/4 neurotoxicity, which may be reflective of a higher cumulative oxaliplatin dose in these patients. Retrospective single-institution data out of Germany reports 11 month PFS in patients receiving maintenance FOLFIRI, delivered after 4 months of FOLFIRINOX.¹⁰³ Importantly, these studies demonstrate that maintenance therapy is a viable treatment strategy that is not associated with inferior clinical outcomes in comparison to continued therapy until disease progression.

The POLO study was a phase III study that addressed the role of maintenance targeted therapy in a biomarker selected patient population. Patients with metastatic PDAC harboring deleterious germline BRCA1 or BRCA2 mutations who had not progressed on first-line platinum-based chemotherapy were randomized to receive either the poly(adenosine diphosphate-ribose) polymerase (PARP) inhibitor olaparib or placebo as maintenance therapy.⁴⁰ Olaparib maintenance therapy improved the median PFS of 7.4 months versus 3.8 months in the placebo group (HR 0.53, p=0.004), but did not improve the median OS.⁴⁰ Although this represents the first trial supporting molecular-guided therapy for PDAC, interpretation of this study is limited by both a low patient number and comparison against a no-treatment control arm, which is not standard of care. However, the POLO trial has led to the FDA approval of olaparib for maintenance therapy in patients with germline BRCA mutant disease.

Precision Oncologic Approaches to Pancreatic Cancer

With the increased analysis of the genomic signatures of PDAC, PDAC has been reliably classified into two distinct molecular subtypes of basal/quasi-mesenchymal or classical subtype. The classical subtype is characterized by an epitheloid phenotype; whereas, the basal/quasi-mesenchymal subtype has a mesenchymal phenotype that has a propensity to metastasize.^104–106 The COMPASS trial demonstrated that patients with the classical subtype demarcated by high levels of GATA6 expression by RNASeq or RNA in situ hybridization (ISH) in baseline tumor specimens have a higher response rate to FOLFIRINOX therapy, leading to an improved mPFS.¹⁰⁷ Consistent with these findings, acquired resistance to FOLFIRINOX was associated with the development of basal phenotype upon FOLFIRINOX exposure.^{108, 109} Prospective studies are now required to confirm these observations. If these observations hold true, the choice of chemotherapy may be determined by the predominant molecular phenotype of the tumor.

Patients with either germline or somatic deleterious mutations resulting in homologous DNA repair deficiency (HRD) represent another important subset of patients likely to require specific therapies. HRD patients comprise upwards of 15 to 20% of patients with metastatic PDAC^{36, 110} As outlined previously, these patients may be specifically sensitive to PARP inhibition. The PARP inhibitor rucaparib has demonstrated clinical activity as a single agent in patients with advanced metastatic PDAC with both germline and somatic mutations in BRCA1/BRCA2.¹¹¹ In addition, there are a number of small molecular inhibitors in development targeting various aspects of the homologous DNA repair pathway. Combination strategies incorporating multiple agents targeting the HRD pathway in combinations with PARP inhibition are in development.¹¹² In addition to these effects, PARP inhibition possesses radiosensitizing properties.¹¹³ Platinum drugs induce double-strand DNA breaks resulting in increased sensitivity of HRD pancreatic tumors to this class of chemotherapy. Consistent with this observation, a randomized phase II study of gemcitabine plus cisplatin with or without veliparib as first-line therapy in HRD patients demonstrated response rates of 74% and 65% and median OS of 16 and 15 months, respectively.¹¹⁴

PDAC harboring microsatellite instability represents approximately 1% of patients afflicted with the disease.¹¹⁵ However, it is important to that all patients be tested for microsatellite instability because 40% of patients who have microsatellite instability respond to immune checkpoint inhibition and derive improved survival from this therapy.^{116, 117} This paradigm of tailoring therapeutic approaches to the molecular vulnerabilities of each cancer is known as precision medicine. In PDAC, where so few agents demonstrate clinically meaningful activity, precision medicine approaches hold substantial promise. In support of precision medicine in PDAC, a recent publication on the Know Your Tumor program, found that patients with actionable molecular alterations who received matched molecular therapies had significantly longer median OS than patients who received unmatched therapies (2.58 years vs 1.51 years, HR 0.42, p=0.0004).¹¹⁸ Furthermore, this study demonstrated the feasibility of molecular assessment on biopsy samples in real time. As capacity for molecular testing improve, such precision approaches may become the standard for the subset of patients with actionable molecular drivers.

Investigational Therapeutic Approaches

Treatments Targeting the Tumor Microenvironment

The PDAC tumor microenvironment (TME) consists of a complex network of cells including cancer associated fibroblasts (CAFs), immune cells, endothelial cells plus a dense extracellular matrix.^119–122 The inhibitory immune checkpoint programmed death ligand-1 (PD-L1) is expressed on myeloid cells within the TME and on tumor cells.¹²³ High expression of PD-L1 expression in PDAC predicts a poor prognosis, yet immune checkpoint blockade (ICB) has been largely an ineffective therapeutic strategy. Anti-CTLA4 inhibition with ipilimumab or tremelimumab either as a single agent or in combination with gemcitabine and anti-PD-L1 (BMS-936559 or durvalumab) demonstrated no evidence of clinical activity.^124–127 Combined CXCR4 and ICB is currently under investigation as an approach to overcome immune exclusion of effector T cells based on preclinical data linking CXCR4 blockade to tumor infiltration of effector T-cells.^{128, 129} CD40 agonism represents another promising approach of ICB. CD40 is a component the costimulatory cascade resulting in upregulation of MHC class I and costimulatory molecules and skewing of myeloid cells to a tumoricidal phenotype.^130–133 A phase I study of the CD40 agonist APX005M in combination with gemcitabine and nab-paclitaxel with or without nivolumab demonstrated early evidence of clinical benefit with a PR rate of 58% and DCR of 91%.¹³⁴ Lastly, ionizing radiation can induce immune priming through the production of toxic nucleotide adducts, enhancing intratumoral antigen presentation by dendritic cells.^{135, 136} An initial evaluation of this hypothesis has been conducted in 25 patients with metastatic PDAC by combining radiation therapy with nivolumab and ipilimumab with overall response rate (ORR) of 14% and median PFS of 2.5 months on interim analysis.¹³⁷ The value of this approach will be further tested in future studies.

The extracellular matrix (ECM), comprised of a collagenous matrix and glycosaminoglycans such as hyaluronic acid (HA), is not inert—specifically, hyaluronic acid promotes metastasis and PDAC initiation.¹³⁸ Recombinant human pegylated hyaluronidase, peg-PH20, was developed as an ECM-targeted strategy. Phase II or III studies evaluating peg-PH20 in combination with both FOLFIRINOX and gemcitabine plus nab-paclitaxel in patients with metastatic PDAC expressing high HA levels by immunohistochemistry failed to demonstrate the benefit of adding peg-PH20 to either chemotherapy backbone.^{139, 140} In fact, the combination of peg-PH20 with FOLFIRINOX led to excessive toxicity.¹⁴⁰ Stroma modification remains a challenging therapeutic approach, however recent work evaluating angiotensin receptor blockade using losartan in patients with locally advanced pancreatic cancer show therapeutic promise (see later).^{141, 142}

Treatments Targeting Metabolism and Autophagy

PDAC arising from PanIN is characterized by activating mutations in KRAS.¹⁴³ The ensuing mitogen activated protein kinase (MAPK) activation results in a hypermetabolic state characterized by increased glycolysis, as well as increased metabolic plasticity via altered glutamine metabolism, dependence on oxidative phosphorylation, and metabolite scavenging through the process of macropinocytosis.^144–147 Autophagy is a cellular process designed to allow cells to utilize cellular components as an alternative fuel source in response to cellular stress. Autophagy is integral for PDAC oncogenesis and proliferative capacity.^148–150 Recently, two independent research teams demonstrated that MAPK activation regulates PDAC dependency on autophagy.^{151, 152} Therefore, phase I clinical trials are now ongoing or in development to evaluate combining either MEK or ERK inhibitors with the autophagy inhibitor, hydroxychloroquine (NCT04132505 and NCT03825289).

Resectable Pancreatic Cancer

Surgery

Surgical resection is currently the only means to achieve long-term survival in patients with PDAC. Although only 15–20% of patients present with resectable disease, the increasing use of neoadjuvant therapies and advances in surgical techniques have broadened the pool of patients who are eligible for surgical resection.⁷³ The goals of care for patients with resectable PDAC are increasing the likelihood of margin-negative (R0) resection, decreasing procedural morbidity and mortality, preventing metastatic spread, and improving the patient’s quality of life. As surgical quality, peri-operative care, and systemic therapies have improved, the ultimate goal is to determine the optimal timing and sequencing of high-quality surgery in the appropriate patient as part of a multi-modality treatment plan.

A step by step review of pancreatectomy is beyond the scope of this review, but key steps to emphasize include: 1) a thorough exploration of the abdomen to rule out metastatic disease, 2) biopsy and examination by frozen section of any suspicious lesions outside of the field of resection, and 3) careful management of the surgical margin to ensure adequate tumor clearance, with the uncinate or retroperitoneal margin the margin most at risk. Careful dissection along the periadvential plane of the SMA mitigates the risk of a positive margin at this location during a pancreaticoduodenectomy (Figure 4).

Figure 4. — The Whipple procedure (pancreaticoduodenectomy) for resectable PDAC in the head of the gland. A. Normal anatomic relationship of the pancreas and surrounding structures. B. Diagram of a pancreatic head mass (white) to surround the pancreatic and bile ducts (shown in green). C. Standard resection for pancreaticoduodenectomy to include the head of the pancreas, duodenum, distal common bile duct, distal stomach, and gallbladder. D. Reconstruction to reconnect the pancreas, common bile duct, and stomach to the gastrointestinal tract.

While mortality rates from pancreatectomy have fallen significantly and are below 2% in numerous high-volume centers world-wide, elevated morbidity remains common and still impacts the delivery of adjuvant therapy in up to 40% of patients.^{153, 154} It is recommended that patients seek out high volume centers with multi-disciplinary expertise to optimize their treatment plan and increase opportunities for clinical trial participation. The use of minimally invasive pancreatic resection has been expanding in the last decade. Excellent outcomes in some studies have been reported, however others cite safety concerns, particularly for the use of minimally invasive pancreaticoduodenenctomy.^155–157 In all cases, it is clear that implementation of rigorous dedicated training programs and experience with open pancreatic surgical techniques are needed to ensure safe use of these techniques in clinical practice and to determine if they will improve longer term oncologic outcomes.

Pathology

Rigorous pathologic assessment is essential for accurate prognosis as well as determining the appropriate adjuvant treatment plan. Despite guidelines from the College of American Pathologists and NCCN on standard protocols for the pathologic analysis of PDAC surgical specimens, this practice remains inconsistent, owing both to vagueness in the protocols and variable degrees of adherence across institutions.¹⁵⁸ Protocols for specimen orientation and inking should be well-established between the surgeon and pathologist to ensure clear definition of key margins, generally using different colored inks (specific recommendations found at^{83, 159}). The definition of involved margins are variable across institutions, with most centers in the US defining a positive margin as “tumor on ink”, whereas in Europe margins are called positive if tumor cells fall within 1 mm of the inked margin.¹⁶⁰ Such inconsistency may help to explain the broad variation in reported R0 resection rates across studies.¹⁶¹ This distinction is significant for determining prognosis, as resection margin distance correlates closely with locoregional failure and survival^{162, 163} The pathology report should also include maximal tumor diameter (for staging), histologic subtype, tumor grade, and the presence of lymphovascular or perineural invasion. For those patients who have undergone neoadjuvant therapy, there are multiple histopathologic systems to evaluate treatment response.^164–166 Unlike other solid tumor types, the relationship between neoadjuvant treatment effect and prognosis is less clear in PDAC, possibly owing to the lack of standardization in this practice (reviewed in¹⁶⁷). The most important prognostic factors for patients who have undergone curative-intent resection are the presence of lymph node metastasis and the ratio of positive lymph nodes to total lymph nodes.^{168, 169} Because the number of negative lymph nodes and total lymph node count influence stage-based survival prediction, the College of American Pathologists recommends microscopic evaluation of at least 12 lymph nodes for pancreatectomy specimens.^{159, 170, 171} Some groups advocate for reporting the involvement of specific lymph node groups, such as the hepatic artery lymph nodes, however the prognostic value of this practice remains controversial in light of contradictory data from retrospective studies.^172–174

Adjuvant Chemotherapy

The median OS of patients with localized PDAC who are treated with surgery alone is 11–20 months.^{7, 8, 175} Currently both NCCN and ASCO recommend 6 months of adjuvant systemic chemotherapy for all patients who undergo pancreatectomy.⁸³ Multiple randomized studies demonstrated consistent improvement in median OS and disease-free survival (DFS) using a variety of fluoropyrimidine-based chemotherapies. In general, the selection of chemotherapeutic agents has closely followed those found superior in the metastatic setting. The evolution of adjuvant chemotherapy regimens is outlined in Table 6. More recently the PRODIGE-24 trial demonstrated an improved DFS for modified FOLFIRINOX treatment of nearly 9 months in comparison to gemcitabine (HR 0.58, CI 0.46 to 0.73, p <0.001), establishing it as standard of care in fit patients.⁷ Unfortunately, the combination of gemcitabine with nab-paclitaxel failed to meet its primary endpoint of improved DFS in the APACT trial.¹⁷⁶ Tolerance of adjuvant therapy remains a limitation with patients commonly receiving < 50% of planned dose.^{7, 8, 176, 177} This observation is a reflection of exposure to significant chemotherapy-related toxicity in patients experiencing substantial post-pancreatectomy morbidity. This has led many centers to move toward a total neoadjuvant approach to systemic therapy, discussed below.

Table 6.

Trials evaluating adjuvant chemotherapy for resectable pancreatic cancer

Trial	Chemotherapy	MS
ESPAC-1⁸	5FU vs Obs	MS 21 vs 15.5 mo. P=0.009
CONKO-001¹⁷⁵	Gemcitabine vs Obs	MS 22.8 vs 20.2 mo. P=0.01
ESPAC-3²⁹⁵	Gem vs 5FU/LV	MS 46 vs 39 mo. (ns)
ESPAC-4¹⁷⁷	Gem vs Gem/Cape	MS 28 vs 25.5 mo. P=0.03
PRODIGE-24⁷	mFOLFIRINOX vs Gem	MS 54.4 vs 35 mo. P=0.003
APACT¹⁷⁶^*	Gem/Abraxane vs Gem	MS 40.5 vs 36.2 mo. P=0.045

Open in a new tab

Presented as abstract

Radiation

Although the role of systemic therapy in the adjuvant setting for resected PDAC has been well-validated, the value of adjuvant (chemo)radiation to improve local control, and in turn survival, is strongly debated. The initial results of the 43 patient GITSG study suggested a substantial benefit to chemoradiotherapy, but the failure of subsequent randomized studies to reproduce this result may be attributable to the inclusion of adjuvant chemotherapy in this study, which would later be established as standard of care in the ESPAC-1 trial (Table 7). Although no modern level 1 evidence supports the role of adjuvant radiation in PDAC, multiple large retrospective observational database studies from both SEER and the National Cancer Database (NCDB) demonstrate improved OS in patients who receive adjuvant radiation therapy, particularly patients with pN+ disease or microscopically positive surgical margins.^178–181 Thus, at many centers adjuvant chemoradiation is reserved for patients who exhibit these risk factors after upfront surgical resection and adjuvant chemotherapy. This practice is being tested again in Radiation Therapy Oncology Group (RTOG) 0848, in which patients with no evidence of disease progression following R0 or R1 resections and multiple cycles of adjuvant chemotherapy were randomized to sequential chemoradiation or systemic chemotherapy alone. This closed trial accrued patients between 2009 and 2018, and results of the primary endpoint (OS) and stratified analyses of patient subgroups (by nodal status, margin status, pre-treatment CA 19–9 levels, adjuvant systemic treatment regimens) are eagerly awaited.

Table 7.

Trials evaluating adjuvant radiotherapy for resectable pancreatic cancer

Trial	Comparison	RT dose	Concurrent Chemo	Adjuvant Chemo	OS
GITSG 9173²⁹⁶	ChemoRT + chemo vs Obs	40 Gy (split-course)	Bolus 5FU	5FU	MS 20 vs 11 mo. P=0.03
EORTC 40891²⁹⁷	ChemoRT vs Obs	40 Gy (split-course)	CI 5FU	--	MS 17.1 vs 12.6 mo. P=0.099
ESPAC-1⁸	ChemoRT vs Obs	40 Gy (split-course)	Bolus 5FU	+/− 5FU	MS 16 vs 18 mo. (ns)
RTOG 0848²⁹⁸	ChemoRT vs Obs	50.4 Gy	CI 5FU or capecitabine	Gemcitabine +/− erlotinib	Not reported

Open in a new tab

To allow for standardized adjuvant radiotherapy techniques, RTOG has developed a contouring guideline in the adjuvant setting.¹⁸² Standard high-risk regions include the anastomoses (hepatico-, pancreatico-, and gastro-jejunostomies) and regional paraaortic nodal volumes near the root of the superior mesenteric and celiac arteries. The focal targeting of microscopic residual disease following pancreatectomy is impossible; thus, the contents of adjuvant radiation volumes consist solely of normal tissue, with maximum prescription doses constrained entirely by the tolerance of radiosensitive endoluminal organs (stomach, duodenum and jejunum). An example of the dose distribution in the adjuvant setting is shown in Figure 5. As a result, standard radiation prescriptions in the adjuvant setting consist of daily treatments over the course of 5–6 weeks to a total dose of 50–54 Gy. Given this and other factors, neoadjuvant regimens consisting of targeted pancreatic tumoral delivery of radiation while maximizing normal tissue avoidance is attractive. These approaches will be discussed in more detail below.

Figure 5. — 76 year-old woman with a large PDAC of the pancreatic tail underwent distal pancreatectomy and splenectomy revealing an 8.5 cm invasive carcinoma with a microscopically positive proximal pancreatic margin and 3 of 14 lymph nodes involved by carcinoma. She underwent 6 months of adjuvant FOLFIRINOX followed by adjuvant chemoradiation therapy to a dose of 45-to-50 Gy in 25 fractions using simultaneous integrated boost technique with concurrent capecitabine. Images show standard adjuvant radiation fields and dose distribution using IMRT in axial (A), coronal (B) and sagittal (C) planes. Scale: blue 17 Gy -> red 50 Gy.

Total neoadjuvant therapy

Patients undergoing upfront surgical resection plus adjuvant chemotherapy experience 5-year OS rates of 25–50% due to high rates of systemic relapse.^{7, 8} Based on this premise, neoadjuvant approaches have been tested in resectable PDAC with the following possible advantages: upfront treatment of occult micrometastases, avoiding unnecessary resection for rapidly progressing tumors, improved likelihood of margin negative resection, and improved chemotherapy delivery compared to post-resection adjuvant therapy. In the absence of clear guidelines, current clinical criteria for considering neoadjuvant therapy in resectable PDAC are large primary tumors, high CA 19–9 levels (>1000 U/ml), peri-pancreatic lymph node involvement or equivocal radiographic features suggestive of more advanced disease.¹⁸³ A number of randomized clinical trials are underway to more comprehensively define the role of neoadjuvant compared to adjuvant therapies in resectable patients, with an ultimate goal to most effectively target micrometastatic disease.

Results from the PREOPANC-1 randomized phase III trial in patients with resectable and borderline resectable PDAC suggests patients had longer survival if gemcitabine–based therapy was given in the neoadjuvant versus the adjuvant setting. In the study, 246 patients were randomized to immediate surgery, then 6 cycles of adjuvant gemcitabine (arm A) versus preoperative chemoradiotherapy, consisting of 3 courses of gemcitabine (second combined with radiotherapy to 36 Gy in 15 fractions), followed by surgery and 4 cycles of adjuvant gemcitabine (arm B). Among patients who underwent resection, this study demonstrated an OS benefit from arm B at a median of 35.2 months vs 19.8 months for arm A (p=.029), although the intention-to-treat analysis revealed only a trend toward improved survival.¹⁸⁴ These results are supported by a propensity score-matched analysis of adult patients with resected stage 1 or 2 PDAC from the National Cancer Database (2006–2012).¹⁸⁵ Neoadjuvant therapy followed by resection had a significant survival benefit compared with upfront resection (median survival 26 months v 21 months, p<0.01), further supporting the consideration of offering neoadjuvant therapy to resectable patients. As in the adjuvant setting, the role of radiation in the neoadjuvant approach to resectable disease remains controversial in the absence of phase III trials directly comparing neoadjuvant treatment approaches with or without radiation. A randomized clinical trial (PREOPANC-2) investigating whether neoadjuvant chemotherapy with FOLFIRINOX improves survival compared to neoadjuvant gemcitabine-based chemoradiotherapy with borderline resectable PDAC is underway.

Using a different approach of neoadjuvant systemic therapy, a randomized prospective trial (SWOG-1505) was conducted evaluating perioperative mFOLFIRINOX vs gemcitabine/nab-paclitaxel (12 weeks pre-, and 12 weeks post-surgery in both arms) in patients with resectable PDAC.¹⁸⁶ The trial has completed accrural, and results should be forthcoming in the near future. The added value of this study will be a direct comparison of the performance of these two regimens and perhaps some interpretable data about genomic signatures that may help predict therapeutic responses.

Borderline Resectable

Treatment paradigm

Patients with borderline resectable pancreatic cancer (BRPC) have no evidence of metastatic disease but are less likely to be resected with negative margins due to close proximity to or direct involvement of venous and/or arterial structures. The goal of the treatment approach is to maximize the chance at a margin-negative resection. Although no randomized studies provide specific guidance regarding the optimal treatment paradigm, several prospective phase II studies demonstrate the feasibility and efficacy of a total neoadjuvant approach consisting of neoadjuvant chemotherapy followed by radiation or chemoradiation. Patients who do not show distant progression or local invasion precluding surgery then undergo surgical exploration and resection.

The single arm Alliance A021101 trial enrolled 23 BRPC patients who underwent neoadjuvant mFOLFIRINOX and neoadjuvant CRT (50.4Gy in 28 fractions) prior to definitive surgical resection. Fifteen of 23 patients (68%) underwent resection with an impressive 93% R0 resection rate, and neoadjuvant treatment was not found to preclude resection.¹⁸⁷ Investigators at Massachusetts General Hospital (MGH) conducted a phase II trial evaluating the R0 resection rates in patients receiving 8-cycles of FOLFIRINOX plus short or long-course neoadjuvant chemoradiation with capecitabine. R0 resection rates for patients who underwent resection was 97% with a median PFS of 48.6 months and 2-year OS of 72%; far surpassing previously published historical outcomes.¹⁸⁸ Further, several retrospective studies have shown high R0 resection rates (88–96.7%) with low toxicity, and one study in particular showed improvement in local control and OS with integration of neoadjuvant chemoradiation compared to neoadjuvant chemotherapy alone.^{189, 190} In a modern cohort of BRPC patients at MD Anderson Cancer Center who underwent either neoadjuvant chemotherapy (n=31) or neoadjuvant chemoradiation (n=227), patients who received neoadjuvant chemoradiation had significantly improved R0 resection rates (91% vs 79%), histologically node positive resection rates (3% vs 23%), and locoregional recurrence (16% vs 33%) with a non-significant OS difference (33.6 vs 26.4 months).¹⁹¹ Although early data suggests the importance of integrating both neoadjuvant chemotherapy and chemoradiation into the treatment paradigm for PDAC, large prospective trial data is lacking.

Chemotherapy

The selection of chemotherapeutic agents in BRPC follows the same rationale used in the resectable and metastatic settings. Typically, patients receive ≥ 2–6 months of neoadjuvant chemotherapy before proceeding to radiation, chemoradiation, or directly to surgery, if there is no evidence of distant metastasis. If the full 6 months of chemotherapy are not delivered preoperatively, the balance is commonly offered after surgery. The goals of therapy are to improve resectability by down-staging the primary tumor, reduce micrometastasis, and avoid surgery in patients with aggressive metastatic biology. The only randomized data supporting this approach come from the PREOPANC-1 study, discussed above, which enrolled both resectable and borderline resectable patients and demonstrated improved OS and disease-free survival when gemcitabine and chemoradiation were delivered prior to surgery as compared with upfront resection.¹⁸⁴ NCCN guidelines now recommend neoadjuvant chemotherapy for all patients with borderline resectable PDAC. The ongoing SWOG 1505 and ALLIANCE A021501 trials are evaluating 3–4 months of upfront chemotherapy, whereas a completed phase II study out of MGH utilized a total neoadjuvant approach in patients who did not progress after 4 months of systemic therapy.^{186, 188, 192} The overall treatment goals and approach are similar among most centers that employ neoadjuvant therapy for these patients, although, in the absence of level 1 data, there is little consensus on optimal timing and sequencing of treatment.

Radiation

The goal of neoadjuvant radiation or chemoradiation in BRPC is to increase the number of patients who can receive R0 resections and improve durable locoregional control. Most commonly, standard fractions of 1.8–2 Gy per day are delivered for 5–6 weeks to the tumor, tumor-vessel interface, and the regional lymph node basins. For patients with BRPC, targeted dose-escalation of regions of the tumor abutting the vessel has been shown to be feasible and safe¹⁹³ and may improve surgical resection rates and survival.^{194, 195} Because there is a definable tumor volume, radiation therapy in the neoadjuvant setting may also allow for the safe delivery of hypofractionated regimens consisting of larger radiation doses per day in fewer treatments, a more convenient and less costly treatment consisting of 5 to 15 fractions. This approach was used in the aforementioned PREOPANC-1 study, which included patients with BRPC.¹⁸⁴ The effectiveness of hypofractionation in PDAC has not been compared against conventional fractionation in a randomized fashion.

The utilization of stereotactic body radiation therapy (SBRT), in which a smaller tumor and tumor-vessel interface volume is treated in 5 or fewer fractions, is increasing nationally.¹⁹⁶ Due to small margins and rapid dose fall off, SBRT requires added measures to ensure target delineation, motion management, and daily imaging to verify that the treatment can be delivered safely and effectively (Figure 6). Initial single institution studies of SBRT in locally advanced disease indicated survival that approximated standard of care treatment and very few grade >3 toxicities.¹⁹⁷ Neoadjuvant SBRT was included as an arm in the ongoing prospective ALLIANCE trial A021501 that is evaluating neoadjuvant FOLFIRINOX for patients with BRPC.¹⁹² The American Society for Radiation Oncology (ASTRO) clinical practice guidelines provide a conditional recommendation to treat patients with borderline resectable pancreatic cancer with either conventional or hypofractionated radiotherapy as part of a neoadjuvant treatment regimen including chemotherapy, generally limiting the use of SBRT to patients with smaller tumors located ≥ 1 cm from a gastrointestinal mucosal margin and with no evidence of nodal involvement.¹⁹⁸

Figure 6. — 48 year-old man with a localized PDAC of the uncinate process encasing the SMA. He underwent 3 cycles of gemcitabine/abraxane, then SBRT guided by fiducial markers to 33 Gy in 5 fractions. He had 2 more cycles of chemotherapy, then underwent exploration revealing retraction from SMA and subsequent Whipple procedure. Images show (A) SBRT tumor volume (red) and duodenal planning organ at risk volume (PRV, purple). Dose distribution in axial (B) and coronal (C) planes. Scale: blue 11 Gy -> red 33 Gy.