PAC: Pancreatic ductal Adenocarcinoma; OS: Overall Survival; DFS: Disease Free Survival; CT: Computed Tomography; MRI: Magnetic Resonance Imaging; PET: Positron Emission Tomography

Pancreatic ductal adenocarcinoma (PAC) is an aggressive malignancy associated with significant morbidity and mortality. There is a paucity of effective screening modalities, and so PAC is often diagnosed when it is either locally advanced or widely metastatic. In the United States (US), PAC is the 4th leading cause of death from cancer in men and women, with an estimated 5-year survival of 9% [1,2]. In 2003, an estimated 31,000 people were diagnosed with, and ultimately succumbed to PAC [3]. The incidence continues to rise, and in 2020, an estimated 57,600 Americans (30,400 males, 27,200 females) will be diagnosed with PAC, with approximately 47,050 (81.7%) (24,640 males, 22,410 females) deaths [4]. In a 4 decade SEER database review of pancreatic cancer trends in the US from 1974 to 2014, the incidence rate of PAC increased by a rate of 1.03% per year, and the overall mortality rate increased at a rate of 2.22% per year [5]. In a more recent SEER review of PAC trends between 2000-2014 in the US, the age adjusted incidence rate rose from 9.96/100,000 to 14.70/100,000, and the incidence-based mortality rate rose from 9.96/100,000 to 12.96/100,000 people [6]. Incidence and mortality rates were higher in white patients, and amongst age groups 20-29 years and >80 years [6]. On a global scale, PAC is the 11th most common reported cancer, and has the 3rd highest reported mortality rate of any malignancy with global 5-year survival between 2-10% [4,5,7,8]. By the year 2040, Globocan estimates that the global incidence of PAC will rise to 815,276 newly diagnosed cases, with an estimated 777,423 deaths [7]. While there have been advances in identification of risk factors, diagnostic techniques, and treatment options, this shocking data indicates an unmet need for better preventative guidance with an emphasis on limiting modifiable risk factors, improved modalities for earlier diagnostic detection, and treatment strategies if the medical community hopes to curb current trends.

There are established risk factors that correlate with an increased risk of PAC. These can be grouped into non-modifiable and modifiable. Non-modifiable risk factors can be further classified by host differences and genetic/inheritable alterations (Tables 1 and 2).

Non-modifiable risk factors/germline mutations

Non-modifiable host risk factors include age, gender, ethnicity, and non-O blood group. PAC is more frequently diagnosed in males and the elderly between the ages of 65-74 years, with a median age of diagnosis at 71 years and death at 72 years [9]. It is theorized that a gender specific variation in cigarette smoking, alcohol consumption, and occupational exposures contributes to this observed difference, although it is also possible that underlying genetic alterations play a role [10]. Race is a well-known risk factor. Based off of SEER review in the US, African-Americans are more likely to be diagnosed with PAC and are frequently diagnosed at a more advanced stage than Caucasians, Asians/Pacific Islanders, Hispanic, and American Indians [9], although a recently published SEER database review indicates that the current rise in incidence and mortality is predominately occurring in Caucasians [6]. While a host of variations in modifiable risk factors including obesity, smoking, and alcohol consumption exist between ethnic groups and may perpetuate this observed difference, population-based studies suggest that these findings do not fully explain this dissimilarity, so it is possible that additional underlying genetic and molecular predispositions exist [11]. In addition, blood types A, AB, and B have been found to be independent risk factors associated with PAC in epidemiologic studies, with odds ratios (OR) of 1.38 [95% confidence interval (95% CI), 1.18–1.62], 1.47 (95% CI, 1.07–2.02), and 1.53 (95% CI, 1.21–1.92), respectively [12,13], with another study showing type A blood conferring the highest risk of the three [14].

There are also various inherited germline mutations and familial cancer syndromes that have been identified as having a positive correlation of developing PAC (Table 2). BRCA2 mutations have the highest known association with inherited familial risk of pancreatic cancer, but many other germline mutations have been discovered that confer a risk of developing PAC, including BRCA1, PALB2, ATM, CDKN2A, APC, MLH1, MSH2, MSH6, PMS2, PRSS1 and STK11 [10, 15, 16) (Table 2).

Modifiable risk factors: Given the rising global incidence of PAC, it is essential to identify modifiable risk factors in at-risk patients. Established modifiable risk factors include smoking, type 2 diabetes mellitus (T2DM), alcohol consumption, diet, obesity, and infections (Table 1) [17].

Smoking: Smoking remains the most significantly associated modifiable risk factor in PAC and has been attributed to 20% of diagnosed PACs [10,18,19]. Smokers are estimated to be twice as likely to develop PAC than nonsmokers, and one meta-analysis found the relative risk of developing PAC in current and former smokers to be 1.74 (95% CI 1.61–1.87) and 1.2 (95% CI 1.11–1.29), respectively [18]. In a large European cohort study, those at highest risk of PAC included those with a ≥ 40 pack year smoking history, smoking for ≥ 50 years, and ≥ 30 cigarettes daily [20]. They also found that the odds of developing PAC decreased to that of a never smoker after smoking cessation for ≥ 15 years suggesting a delayed or late carcinogenic effect [20].

Diabetes Mellitus: Diabetes has been linked with developing PAC since the late 1950s [21,22]. It is possible that up to 1% of patients will be diagnosed with PAC following the diagnosis of T2DM [23]. In an Italian based epidemiologic study, the attributable risk of T2DM was 9.7% (95% CI, 5.3-14.1), a significant association in patients diagnosed with PAC [24]. Utilization of insulin may further increase the risk of developing PAC [25] with data showing a greater RR of developing PAC with insulin (OR 5.60, 95% CI 3.75–8.35) compared to a lower risk with oral hypoglycemic agents (OR 0.31, 95% CI 0.14–0.69). While the risk may decline with the duration of T2DM, a significant risk may persist for more than 2 decades after the diagnosis is made (OR 1.30, 95% CI 1.03-1.63) [27]. A recent risk model called the Enriching New-Onset Diabetes for Pancreatic Cancer (ENDPAC) model has been made using retrospective data and found that the change in weight overtime, average range of blood glucose levels, and age at diagnosis of T2DM are associated with a higher likelihood of PAC. The model subcategorizes patients into low risk (<0), intermediate risk (1-2), and high risk (≥ 3). Sensitivity and specificity for a high-risk patient was at least 80%. While this may be a useful risk model to predict the likelihood of developing T2DM, prospective studies are needed to validate this risk model [28].

Diet-resume: Dietary habits may impart a risk of developing PAC. In Northern Italy, lack of adherence to a Mediterranean diet had an attributable risk of 11.9% [24]. Regular intake of foot mutagens found in cooked and well-done red meats, including 2-amino-3,4,8-trimethylimidazo[4,5-f] quinoxaline (DiMeIQx) and benzo(a)pyrene (BaP), have been studied and considered an independent risk for PAC [29,30]. Consumption of ≥ 2-3 sugar-sweetened soft drinks or juices on a weekly basis [31-33] and high alcohol/liquor consumption has also been linked to a higher associated risk of developing PAC [34], with the latter further increased with concomitant smoking of tobacco products [35]. While several studies have found low-to-moderate alcohol intake by itself does not appreciably impart a risk of PAC, it may become a risk with concomitant smoking [34,35].

Body weight: Obesity (defined as a body mass index [BMI] of ≥ 30kg/m2) has a well-established link with various malignancies, most notably esophageal and endometrial carcinoma, though modest risk elevations are also reported with PAC [36,37]. In a 2011 pooled analysis of PAC, patients that were overweight (BMI ≥ 25mg/kg2) or obese in early adulthood had a 54% and 47% higher likelihood of developing PAC, respectively. Young adults that had a steady weight gain with BMI ≥ 10kg/m2 over time compared to a stable weight over time were 40% more likely to develop PAC. In addition, a higher waist to hip ratio had a pooled relative risk of 1.35 (95% CI 1.03-1.78) [38]. Many of these epidemiologic studies primarily consisted of Caucasian patients, so statistical interpretation and extrapolation to other races/ethnicities has not yet been definitively determined [38].

Infection: Infection, particularly gastric colonization with Helicobacter Pylori, has a population attributable risk of 4-25% [39]. This association has been postulated to be related to the augmentation of carcinogenic mutagens related to dietary intake or tobacco product use [10,40]. Alternatively, other studies have not found an associated risk, so additional long term follow up of these patients is warranted [41]. Infection with Hepatitis B virus and Hepatitis C virus have also been correlated with a positive risk, with meta-analysis suggesting a relative risk ranging from 1.2-3.8 and 1.2, respectively [42-44]. These analyses were limited by a small number of available observational studies in the literature, and a large proportion of patients were of Asian ethnicity, which may limit the capacity by which these results can be inferred to the global population.

PAC risk reduction

Clinicians should be mindful of identifying modifiable risk factors in their patients. Encouraging long term smoking cessation, reduction in dietary intake of red meats, sweetened drinks, and alcohol consumption, reduction of risk factors that may precipitate T2DM or steady weight gain, and high risk practices that may lead to Hepatitis B or C infections may be a possible strategy to reduce this trend over time. While still under investigation and not clearly understood, large epidemiologic studies have found other factors that may reduce the risk of developing PAC or otherwise prolong the survival in patients already diagnosed with PAC. There have been several studies reporting an inverse relationship between respiratory allergies such as hay fever and allergies to plants, animals, or pollen and the risk of developing PAC [45,46]. Asthma has been suggested to have a lower risk of PAC, particularly in patients with asthma for ≥ 17 years [47]. Statin (HMG-coreductase inhibitors) therapy may also play a protective role. One meta-analysis indicated that ever-use of statin reduced the risk of developing PAC by 34% (OR, 0.66; 95% CI, 0.47-0.92), although sex-stratified analysis indicated this finding was only significant in males [48]. Statins are known to have anti-inflammatory and immunomodulatory effects [49]. In addition, they may confer anti-neoplastic properties through the inhibition of key proteins directly involved in tumor proliferation and metastasis [50,51]. In a SEER review, use of a statin was associated with improved survival in stage I-II PAC (Hazard ratio (HR) = 0.79, 95% confidence interval (CI) 0.67, 0.93) [50]. The oral diabetic medication metformin may also offer a benefit. In obese mouse models, metformin inhibited pancreatic tumor growth [52]. In patients with T2DM and PAC, treatment with metformin improved survival compared to diabetics on other oral hypoglycemic regimens [53]. Alterations in diet, including frequent consumption of citrus fruits and flavonoids [54,55], as well as folate supplementation have also been suggested to have a modest protective role [56]. Further studies are needed to determine whether these factors are preventative or provide a survival benefit in patients at-risk for, or diagnosed with PAC (Table 1).

PAC carcinogenesis follows a series of stepwise mutations resulting in the formation of pre-neoplastic lesions with eventual transformation into an invasive malignancy [57]. It can be grouped into three broad yet distinct phases: Acquisition of driver mutations, clonal expansion into a multicellular neoplasm, and introduction of the neoplastic cells into local and distant microenvironments [58,59]. On average, there are 63 genetic alterations per PAC, with Kirsten rat sarcoma viral oncogene homolog (KRAS) activation occurring in 90% of cases [60,61]. KRAS activation leads to uncontrolled cellular proliferation resulting in exponential clonal expansion. This often results in additional mutations in progeny cells, including the inactivation of tumor suppressor genes cyclin-dependent kinase (CDK) inhibitor 2A, TP53, and Mothers Against Decapentaplegic Homolog 4 (SMAD4), the latter being a late carcinogenic event and associated with the development of metastatic disease [62]. 
In general, mutational analysis suggests metastasis is a late event in PAC [63]. Upon acquisition of the initial driver mutation, it may take up to 2 decades for the development of metastatic disease. One study found that it took 11.7 years from the time the initial driver mutation was acquired to the formation of the malignant parental cell. It required an additional 6.8 years thereafter for the index lesion to form [63]. While this is the typical pathogenesis by which stromal invasion and eventual disease metastasis occurs, cancer cells do not always follow a pre-conceived dogma [59,64]. In fact, mouse models have demonstrated that metastatic dissemination can occur early in carcinogenesis, and although most of these cells do not survive, very few cells are actually needed to establish a metastatic foothold [65,66]. Therefore, it is possible that these metastatic cells can become fully malignant prior to the primary tumor site invading into the pancreatic stroma. It is known that some malignancies have the capability to disseminate early in the disease course and may occur years before the primary malignancy is detected. This has been documented in malignant melanomas, as well as cancers involving the breast, prostate, lungs, colon, and kidneys [67]. This has also rarely been reported in PAC. The authors of the present review published a case in which an adenocarcinoma of unknown primary was diagnosed in the esophagus months before the primary pancreatic body lesion manifested radiographically, suggestive of a profound metastatic potential, and further adding to the diagnostic challenge associated with PAC [59]. While PAC frequently follows a predictable time course, the potential ability of PAC to metastasize early carries profound implications on the treatment selection, outcome, and mortality of these patients. Since PAC screening is only recommended for a very specific subset of patients (see section “PAC Screening”), clinicians should have a low index of suspicion when patients present with concerning symptoms to institute a diagnostic workup for PAC, including when the malignancy is of unknown origin. 

PAC accounts for 90% of all pancreatic cancers, with the majority arising in the pancreatic head, and one-third occurring in the pancreatic body or tail [68]. While presenting symptomatology depends on the anatomic location of the primary pancreatic lesion and distant metastatic sites, patients have a range of presenting features including abdominal discomfort, pain, bloating, dyspepsia, nausea, weight loss, and jaundice. 
Diagnosis of PAC requires imaging and ultimately tissue sampling. A range of radiographic imaging techniques are available, including computed tomography (CT), magnetic resonance imaging (MRI), Positron Emission Tomography (PET), and endoscopic ultrasonography (EUS) (Table 3). 

Computed tomography
CT is the most validated, widely available, and most inexpensive imaging modality [69]. PAC typically manifests on CT as an ill-defined mass that has poor enhancement compared to adjacent pancreatic parenchyma, and typically appears as a hypodense lesion, although isodense lesions may also be seen [69]. The combined sensitivity and specificity of PAC is 89% and 90% respectively [70], although these are subject to changed based on various CT findings. When a mass <2 cm is present, the sensitivity drops to 67-77%, and additional imaging techniques should be considered [71-73]. 

Magnetic resonance imaging
On MRI, PACs are commonly hypointense on pre-contrast T1-weighted images and hypointense or isointense on post-contrast T1-weighted images [74]. Sensitivity and specificity approach 89%, and while more expensive and not as widely available as CTs, there may be more benefit for characterization of smaller lesions <2 cm and better delineation of hepatic metastasis with undetermined malignant potential on CT [70,75]. 

Positron emission tomography
Eighteen-fluorodeoxyglucose ([18]FDG)-PET and PET/CT is not routinely recommended as an initial diagnostic modality. While the pooled sensitivity and specificity of PET/CT is higher than CT alone, PET/CT performs similarly to CT alone and adds no additional initial diagnostic benefit [69,76-78]. However, data suggests that using PET/CT when assessing tumor response to chemoradiation confers a benefit when monitoring for treatment response or progressive disease [77]. 

Endoscopic ultrasonography
EUS involves an upper gastrointestinal endoscopy examination under conscious sedation with an echoendoscopy. In general, this is a well-tolerated procedure with the added benefit of having the ability to extract image guided tissue sampling via fine needle aspiration (FNA). This is limited by operator expertise, but studies indicate diagnostic accuracy markedly improves over time likely reflective of operator proficiency with repetition [69]. Sensitivity and specificity for identifying PAC via EUS are >85% and 96%, respectively [60,79]. When EUS combined with FNA, pooled specificity was 95.8% (95% CI, 94.6-96.7), with a positive likelihood ratio of 15.2 (95% CI, 8.5-27.3), and negative likelihood ratio of 0.17 (95% CI, 0.13-0.21) [79]. 

At the time of diagnosis, approximately 80-90% of patients have either locally advanced lesions or metastatic disease that are not amenable to surgical resection [80,81]. In the 10% of patients that are candidates for operative resection, nearly 70% are found to have pathologic evidence of disease [17,82]. Metastasis to virtually every organ system has been reported, with the most common sites being the liver, lungs, abdominal lymph nodes, peritoneum, thoracic lymph nodes, and adrenal glands [83]. In rare cases, evidence of metastatic disease may be found prior to radiographic evidence of the primary lesion [59]. While there is ample evidence of the benefit of performing screening colonoscopy, mammography, and Papanicolaou smear in the general population, wide-spread screening for PAC is not realistic given the low population-based prevalence, irrespective of rising trends. However, it has been suggested that a screening system may be feasible for individuals identified as being at high risk (>5% life time risk of developing PAC). These situations include family history of PAC and/or inherited germline mutations associated with elevated risks of PAC [84,85]. In patients with hereditary pancreatitis, a consensus conference agreed that screening for PAC should be offered starting at age 40 years [86]. For patients with familial PAC, screening should be offered at age 50 years [85]. Given the younger median age of PAC diagnosis in patients with Peutz-Jeghers syndrome [87], screening should be offered starting at 30 years of age [88]. 

Staging

Non-metastatic PAC is subdivided into resectable, borderline resectable, and locally advanced. The crucial features that need to be radiographically assessed are the contact of the tumor with the superior mesenteric vein (SMV) or portal vein (PV) as venous structures, and the superior mesenteric artery (SMA), common hepatic artery (CHA), and celiac axis (CA) as major surrounding vasculature [89]. Several criteria have been proposed over the past 12 years to aid in defining resectability status, summarized in Table 4. Commonly used criteria include the NCCN guidelines (updated in 2019) [90], MD Anderson Cancer Center (MDACC) guidelines [91,92], the Americas HepatoPancreato-Biliary Association/Society of Surgical Oncology/Society for Surgery of the Alimentary Tract 2009 consensus recommendations (AHPBA/SSAT/SSO) expert consensus guidelines [93], and the International Study Group of Pancreatic Surgery (ISGPS) criteria [94] (Table 4). While these criteria share commonalities, there remains no uniform or standardized criteria. Furthermore, at a consensus meeting in 2016, biological and functional risk factors were added to resectability criteria. Biological factors include elevated Carbohydrate Antigen (CA) 19.9 levels above 500 units/mL, regional lymph node metastases, and suspicion of distant metastases without the possibility for pathological proof. Functional factors include PS (ECOG>3) and comorbidities [89,90]. These additional recommendations have resulted in fewer patients classified with resectable disease [89,90].

Surgery

Surgical resection remains the only curative option for PAC. At the time of diagnosis, patients with TNM stages IA-IIB may be considered for resection [95]. Unfortunately, only 10-20% of patients meet this surgically acceptable category at diagnosis (Figure 1). Pancreatico-duodenectomy (Whipple’s procedure), distal pancreatectomy, or total pancreatectomy are the various surgical approaches for PAC. The pathologic goal of surgical resection is to have microscopic margins without evidence of cancer (R0), which improves disease free survival (DFS) and OS compared to tumor resection with microscopic evidence of retained tumor margins (R1) [96]. Prognostic factors used to predict a poor postoperative surgical outcome include tumor size [97,98], vascular invasion [97], number of units of packed red blood cells transfused during surgery [97,98], regional lymph node metastasis [97-99], tumor grade [98], aneuploid karyotype [100], anatomic location [101], the presence of KRAS/TP53 mutations [97,100], and inability to achieve R0 resection [96-99] (Table 5). Contraindications to a surgical intervention include distant metastatic disease, gross vascular tumor invasion into the mesenteric root or celiac axis (stage III-IV) (Table 4), and poor performance status (PS) [95,102].

Figure 1: Treatment outcomes in patients with PAC. OS (overall survival) is reported in months. NAT: Neoadjuvant Therapy.

Neoadjuvant therapy

The utilization of neoadjuvant therapy (NAT) is being studied in ongoing trials. For locally advanced or borderline resectable disease, NAT may provide a survival benefit. In patients who have surgical exploration and are initially deemed unresectable, initiation of NAT was associated with improved TNM staging, fewer positive lymph nodes, improved survival (24 vs.13 months, P=0.044), and improved resection margins on re-exploration [103]. However, utilization of NAT for resectable PAC remains an emerging notion still being studied. While it is a theoretically reasonable approach to reduce the size of the primary tumor, reduce the high likelihood of preoperative micrometastasis, and increase the likelihood of an R0 resection margins, up to 30% of patients may develop significant AE or disease progression leading to delay or inability to tolerate surgical resection [57,104-106]. Furthermore, it has been difficult to draw sound conclusions due to the inability to achieve target recruitment numbers in randomized trials [107]. In patients with resectable PAC at the time of diagnosis, data is conflicting. In a 2016 meta-analysis of resectable PAC with NAT prior to tumor resection, there was an improvement in median OS (26 months vs. 21 months, P<0.01; hazard ratio, 0.72; 95% CI, 0.68 to 0.78), T stage, lower number of involved regional lymph nodes, and lower positive resection margins. However, the study was limited by the inability to discern the number of patients that received NAT but did not proceed with surgery [108]. In a more recent meta-analysis, Verstaijne and colleagues reported a significant survival benefit when patients with resectable or borderline resectable disease were treated with NAT compared to upfront surgery in an analysis of intention to treat populations. The OS for the total intention to treat population was 18.8 in the NAT group vs 14.8 months in the upfront surgery group, but 17.8% of patients in the NAT group were not able to undergo surgery, primarily due to reported disease progression. For the NAT patients that underwent surgery, the OS compared to the upfront surgery group was 26.1 vs 14.8 months, respectively. The findings also reported a greater R0 resection rate and lower rate of lymph node metastasis [109]. Alternatively, in the first randomized clinical trial assessing chemoradiotherapy (gemcitabine and cisplatin) in the NAT setting, Golcher and colleagues failed to find a statistically significant survival benefit, and the trial was terminated early [110].

Other meta-analyses have reported similar findings. In one analysis assessing per protocol and intention to treat populations, the risk for overall mortality in patients with resectable PAC was lower with NAT compared to upfront surgery in resectable PAC (HR 0.80, 95% CI 0.70–0.92, P < 0.01), with improved R0 resection rates (83.7% in neoadjuvant vs 76.8% in upfront surgery), and less frequently occurring lymph node metastasis (45.0% in neoadjuvant therapy vs. 69.3% in US) [111]. However, in the intention to treat population, these results were not statistically significant and attributed to higher presurgical attrition rate in the neoadjuvant group (36.3% versus 17.3%) [111]. In another recent study, NAT for resectable PAC indicated an increased R0 resection rate (OR=1.89; 95% CI=1.26–2.83) and a reduced positive lymph node rate (OR=0.34; 95% CI=0.31–0.37), but no significant difference in overall survival (OS) time (HR=0.91; 95% CI=0.79–1.05) [111]. Zhan and colleagues assessed NAT vs upfront surgery in resectable, borderline resectable, and locally advanced disease. In patients with resectable disease, 73.0% went on to have surgery, whereas 40.2% of patients with previously borderline resectable or advanced surgery were able to have resection. While survival did not significantly differ for resectable disease, patients with borderline resectable or locally advanced disease may have a survival benefit when NAT is given prior to surgery [104].

In the recently reported results of the randomized, phase III PREOPANC trial, 246 eligible patients with either resectable or borderline resectable disease were randomized into NAT with 3 courses of gemcitabine, the second combined with 15 × 2.4 gray radiotherapy, followed by surgery and 4 courses of adjuvant gemcitabine or to immediate surgery and 6 courses of adjuvant gemcitabine (Table 6). Results from this trial showed that patients randomized into the NAT group had improved R0 resection rates, lower proportion of involvement regional lymph nodes and vascular invasion, better disease free survival, and improved survival if patients that received NAT also had adjuvant chemotherapy (35.2 vs 19.8 months; P=0.029). However, the intention to treat population did not have a significant OS benefit [113]. Further clinical trials including the NEOPA trial [114] assessing the impact of neoadjuvant chemoradiation in PAC of the pancreatic head, NEONAX trial [115] assessing intensified perioperative treatment with nab-paclitaxel plus gemcitabine in resectable PAC), the randomized phase II/III NEPAFOX trial [116] assessing NAT with FOLFIRINOX, phase III NorPACT- 1 randomized trial [117] assessing neoadjuvant FOLFIRINOX are ongoing, and phase 2/3 randomized Prep-02/JSAP05 [118] assessing gemcitabine/S1 for resectable/borderline resectable PAC are ongoing.

At this time, the 2019 National Comprehensive Cancer Network (NCCN) guidelines [90] recommend upfront surgery followed by adjuvant therapy for resectable PAC, but recommend the consideration of NAT in patients with imaging findings suspicious of advanced or metastatic disease, significantly elevated CA 19-9, large primary tumors or evidence of regional lymph node involvement, excessive weight loss, and significant pain [90,119]. NAT with either FOLFIRINOX or gemcitabine/nab-paclitaxel is recommended for patients with borderline resectable disease, with or without subsequent chemoradiotherapy [90,119]. The 2019 American Society of Oncology (ASCO) Clinical Practice Guidelines recommends upfront surgical resection for PAC without radiographic evidence of metastasis, with no tumor invasion of surrounding vasculature, CA 19.9 level suggestive of curable disease, and a PS that supports surgical intervention [120]. NAT is reserved for patients who do not meet these criteria, but no specific treatment recommendations are offered [119,120].

Adjuvant chemotherapy

Even with patients that are candidates for curative surgical resection, OS and DFS remain poor, indicating a significant need to find additional strategies to improve outcomes. There is ample data supporting superior outcomes in OS and DFS associated with adjuvant chemotherapy following surgical resection. In the landmark Charité Onkologie 001 (CONKO 001) phase III randomized controlled trial, adjuvant gemcitabine alone for 6 months after macroscopic resection versus no adjuvant therapy doubled 5-year OS (20.7 vs 10.4 months) and DFS (13.4 vs 6.7 months) (Table 6) [121]. In the final analysis presented at American Society of Oncology in 2016, survival at 3 years and 5 years for adjuvant gemcitabine vs surgery alone was 36.5% and 21.0% vs. 19.5% and 9.0%, respectively [122]. Despite these results, median OS while statistically significant was 22.8 months with gemcitabine vs. 20 months with operative resection alone [122]. Adding to these findings, the phase III randomized multicenter ESPAC 4 trial found further OS and DFS benefit by adding capecitabine to gemcitabine. Results from this study found that median OS for capecitabine + gemcitabine vs. gemcitabine alone was statistically improved at 28.0 vs 25.5 months, respectively, with the caveat that a larger proportion of patients on the combined regimen reported grade 3-4 adverse events (Table 6) [123]. In 2018, the phase III randomized PRODIGE 24/CCTG PA.6 trial compared gemcitabine with modified oxaliplatin, leucovorin (LV), irinotecan, and 5-Flourouracil (5FU) (mFOLFIRINOX) in R0/R1 PAC resection patients with a PS ≤1 for 6 months. Patients that received mFOLFIRINOX were found to have a significantly longer DFS and OS of 21.6 and 54.4 months, respectively, compared to gemcitabine alone with a DFS and OS of 12.8 and 34.8 months, respectively. A larger proportion of grade 3 and 4 AE occurred with mFOLFIRINOX, but it remains the adjuvant therapy standard of care in patients with good PS following surgical resection [124,125]. In the 2019 phase 3 APACT trial, the use of adjuvant nab-paclitaxel and gemcitabine was compared to gemcitabine alone for 6 months following surgical resection. OS and DFS in the nab-paclitaxel and gemcitabine group compared to gemcitabine alone was 40.5 vs 36.2 months (HR, 0.82; 95% CI, 0.680-0.996; P=0.045), and 16.6 vs. 13.7 months ((HR, 0.82; 95% CI, 0.694 - 0.965; nominal P=0.0168), respectively [126] (Table 6).

Metastatic disease

Treatment for metastatic disease involves chemotherapy with palliative intent and is based strongly off the patient’s PS and symptom control given the high risk for systemic toxicity. With current standard of care systemic chemotherapy, median PFS is approximately 6 months and less than 10% are alive at 5 years (Figure 1) [10,127]. In general, FOLFIRINOX is the preferred first line regimen. In the 2011 PRODIGE4/ ACCORD11 trial comparing between FOLFIRINOX and gemcitabine, patients treated with FOLFIRINOX had a significantly longer survival (11.1 months) than Gemcitabine (6.8 months) (HR: 0.57; 95% confidence interval [CI], 0.45 to 0.73; P<0.001), improved progression free survival (PFS) of 6.4 vs. 3.3 months, respectively (HR: 0.47; 95% CI, 0.37 to 0.59; P<0.001), and a better objective response rate (ORR) of 31.6% compared to 9.4%, respectively (Table 6). While more patients in the FOLFIRINOX group had significant toxicities reported, fewer reported a degradation in the quality of life (HR: 0.47; 95% CI, 0.30 to 0.70; P<0.001) [127]. In a recent phase 2 trial, the safety and efficacy of mFOLFIRINOX (no 5FU bolus) was assessed. In this group of 69 patients in Japan, median OS was 11.2 months (95%CI 9.0-), with a median PFS of 5.5 months (95% CI 4.1-6.7). The ORR was 37.7% (95% CI 26.3-50.2). There were 47.8% of patients with a grade 3 or higher AE, with 1 treatment related death. The authors concluded mFOLFIRINOX had an improved safety profile with similar median survival [128].

In patients that previously received first-line treatment with FOLFIRINOX and have an Eastern Cooperative Oncology Group PS of 0-1 with disease progression, the 2019 ASCO clinical practice guidelines recommend second line therapy with gemcitabine plus nanoparticle albumin-bound (nab)-paclitaxel [120]. In the phase 3 MPACT trial assessing the combination regimen, the median OS for nab-paclitaxel and gemcitabine vs gemcitabine alone was 8.6 months and 6.7 months, respectively (0.72; 95% confidence interval [CI], 0.62 to 0.83; P<0.001) (Table 6). The median PFS was 5.5 months for nab-paclitaxel and gemcitabine and 3.3 months for gemcitabine alone (HR: 0.69; 95% CI, 0.58 to 0.82; P<0.001) [129]. In the updated survival analysis, 1, 2, and 3 year survival was superior in the gemcitabine and Nab-paclitaxel (35%, 10%, 4%, respectively) treatment arm compared to gemcitabine alone (22%, 5.0%, and 0%, respectively) [130]. Whether FOLFIRINOX or Nab-paclitaxel and gemcitabine should be used as first line therapy remains uncertain. In a recent retrospective cohort study, patients that received FOLFIRINOX first line were younger and had a better PS. However, the median OS (10.7 vs 12.1 mo; P = 0.157), PFS (8.0 vs 8.4 mo; P = 0.134), and ORR (33.7% vs 46.9%; P=0.067) were not significantly different [131]. In a recent propensity score matched analysis comparing FOLFIRINOX and gemcitabine/Nab-Paclitaxel, median OS was significantly better in the FOLFIRNOX group (14 vs 9 months respectively, p=0.008), with the authors concluding that gemcitabine/Nab-Paclitaxel following FOLFIRINOX treatment failure is a reasonable approach [132]. In another retrospective study of 83 patients treated with either first line FOLFIRINOX followed by Gemcitabine/Nab-Paclitaxel or vice-versa, there was no significant difference in OS (13.7 vs. 13.8 months, p=0.9), leading to the authors to conclude either sequence leads to a similar outcome in OS [133]. In patients that received mFOLFIRINOX compared to Gemcitabine/Nab-Paclitaxel, patients in the Gemcitabine/Nab-Paclitaxel groups had significantly longer 1 year median OS compared to the mFOLFIRINOX (67% vs 44%, p=0.0006) and better ORR (39% vs 27%, p=0.02), suggesting Gemcitabine/Nab-Paclitaxel had favorable efficacy and survival compared to mFOLFIRINOX [134]. While FOLFIRINOX is the generally accepted first line therapy, to truly compare the efficacy and survival benefit between FOLFIRINOX and Nab-Paclitaxel/Gemcitabine, additional prospective studies are needed.

In patients previously treated with a gemcitabine based chemotherapy who have progressive disease and a PS of 0-1, it is recommended patients be offered nanoliposomal irinotecan (nal-IRI) with either 5FU or oxaliplatin [135]. These recommendations are based on significantly improved median overall survival compared with 5-FU/LV alone (6.1 vs 4.2 months; HR: 0.67, 95% CI 0·49–0·92; p=0.012) in the global phase 3 NAPOLI-1 trial (Table 6). In the final OS survival analysis, estimated one-year overall survival rates were 26% with nal-IRI+5-FU/LV and 16% with 5-FU/LV [136]. As is frequently seen in the real-world clinical setting, utilization of this regimen may be more used in a sicker population with poorer PS and comorbidity profile. In the initial results of a multicenter, retrospective real-world chart review, 26 patients given liposomal irinotecan were older (median age, 68 years), sicker (81% had an ECOG PS ≥ 1), and 65% had 2 or more lines of previous chemotherapy, but had a similar median OS of 4.9 months [137]. In areas that lack availability of 5FU and nan-IRI, fluorouracil plus irinotecan or fluorouracil plus oxaliplatin may be offered [135], recommendations that stem from improved OS (5.9 months, HR: 0.66; 95% CI, 0.48 to 0.91; log-rank P = .010) found in the CONKO-003 trial [138]. In patients with an ECOG PS of 2 or with significant comorbidities, gemcitabine or fluorouracil monotherapy can be considered [135].

At the time of diagnosis, the NCCN recommends that patients with PAC should undergo germline testing using a comprehensive gene panel to assess for hereditary cancers [90]. The 4.6% of patients that are found to have a loss of function mutation in BRCA1/2 may benefit from the oral poly (ADP-ribose) polymerase (PARP) inhibitor Olaparib, which has been shown to have activity as a monotherapy [90,139-141]. In a 2015 multicenter phase 2 study of Olaparib in various malignancies with BRCA1/2 mutation, the tumor response rate for the 23 patients included with PAC at ≥ 8 weeks was 35% (95% CI, 16.4 to 57.3) [140]. In the recent phase 3 POLO (Pancreas Cancer Olaparib Ongoing) trial assessing efficacy of Olaparib in patients who previously received a first line platinum base chemotherapy regimen without disease progression, patients in the treatment arm had significantly improved PFS compared to placebo (7.4 months vs. 3.8 months; HR 0.53; 95% CI 0.35-0.82; P=0.004) (Table 6). At 2 years, 22.1% of patients in the Olaparib group did not have progression or death compared to 9.6% in the placebo arm [141]. As a result, the NCCN recommends Olaparib as maintenance therapy in patients with metastatic PAC, a germline BRCA1/2 mutation, no evidence of disease progression in the first 16 weeks of first-line platinum-based chemotherapy, and a good PS [90].

Immune check point inhibitor therapy

Immune checkpoint inhibitor (ICI) is an emerging therapeutic option for a broad range of malignancies. This is particularly true for patients with deficiencies in mismatch repair (dMMR) or confirmed microsatellite instability (MSI). In 2017, the U.S. Food and Drug Administration approved pembrolizumab for dMMR that can lead to high levels of MSI, regardless of disease site [142]. As such, routine testing for dMMR or MSI is recommended [135]. One study found that in PAC, only 0.8% (7/833) of patients had dMMR and high MSI [142]. Of these, 4 were treated with checkpoint inhibitor therapy, and all 4 benefitted from treatment (1 complete response, 2 partial responses, 1 stable disease) [143]. In general, the currently available studies that utilize ICIs in PAC comprise relatively small sample sizes, which make it difficult to draw definitive conclusions on recommended use. A recent review article by Henriksen and colleagues provides a comprehensive report on the available ICI used in PAC [144].

There have been several trials looking at both ICI monotherapy and combination therapy in the adjuvant setting and in patients with either locally advanced or metastatic disease. In a phase II study using monotherapy with the CTLA-4 inhibitor ipilimumab in 27 patients with advanced pancreatic cancer, no response was noted, with the exception of a delayed response in one patient with regression of the primary tumor and the metastatic sites. Median OS in this small cohort was 4.5 months [145]. In a randomized phase II trial, 65 patients with metastatic PAC pre-treated with chemotherapy that were given durvalumab monotherapy with or without tremelimumab had a modest improvement in OS of 3.1 vs 3.3 months respectively, and the same PFS of 1.5 months [146].

There are several treatment combination regimens that have been studied in the neoadjuvant and palliative settings. In a phase Ib/II study comparing neoadjuvant chemoradiation therapy alone vs in combination with pembrolizumab, the initial results of the first 22 patients have been reported. Of the 14 enrolled in the pembrolizumab treatment arm, 10 (71.4%) underwent surgical resection, and the combination was safely tolerated [147]. In a prospective pilot study assessing durvalumab or durvalumab plus tremelimumab in combination with stereotactic body radiation therapy in chemorefractory, metastatic PAC, 24 patients were enrolled. While 21 (87.5%) had stable disease, none of the patients had an objective response [148]. In the Canadian Cancer Trials PA.7 phase II trial, 11 patients with previously untreated metastatic PAC were treated with Gemcitabine, nab-Paclitaxel, durvalumab, and tremelimumab. In the safety analysis group, 8 (72.7%) patients had a partial response, with a median response duration of 7.4 months, and 6 month survival rate of 80% (95% C.I 40.9%-94.6%). The randomized phase II trial is ongoing and international phase III study is planned [149]. While there have been other reported trials and ongoing studies assessing ICI in various combination setting, low sample size may preclude definitive conclusions.

PAC remains an aggressive malignancy with diagnosis often late in the disease course and poor outcomes, even when curative surgical resection is plausible. Improved understanding in modifiable risk factors over the past decade and educating the medical community of these risks so that appropriate recommendations may be given to patients may provide an outlet by which the long term rising incidence of PAC is curbed. In patients diagnosed with PAC, OS continues to slowly rise with strides in surgical resection, adjuvant chemotherapy, and palliative chemotherapy. Additional research is ongoing for neoadjuvant therapy and additional lines of therapy including checkpoint inhibitors for metastatic disease.