Neoadjuvant therapy for pancreatic cancer: an ongoing debate

2016 Gastrointestinal Cancers Symposium summary of plenary session, 21–23 January 2016, San Francisco, CA, USA

The 2016 Gastrointestinal Cancers Symposium pancreatic cancer session focused on current controversies in the management of pancreatic cancer and novel treatment approaches that hold promise for the future. The session began with a thought-provoking debate of the pros and cons of neoadjuvant therapy for resectable pancreatic cancer. Robert Wolff MD from MD Anderson Cancer Center emphasized that pancreatic cancer is both a locally invasive disease and a systemic disease, and reviewed adjuvant therapy data that demonstrate little to no improvement in outcome over the last 25 years. He pointed out that results from adjuvant trials only apply to the 50–60% of surgical patients who are not found to have occult metastatic disease [Herman et al. 2008; Parmar et al. 2014], undergo R0-1 resections and are medically fit to undergo postoperative treatment. The locally invasive nature of resectable pancreatic cancer was emphasized using surgical data demonstrating R1 resection rates ranging from 17% to 42% [Richter et al. 2003; Takai et al. 2003; Winter et al. 2006; Oettle et al. 2007; Regine et al. 2008; Neoptolemos et al. 2010] associated with inferior median survival rates compared with R0 resections [Richter et al. 2003; Fatima et al. 2010; Neoptolemos et al. 2010]. He also reviewed autopsy data showing 75–85% local and over 80% distant recurrence rates in patients who underwent curative resection [Hishinuma et al. 2006; Iacobuzio-Donahue et al. 2009]. He emphasized that delay in systemic treatment results in rapid radiology progression of initially occult metastatic disease [Glant et al. 2009] and showed data revealing that 16% of patients thought to have resectable disease develop radiographic metastatic disease at the end of neoadjuvant therapy [Evans et al. 1992; Hoffman et al. 1998; Pisters et al. 1998; White et al. 2001; Pisters et al. 2002]. He argued that neoadjuvant therapy provides early treatment of micrometastatic disease, avoids surgery in patients with rapidly progressive disease and provides an opportunity to observe patient tolerance to therapy. Dr Wolff reviewed data from MD Anderson using the neoadjuvant approach resulted in 66% of patients completing all prescribed therapy resulting in a lower R1 resection rate (8%) and improved median survival rates (31–34 months for resected patients). [Evans et al. 2008; Varadhachary et al. 2008]. He proposed that the neoadjuvant approach provides a platform for clinical investigation that will allow more patients to receive new systemic therapies which may result in improvements in outcome.

In contrast, Nipun Merchant MD from University of Miami Medical Center delved into the purported benefits of neoadjuvant therapy for potentially resectable pancreatic cancer. He argued that the neoadjuvant approach may prevent curative treatment in initially resectable pancreatic cancer by allowing progression of disease in patients whose tumors do not respond to therapy, or preventing potentially curable surgery if the neoadjuvant treatment is too toxic. He reviewed data for neoadjuvant therapy from multiple institutions which demonstrated that the percentage of patients who undergo resection ranges from 38% to 85%, resulting in median survival rates on average of 20 months [Yeung et al. 1993; Staley et al. 1996; Spitz et al. 1997; Pisters et al. 1998, 2002; Talamonti et al. 2006]. He emphasized that survival rates following neoadjuvant therapy in these studies are similar to survival demonstrated in multiple phase III randomized trials of adjuvant therapy for pancreatic cancer [Kalser and Ellenberg, 1985; Klinkenbijl et al. 1999; Oettle et al. 2007; Neoptolemos et al. 2010]. He also pointed out that the definition of resectability varies widely and many neoadjuvant studies included patients with resectable and borderline resectable tumors and cited the data from a prospective phase II study of neoadjuvant therapy that included only patients with resectable pancreatic cancer whose outcome did not appear to improve compared with adjuvant therapy trials, with 19-month overall survival and 9-month disease-free survival [Heinrich et al. 2008a]. He reviewed data from the only trial attempting to randomize patients with resectable pancreatic cancer to neoadjuvant versus adjuvant therapy was stopped after 73 of a planned 254 patients due to slow accrual, with 66 eligible for analysis. In this trial the R0 and survival rates were not significantly different, failing to demonstrate an advantage for the neoadjuvant approach once again [Golcher et al. 2015]. Dr Merchant cited additional data from MD Anderson demonstrating that radiographic downstaging is rare after neoadjuvant therapy and does not appear to correlate with survival [Katz et al. 2012], but an objective response metric is currently lacking and consensus of the definition of resectability and staging algorithms is necessary across institutions to make accurate observations regarding the effects of neoadjuvant therapy on surgical resectability. In that regard, Dr Merchant shared data demonstrating that there appears to be a robust histologic response following neoadjuvant therapy, but this also does not appear to correlate with improvements in disease-free or overall survival [Heinrich et al. 2008a]. In contrast, an apparent survival benefit has been reported for patients who achieve a pathologic complete response following neoadjuvant therapy, however this is a rare occurrence (2.5%) [Zhao et al. 2012]. Finally, Dr Merchant also refuted that the neoadjuvant approach improves delivery of therapy to all patients. He cited data from 2 modern randomized trials (RTOG 97-04 and ESPAC-3) in which approximately 90% of patients were able to receive the prescribed adjuvant therapy following surgery [Regine et al. 2008; Neoptolemos et al. 2012]. These results were compared with MD Anderson data showing that approximately 20% of patients with potentially resectable pancreatic cancer who received neoadjuvant therapy did not undergo planned surgery.

The pros and cons of neoadjuvant treatment for resectable pancreatic cancer were well presented, and it is clear that further evidence will be necessary to recommend one approach over another. Dr Joseph Herman MD followed this discussion with data from Johns Hopkins University using a novel neoadjuvant approach incorporating induction chemotherapy followed by stereotactic body radiotherapy (SBRT) as an aggressive approach to get more patients to potentially curative resection. He also stressed the need for consensus of the definition of resectability and staging algorithms for patient selection. The study from Hopkins utilized current chemotherapy regimens used to treat metastatic disease (gemcitabine based and FOLFIRINOX) and SBRT, which precisely targeted high-dose short-course radiation to enhance local response prior to surgery. Dr Herman pointed out that this radiation approach appears to be more tolerable than longer standard fractionation radiation treatment regimens improving quality of life, and is easily combined with other modalities by limiting delays in chemotherapy and time to surgery. He presented data from the phase II multi-institutional trial evaluating gemcitabine and SBRT in patients with locally advanced unresectable pancreatic cancer [Herman et al. 2015]. These data demonstrate that 10% of these patients who would not have been candidates for initial surgery were deemed to have resectable tumors following therapy and 8% ultimately underwent R0 and node negative resection. He reiterated that most tumors do not shrink in response to neoadjuvant therapy and radiographic response does not correlate well with histologic response [Dholakia et al. 2013]. Dr Herman reviewed current data incorporating aggressive multiagent chemotherapy and radiation used in the neoadjuvant setting, which resulted in a high R0 resection and survival rates for those patients achieving an R0 resection approaching that of patients with initially resectable disease [Mellon et al. 2015; Moningi et al. 2015]. He then presented retrospective data from Johns Hopkins comparing neoadjuvant chemotherapy with neoadjuvant chemotherapy followed by SBRT which resulted in a doubling of the R0 resection rate, which appeared to be similar for patients with initially borderline resectable and locally advanced unresectable disease. A 9% pathological complete response rate was observed for patients who received SBRT compared with 0% for those who received chemotherapy alone. He concluded from these data that induction multiagent chemotherapy followed by SBRT appears to improve surgical outcomes and select locally advanced unresectable patients may be able to undergo potentially curative surgery. He presented the concept for Alliance A021501: Preoperative Extended Chemotherapy versus Chemotherapy plus Hypofractionated Radiation Therapy for Borderline Resectable Pancreatic Cancer of the Head of the Pancreas which will prospectively compare chemotherapy (mFFX 4) alone with induction chemotherapy (mFFX 4) followed by SBRT prior to resection and adjuvant chemotherapy (FOLFOX 4) as neoadjuvant treatment for borderline resectable pancreatic cancer. At present, there are no randomized data favoring neoadjuvant over adjuvant therapy. However, multiple phase I and II clinical trials have evaluated the role of neoadjuvant radiation, chemotherapy, and chemoradiation regimens in pancreatic cancer. Different neoadjuvant treatment strategies have been investigated which utilize systemic therapy, local radiation, or combined approaches. Older trials demonstrate an improvement in resection rate and local control rates using neoadjuvant radiation [Pilepich and Miller, 1980], and current trials focus on neoadjuvant chemotherapy or chemoradiation approaches. The use of neoadjuvant chemoradiation is associated with higher R0 resection rates in a subset of patients, ranging from 87% to 100% [Mehta et al. 2001; Shinchi et al. 2002; Massucco et al. 2006; Huguet et al. 2007; Krishnan et al. 2007; Small et al. 2008; Tinkl et al. 2009; Landry et al. 2010; Polistina et al. 2010; Patel et al. 2011; Sahora et al. 2011a, 2011b; Stokes et al. 2011; Pipas et al. 2012; Kim et al. 2013; Leone et al. 2013; Mukherjee et al. 2013]. Others have investigated more aggressive neoadjuvant chemotherapy regimens without concurrent radiation in an attempt to maximize systemic dosing without added toxicity of local radiation, theoretically leading to more effective elimination of distant micrometastasis and potentially improved long-term outcome [Sahora et al. 2011a, 2011b; Hosein et al. 2012; Lee et al. 2012; Parmar et al. 2014]. To date, no clinical trial has compared neoadjuvant chemotherapy with chemoradiation. Longer follow up is needed to determine if this strategy improves survival. Distant failures continue to limit survival in patients who receive neoadjuvant chemoradiation.

Surgical resection is considered the only treatment modality that impacts outcome in pancreatic cancer [Griffin et al. 1990]. Historically the lack of a standard definition of surgical resectability has confounded surgical outcome data and it is recognized that the ability to accurately stage patients is essential for the development and evaluation of stage-specific therapies to maximize outcome and quality of life for all patients. Standardization of definitions of resectability have recently been incorporated into staging systems for pancreatic cancer [Vauthey and Dixon, 2009]. Many neoadjuvant trials incorporate patients with borderline and unresectable tumors, making data difficult to interpret.

Conflicting data exist regarding the benefit of neoadjuvant therapy in patients with initially resectable pancreatic tumors [Palmer et al. 2007; Evans et al. 2008; Heinrich et al. 2008b; Varadhachary et al. 2008; Van Buren et al. 2013]. A meta-analysis including initially resectable disease demonstrated that resection rates did not differ between those patients with resectable tumors who received neoadjuvant therapy compared with those treated with adjuvant therapy, but better response rates were observed in patients treated with chemoradiation compared with chemotherapy alone [Gillen et al. 2010]. Another meta-analysis evaluated gemcitabine-based neoadjuvant regimens and found only marginal survival benefits for patients with resectable cancer whether they received radiation or not [Andriulli et al. 2012]. Further randomized studies are necessary to confirm the benefit of neoadjuvant therapy in patients with initially resectable pancreatic cancer. The ongoing NEOPA trial [ClinicalTrials.gov identifier: NCT01900327] is a prospectively randomized phase III trial of patients receiving neoadjuvant chemoradiation followed by curative surgery versus primary surgery followed by adjuvant therapy with a primary endpoint of 3-year overall survival [Tachezy et al. 2014]. In addition, the NEOPAC study is currently evaluating neoadjuvant gemcitabine versus neoadjuvant gemcitabine and oxaliplatin (GEMOX) in patients with resectable disease.

Data suggest that patients with borderline resectable tumors are most likely to benefit from downstaging using the neoadjuvant approach [Turrini et al. 2009; Landry et al. 2010; McClaine et al. 2010; Kang et al. 2012; Katz et al. 2012]. The most significant prognostic factor for survival in patients with pancreatic cancer is an R0 resection and it has been demonstrated that resection with positive margin is independently associated with prognosis similar to inoperable disease [Willett et al. 1993; Sohn et al. 2000; Evans, 2001; Chandler et al. 2003; Hosein et al. 2012; Russo et al. 2012; Satoi et al. 2012]. Compared with initially resectable pancreatic cancer, borderline resectable pancreatic cancer is associated with a higher risk for positive resection margin. Several studies suggest that neoadjuvant chemoradiation may enhance margin-negative resectability rates and improve local control [Mehta et al. 2001; Shinchi et al. 2002; Massucco et al. 2006; Huguet et al. 2007; Krishnan et al. 2007; Small et al. 2008; Tinkl et al. 2009; Landry et al. 2010; Polistina et al. 2010; Patel et al. 2011; Sahora et al. 2011a, 2011b; Stokes et al. 2011; Pipas et al. 2012; Kim et al. 2013; Leone et al. 2013; Mukherjee et al. 2013; Yeung et al. 1993]. Unfortunately, many of the studies are confounded by inclusion of patients with locally advanced unresectable tumors and lack of strict definition of borderline resectable disease. We await the results of recently completed trials evaluating neoadjuvant gemcitabine-based regimens in borderline resectable and locally advanced pancreatic cancer including GAIN-1 trial, Gemcitabine With Abraxane and Other Investigational Therapies in Neoadjuvant Treatment of Pancreatic Adenocarcinoma [ClinicalTrials.gov identifier: NCT01470417] and Regional Chemotherapy in Locally Advanced Pancreatic Cancer: RECLAP trial [ClinicalTrials.gov identifier: NCT01294358] which explores intra-arterial gemcitabine delivered to the tumor through an indwelling subcutaneous port.

It is clear that systemic progression remains a significant problem and the impact of local modalities might be better understood if more effective systemic therapies are available. Historically, limited knowledge of tumor biology has posed challenges for development of effective therapies. Pancreatic cancer microenvironment consists of multiple components, including stellate, endothelial, and immune cells, fibroblasts, and an extensive paracrine and autocrine network that reside in a stiff extracellular environment, associated with desmoplastic stroma and disturbed microvasculature and lymphatics, which may increase the potential for early metastases. To date, adjuvant trials have investigated the role of gemcitabine and 5-fluoruracil (5-FU) with and without radiation with limited impact on outcomes. Recent trials investigate the role of chemotherapy in combination with immunotherapy or vaccines in attempt to overcome some of the biological challenges inherent to pancreatic cancer [Hardacre et al. 2013]. Dr Robert Vonderheide from the University of Pennsylvania discussed inflammatory networks and immune surveillance in cancer and new treatment strategies for pancreatic cancer using biologic strategies during his Keynote. He described a novel approach using gemcitabine combined with an antibody against CD40 used to stimulate the immune system that has demonstrated promising results in mouse models [Beatty et al. 2011, 2013; Vonderheide and Glennie, 2013]. He also discussed a novel strategy for cancer prevention using a universal tumor antigen target (hTERT immunogenic protein expressed in nearly all human cancers with a functional role in oncogenesis) using DNA vaccine platform and electroporation applied at the site of the vaccine to allow the vaccine to be taken up (Penn technology licensed to Inovio), and a DNA cytokine (interleukin 12) to provide molecularly defined adjuvant [Yan et al. 2013]. He presented data using this approach showing significant response in mice and nonhuman primate models. The technology is currently being tested in a phase I trial for high-risk patients for breast lung and pancreatic cancer at University of Pennsylvania (UPCC 22913). He discussed a small clinical trial that is underway at Stanford University using an antibody targeting CD47 on the surface of cancer cells which results in activation of macrophages to attack cancer cells [Yan et al. 2013].

In conclusion, there are a number of unresolved controversies surrounding the current management of pancreatic cancer and we continue to collect data to help determine optimal treatment strategies, especially for those patients who may be candidates for surgery. As systemic disease continues to be a problem and available drug therapies have limited impact, we continue to strive for improved therapies using novel approaches so that we may continue to improve outcomes for patients with this deadly disease.