Successes and Failures of Combined Modalities in Upper Gastrointestinal Malignancies: New Directions

Abstract

Upper gastrointestinal malignancies generally have moderate to poor cure rates, even in the earliest stages, thereby implying that both local and systemic treatments have room for improvement. Therapeutic options are broadening, however, with the development of new immunotherapies and targeted agents, which can have synergistic effects with radiotherapy. Here we discuss the current state of combined modality therapy for upper gastrointestinal malignancies, specifically recent successes and setbacks in trials of radiation therapy with targeted therapies, vaccines, immunotherapies, and chemotherapies.

Introduction

Recent data estimate that out of 1,665,540 new cancers diagnosed in the United States in 2014, upper gastrointestinal (UGI) malignancies comprised 130,650 (7.8%), with approximately 36% pancreatic, 33% hepatobiliary, 17% gastric, and 14% esophageal.¹ The estimated mortality thereof was 92,660, of which approximately 43% were pancreatic, 29% were hepatobiliary, 16% esophageal, and 12% gastric. Pancreatic cancer in particular is the fourth-leading cause of death among all malignancies for both men and women in the United States. Overall, UGI cancers are often advanced, if not metastatic, at time of diagnosis,^2–4 which in turn increases morbidity and mortality.

The aggressiveness of UGI cancers has led researchers to evaluate multimodal approaches to treatment involving combinations of surgery, radiotherapy, and systemic agents. Other authors have discussed the synergy between radiation and systemic therapies, for instance the radiosensitizing properties of systemic therapies⁵ as well as the potential “abscopal effect” of immunotherapy wherein local radiation facilitates distal control.⁶ We present findings from recent studies on combined modality regimens for UGI malignancies and the ramifications thereof on treatment.

Targeted Therapies

Monoclonal Antibodies

Targeted therapies have been a mainstay in treating certain cancers because of their ability to selectively target tumor cells while sparing normal cells. Monoclonal antibodies (MABs) are a category of targeted drugs that generally target specific proteins in growth factor and angiogenic pathways, which are often overactive or otherwise dysfunctional in tumor cells. MABs have been evaluated in combination with chemotherapy or with radiation therapy⁷ because of evidence that MABs can enhance radiation-induced apoptosis⁸ and promote cell cycle arrest,⁹ thereby preventing tumor repopulation. Recent studies aimed at using MABs with radiation therapy have shown promising results in treating UGI malignancies (Table 1).

Table 1.

Trials Combining Radiotherapy With Monoclonal Antibodies (MAB)

Primary Investigators	Site (#Accrued)	Regimen	Agent and Target	Key Findings	Tolerability and Toxicity
Meng et al14	Locally advanced esophageal SCC (55)	Paclitaxel, cisplatin, and 59.4 Gy (33 fractions) with cetuximab	Cetuximab, anti-EGFR MAB	80% partial or complete response and 1 year OS 93%.	Grade 3 neutropenia and mucositis in 18/55 (33%) and 7/55 (13%), respectively. No Grade 4 toxicities observed.
Becerra et al16	Esophageal ACA and GEJ (39)	Neoadjuvant cetuxiumab and 50.4 Gy (28 fractions)	Cetuximab, anti-EGFR MAB	48% complete response rate 3 months after surgical resection.	Grade 5 aspiration seen in 1/39 (2%). Grade 3 dysphagia in 7/39 (17%). Grade 4 toxicities did not occur in 5% or more of patients.
Lee et al18	Locally advanced esophageal ACA (19)	Neoadjuvant irinotecan, cisplatin, and cetuximab with 50.4 Gy (28 fractions)	Cetuximab, anti-EGFR MAB	Median OS 31 months and median PFS 10 months with 16% CR following surgery.	Grade 3–4 neutrophilia and dysphagia in 9/19 (47%) and 6/19 (31%), respectively. Grade 3 diarrhea also in 9/19 (47%).
Tomblyn et19	Esophageal ACA (21)	Neoadjuvant irinotecan, cisplatin, and cetuximab with 50.4 Gy (28 fractions)	Cetuximba, anti-EGFR MAB	2-years OS 33% and PFS 24% with overall response rate 18%.	Grade 3–4 toxicity in 76% (16/21), including 11/21 (52%) from hematologic derangements. GI necrosis and sudden death each occurred in 1/21 (5%).
Crosby et al21	Esophageal ACA and SCC (258)	Cisplatin, capecitabine, and 50.4 Gy (28 fractions) with and without cetuximab	Cetuximab, anti-EGFR MAB	Cetuximab group had higher rate of treatment failure at 6 months (77%) and lower median OS (22 months).	Cetuximab group had more (102/129; 79%) nonhematological Grade 3–4 toxicities (eg, dysphagia, rash) vs control group (81/129; 63%). Control group had more (36/129; 28%) Grade 3–4 hematological toxicity vs cetuximab group (27/129; 21%).
Ilson et al22	Esophageal ACA and SCC (328)	Capecitabine, paclitaxel, and 50.4 Gy (28 fractions) with and without cetuximab	Cetuximab, anti-EGFR MAB	No difference between cetuximab and control groups in terms of efficacy or safety.	Nearly equivalent rates of Grade 3 (45%−49%), 4 (17%−22%), and 5 (1%−4%) toxicities between groups.
Esnaola et al23	Unresectable pancreatic ACA (37)	Gemcitabine, oxaliplatin, cetuximab, capecitabine, and 54 Gy (30 fractions)	Cetuximab, anti-EGFR MAB	Median OS and PFS of 12 months and 10 months, respectively. PFS after 6 months of 62%. No survival benefit seen	Grade 3–4 leukopenia in 5/37 (13%). Overall, well-tolerated regimen.
Zhao et al25	Esophageal SCC (11)	Cisplatin, 5-FU, and 60.2 Gy (34 fractions) with nimotuzumab	Nimotuzumab, anti-EGFR MAB	6-month OS 78%. 1-year OS and PFS of 67% and 100%, respectively.	No dose-limiting toxicity observed. Grade 3–4 esophagitis and leukopenia each seen in 2/11 (18%).
Ramos-Suzarte et al26	Esophageal ACA and SCC (63)	Cisplatin, 5-FU, and 45–50 Gy (25–28 fractions) with and without nimotuzumab	Nimotuzumab, anti-EGFR MAB	Nimotuzumab group had better median OS and response rates.	Grade 3–4 diarrhea more common in nimotuzumab group (6/33; 18%) than in control (2/30; 7%). Well-tolerated regimen.
Lockhart et al96	Locally advanced esophageal ACA (70)	Neoadjuvant docetaxel, cisplatin, panitumumab, and 50.4 Gy (28 fractions)	Panitumumab, anti-EGFR MAB	pCR rate 33% and near-pCR rate (ie, 10% or fewere viable cancer cells) another 20%.	Grade 3–4 esophagitis and lymphopenia seen in 13/70 (18%) and 30/70 (43%), respectively. Overall, Grade 4 toxicity in 48% (34/70) of patients.
Kordes et al97	Locally advanced esophageal SCC and ACA (90)	Neoadjuvant carboplatin, paclitaxel, panitumumab, and 41.4 Gy (23 fractions)	Panitumumab, anti-EGFR MAB	Low CR rate of 22% overall with 14% in ACA and 47% with SCC.	Grade 3–4 rash and neutropenia in 10/90 (11%) and 9/90 (10%), respectively.
Bendell et al38	Esophageal and GEJ ACA (62)	Neoadjuvant paclitaxel, carboplatin, 5-FU, and 45 Gy (25 fractions) with bevacizumab and erlotinib	Bevacizumab, anti-VEGF MAB Erlotinib, anti-EGFR TKI	CR = 29% with median OS and PFS of 30 months and 29 months, respectively.	Grade 3–4 leukopenia (40/62; 64%) and mucositis (26/62; 42%) were most common adverse effects. Grade 3–4 diarrhea and esophagitis each seen in 17/62 (27%).
Crane et al39	Pancreatic ACA (82)	Capecitabine, bevacizumab, and 50.4 Gy (28 fractions)	Bevacizumab, anti-VEGF MAB	No survival benefit over standard therapy.	Overall, 29/82 (35%) had Grade 3 or higher GI toxicity.
Small et al40	Pancreatic ACA (32)	Gemcitabine, bevacizumab, and 36 Gy (15 fractions)	Bevacizumab, anti-VEGF MAB	Improved PFS, but no survival benefit over standard therapy.	Grade 3 leukopenia (6/32; 19%) and nausea (5/32; 16%) were most common adverse events.

Cetuximab

Cetuximab is an epidermal growth factor receptor (EGFR) inhibitor that has been used in colorectal cancer,¹⁰ as well as in squamous cell carcinomas of the head and neck.¹¹ Recently, researchers have investigated its applicability in esophageal cancers, particularly since EGFR is widely expressed therein¹² and generally portends a poor outcome.^8,13 Of note, most of the studies involving EGFR therapies did not explicitly state whether EGFR status had been evaluated before enrollment.

Meng et al¹⁴ conducted a Phase II trial that combined cetuximab with paclitaxel, cisplatin, and 59.4 Gy delivered in 33 fractions to 55 patients with unresectable, locally advanced squamous cell carcinoma (SCC) of the esophagus. They found that 44/55 (88%) had a clinical response, including 29/55 (53%) with complete response and another 15/55 (27%) with partial response. The 1-year progression-free survival (PFS) and overall survival (OS) for the 45 patients who responded were 84.2% and 93.3%, respectively; at 2 years they were 74.9% and 80.0%, respectively. Their efficacy and safety outcomes were favorable (Table 1) and their trial was notable in that it included only patients with SCC histology, whereas other studies have also included adenocarcinoma (ACA) types.¹⁵

A Phase II trial of 39 patients (78% ACA and 37% gastroesophageal junction (GEJ) cancers) involved 2 weeks of cetuximab followed by neoadjuvant cetuximab with concurrent 50.4 Gy in 28 fractions, which in turn preceded surgical resection by 6–8 weeks.¹⁶ The pathological complete response (pCR) rate was 36.6% by intent-to-treat and 48% at 3 months in those patients who had an esophagectomy. Although there was no statistically significant difference in pCR between esophageal and GEJ sites, only 9/32 (28%) ACA patients achieved pCR whereas 6/9 (67%) SCC patients did, thereby suggesting that histology may predict response. pCR was also higher for those with IIA disease (70%) compared to those with IIB and III (28%–29%) disease. Toxicities and efficacy were comparable to what others found when evaluating neoadjuvant chemotherapy and chemoradiation in esophageal cancer.¹⁷

There have been, however, at least 2 studies that included cetuximab with preoperative irinotecan, cisplatin, and 50.4 Gy (28 fractions) for locally advanced esophageal cancer that produced less favorable results (Table 1).^18,19 Lee et al¹⁸ enrolled 19 patients with IIA-IVA esophageal and GEJ cancer, most with ACA histology (84%) and 95% proceeded to surgery. Median PFS was 10 months and median OS was 31 months. Only 3/19 (16%) patients had a pCR, in contrast to the 25%–40% complete response rate seen in other studies using standard neoadjuvant chemoradiation with surgery.²⁰ Tomblyn et al¹⁹ used a similar treatment schedule on 21 enrolled patients—48% ACA—with unresectable disease and found 2-years OS = 33.3% and PFS = 23.8% with overall response rate 17.6%. This was, however, a poorly tolerated regimen (Table 1). Thus, adding cetuximab to irinotecan, cisplatin, and radiotherapy led to worse efficacy and toxicity.

The SCOPE-1 trial by Crosby et al²¹ investigated the addition of cetuximab to definitive cisplatin-capecitabine-50 Gy in patients with nonmetastatic esophageal cancer. This was a multicenter Phase II-III randomized control trial wherein 129 patients were assigned to receive cetuximab with standard cisplatin-capecitabine with 50 Gy (25 fractions) whereas another 129 received standard therapy alone. Within each arm, approximately 70%–75% of patients had SCC histology, 25% had ACA type, and 60% had stage III disease. The study was terminated early because the experimental group had more treatment failures at 6 months (33.6% vs 23.1%) and grade 3–4 toxic events (79% vs 63%, P = 0.004), plus shorter median OS (22.1 months vs 25.4 months). Moreover, fewer patients in the experimental arm were able to complete full or reduced doses of cisplatin (77% vs 90%), capecitabine (69% vs 90%), or radiotherapy (81% vs 92%) due to toxicity or worsened disease. As such, the authors did not recommend incorporating cetuximab into cisplatin-capecitabine-based chemoradiation for localized esophageal cancer.

The radiation therapy oncology group (RTOG)-0436 multicenter Phase III trial by Ilson et al²² likewise showed no benefit when cetuximab was added to concurrent chemoradiation in nonresectable esophageal cancer. A total of 328 patients (62% ACA and 38% SCC) were randomized to receive cisplatin-paclitaxel-50.4 Gy (28 fractions) with or without cetuximab with OS as primary endpoint. CR was equivalent between the groups with 56% for those who had received cetuximab and 59% for those in the control arm. Moreover, the 1- and 2-year OS for the cetuximab group were 64% and 44%, respectively, compared to 65% and 42%, respectively, for the control group (P = 0.70). These OS rates remained largely similar when stratified for histology. The authors concluded that adding cetuximab to chemoradiation improved neither OS nor CR, thereby calling into question its role in esophageal cancer treatment.

Cetuximab has also been studied in both gastric and pancreatic cancer, but as of this writing we did not come across recent studies that investigated its use with radiotherapy in gastric cancer. One recent Phase II trial by Esnaola et al²³ evaluated the safety and efficacy of induction gemcitabine-oxaliplatin with cetuximab followed by chemoradiation with capecitabine and 50.4 Gy (30 fractions) in locally advanced, unresectable pancreatic cancer. In the 37 patients involved, 30% underwent surgical resection; median OS was 11.8 months and median PFS was 10.4 months with 6-month PFS at 62%. Grade 3–4 toxicity was seen in around 30% after induction chemotherapy and less than 10% after chemoradiation. Although the treatment was well tolerated, neither PFS nor OS was ultimately improved that prompted the authors to call for more prospective studies.

Nimotuzumab

Nimotuzumab is another anti-EGFR MAB that has been evaluated with chemoradiation in squamous cell carcinomas of the head and neck.²⁴ Zhao et al²⁵ carried out a Phase I study to assess the maximum tolerated dose (MTD) and efficacy of nimotuzumab with concurrent cisplatin, 5 fluorouracil (5-FU), and 61.2 Gy (34 fractions) in locally advanced esophageal SCC. This regimen was well tolerated (Table 1) with 1-year OS and local PFS of 67% and 100%, respectively, which compared favorably with standard treatment.²⁰

Ramos-Suzarte et al²⁶ conducted a Phase II trial wherein 30 patients with unresectable Stage III–IV esophageal cancer were randomized to receive cisplatin, 5-FU, and 45.0–50 Gy (25–28 fractions of 1.8 Gy) and another 33 were assigned to have the same regimen plus nimotuzumab. EGFR expression was confirmed in a subset of patients (5 from control group and 13 from nimotuzumab group). The nimotuzumab group had a higher objective response rate (47.8%–15.4%), better local control rate (60.9%–26.9%), longer median survival (8.1 months to 2.9 months) and similar toxicity (Table 1). Thus, the nimotuzumab group had improved efficacy and comparable safety to standard treatment.

Nimotuzumab has been tested in both gastric and pancreatic cancer, but we could not find recent trials that included radiotherapy in the experimental regimens.

Trastuzumab

In addition to its prevalence in breast cancer, Human Epidermal Growth Factor Receptor-2 (HER-2) has been found to be overexpressed in at least 20% of patients with GEJ and both esophageal and gastric ACA tumors.^27,28 There is a mortality benefit when HER-2-positive patients with both gastric and GEJ ACA receive trastuzumab, an anti-HER-2 MAB, with concurrent cisplatin-based chemotherapy.²⁹ Trastuzumab also has radiosensitizing effects in esophageal cancer cell lines,³⁰ which further enhances its appeal in combined modality therapy. The utility of HER-2 targeting in pancreatic and hepatobiliary malignancies, however, is less promising; a recent review by Omar et al³¹ found that HER-2 amplification in pancreatic and biliary malignancies ranged from 2%–24% and only 5%–8%, respectively. Xian et al³² had previously noted that HER-2 was amplified in just 1/868 (0.11%) of their patients with primary hepatocellular carcinoma (HCC).

Trastuzumab has not been well studied in conjunction with radiotherapy for UGI malignancies. Safran et al³³ initially carried out a Phase I–II study wherein 19 patients with HER-2 positive, nonmetastatic esophageal ACA received cisplatin, paclitaxel, and 50.4 Gy (28 fractions) with trastuzumab. None of the patients had undergone surgical resection before enrollment and only one had surgical resection during the trial. A total of 74% (14/19) of the patients had celiac, retroperitoneal, portal, or scalene lymphadenopathy at enrollment, yet the authors found that their study group had a 2-years OS of 50% and median survival of 24 months, which was similar to what they had previously found without trastuzumab and without a notable difference in toxicity.³⁴

The authors conceded that their small sample size necessitated further studies to evaluate efficacy and safety, which then led them to carry out the ongoing NCT01196390 Phase III trial involving patients with HER-2–positive esophageal and GEJ ACA. At the time of this writing, they have registered 194 patients to receive neoadjuvant paclitaxel, carboplatin, and 50.4 Gy (28 fractions) with or without trastuzumab. Results from this trial are still pending.

We did not find recent trials investigating the utility of trastuzumab with radiation in hepatobiliary or pancreatic cancer, which may be due to the low amplification levels cited earlier.

Bevacizumab

Vascular endothelial growth factor (VEGF) levels have been associated with esophageal and gastric cancer aggressiveness and mortality.^35,36 As such, MABs that target VEGF, such as bevacizumab, have been evaluated for efficacy in various malignancies, including metastatic gastric cancer.³⁷

Bendell et al³⁸ tested the efficacy of bevacizumab and erlotinib with standard neoadjuvant paclitaxel, carboplatin, 5-FU, and 45.0 Gy (25 fractions) in treating localized esophageal and GEJ cancer. In their Phase II trial, 62 patients with resectable disease (93.5% with ACA) underwent treatment; 29% had pCR and the median OS and PFS were 30.2 months and 28.6 months, respectively. VEGF expression was not mentioned in the inclusion criteria. Although overall efficacy and toxicity were similar to those seen in standard neoadjuvant chemoradiation (Table 1), the authors found no evidence of improved efficacy and so did not feel further study of the specified regimen in esophageal cancer was warranted.

At least 2 other studies have investigated the applicability of bevacizumab to chemoradiation in localized pancreatic cancer. Crane et al evaluated the efficacy and safety of bevacizumab with capecitabine and 50.4 Gy (28 fractions) in patients with locally advanced, nonmetastatic pancreatic cancer.³⁹ Among the 82 patients, the median and 1-year survival rates were 11.9 months and 47%, respectively, whereas the median and 1-year PFS were 8.6 months and 30%, respectively. Although toxicity was modest (Table 1), there was no survival benefit observed, so further evaluation was not advised.

Small et al⁴⁰ later assessed the response rate, survival, and toxicity of gemcitabine-bevacizumab with 36 Gy of radiation in 15 fractions in 32 patients with nonmetastatic pancreatic cancer. At 10 weeks into treatment, 66% (21/32) had stable disease whereas another 9% (3/32) had partial or complete response; the remaining 4/32 (12%) had progressive disease. Overall, the authors found median PFS to be 9.9 months with PFS at 6 months, 1 year, and 2 years to be 62%, 38%, and 13%, respectively; median OS was 11.8 months with OS at 6 months, 1 year, and 2 years at 86%, 46%, and 23%, respectively. The regimen was well tolerated (Table 1), but while median PFS and 1-year PFS were moderately improved from that seen by the Crane et al study, no survival benefit was seen.

Tyrosine Kinase Inhibitors

Tyrosine kinase inhibitors (TKIs), like MABs, are typically engineered to target growth factor and angiogenic pathways. They differ in that while MABs are designed to prevent extracellular ligand-receptor interactions and subsequent receptor activation, the TKIs target adenosine triphosphate (ATP) binding sites on the cytoplasmic side of the receptor protein and thereby block downstream cascades.⁴¹ Much like MABs, TKIs have been found to exhibit immunomodulatory properties⁹ with radiotherapy, so their utility in combined modality approaches is of significant interest (Table 2).

Table 2.

Trials Combining Radiotherapy With Tyrosine Kinase Inhibitors (TKI)

Primary Investigators	Site (#Accrued)	Regimen	Agent and Target	Key Findings	Tolerability and Toxicity
Li et al44	Esophageal SCC (24)	Cisplatin, paclitaxel, erlotinib, and 60 Gy (30 fractions)	Erlotinib, anti-EGFR TKI	Improved OS, local control, and toxicity profile compared to RTOG-8501	No Grade 4 toxicities. Most common Grade 3 toxicities were esophagitis and leukopenia, which were seen in 5/24 (21%) and in 4/24 (17%) patients, respectively.
Iyer et al46	Esophageal ACA and SCC (17)	Erlotinib with 50.4 Gy (28 fractions)	Erlotinib, anti-EGFR TKI	Well-tolerated regimen with 1-year OS 29%	Grade 3–4 toxicities occurred in 5/17 (29.4%) patients; similar to RTOG-8501 trial.
Herman et al47	Pancreatic ACA (48)	Adjuvant capecitabine, gemcitabine, erlotinib, and 50.4 Gy (28 fractions)	Erlotinib, anti-EGFR TKI	OS 93% at 1 year and 51% at 2 years. Recurrence-free survival 87% at 1 year and 30% at 2 years.	Grade 3–4 toxicities, mostly neutropenia, occurred in 26/48 (54%) patients with 16/26 (61%) during chemoradiation and 10/26 (48%) after maintenance chemotherapy. Unplanned hospitalizations needed in 16/48 (33%).
Konski et al50	Pancreatic ACA (26)	Gemcitabine, erlotinib, and 28.0–28.8 Gy (0.5–0.6 Gy bid)	Erlotinib, anti-EGFR TKI	MTD not reached; median survival = 9.1 months and stable to improved disease in 88%	Grade 3–4 leukopenia seen in 14/26 (54%) patients. One patient had Grade 5 toxicity (bowel perforation).
Rodriguez et al52	Esophageal and GEJ SCC and ACA (80)	Neoadjuvant cisplatin, 5-FU, geftinib, and 30 Gy (1.5 Gy bid)	Geftinib, anti-EGFR TKI	Geftinib showed superior OS compared to historic controls, but no difference in local or distal control. Gefitinib poorly tolerated.	Gefitinib-treated patients experienced less Grade 3–4 mucositis and (2/80; 2.5%), neutropenic fever (4/80; 5.0%), and fewer unplanned hospitalizations (11/80; 13.7%) compared with historic controls (14%, 17%, 28%, respectively).
Olsen et al53	Pancreatic ACA (11)	Paclitaxel, geftinib, and 50.4 Gy (28 fractions)	Geftinib, anti-EGFR TKI	Poor survival	No Grade 4 toxicities seen. Diarrhea occurred in 10/11 (91 %), of which 3/10 (30%) were Grade 3. Fatigue seen in 9/11 (82%), of which 3/9 (33%) were Grade 3.
Chen et al55	Hepatocellular carcinoma (40)	Sorafenib with 50–60 Gy (fractions of 2.0–2.5 Gy) then sequential sorafenib	Sorafenib, anti-VEGF-R TKI	2-year OS 32% and in-field local control 39%. Increased toxicity with sequential sorafenib.	Grade 3–5 hepatic toxicity seen in 9/40 (22%) during sequential sorafenib, but during concurrent sorafenib-XRT only 4/40 (10%) had Grade 3 hepatotoxicity. No Grade 4 toxicities with concurrent sorafenib. Grade 1–2 GI ulcers and rashes very common (>90%).
Chiorean et al58	Pancreatic ACA (27)	Gemcitabine, sorafenib, and 50 Gy (25 fractions)	Sorafenib, anti-VEGF-R TKI	MTD not reached, but no OS benefit.	No dose-limiting toxicity observed. Grade 3–4 GI bleeding and ulcers each in 7/27 (26%).
Aparicio et al60	Pancreatic ACA (12)	Gemcitabine, sorafenib, and 45 Gy (25 fractions)	Sorafenib, anti-VEGF-R TKI	MTD not reached. No survival benefit.	Grade 4 posterior leukoencephalopathy and Grade 3 thrombocytopenia and fatigue each seen in 1/12 (8%). No other Grade 3–4 toxicities observed.

Erlotinib

Erlotinib is a TKI that reversibly interacts with the adenosine triphosphate binding site of EGFR and has been shown to be an effective radiation sensitizer.⁴² Its favorable safety profile in esophageal carcinoma patients undergoing chemoradiation with cisplatin and 5-FU has been previously described.⁴³ A 2010 Phase II study by Li et al⁴⁴ evaluated its efficacy and toxicity with concurrent paclitaxel, cisplatin, and 60 Gy (30 fractions) for locally advanced esophageal SCC. They enrolled 24 patients and administered 6 weeks’ of treatment; a total of 11 (46%) had complete response whereas another 11 (46%) had partial response. At 2 years, their OS was 70.1% with disease-free survival (DFS of 57.4% and locoregional control rate of 87.5%; both OS and locoregional control rates were improved from those seen in RTOG-8501.⁴⁵ Although 29% of the RTOG-8501 patients developed grade 3–4 toxicities, only 16.7% in the later study had similar events (Table 2).

Iyer et al⁴⁶ also had encouraging results for their Phase II trial that looked at 17 (94% ACA) esophageal cancer patients ineligible for standard chemoradiation and whom they treated with concurrent erlotinib and 50.4 Gy in 28 fractions. Surgery was not included in the treatment. The median OS was 7.3 months with 1-year OS of 29% whereas median PFS was 4.5 months. Toxicity was similar to that seen for the RTOG-8501 trial (Table 2).

Erlotinib has also been studied in pancreatic cancer treatment. Herman et al⁴⁷ published a Phase II trial that involved dosing erlotinib with concurrent adjuvant capecitabine and 50.4 Gy (28 fractions) followed by a course of erlotinib-gemcitabine, all following resection of stages I-II pancreatic ductal ACA. Among the 48 patients enrolled, OS and recurrence-free survival at 1 year were 93% and 87%, respectively. At 2 years, OS and recurrence-free survival dropped to 51% and 30%, respectively. Median OS was 24.4 months and median recurrence-free survival was 15.6 months. Their median OS and RFS as well as toxicity rates (Table 2) were similar to what other authors had found.⁴⁸ A Phase I dose-escalation study of capecitabine-erlotinib with concurrent 50.4 Gy further supported the safety and feasibility of capecitabine-erlotinib with radiotherapy in locally advanced pancreatic cancer, although both median OS and PFS were only 11–12 months and 7 months, respectively.⁴⁹

In a 2014 Phase I study, Konski et al⁵⁰ attempted to use low-dose ultra-fractionated radiation to induce chemosensitization to gemcitabine and erlotinib in locally advanced pancreatic cancer. Patients were assigned to receive 0.4 Gy twice daily to a total of 28.8 Gy, 0.5 Gy twice daily to a total of 28.0 Gy, or 0.6 Gy twice daily to a total of 28.8 Gy. Median survival was 9.1 months with 8/26 having partial response and 15/26 having stable disease; the treatment was generally well tolerated and MTD was not reached (Table 2). The authors therefore encouraged further investigation of low-dose ultra-fractionated radiation in pancreatic cancer.

Overall, erlotinib has shown promising safety and efficacy in combined modality trials for both esophageal and pancreatic cancer, although further investigation with larger studies is needed.

Gefitinib

Gefitinib has a mechanism of action similar to that of erlotinib in that it targets EGFR and prevents downstream activation of antiapoptotic and proliferative signaling cascades; it differs mainly in having a different molecular structure.⁵¹ A Phase II study based in the Cleveland Clinic examined the utility of neoadjuvant gefitinib, cisplatin-5-FU, and 30 Gy (1.5 Gy twice daily) followed by surgical resection and then adjuvant gefitinib with 30 Gy (1.5 Gy twice daily) in locally advanced esophageal and GEJ cancer.⁵² Among the 80 patients enrolled, the authors found superior results in terms of OS (42% vs 28%) compared with a historic control group of patients who had received the same regimen, but without gefitinib (Table 2). There was no statistically significant difference, however, regarding distant metastatic control (40% vs 32%) and locoregional control (76% vs 77%). Thus, the authors advocated further assessment in light of the improved survival rate, but they conceded that almost half the patients could not tolerate maintenance gefitinib, which they attributed to the aggressive multimodality approach.

A smaller Phase I study investigated the combination of gefitinib with paclitaxel and 50.4 Gy (28 fractions) in 11 patients with locally advanced pancreatic cancer.⁵³ Mean survival was 9 months from time of enrollment (range from <1 to 36 months) and toxicity fairly common (Table 2). Given the poor survival outcomes, the gefitinib-paclitaxel and radiation regimen is unlikely to merit continued evaluation despite having been well tolerated.

Sorafenib

Sorafenib is a TKI that targets VEGF receptor (VEGF-R), plasma-derived growth factor receptor, and numerous kinases commonly used in HCC.⁵⁴ A 2013 Phase II study by Chen et al⁵⁵ looked at concurrent sorafenib with radiation (50–60 Gy in fractions of 2.0–2.5 Gy) followed by sequential sorafenib in patients with unresectable, nonmetastatic HCC. A total of 40 patients with underlying cirrhosis were enrolled; 22/40 (55%) achieved partial to complete response whereas the other 18/40 (45%) had stable or worsened disease within a month of completing treatment. A 2-years OS and in-field progression free survival were 32% and 39%, respectively. Median OS was 14.0 months (95% CI: 9.1–18.9 months) whereas median in-field progression free survival was 8.9 months (95% CI: 2.4–15.3 months). Over 90% of patients developed Grade 1–2 gastroduodenal ulcers and hand-foot-skin reactions and some developed Grade 3 hepatotoxicity as well (Table 2).

In comparison, the 2013 Phase I–II trial by Bujold et al⁵⁶ wherein 102 patients with locally advanced HCC received 30–54 Gy (6 fractions) via stereotactic body radiation therapy (SBRT) (without sorafenib) resulted in median OS = 17.0 months (95% CI: 10.4–21.3 months) and median time to progression 6.0 months (95% CI: 3.4–6.4 months). Around 36% of patients in the SBRT study developed at least Grade 3 toxicity, most of which (54%) were Grade 3 liver function panel derangements or thrombocytopenia. Indeed, a recent article by Pollom et al⁵⁷ noted that VEGF inhibition may predispose patients to poor intestinal healing, ischemia, and even thromboembolism following injury from SBRT. Thus, while Chen et al highlighted the efficacy of sorafenib with radiation, the safety requires further assessment.

Two Phase I studies later evaluated sorafenib with simultaneous gemcitabine and radiotherapy in unresectable pancreatic ACA. Chiorean et al⁵⁸ administered 50 Gy (25 fractions) along with gemcitabine (600 mg/m² weekly) and sorafenib to 27 patients; no dose-limiting toxicity was observed (Table 2) and median OS was 12.6 months with median PFS 10.6 months. Unfortunately, the OS was not improved from standard protocols and about half the patients stopped treatment early.⁵⁹ Aparicio et al⁶⁰ carried out a similar study on 12 patients with gemcitabine (300 mg/m² weekly), 45 Gy (25 fractions), and sorafenib. No dose-limiting toxicity was observed and median OS was 10 months whereas median PFS was just 4.4 months (Table 2). Thus, no survival benefit has been established yet for sorafenib with concurrent radiation in treating pancreatic cancer.

Nanoparticles and Conjugates

Nanotechnology is a bourgeoning field in oncology since the ability of nanoparticles to accumulate preferentially within tumors and deliver chemotherapeutic agents in slow, controlled fashion can potentially augment existing therapies.⁶¹ Recent studies, detailed in this section, have looked into the feasibility of using such molecules in conjunction with radiotherapy, especially since they can extravasate through the leaky tumor vasculature^62,63 and exert radiation sensitivity effects within treatment fields.⁷

TNFerade

TNFerade is engineered from replication-deficient adenovirus containing tumor necrosis factor alpha (TNF-α) DNA ligated to radiation-inducible Egr-1 gene promoter.⁶⁴ TNF-α helps induce apoptosis, cytotoxic free radical production, and immune cell recruitment, all of which can facilitate tumor cell destruction. Still, TNF-α produces a shock-like syndrome when infused systemically or directly into tumors, hence its incorporation into a virus, which can then pass through highly permeable tumor vessels and exert cytotoxic effects selectively within tumor.

A 2012 Phase I dose-escalation study by Chang et al⁶⁵ to evaluate the efficacy of intratumoral TNFerade with neoadjuvant cisplatin, 5-FU, and concurrent 45 Gy (25 fractions) in locally advanced esophageal cancer resulted in median OS = 47.8 months and median DFS = 26.4 months. A total of 24 patients were enrolled with 20/24 (83%) ACA, 4/24 SCC (17%), and 20/24 (83%) having GEJ lesions. Ultimately, 21 patients were evaluated for clinical response with 20/21 (95%) having undergone surgical resection within 4–10 weeks of completing radiation treatment and 1/21 (5%) having an autopsy. The authors observed complete response in 6/21 patients (29%; 2/6 with SCC, and 4/6 with ACA) with 1-year and 3-years DFS rates 63% and 38%, respectively. The 1-year, 3-years, and 5-years OS rates were 88%, 54%, and 41%, respectively. The regimen was generally well tolerated, although the highest dose of TNFerade was associated with thromboembolic events. The authors concluded that TNFerade delivered endoscopically with concurrent 5-FU-cisplatin chemoradiation is efficacious and generally safe, but does not appear to change survival outcomes.

Another Phase I-II dose-escalation study that evaluated intratumoral TNFerade with concurrent 5-FU and 50.4 Gy (28 fractions) in locally advanced pancreatic cancer underscored the efficacy of TNFerade, with median OS of 9.8 months and median DFS of 3.6 months, both similar to standard 5-FU chemoradiation.⁶⁶ Dose-limiting toxicities of pancreatitis and cholangitis were seen (3/50) in the highest-dose group, but overall the treatment was well tolerated. A Phase III trial by Herman et al⁶⁷ had more sobering results; patients with locally advanced pancreatic cancer were randomly assigned to have standard 5-FU with 50.4 Gy either with or without TNFerade. Both groups had equivalent median OS of 10.0 months as well as median PFS of about 7 months. They also had almost similar survival rates at 1 year (41% in TNFerade group vs 36.7%) and at 2 years (11.3% in TNFerade group vs 10.3%). Median DFS was also similar at 11.6 months for the TNFerade group and 10.8 months for control group. The authors deemed TNFerade safe with mostly grade 2–3 toxicity, but ineffective in prolonging survival in locally advanced pancreatic cancer.

Liposomal Cisplatin (Lipoplatin)

Lipoplatin is a liposomal formulation of cisplatin; it is engineered to allow the large cisplatin molecule to be taken up by tumor cells and thereby concentrate within neoplastic masses.⁶⁸ A 2010 Phase I–II study by Koukourakis et al⁶⁹ involved treating 11 patients with locally advanced gastric cancer with 5-FU, lipoplatin, and radiotherapy to assess efficacy and toxicity. Grade 2–4 toxicity was uncommon (3/11), but the short median follow-up of 9 months could have precluded seeing evidence of longer-term morbidity or adverse events. Local control at 1 year was 51% and 1-year OS was 70%. Overall, the treatment was well tolerated, albeit in a small treatment group, and the authors called for further testing of lipoplatin with chemoradiation for gastric cancer.

CO-101 Lipid

CO-101 is a lipid-drug conjugate of gemcitabine that is designed to allow the medication to enter cells independently of levels of human equilibrative nucleoside transporter-1 (hENT-1) that can be down-regulated by tumor cells.⁷⁰ Poplin et al⁷¹ evaluated its efficacy in metastatic pancreatic cancer in a randomized Phase II trial aimed at determining whether or not CO-101 could outperform gemcitabine depending on hENT-1 expression. They enrolled 367 patients and found no difference in OS within the gemcitabine arm between patients with low and high hENT-1 expression. Furthermore, differences in OS and toxicity profiles between the gemcitabine and CO-101 arms were not statistically significant. The authors deduced that hENT-1 expression is unlikely to predict gemcitabine sensitivity and, in any case, that CO-101 is not superior to gemcitabine in stage IV pancreatic cancer.

Paclitaxel-Poliglumex

Paclitaxel-poliglumex (PPX) is a conjugate of paclitaxel with biodegradable poly-l-glutamate to render it water-soluble in circulation, to facilitate tumor cell uptake via hyper-permeable neoplastic vasculature, and to enhance its radiation sensitizing effects.⁷² In a 2012 Phase II trial at Brown University, 40 patients with esophageal cancer (37 with ACA and 3 with SCC) were given PPX, cisplatin, and 50.4 Gy concurrent radiation to elucidate safety and treatment response.⁷³ Three patients refused surgery, including 2 with SCC histology, but the remaining 37/40 (92%) underwent resection. No grade 4 toxicity was observed and the authors reported a 32% (12/37) complete response rate, all of whom had ACA. Thus, the study suggested that PPX could be a valuable radiation sensitizer in esophageal cancer treatment and that larger prospective studies would need to be undertaken to elucidate safety and efficacy.

Immunotherapies

Immunotherapy, specifically the direct recruitment of the immune system to attack tumors, has become an area of intensive research in recent years. Its incorporation into combined modality trials with radiation is of particular interest given the inherent immunomodulatory properties of radiotherapy.⁶ Indeed, others have written about the “abscopal effect” whereby localized radiation can induce regression or control of distant metastases because tumor antigens released from dying neoplastic cells potentiate an antibody-like response by the immune system.⁷⁴ Although numerous trials of immunotherapeutic agents in UGI cancers are underway or being planned,⁷⁵ we report the latest results of combined modality approaches using radiation and immunotherapy below.

Clivatuzumab Tetratexan

Clivatuzumab tetratexan is an MAB conjugated to yttrium-90 (Y90-hPAM4) that targets mucin-1 glycoprotein (MUC-1), which is produced by over 85% of pancreatic ACAs and which is largely absent from normal pancreatic tissue.⁷⁶ Pancreatic cancer cells have shown sensitivity to radioimmunotherapy (RAIT), particularly when combined with gemcitabine, hence the interest in clivatuzumab tetratexan.⁷⁷ In a 2012 Phase I trial, Ocean et al⁷⁸ tested gemcitabine along with a fractionated schedule of clivatuzumab tetratexan in 38 patients with advanced pancreatic cancer. Median OS was 7.7 months, although 13 patients who received a repeat treatment cycle had a median OS of 11.8 months. Around 16% (6/38) had partial responses and another 42% (16/38) had disease stabilization. A total of 74% (28/38) developed grade III–IV thrombocytopenia or neutropenia after 1 cycle; this rate rose to 100% (38/38) if the patient opted for a repeated cycle. The authors felt that patients could tolerate fractionated radioimmunotherapy together with gemcitabine and called for further studies to investigate the promising therapeutic and safety indices.

GM-CSF Vaccine

Granulocyte-macrophage colony stimulating factor (GM-CSF) is an immunostimulant,⁷⁹ so vaccine cells engineered to secrete GM-CSF have been evaluated as a potential adjunct to pancreatic cancer treatment. Jaffee et al⁸⁰ designed irradiated tumor cells to secrete GM-CSF and thereby facilitate tumor antigen presentation by recruited dendritic cells and coordination among CD4, CD8, and NK cells. Lutz et al⁸¹ conducted a single institution Phase II study wherein 60 patients with resected pancreatic ACA received adjuvant 5-FU chemoradiation with GM-CSF immunotherapy to evaluate safety and efficacy. GM-CSF was delivered via irradiated and transfected allogeneic whole cell tumor lines. They reported median survival of 24.8 months and median DFS of 17.3 months; 1-year OS and DFS were 85.0% and 67.4%, respectively. Their median and 1-year OS was not significantly different from those seen in a historical cohort of patients treated with adjuvant chemoradiation alone at their institution. No dose-limiting toxicity was observed and most adverse effects consisted of dermatitis-type reactions that usually resolved within a week. The authors concluded that adjuvant GM-CSF immunotherapy was safe and efficacious; they also noted that mesothelin-specific T-cell frequencies induced by immunotherapy may be associated with improved DFS, and called for broadening scope of treatment in a multicenter study.

Checkpoint Inhibitors

Immune checkpoint inhibitors are among the most prominent immunotherapeutic agents within the past few years. These molecules target important stimulatory and inhibitory pathways in immune cells to modulate their tumor detection and cytotoxic abilities. Recent agents that target programmed death receptor-1 (PD-1)⁸² and programmed death receptor ligand-1 (PD-L1)⁸³ have shown demonstrated efficacy in melanoma and other malignancies. Several authors have discussed the prevalence of PD-1 and PD-L1 overexpression in gastric and GEJ ACA⁸⁴ as well as their negative effect on prognosis, which underscores the importance of testing checkpoint inhibitors in UGI malignancies. Some researchers have found that the anti-PD-1 MAB pembrolizumab has efficacy without severe toxicity in advanced gastric ACA⁸⁵ as well as in both esophageal SCC and ACA.⁸⁶ Others have evaluated ipilimumab that binds the inhibitory T-cell protein CTLA-4, together with gemcitabine in pancreatic ACA and found significant Grade 3–4 hematological adverse events (69%).⁸⁷

Still, while the systemic efficacy of checkpoint inhibitors has been explored, we have not come across publications on checkpoint inhibitors with concurrent radiation for UGI malignancies as of this writing.

Other Agents

Eniluracil

Eniluracil is an irreversible inhibitor of dihydropyrimidine dehydrogenase, which usually breaks down 5-FU. Its utility in treating locally advanced pancreatic and biliary cancers with concurrent neoadjuvant 5-FU and 45 Gy (28 fractions) was assessed in a Phase I study by Czito et al.⁸⁸ The patients received an additional 5.4 Gy (3 fractions) to gross tumor volume following initial radiation treatment. Among the 13 patients enrolled, MTD was not reached, only 1 dose-limiting adverse event was observed, and 3/13 (23%) had a pCR or CR. The authors showed that neoadjuvant eniluracil-5-FU and radiation was well tolerated and feasible, although further studies are lacking.

Triapine

Triapine is a radiosensitizer and ribonucleotide reductase inhibitor that prevents conversion of ribonucleoside diphosphates to deoxyribonucleotides, which in turn are used for DNA synthesis.⁸⁹ In a 2012 Phase I dose-escalation study (n = 12), Martin et al⁹⁰ investigated 3 levels of triapine with 50.4 Gy in patients with locally advanced pancreatic cancer. MTD was not reached, 6/12 (50%) had stable disease, 2/12 (17%) had partial response, and 11/12 (92%) had no local disease progression. The median time to first documented progression was 6.1 months with median OS of 6.8 months. There were 5 grade 4 hematologic events (lymphopenia), but no grade 4 nonhematologic toxicities. Moreover, ribonucleotide reductase levels were not predictive of outcomes. The authors concluded that triapine is well tolerated and may be useful with chemoradiation in future studies.

Amifostine

Amifostine is a cytoprotective agent that selectively protects normal tissue and scavenges free radicals and other reactive metabolites.⁹¹ Its utility in intrahepatic cancer was evaluated in a 2012 Phase I study by Feng et al⁹² to determine whether or not radiation dosing could be increased without significantly worsening radiation-induced liver disease. The authors enrolled 23 patients with diffuse intrahepatic malignancy—both primary and metastatic—and treated them with whole liver radiation to maximum dose 40 Gy with concurrent amifostine. They found that amifostine conferred a modest radioprotective effect in patients, particularly in primary liver cancer, up to 38 Gy; they could not estimate MTD in patients with liver metastases because of a few patients who previously received alkylating chemotherapy. Though they stopped short of saying amifostine would provide substantial benefit for whole liver radiation, the authors suggested that it could potentially allow higher radiation dosing for focal treatment.

IFN-α_2b

Interferon-alfa-2b (IFN-α_2b) has demonstrated radiosensitizing effects and synergy with other chemotherapeutic agents.⁹³ In a 2011 Phase II trial of adjuvant cisplatin, 5-FU, and IFN-α_2b with 50.4 Gy in 84 patients with pancreatic cancer, Picozzi et al⁹⁴ reported median OS and DFS of 25.4 months and 14.1 months, respectively. The OS at 18 months was 69% and it dropped to 59% at 2 years. Both median and 2-years survival rates compared favorably with those seen in other trials of chemotherapy or chemoradiation for pancreatic cancer.⁹⁵ No long-term toxicity or treatment-related deaths were observed, but they stated that 80/84 (95%) developed grade 3–4 toxicity events, including neutropenia and nausea. Thus, despite the encouraging efficacy seen in the trial and the absence of irreversible long-term toxicity, the safety profile of IFN-α_2b was of concern to the authors and they noted that future investigations should explore quality-of-life measures.

Ongoing Trials or Future Directions

In summary, there have been numerous recent studies on combined modality approaches involving radiotherapy in treating UGI malignancies, mostly in esophageal and pancreatic tumors, and results have been varied. Among the targeted therapies, nimotuzumab, trastuzumab, and erlotinib have shown promise in esophageal cancers and the last of these has had encouraging results in pancreatic ACA as well. Unfortunately, cetuximab, bevacizumab, panitumumab, and gefitinib have not shown survival benefits and can even have higher toxicity risks. Sorafenib has shown efficacy, albeit with poor safety in hepatic tumors. TNFerade has been shown to be safe, but without survival benefit, in both esophageal and pancreatic cancers. Lipoplatin, paclitaxel, and IFN-α_2b likely require further evaluation to determine safety and efficacy. The other agents—clivatuzumab, GM-CSF vaccine, and the rest—have shown good results in hepatobiliary and pancreatic tumors, but larger studies are needed to determine utility.

A recurrent theme in the literature is the failure to accrue enough patients to draw results that can influence current practice. This may be a reflection of fragmentation in the oncology community regarding trials for combined modality regimens as well as of the sheer array of novel agents to test. Though we concede that multicenter trials are challenging to initiate and carry out, it does appear that the time has come for larger, more collaborative endeavors to test the safety and efficacy of combined modality approaches. Moreover, we encourage further exploration of treatments for hepatobiliary and gastric malignancies as these were relatively underrepresented among the studies we reviewed compared to pancreatic and esophageal cancers. Even esophageal cancers can produce different outcomes based on histology, but all too often studies have not successfully included comparable numbers of SCC and ACA patients, perhaps from slow accrual. It may be easier to investigate these populations separately and thereby establish more definitive results for future reference. We also expect that the future would bring new studies investigating radiotherapy in conjunction with immunotherapies and checkpoint inhibitors for UGI neoplasms given the exciting developments seen in melanoma and non–small cell lung cancer. These agents represent a largely unknown quantity in UGI cancers and should be tested with radiation. Finally, we urge researchers to test other modalities of delivering radiation—proton beams, brachytherapy, or hypofractionation—in conjunction with systemic agents and different localized treatments like thermoablation, as these may yet yield surprising effects.

There are newer combined modality trials for UGI cancers (Table 3) besides the aforementioned NCT01196390. For instance, there is NCT01342224 at New York University wherein human telomerase vaccine is administered with GM-CSF, tadalafil, and 50.4 Gy (28 fractions) in locally advanced pancreatic ACA. The recently approved NCT02530437 trial at M.D. Anderson Cancer Center would examine the Hedgehog inhibitor taladegib with paclitaxel, carboplatin, and 50.4 Gy (28 fractions) in localized esophageal ACA and GEJ tumors. Overall, there is much optimism in combined modality approaches and we look forward to the results of the latest investigations.

Table 3.

Ongoing Trials Involving Systemic Approaches With Radiotherapy

Clinical Trial Identifier	Cancer Type	Regimen	Agent and Target	Outcomes of Interest
NCT01196390; ongoing, but no longer recruiting	HER-2 (+) esophageal ACA	Neoadjuvant paclitaxel, carboplatin, and 50.4 Gy (28 fractions)with or without trastuzumab	Trastuzumab, anti-HER-2/Neu MAB	Disease-free survival, pCR, and toxicity over 8 years. Estimated 591 people enrolled.
NCT02530437; not yet recruiting	Esophageal and GEJ ACA	Paclitaxel, carboplatin, and 50.4 Gy (28 fractions) with either concurrent or preceding taladegib	Taladegib, Hedgehog inhibitor	Toxicity and pCR over 10 weeks. Goal enrollment 66 patients.
NCT02375581; currently recruiting	Esophageal SCC and ACA	Concurrent icotinib with 50–60 Gy (25–30 fractions)	Icotinib, anti-EGFR TKI	Mortality and toxicity over 2 years. Goal enrollment 130 patients.
NCT02545751; not yet recruiting	Metastatic esophageal SCC and ACA	Concurrent thymalfasin with 25 Gy (5 fractions)	Thymalfasin, stimulates immune system proliferation and differentiation	Tumor response over 8 weeks and toxicity and OS over 2 years. Goal enrollment 29 patients.
NCT02381561; not yet recruiting	Advanced UGI and lower GI cancers	Ropidoxuridine with concurrent radiotherapy (dosing dependent on cancer site)	Ropidoxuridine, prodrug for nucleoside analog idoxuridine	MTD over 28 days. Goal enrollment 30 patients.
NCT02425605; currently recruiting	Advanced HCC	Intra-arterial 5-FU with 45 Gy (25 fractions) followed by sorafenib	Sorafenib, anti-VEGF-R TKI	OS, PFS, and pCR over 36 months. Goal enrollment 47 patients.
NCT01342224; active, but no longer recruiting	Locally advanced pancreatic ACA	Neoadjuvant telomerase vaccine (GV1001), followed by GM-CSF with gemcitabine and 50.4 Gy (28 fractions). Tadalaf l given throughout treatment and after surgery.	Telomerase vaccine (GV1001), peptide subunit of telomerase reverse transcriptase that can generate T-cell response to telomerase Tadalafl, phosphodiesterase inhibitor	Toxicity and tumor response over 180 days. 11 enrolled.
NCT02439593; not yet recruiting	Locally advanced pancreatic ACA	Neoadjuvant FOLFIRINOX followed by concurrent gemcitabine and 50.4 Gy (28 fractions) with or without weekly hyperthermia (40–41C)	Localized hyperthermia to 40–41C	1-year OS and safety. Goal 39 patients in each arm.
NCT02405585; currently recruiting	Borderline resectable pancreatic ACA	Neoadjuvant FOLFIRINOX followed by concurrent gemcitabine and 50.4 Gy (28 fractions) with monthly algenpantucel-l immunotherapy	Algenpantucel-L, immunotherapy with human pancreatic cancer cells that contain mouse gene, thereby inducing immune system to recognize and attack cancer cells as foreign	18-month OS, PFS, CR, and safety. Goal 48 patients total.
NCT02349867; currently recruiting	Untreated pancreatic ACA	Neoadjuvant gemcitabine and nab-paclitaxel followed by concurrent gemcitabine, vorinostat, sorafenib, and 50.4 Gy (28 fractions)	Vorinostat, histone deacetylase inhibitor Sorafenib, anti-VEGF-R TKI	Determine optimal dose and schedule of vorinostat and sorafenib over 18–36 months. Goal 36 patients total.