One of the major difficulties in allocating and comparing treatment modalities for patients with esophageal cancer is the lack of precise preoperative staging with resultant variations in the extent of disease and varying response criteria. Compared with adenocarcinoma, SCC appears to be more sensitive to chemotherapy, chemoradiation, or radiotherapy (RT); however, the long-term outcome is not different for patients with the two histologic types.^3,4 Unfortunately, most studies do not differentiate between patient groups with squamous versus adenocarcinoma of the esophagus, although it appears that the clinical, histologic, and oncologic differences between SCC and adenocarcinoma justify a differentiated therapeutic concept for these two tumor entities and distinct consideration in clinical reports.³

Over the past three decades, multiple studies evaluated the role of single-agent chemotherapy in the treatment of esophageal cancer.⁵ Cisplatin has been considered one of the most active agents, with a single-agent response rate consistently in the range of 20% or greater.⁶ Older agents considered to be active include 5-fluorouracil (5-FU), mitomycin, bleomycin, methotrexate, vindesine, mitoguazone, and doxorubicin.⁷ Newer agents that have shown activity include paclitaxel, docetaxel, etoposide, vinorelbine, irinotecan, capecitabine, and S-1.^{8,9,10,11,12,13,14}

Given that the response to single-agent chemotherapy is generally short-lived at 4 to 6 months, combination chemotherapy for metastatic carcinoma of the esophagus has been extensively studied over the past decades and continues to evolve especially with the inclusion of targeted therapies. Despite conflicting reports on the survival benefit of combination chemotherapy as compared with single-agent chemotherapy, a meta-analysis found a modest but statistically significant survival benefit of combination chemotherapy when compared with single-agent therapy.¹⁰

The combination of 5-FU plus cisplatin is considered the historic “standard” therapy for patients with carcinoma of the esophagus with reported response rates between 20% and 50%.⁷ Efforts have been made to improve upon this regimen by adding a third agent or substituting 5-FU or cisplatin. In a phase III trial (n = 580), Ross et al.¹⁵ found that the addition of epirubicin to cisplatin and protracted 5-FU (ECF) in metastatic and locally advanced, nonmetastatic disease patients with SCC and adenocarcinoma of the esophagus was superior to mitomycin, cisplatin, and protracted 5-FU combination in improving the quality of life but was similar in response rates (42.4% vs. 44.1%; P = .692) and median OS (9.4 vs. 8.7 months; P = .315). ECF resulted in more grade 3/4 neutropenia. A meta-analysis showed a modest survival benefit of around 1 month for adding epirubicin to cisplatin and fluorouracil.¹⁰ Van Cutsem et al.,¹⁶ in a phase III trial (n = 445) in patients with advanced gastric cancer, compared the combination of docetaxel, cisplatin, and a shorter infusional course of 5-FU (DCF) to cisplatin and 5-FU (CF). The addition of docetaxel significantly improved time-toprogression (TTP) and OS but at the expense of toxicity with grade 3/4 neutropenia seen in 82% versus 57% of patients. In the phase III REAL-2 study (n = 1,000), Cunningham et al.¹⁷ compared four regimens (triplet therapy with ECF or ECX, or triplet therapy with epirubicin and oxaliplatin plus either fluorouracil (EOF) or capecitabine (EOX) in patients with untreated advanced esophagogastric cancer and demonstrated that EOX was noninferior to ECF with a statistically nonsignificant improvement in median OS of 1 month compared with the other treatment arms. Nonfluorouracil-containing regimens include irinotecan combined with weekly cisplatin (response rates 57%, median actuarial survival was 14.6 months)¹⁸ or combined with docetaxel and cisplatin (overall response rates 54% and median survival was 11.9 months).¹⁹ Nonplatinumcontaining regimens include irinotecan-containing regimens in combination with a taxane (response rates up to 30%).^20,21 The combination of docetaxel and vinorelbine was also found to be effective in patients with esophageal cancer (overall response rate was 60%).²²

The benefit obtained from combination therapy has been limited. This has led to a search for more effective therapies
based on the understanding of the underlying molecular mechanisms of esophageal cancer. Given the success of targeted therapies in other cancers, a number of targeted therapies are currently under study.

Agents targeting epidermal growth factor receptor (EGFR) include small-molecule tyrosine kinase inhibitors (TKI such as erlotinib and gefitinib) and monoclonal antibodies (cetuximab). In phase II trials, TKIs showed modest activity but on subanalysis the outcome was significantly better in female patients, patients demonstrating high EGFR expression, or squamous histology.²³ In the FOLCETUX phase II study (n = 38), cetuximab was added to 5-FU, leucovorin, and irinotecan (FOLFIRI) in patients with advanced gastric or gastroesophageal adenocarcinoma. The overall response rate was 44.1% and median TTP was 8 months. Grade 3/4 toxicity included neutropenia (42.1%), acne-like rash (21.1%), and diarrhea (7.9%).²⁴ Agents targeting vascular endothelial growth factor (VEGF), such as bevacizumab, have been added to combination chemotherapy in a number of phase II trials. The addition of bevacizumab to irinotecan/cisplatin was associated with an overall response rate of 65% and a median survival of 12.3 months,²⁵ and the addition to modified docetaxel, cisplatin, and fluorouracil was associated with a response rate of 67% and a median OS of 16.8 months.²⁶ Targeting the HER2/neu receptor with agents, such as the monoclonal antibody trastuzumab, was associated with a better response rate and median OS (13.8 vs. 11.1 months; P = .0046), when studied in combination with chemotherapy for first-line treatment of HER2-positive advanced gastric or gastroesophageal junction (GEJ) cancers in the phase III ToGA trial (n = 594).²⁷

Multiple management modalities have been utilized for the treatment of esophageal cancer because of the overall poor survival rates and an operative mortality rate of up to 10% of patients who have been treated with either resection alone^28,29 or RT alone.^30,31 As a result, the focus of many current clinical trials involves the use of chemotherapy in conjunction with RT and surgical resection. At least three combined modality alternatives have been explored, including induction or neoadjuvant chemotherapy followed by surgery, preoperative chemoradiotherapy, and chemoradiotherapy without surgery.

Neoadjuvant chemotherapy has been studied in a number of clinical trials. The results of an intergroup trial (Intergroup 0113), in which patients with potentially resectable carcinoma of the esophagus of both histologic types were randomly assigned to receive either preoperative chemotherapy with 5-FU plus cisplatin or surgery alone, did not demonstrate any survival benefit among the patients who received preoperative chemotherapy.⁴ The Medical Research Council (MRC) published their trial, which involved 802 patients with potentially resectable carcinoma of the esophagus.³² In this trial, patients were randomly assigned to receive either two cycles of preoperative 5-FU (1,000 mg/m²/day by continuous infusion for 4 days) and cisplatin (80 mg/m²/day on day 1) repeated every 21 days followed by surgery, or surgery alone. However, this trial had several clinical methodology problems. Approximately 10% of patients received off-protocol preoperative RT, and patients accrued in China were excluded. At a short median follow-up time of 2 years, there was a 3.5-month survival time advantage (16.8 vs. 13.3 months) for the group treated with preoperative chemotherapy. At a median follow-up of 6 years, the OS benefit of the chemotherapy arm had fallen but remains statistically significant. Tumor histology did not influence the effect of chemotherapy.³³ In the MAGIC trial (n = 503), patients with adenocarcinoma of the stomach and GEJ or distal esophagus were randomly assigned to perioperative ECF, three cycles pre- and postoperatively, or to surgery alone. The 5-year OS was improved by 13% (36% vs. 23%; P < .001).³⁴ The resected tumors were significantly smaller and less advanced in the perioperative-chemotherapy group but there was no improvement in the rate of curative resection, and no pathologic complete responses were observed.

Two recent trials evaluated the role of preoperative cisplatin/fluorouracil followed by surgery or to surgery alone. The results were mixed with only one of them showing a benefit of the chemotherapy arm.^35,36 However, among all these mixed results, a recent meta-analysis (n = 1724)³⁷ have shown that the hazard ratio (HR) for neoadjuvant chemotherapy was 0.90 (0.81 to 1.00; P = .05), with a 2-year absolute survival benefit of 7%. There was no significant effect on all-cause mortality of chemotherapy for patients with squamous histology (HR = 0.88 [0.75 to 1.03]; P = .12), but there was a significant benefit for those with adenocarcinoma (0.78 [0.64 to 0.95]; P = .014).

Currently, there is no standard postoperative systemic chemotherapy for patients who have undergone an R0 resection.

In an attempt to avoid perioperative mortality and to relieve dysphagia, definitive radiation therapy (RT) in combination with chemotherapy has been studied in phase II and III trials. An Intergroup randomized trial, Radiation Therapy Oncology Group (RTOG) 85-01, of chemotherapy and RT versus RT alone resulted in an improvement in 5-year survival for the combined modality group (27% vs. 0%). Eight-year follow-up of this trial demonstrated an OS rate of 22% for patients receiving chemoradiation therapy.³⁸ In an attempt to improve upon the results of RTOG 85-01, Intergroup 0123 (RTOG 94-05) randomized 236 patients with localized esophageal tumors to chemoradiation with high-dose RT (64.8 Gy), and four monthly cycles of 5-FU and cisplatin versus conventional-dose RT (50.4 Gy) and the same chemotherapeutic schedule. For the 218 eligible patients, no significant difference was observed in median survival (13.0 vs. 18.1 months), 2-year survival (31% vs. 40%), and local/regional failure or local/regional persistence of cancer (56% vs. 52%) between the high-dose and standard-dose RT arms.³⁹ Recent phase II trials have used more novel agents such as paclitaxel-based chemotherapy,^40,41 docetaxel-based,⁴² oxaliplatin-based,⁴³ or irinotecan-based^44,45 chemotherapy.

A number of phase III randomized studies compared preoperative chemoradiotherapy with surgery for esophageal cancer with mixed results.^{46,47,48,49,50,51} Two recent meta-analyses assessed 9 randomized controlled trials (n > 1,000) and 10 randomized controlled trials (n = 1,209) comparing neoadjuvant chemoradiation and surgery with surgery alone for esophageal cancer. The first meta-analysis found that neoadjuvant chemoradiation
and surgery improved 3-year survival and reduced local-regional cancer recurrence when compared with surgery alone.⁵² The second meta-analysis found that the HR for all-cause mortality with neoadjuvant chemoradiotherapy versus surgery alone was 0.81 (95% CI, 0.70 to 0.93; P = .002), corresponding to a 13% absolute difference in survival at 2 years, with similar results for different histologic tumor types: 0.84 (0.71 to 0.99; P = .04) for SCC, and 0.75 (0.59 to 0.95; P = .02) for adenocarcinoma.³⁷ In conclusion, for esophageal and GEJ cancers local control appears to be better with preoperative chemoradiotherapy compared with surgery alone. In addition, concomitant chemoradiotherapy followed by surgery appears to provide a survival benefit compared with surgery alone.

Advanced gastric cancer is incurable, but small studies have suggested that quality of life and OS can be modestly improved with chemotherapy.^53,54,55,56 A meta-analysis confirmed the OS benefit of chemotherapy over best supportive care (HR = 0.39; 95% CI, 0.28 to 0.52).¹⁰ There are few adequately studied chemotherapeutic agents that have demonstrated a significant single-agent activity in advanced gastric cancer; these agents include 5-FU,⁵⁷ mitomycin,^58,59 etoposide,⁶⁰ and cisplatin.⁶¹ Therefore, in an attempt to improve response rates, several 5-FU-based combination regimens have been studied in advanced gastric cancer. In the early 1980s, the FAM (5-FU, doxorubicin, and mitomycin) regimen was considered to be the gold standard therapy for patients with advanced gastric carcinoma.⁶² In a pivotal study performed by the North Central Cancer Treatment Group (NCCTG),⁶³ the FAM regimen was compared with 5-FU as a single agent and 5-FU plus doxorubicin. No significant survival difference was detected among patients treated with these three regimens. However, response rates were higher with combination chemotherapy than with 5-FU alone. Several randomized studies comparing FAM with FAMTX (5-FU, Adriamycin, and methotrexate with leucovorin rescue),⁶⁴ FAMTX with ECF,⁶⁵ and FAMTX with ELF (etoposide, leucovorin, and 5-FU) versus 5-FU plus cisplatin⁶⁶ have been reported in the 1990s.

Over the past decade, newer regimens have emerged showing benefit with combination chemotherapy. In a phase III trial (n = 445), the addition of docetaxel to cisplatin and 5-FU (DCF) significantly improved TTP, OS (9.2 vs. 8.6 months; P = .2), and response rate in untreated gastric cancer patients, but resulted in increased toxicity.¹⁶ The combination of irinotecan and 5-FU was compared with cisplatin and 5-FU and did not yield a significant TTP or OS difference.⁶⁷ In an attempt to decrease the toxicity of regimens containing cisplatin, Al-Batran et al.⁶⁸ compared the combination of 5-FU, leucovorin, and oxaliplatin (FLO) with 5-FU, leucovorin, and cisplatin (FLP) in patients with advanced gastric cancer. The main benefit was seen in the subset of patients older than 65 years, where treatment with FLO resulted in significantly superior PFS (6.0 vs. 3.1 months; P = .029) and OS (13.9 vs. 7.2 months) as compared with FLP. As described earlier (esophageal cancer), the phase III REAL-2 study, comparing four regimens (ECF, ECX, EOF, and EOX) in patients with untreated advanced esophagogastric cancer, has demonstrated that EOX was noninferior to ECF with a statistically nonsignificant improvement in median OS of 1 month compared with the other treatment arms.¹⁷ Kang et al.⁶⁹ conducted a phase III noninferiority study (n = 316), where capecitabine was compared with continuous infusion 5-FU, both in combination with cisplatin. The combination of capecitabine/cisplatin was noninferior to cisplatin/5-FU in PFS and OS. In the phase III ToGA trial (n = 594), the monoclonal antibody trastuzumab was associated with a better response rate and medial OS (13.8 vs. 11.1 months; P = .0046), when studied in combination with chemotherapy for first-line treatment of HER2-positive advanced gastric cancer.²⁷

In conclusion, combination chemotherapy is superior to best supportive care and single-agent chemotherapy, but the benefit is limited with a median survival of <12 months. The addition of targeted therapy in the subset of patients with HER2/neu positive tumors improved the median survival to beyond 12 months. Outside of clinical trials, the chemotherapy backbone for advanced gastric carcinoma remains cisplatin/5-FU-based combination chemotherapy. However, several other drugs, such as docetaxel, capecitabine, or oxaliplatin, can be used with noninferior results.

Until 1999, studies had not shown better outcomes with the addition of chemotherapy and radiation to surgery.⁷⁰ However, nonrandomized data have suggested that preoperative induction chemotherapy followed by chemoradiotherapy yields a substantial pathologic response that results in durable survival time.⁷¹ The landmark trial is the Intergroup trial INT-0116.⁷² Eligibility included patients with T3 and/or node positive (N+) adenocarcinoma of the stomach or GEJ. After a resection with negative margins, 603 patients were randomly assigned to either observation alone or postoperative-combined modality therapy consisting of five monthly cycles of bolus chemotherapy with 45 Gy concurrent with cycles two and three. There was a significant decrease in local failure as the first site of failure (19% vs. 29%) as well as an increase in median survival (36 vs. 27 months), 3-year relapse-free survival (48% vs. 31%), and OS (50% vs. 41%; P = .005) with combined modality therapy. A criticism of this study was that many patients received suboptimal lymph node dissections. Nevertheless, adjuvant chemoradiation has become the standard of care in US patients. In 2006, Cunningham et al.³⁴ reported the results of the MAGIC trial (n = 503), where patients with adenocarcinoma of the
stomach and GEJ or distal esophagus were randomly assigned to perioperative ECF, three cycles pre- and postoperatively, or to surgery alone. The 5-year OS was improved by 13% (36% vs. 23%; P < .001) but despite having significantly smaller resected tumors, there was no improvement in the rate of curative resection in the perioperative chemotherapy group, and no pathologic complete responses were observed. In conclusion, patients with locally advanced disease should receive some form of perioperative therapy, either perioperative chemotherapy or postoperative chemoradiation.

Chemotherapy has generally been ineffective in extensive hepatocellular carcinoma and this fact is well reflected by the large number of available phase I and II protocols for unresectable hepatic cancer. Single-agent doxorubicin may produce response rates of <20% but no prolongation of survival.⁷³ In a phase III trial (n = 188), single-agent doxorubicin has been compared with combination chemotherapy of cisplatin, interferon, doxorubicin, and 5-FU (PIAF). Despite increased treatment-related toxicity, PIAF had a statistically nonsignificant improved response rate and survival rate. The response rate in the single-agent doxorubicin arm was 10.5%.⁷⁴ Other agents that have produced marginal responses include docetaxel, gemcitabine, 5-FU, interferon α, mitoxantrone, cisplatin, and etoposide.^75,76,77,78 Combination chemotherapy generally did not augment single-agent cytotoxic efficacy but only increased toxicity.

Following the disappointing results from studies evaluating the role of systemic chemotherapy, investigators focused their efforts toward targeting the underlying molecular pathways implicated in the pathogenesis of hepatocellular carcinoma. The SHARP trial was a multicenter, phase III, double-blind, placebo-controlled trial (n = 602), where patients with advanced hepatocellular carcinoma and Child-Pugh A cirrhosis where randomized to receive either sorafenib (a multityrosine kinase inhibitor) at a dose of 400 mg twice daily or placebo. The study showed an improvement in survival (10.7 vs. 7.9 months for placebo; HR = 0.69; P = .00058). The study did not show any difference in time to symptomatic progression (P = .77). Grade 3 and 4 toxicities include diarrhea and hand-foot skin reaction.⁷⁹

Intraarterial fluorodeoxyuridine (FUDR) with mitomycin or leucovorin, doxorubicin, and cisplatin yielded response rates of 40% and median survivals of >1 year.^80,81 Nevertheless, the true impact of regional chemotherapy in these uncontrolled trials is difficult to determine as many of these studies also involved embolization, ethanol injection, or ligature of the hepatic artery. A variety of treatment approaches are under active investigation, including intraarterial radiolabeled lipiodol, to prevent tumor recurrence after curative resection.^82,83 In conclusion, for patients who are not eligible for resection or liver transplantation because of the severity of underlying liver disease, other treatment options include local nonsurgical methods of tumor ablation (RFA, PEI), transarterial chemoembolization, RT, and systemic therapy. A general approach for selecting patients can be based on Child-Pugh scoring system according to underlying liver disease.

Cancer of the exocrine pancreas is rarely curable and has a 5-year relative survival of 6%.¹ The highest cure rate occurs if the tumor is truly localized to the pancreas; however, this stage of the disease accounts for <20% of cases and usually results from a tumor in the pancreatic head, causing early jaundice, or the incidental finding of a tumor in the body or tail on a computed tomography (CT) scan done for other reasons. Hence, patients with any stage of pancreatic cancer can appropriately be considered candidates for clinical trials because of the poor response to chemotherapy, RT, and surgery as conventionally used. The identification of active agents is complicated by a variety of patient-related factors, such as nutritional deficiencies and poor performance status, which may make the administration of aggressive chemotherapy difficult and hamper attempts to define and compare response rates.