Cocktail without hangover: in search for the optimal chemotherapy in the combined management of non-operable esophageal carcinomas

Abstract

Background: The worldwide incidence of esophageal cancer has greatly increased over the past few decades making it the sixth deadliest cancer. The disease is often detected in advanced stages when surgery is no longer an option. The standard treatment in these situations is combined chemoradiotherapy, by employing drug cocktails that lead to optimal treatment outcomes both from the perspective of tumor control and normal tissue toxicity.

Methods: The aim of this work was to collate the existing trials and clinical studies reported on non-operable esophageal cancer and to analyze the results based on treatment outcomes after various drug combinations.

Results: Of all drug combinations, cisplatin/5-FU is the most well established chemotherapy regimen for esophageal cancer as both neoadjuvant therapy, an alternative option to surgery, and for palliative purposes. Although this regimen is associated with the most toxicity, it also appears to have the best survival benefit and relief of symptoms.

Conclusions: More research is warranted to further increase the therapeutic ratio in non-operable esophageal cancers.

Introduction

According to the GLOBOCAN 2012 report, the worldwide incidence of esophageal cancer has increased greatly over the past few decades, thus becoming the eighth most common cancer worldwide [1]. The two types of esophageal cancer: adenocarcinomas and squamous cell carcinomas have different risk factors and incidences. While adenocarcinomas are predominantly found in developed countries (Western Europe, Australia and North America), with a typical patient being a bariatric white male with reflux, squamous cell carcinomas are more common in poorly nourished smokers with a history of alcohol abuse or betel quid chewing, of whom there is a higher prevalence in developing countries (South-eastern Africa, Southern Russia and Asia) [2]. Other risk factors associated with squamous cell carcinomas are chronic irritation and inflammation, or a diet high in salt and nitrosamine carcinogens. There are also some predisposing genetic as well as medical conditions that can lead to esophageal cancer such as: achalasia, human papillomavirus, Plummer–Vinson syndrome and tylosis. Adenocarcinomas are also associated with Barrett’s esophagus. Another plausible risk factor is radiotherapy to the mediastinum for the treatment of breast cancer 10 years beforehand [3,4].

Esophageal cancer has a very poor prognosis and 5-year survival rarely exceeds 40% [5]. Predictions for locally advanced carcinomas are even worse, as survival is usually around 15% [6]. This is due to the fact that many carcinomas are not detected until late stages, allowing the tumor to advance locally, invade nearby organs, metastasize into lymph nodes, or spread hematologically into distant organs such as the liver and lung.

In advanced stages, the esophageal cancer is often non-operable, thus the patient relies on the success of combined chemoradiotherapy. As shown by the long-term follow-up analysis of a prospective randomized trial, combined therapy significantly improves overall survival as compared to radiotherapy alone (26% vs 0% 5-year overall survival) [7].

While several chemotherapeutic agents have shown successful responses when combined with radiation, side effects are often limiting factors. It is, therefore, crucial to design treatments that limit such toxicities by using the optimal drug cocktail of all available combinations trialed to date.

Aim of the article and search strategy

The aim of this work is to collate the existing trials and clinical studies reported for non-operable esophageal cancer and to analyze the results based on treatment outcome after various drug combinations. Median overall survival is the main outcome examined. While the authors aimed to present 5-year survival rates, many studies included only two or three-year survival rates in their reports. Response rates, progression-free survival and toxicity levels are also considered when reviewing the treatments.

A search on Medline was conducted, and the search strategy and results are presented in Supplementary Table 1. Search limits included English publications from 2000 onwards, and clinical trials only. Furthermore, since this article was written with the intent of examining multimodal therapies, the search strategy did not include radiotherapy alone as a treatment modality. As the aim of this review was to tackle the challenge of non-operable esophageal cancers, trials involving surgery have not been considered.

Table 1. Trials and outcomes for cisplatin/5-fluorouracil therapies.

Study, year, phase	Chemotherapy	Radiotherapy	No. of patients(m-male;f-female)	Histology	Median survival	Overall survival	Progression-free survival	Comments
Gaspar (2000) [10] Phase 1/2	CDDP 75 mg and 5-FU 1000 mg once a week on weeks 1, 5, 8 and 11	External beam radiation 50 Gy/25 fractions weeks 1–5, brachytherapy 15–20 Gy week 8	49 32 m 17 f	45 SCC 4 AC	11 months	1 year: 49% 2 years: 31% 3 years: 29%	–	Life-threatening toxicity occurred in 24% patients; treatment-related death in 10% patients. Caution is required with brachytherapy.
Nemoto (2001) [11] Pilot study	CDDP 5 mg daily until radiotherapy was completed	60 Gy	38	Stage I: 3 Stage II: 13 Stage III: 13 Stage IV: 9	12 months	1 year: 51% 2 years: 19% 5 years: 8%	–	Manageable toxicity. Probably effective. More trials needed, dose needs to be raised.
Waters (2002) [13] Phase 2	CDDP 60 mg once a week on week 1, 4, 7, 10 and 13 5-FU 300 mg every day for week 1–12, then 200 mg/day for week 13–18	Week 13–18 54 Gy/30 fractions	72 51 m 21 f	43 SCC 27 AC	14.6 month	1 year: 58.7% 2 years: 44.4%	–	Some patients underwent surgery. Grade 3 or 4 toxicity in 38% of patients. Feasible and active treatment for palliation.
Harney (2003) [13] Phase 1/2	CDDP 75 mg day 1 5-FU 1000 mg day 2–5 One cycle only	Concurrent CHARTWEL (continuous hyperfractionated accelerated RT weekend-less) from 40.5 to 49.5 Gy/1.5 Gy fractions thrice daily	19 10 m 9 f	13 SCC 6 AC	10.7 months	1 year: 50% 2 years: 26%	9.8 months 1 year: 41% 2 years: 28%	Feasible, well tolerated. Relieves dysphagia. Needs dose escalation to 51 Gy.
Crosby (2004) [14] Retrospective	CDDP 60 mg once a week on weeks 1, 4, 7 and 10 5-FU 300 mg/day on week 1, 4, then 225 mg on week 7, 10	External beam 50Gy/25 fractions on week 7–11	73 49 m 41 f	43 SCC 42 AC	26 months	2 years: 51% 5 years: 26%	18 months	50% had grade 3 or 4 toxicity; manageable, feasible and tolerable.
Ishida (2004) [15] Phase 2	CDDP (70 mg on days 1, 29) 5-FU (700 mg on days 1–4 and 29–32)	60 Gy/30 fractions on days 1–21 and 29–49	60	SCC	10.2 months	2 years: 31.5%	–	Response rate is not promising, but there are fewer toxicities
Brunner (2008) [16] Phase 2	CDDP 20 mg day 1–5, 29–33 5-FU 800 mg day 1–5, 29–33	Concurrent 50.4 Gy/28 fractions, 14.4Gy/12 fractions week 4–6, 12 Gy/2 fractions brachytherapy week 4, 6	50 40 m 10 f	45 SCC 3 AC	16 months	1 year: 61% 2 years: 29% 3 years: 18%	15.6 months 1 year: 58% 2 years: 33% 3 years: 17%	Acceptable toxicity, excellent survival for node-free disease
Nishimura (2009) [17] Phase 2	CDDP 70 mg day 1, 29 5-FU 700 mg day 1–5, 29–33	60 Gy/30 fractions weeks 1–3, 5–7	46 41 m 5 f	45 SCC 1 AC	–	2 years: 46% 5 years: 35%	2 years: 30% 5 years: 30%	Was later followed up in 2012.
	CDDP 7 mg days 1–5, 8–12, 29–33, 36–40 5-FU 250 mg days 1–14, 29–42	60 Gy/30 fractions weeks 1–3, 5–7	45 41 m 4 f	45 SCC 0 AC	–	2 years: 44% 5 years: 24%	2 years: 29% 5 years: 12%	Low dose protracted infusion – no benefit.
Kato (2009) [18] Phase 2	CDDP 70 mg day 1, 29 5-FU 700 mg day 1–4, 29–32	60 Gy/30 fractions	72 66 m 6 f	SCC Stage I	Unavailable >48 months	4 years: 80.5%	4 years: 68%	Esophagectomy for those with residual tumor. No grade 4 toxicity. Suitable alternative to surgery.
Kato (2011) [19] Phase 2	CDDP 40 mg days 1–8 5-FU 400 mg days 1–5, 8–12 Repeated every five weeks.	60 Gy/30 fractions	76 68 m 8 f	SCC	29 months	3 years: 44.7% 5 years: 36.8%	12 months 3 years: 32.9% 5 years: 25.6%	Effective non-surgical option. Several late toxicities – must be monitored.
Nishimura (2012) [20] Phase 2	CDDP 70 mg days 1, 29 5-FU 700 mg days 1–5, 29–33	60 Gy/30 fractions	46 41 m 5 f	45 SCC 1 AC	–	2 years: 46% 5 years: 35%	2 years: 38% 5 years: 30%	This was a long-term follow-up study. 17% late grade 3/4 toxicity
	CDDP 7 mg days 1–5, 8–12, 29–33, 36–40 5-FU 250 mg days 1–14, 29–42	60 Gy/30 fractions	45 41 m 4 f	45 SCC 0 AC	–	2 years: 44% 5 years: 22%	2 years: 34% 5 years: 21%	Eighteen percent late grade 3/4 toxicity. Low-dose protracted infusion is not superior.
Yamashita (2014) [21] Phase 2	CDDP 80 mg on day 1 5-FU 800 mg on day 1–4	IFRT 50.4 Gy/25 fractions	63 55 m 8 f	59 SCC 4 AC Stage I–IV	38.4 months	3 years: 51.1%	34.6 months	Effective treatment.
Ma (2014) [22] Phase 2	–	62–70 Gy/31–35 fractions	49 32 m 17 f	SCC	19 months	1 year: 69.4% 3 years: 28.6% 5 years: 8.2%	11.4 months	No difference in 5YSR or late toxicity.
	CDDP 30 mg weekly for five weeks	62–70 Gy/31–35 fractions	49 35 m 14 f	SCC	35 months	1 year: 85.7% 2 years: 46.9% 5 years: 14.3%	12.2 months	Greater response, more acute toxicity
Shinoda (2015) [23] Phase 2/3	CDDP 70 mg days 1, 29 5-FU 700 mg/m^2 days 1–4, 29–32	60 Gy/30 fractions	71 67 m 4 f	SCC	14.4 months	1 year: 56.3% 3 years: 25.7%	–	T4 cancer or non-operable metastatic lymph node
	CDDP 4 mg 5-FU 200 mg days 1–5 every week	60Gy/30 fractions	71 61 m 10 f	SCC	13.1 months	1 year: 55.9% 3 years: 25.9%	–	Low-dose CRT was not tested for Phase 3 trial.
Minsky (2002) [24] Phase 3	CDDP 75 mg and 5-FU 1000 mg day 1–4, week 1, 7 and 11	Concurrent 64.8 Gy/36 fractions	109 71 m 38 f	95 SCC 14 AC	13.0 months	2 years: 31%	–	No difference in survival, but more treatment mortality.
	CDDP 75 mg and 5-FU 1000mg day 1–4, week 1, 5 and 9	Concurrent 50.4 Gy/28 fractions	109 83 m 26 f	92 SCC 17 AC	18.1 months	2 years: 40%	–	50.4 Gy should remain standard.
Zhao (2005) [25] Phase 3	–	Concurrent 41.4 Gy/23 fractions, followed by LCAF 27 Gy/18 fractions twice daily	57 36 m 21 f	SCC	23.9 months	1 year: 77% 3 years: 39% 5 years: 28%	–	Grade 3 or 4 toxicity in 46%
	CDDP 25 mg and 5-FU 600 mg on days 1–3, weeks 1, 5, 8 and 11	Concurrent 41.4 Gy/23 fractions, followed by LCAF 27 Gy/18 fractions twice daily	54 42 m 12 f	SCC	30.8 months	1 year: 67% 3 years: 44% 5 years: 40%	–	Grade 3 or 4 toxicity in 25%. Better survival but higher mortality rate from chemotherapy toxicities
Crehange (2007) [26] Phase 3	CDDP 15 mg day 1–5 or 75 mg day 2 Fluorouracil 800 mg day 1–5 Two cycles on week 1 and week 3	46 Gy/23 fractions over 4.5 weeks	161 148 m 13 f	72 SCC 16 AC	18.4 months	2 years: 37.1%	2 years: 76.7%	Protracted RT had better control.
	CDDP 15 mg day 1–5 or 75 mg day 2 Fluorouracil 800 mg day 1–5 Two cycles on week 1 and week 3	30 Gy/10 fractions week 1, 4	285 264 m 21 f	158 SCC 13 AC	15.6 months	2 years: 30.5%	2 years: 56.8%	Split course RT not as effective
Kumar (2007) [27] Phase 3	–	50 Gy/25 fractions followed 1–2 weeks later by 12 Gy/2 fractions brachytherapy	60 49 m 11 f	SCC	7.1 months	1 year: 32.3% 2 years: 22.8% 5 years: 13.7%	–	Later RT was modified to 66 Gy/33 fractions
	CDDP 35 mg weekly on week 1–6 or 1–7	50 Gy/25 fractions followed 1–2 weeks later by 12 Gy/2 fractions brachytherapy	65 43 m 22 f	SCC	13.4 months	1 year: 57.6% 2 years: 38.9% 5 years: 24.8%	–	Improvement in survival – manageable toxicities.

The search strategy is shown in the appendix (Supplementary Table 1). Three hundred and forty seven abstracts have been reviewed. Excluding manuscripts that reported results on Phase 1 trials only, 155 articles have been extracted and read.

The trials are categorized based on the drug combinations used and the results presented in the tables are listed in chronological order for a better follow-up of dosage and schedule changes over the years.

Because of the large variations in the design and outcome reporting of reviewed clinical trials, statistical analysis of data has not been possible. Consequently, this manuscript is an integrative research review that summarizes and evaluates the accumulated current status of knowledge for treatment of non-operable esophageal cancer.

Chemoradiotherapy of non-operable esophageal cancer

Chemoradiotherapy is the most common treatment for non-operable cancers. The role of chemotherapy is to radio sensitize the tumor and also to manage the systemic disease. Cancer chemotherapy strives to arrest tumor progression by causing a lethal cytotoxic event or apoptosis in the cancer cell. Ideally, chemotherapy drugs should interfere only with cellular processes that are unique to malignant cells. Unfortunately, most currently available anticancer drugs affect both normal and abnormal proliferating cells, resulting in a steep dose-response curve for both toxic and therapeutic effects.

The dose prescribed for esophageal cancer patients undergoing external radiotherapy alone is generally 65 Gy. This dose decreases to 50 Gy when radiation is combined with chemotherapy due to tissue sensitization. There is a limited number of studies that have examined the efficacy of radiation techniques in esophageal cancer, as opposed to the efficacy of new chemotherapy drugs [8,9]. Thus, more research is needed in order to reach a conclusion about the effectiveness and safety of modern radiotherapy techniques.

Cisplatin/5-fluorouracil

CDDP/5-FU is a well-established drug combination for both curative and palliative purposes. The studies presented in Table 1 have investigated the efficacy of CDDP/5-FU with or without RT in non-operable esophageal cancers.

While cisplatin is one of the most potent chemotherapeutic agents when combined with radiation, normal tissue toxicity remains a dose-limiting factor. Commonly, cisplatin is administered in short-term high doses (60–75 mg/m²) rather than daily low doses (4–5 mg/m²). Nevertheless, the latter schedule is an alternative due to the potentially decreased adverse events (supported by the low dose) and increased tumor control (supported by the radiosensitizing effect of daily cisplatin) [28,29].

Several Phase 2 studies examined the effect of the combined CDDP/5-FU and RT and found that cisplatin-based chemoradiotherapy was effective in locally advanced non-operable tumors [10–16,18,19,21]. Median survival ranged from 10 to 38 months, with manageable toxicities. However, dosing schedules of both chemotherapy and RT varied greatly, as the optimum dose has not been established and some patients lowered or deferred their doses due to toxicities.

Nishimura et al. [20] conducted a randomized Phase 2 trial with long-term follow-up comparing low-dose protracted infusion CDDP/5-FU against short-term higher doses. One group received CDDP 70 mg/m² on days 1 and 29, the other CDDP 7 mg/m² from days 1–5, 8–12, 29–33 and 36–40. There was no difference in late toxicities (17% vs 18%) and 5-year survival rate was inferior (35% vs 22%) [20]. Similarly, Shinoda et al. [23] confirmed in a Phase 3 trial that low dose protracted chemotherapy (4 mg/m²) offers neither survival advantage nor decreased toxicity when compared to standard high dose short-term (70 mg/m²) chemotherapy. In their study, there was no difference in either median overall survival (14.4 months vs 13.1 months) or 3-year survival rate (25.7% vs 25.9%, HR 0.94, 95% CI), and the trial was ended before it could be suitably completed [23].

In a Phase 3 study Zhao et al. [25] has randomized patients with squamous cell carcinomas of the esophagus to receive either late course accelerated hyperfractionated radiotherapy alone or radiotherapy combined with CDDP and 5-FU. Fifty seven patients received a total of 68.4 Gy of radiation, while another 54 have been administered an additional dose of CDDP 25 mg/m² and 5-FU 600 mg/m² on days 1–3 during weeks 1, 5, 8 and 11. This resulted in an increased median survival from 23.9 months to 30.8 months, and a 5-year survival rate from 28 to 40% (p = .31). Larger randomized studies are needed to confirm the findings and additional measures are necessary to manage treatment-related toxicities in the chemotherapy group [15]. Similarly, Kumar et al. [27] conducted a Phase 3 study of radiotherapy compared with cisplatin-based chemoradiotherapy for non-operable SCC patients. The addition of CDDP (35 mg/m² weekly from weeks 1–6) to a total of 62 Gy resulted in an increased median survival from 7.1 months to 13.4 months, and a 5-year survival rate from 13.7 to 24.8% (HR 0.65, p = .038). Toxicities were manageable [27].

Overall, CDDP/5-FU is the most well established chemotherapy regimen for esophageal cancer. Although this regimen is associated with the most toxicity, it also appears to have the best survival benefit and relief of symptoms (particularly dysphagia). Most of these studies have focused on SCC, and dosing schedules have varied widely. Better-designed trials are required for more conclusive results.

Carboplatin/paclitaxel

Despite cisplatin’s potential for radiosensitization toward better tumor control, normal tissue toxicity is often a limiting factor. Carboplatin is part of the same platinum drugs family: however, it has less adverse effects than cisplatin. Often administered in combination with paclitaxel, carboplatin-based chemotherapy appears to have high treatment compliance (see Supplementary Table 2). Therefore, treatment guidelines in Europe and the United States recommend the administration of cisplatin/5FU as standard chemotherapy and the use of carboplatin/paclitaxel as an alternative for patients with extensive comorbidity [30].

Most trials have found that carboplatin and paclitaxel are well tolerated, with only few grade 3 toxicities and almost no grade 4 toxicities. Few studies found unacceptably severe toxicities, which may have been due to the high doses in Jatoi et al.’s trial [31], or to the added toxicity when combined with continuous 5-FU in Seung et al.’s [32].

In a retrospective study, Honing et al. [33] compared two groups of patients: one group consisting of 47 patients received CDDP/5-FU, and the other group of 55 patients received carboplatin and paclitaxel. Both groups underwent 46.8–70 Gy of concurrent radiotherapy in 1.8–2 Gy fractions. While the carboplatin/paclitaxel group had a slightly decreased median overall survival (13.8 months as opposed to 16.1 months) the difference is not significant and it is surpassed by the fact that more patients completed the treatment due to lower toxicities [33]. Jatoi et al. [31] found the median overall survival to be 21.2 months, which is still poorer than CDDP/5-FU.

Induction chemotherapy with this drug combination has not been particularly beneficial. Ajani et al. found that inducing a patient with paclitaxel, CDDP and 5-FU somewhat improves survival, but the one-year survival rate fell short of the target of 77.5% [34]. Another trial found that an induction cycle of docetaxel, CDDP and 5-FU, followed by a chemoradiotherapy regimen consisting of carboplatin, did not improve survival rates and there were too many side effects [35].

Thus, carboplatin/paclitaxel does not increase survival in patients with non-operable esophageal cancer, but is associated with less adverse effects. Nevertheless, for resectable tumors, according to the CROSS trial (Chemoradiotherapy for esophageal cancer followed by Surgery study) neoadjuvant chemoradiotherapy (carboplatin + paclitaxel) followed by surgery should be considered as standard of care [5]. This recommendation is based on the superior results obtained for the neoadjuvant chemoradiotherapy group as compared with the surgery-only group (48.6 months vs 24 months median overall survival).

To date, carboplatin failed to replace cisplatin in the management of esophageal cancer. The lack of randomization can bias the results and the objective analysis of the findings thus there is need for Phase 3 trials for more conclusive results [33,36].

Oxaliplatin/5FU/leucovorin

There is a limited number of trials that have studied the combination of oxaliplatin with 5-FU [37,38] (Supplementary Table 3).

While Khushalani et al. found this drug combination to be safe and tolerable [38], Burmeister et al. concluded that the number of grade 3 and 4 toxicities is excessive [37]. However, Burmeister’s trial only included few patients [37].

More recently, Conroy et al. conducted Phase 2 and 3 trials in 2010 and 2014, respectively, comparing the traditional CDDP/5-FU combination against FOLFOX4 [39,40]. This cocktail consists of oxaliplatin (85 mg/m²) and leucovorin (200 mg/m²) both given on day 1, combined with 5-FU (1000 mg/m²) over the first two days of a two-week cycle. Initial investigations found an improvement in overall survival from 15.1 months in traditional CDDP/5-FU to 22.7 months with FOLFOX4. Acute toxicity was acceptable, as the incidence of adverse effects was not significantly different between the two groups. While FOLFOX4 did not increase progression-free survival, this treatment option may be more convenient since fewer days are needed to complete chemotherapy and it can be administered in the outpatient setting [40].

This logistical advantage of FOLFOX4 over cisplatin-based chemotherapy has determined further investigations in order to increase its therapeutic efficiency. Consequently, in a recent report of a Phase 2 trial (ERaFOX trial) that enrolled Stage III esophageal cancer patients, FOLFOX was combined with the EGFR-inhibitor cetuximab, with clinical overall response rate as the primary endpoint [41]. Complete response was achieved in 40% of patients, while overall response rate was 77%, well above the set target of 50% clinical response. Despite the promising results obtained by this study [41], the results could not be confirmed by Phase 3 trials that employed cetuximab, thus the authors do not recommend this drug combination in patients with esophageal cancer.

Nedaplatin/docetaxel

An alternative chemotherapeutic treatment involves nedaplatin or docetaxel. Nedaplatin is administered in similar doses to cisplatin, ranging from 50 to 100 mg/m², and combined with docetaxel or 5-FU.

Trials conducted on docetaxel or nedaplatin have mostly been Phase 2 trials, and have generally found favorable results in terms of overall survival and toxicity levels (Supplementary Table 4). However, these studies have little significant power, thus randomized trials are needed for conclusive results.

Literature results regarding normal tissue toxicity are contradictory. While there are reports stating that nedaplatin is safe and well-tolerated [42–45], other studies have demonstrated unacceptable levels of Grade 3 or 4 toxicities, including myelosuppression and/or esophagitis [46–49]. These conflicting results can be reconciled with the fact that though toxicities may be severe, they can be managed by lowering the dose without decreasing anti-tumor effects [50]. The efficacy of nedaplatin as opposed to cisplatin is yet to be established in large clinical trials.

To evaluate the efficacy of docetaxel vs 5-FU in non-operable esophageal cancer patients, a Phase 3 randomized trial reported by Zhao et al. compared a standard CDDP/5-FU treatment to a CDDP/docetaxel treatment, both combined with radiotherapy [51]. Median overall survival was improved from 22.3 months to 43.2 months for a total of 90 patients in the docetaxel group (p < .05). There were no significant differences in toxicities (p > .05) [51]. The good results obtained by this trial could be due to the efficient pharmacokinetic interactions on a cellular level between docetaxel and cisplatin, both being potent radiosensitizing agents.

Other chemotherapy and biological drugs

The paragraphs below present the results of clinical studies employing chemotherapy and/or biological drugs that to date have limited clinical use, either because of inconclusive findings or because of their relative novelty.

Two Phase 2 studies have examined cisplatin combined with irinotecan, with or without radiotherapy. Both studies have found that irinotecan is active and well tolerated [52,53]. Ilson et al. administered irinotecan (65 mg/m²) and CDDP (30 mg/m²) with no RT; median overall survival was 14.6 months with low toxicity rate, but there was a modest complete response (6%) [52]. Michel et al. found that irinotecan (60 mg/m²), CDDP (30 mg/m²) and concurrent 50 Gy of radiation gave a 2-year survival rate of 27.9% with a complete clinical response of 58.1%, and only 23.3% of patients suffered Grade 3 toxicity or above [53]. Irinotecan in combination with cisplatin appears to be a promising combination that yields a good response rate and is well tolerated.

A promising Phase 2 study combined pemetrexed (400 mg/m²) on day 1 and 22, and low-dose cisplatin (10 mg/m²) on days 1–5 and 22–26. Five-year survival was 44%, which was more favorable than other trials on inoperable locally advanced esophageal cancer. However, there were severe acute side effects, including 12% rate of grade 4 or 5 toxicities [54].

Vinorelbine, a semi-synthetic alkaloid, appears to be active in patients with CDDP/5-FU resistant cancers. Bidoli et al. administered vinorelbine alone (30 mg/m²) to 17 patients who had advanced or relapsed SCC, and partial response rates were found in 25% with significant improvement of dysphagia [55]. Toxicities were mild, with only one episode of grade 4 neutropenia. More promising results were found when vinorelbine (25 mg/m²on day 1 and 8) was combined with carboplatin (AUC4 on day 1, and then weekly during radiation) and 70 Gy of radiation [56]. In this study, complete response rate was 5.7%, while partial response rate reached 71.4%. This combination was considered a feasible and active treatment in older patients; however, toxicities were significant with several grade 4 events [56].

Etoposide is another plant-based alkaloid drug that has been studied in esophageal cancers due to promising results in other sites. One small study found that adding etoposide (125 mg/m² on day 1, 200 mg/m² on day 3 and 5) to cisplatin (80 mg/m² on day 1) and 5-FU (375 mg/m² every day) with no radiotherapy gave a median survival of 9.5 months. Side effects were manageable, with few grade 4 toxicities. This combination was concluded to be safe and active for advanced carcinomas [57], though the low median survival might be due to the avoidance of radiotherapy.

While the use of etoposide is very limited in esophageal cancers, this drug has some possible advantages when combined with the more common cisplatin-radiotherapy schedule: (1) the combination of etoposide with cisplatin does not lead to overlapping toxicities and (2) the most important cytostatic action of etoposide is cell arrest in the most radiosensitive phase of the cell cycle (G2/M) that potentiates the effect of radiotherapy.

Uracil–tegafur is a derivative of 5-FU. Iwase et al. administered cisplatin (70 mg/m² on days 8 and 36) combined with uracil–tegafur (200 mg/m² every day) and 70 Gy concurrent radiotherapy, and found that toxicities were manageable [58]. Five-year survival was poor, at 7% for those with Stage III cancer, and 0% for Stage IV. This treatment was stated to be suitable for palliation [58].

More recently, the focus is on the development of biological drugs, particularly drugs targeting epidermal growth factor such as gefitinib and erlotinib. Iyer et al. conducted a study where patients received a daily dose of erlotinib, combined with 50.4 Gy radiation dose in 28 fractions [59]. The low rate of side effects suggested that this is a viable palliative treatment for those where chemotherapy is contra-indicated [59]. Another study that combined erlotinib with cisplatin, paclitaxel and 60 Gy radiotherapy with promising result was reported by Li et al. [60]. When oxaliplatin was combined with gefitinib the median overall survival was 10.8 months. Once again, this was tolerable, but there was limited clinical activity as demonstrated by the poor survival [61].

The epidermal growth factor receptor drug cetuximab was already mentioned before in combination with FOLFOX. Several other studies have employed combined therapy with cetuximab with varied success. While Chen et al. found that cetuximab combined with cisplatin, paclitaxel and 59.4 Gy gave a good response rate and acceptable safety [62], others found that cetuximab was poorly tolerated [63,64]. One such trial used cetuximab, cisplatin, irinotecan and 50.4 Gy [64] while another administered a combination of cetuximab, cisplatin, capecitabine and 50 Gy [63]. Statistics on all studies are unreliable as there were few patients with short follow-up.

HER2 receptor antibodies have also been trialed. Although usually used in breast cancer, Safran et al. investigated the addition of trastuzumab to a cisplatin/paclitaxel treatment with 50.4 Gy [65]. Two-year survival rate was 50%, which was similar to known therapies, and there were few side effects [65]. Trastuzumab for gastric cancer (ToGA), a large Phase 3, randomized controlled trial for HER2-positive advanced gastric or gastro-esophageal junction cancer undertaken in 122 centers has randomized patients into trastuzumab + chemotherapy group vs chemotherapy alone [66]. The results of the trial showed that trastuzumab combined with chemotherapy led to higher median overall survival (13.8 months vs 11.1 months, HR 0.74, p = .0046) than chemotherapy alone, without notable differences in toxicities, thus it can be considered a treatment alternative for this patient group.

There have been many other small-scale studies on different chemotherapy combinations. Mitomycin-C, a DNA cross-linker, was used for palliative purposes and improved dysphagia and esophagitis in up to 68% of patients [67,68]. S-1, a new chemo-agent that contains tegafur and two types of enzyme inhibitors, combined with cisplatin was found to have good safety and efficacy [69,70]. TS-1 is a novel oral anti-cancer drug, consisting of a cocktail of tegafur, gimeracil and oteracil, and has radio-sensitizing and cytotoxic effects. When combined with docetaxel, TS-1 had acceptable toxicity and a satisfactory median survival of 15.2 months [71]. Celecoxib, a selective COX-2 inhibitor NSAID, administered at 400 mg/m² twice daily and combined with CDDP/5-FU was safe and tolerable. However, the associated increased cardiovascular risk is likely to negate further investigations in the future [72].

Many of these drug combinations are used as second or third-line treatments after the failure of traditional CRT regimes, with poor response rates. Hence, it is difficult to discern the efficacy of these drugs, and it is yet unconfirmed if there is any significant survival benefit.

Conclusions and future directions

Despite many treatment attempts with the aim to increase efficacy, prognosis for esophageal cancer remains grim. To date, there is no general consensus on a treatment regimen, which might be due to large variations in trial designs employing different drugs, doses and schedules of both chemotherapy and radiotherapy [73]. Although it is widely accepted that chemoradiotherapy is more effective than radiotherapy alone [74], the exact doses and indications are yet to be determined.

Usually the choice of chemotherapeutic agents is common for squamous cell carcinomas and adenocarcinomas. Nevertheless, there are some particularities that can differentiate between treatments. One such example is the expression of EGFR which is higher in SCC, thus EGFR-targeting agents like gefitinib or cetuximab could be more beneficial in SCC than in adenocarcinoma patients [75].

Supported by the results of Phase 3 randomized trials, chemotherapy for esophageal carcinoma has progressed from single-agent treatment to multiple-agent chemotherapy. This evolvement of treatment strategies led to modest improvements in response rate and overall survival; however, with greater normal tissue toxicity. The currently acknowledged most potent agent is cisplatin, reason why the vast majority of drug cocktails are based on this platinum compound. Despite several combination regimens trialed to date, cisplatin/5-FU with concurrent radiotherapy is considered standard of care for patients with esophageal cancer.

Future randomized trials employing novel or well-established agents should have better designs in order to identify the most effective therapy, while maximizing quality of life. Patient stratification should play a key role in future trials, aiming toward a more individualized, patient-centered approach. Since distant failure remains the main cause of death, novel therapies are needed to manage distant metastasis without additive adverse events.

Disclosure statement

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this article.