27: Radiation Therapy in the Management of Esophageal Cancer

Radiation therapy plays an important role in the multimodality treatment of esophageal cancer. Only 30% to 40% of patients with esophageal cancer have potentially resectable disease at presentation. In combination with chemotherapy in the neoadjuvant setting, radiation has been shown to improve survival over surgery alone. Radiation also can be very effective in the palliative setting. In this chapter, we discuss the general background and data on the use of radiation in the management of esophageal cancer.

There are few studies of radiation alone for esophageal cancer, and these have generally had poor outcomes. Historically, radiation was used for patients with extensive disease or those unfit for surgery. In a review of 49 early series of patients treated with radiation alone for esophageal squamous cell carcinoma, the overall survival rate at 5 years was 6%.¹ In a British study of 101 patients with clinically localized esophageal cancer, who received definitive radiation doses of 45 to 53 Gray (Gy, unit of absorbed radiation dose), the 5-year survival rate was 21%.² In a randomized trial (Radiation Therapy Oncology Group [RTOG] 85-01) of combined chemoradiation versus radiation alone to 64 Gy for 121 patients with mostly esophageal squamous cell carcinoma, 5-year overall survival in the radiation alone arm was 0.^3,4 For this reason, radiation alone for esophageal cancer is generally considered palliative rather than curative in intent. However, there are some studies exploring whether early-stage (stage I) esophageal cancer could be effectively treated with radiation alone.⁵

Definitive concurrent chemotherapy and radiation therapy have been studied as an alternative to operative management for esophageal cancer, particularly for patients who are not good surgical candidates. RTOG 85-01 was a prospective, randomized phase III trial of patients with nondisseminated adenocarcinoma or squamous cell carcinoma of the thoracic esophagus.^3,4,6 Patients were randomized between radiation alone (64 Gy in 32 fractions over 6.5 weeks) versus concurrent chemoradiation (2 cycles of infusional 5-FU plus cisplatin and radiation [50 Gy in 25 fractions over 5 weeks]). The trial closed early because a planned interim analysis showed a significant survival advantage for chemoradiation (5-year survival 27% vs. 0%) as well as a reduction in locoregional and distant failures. However, 46% of patients in the chemoradiation arm had persistent or locally recurrent disease in the esophagus at 12 months. Severe acute but not late toxicity was significantly higher in the chemoradiation arm.

The follow-up trial to RTOG 85-01 asked the question whether radiation dose escalation may improve local control rates. In the Intergroup 0123 study (RTOG 94-05), 236 patients with nonmetastatic squamous cell carcinoma or adenocarcinoma of the thoracic esophagus received concurrent cisplatin and 5-FU, and the patients were randomized between two different radiation doses: 50.4 Gy in 28 fractions or 64.8 Gy in 36 fractions.⁷ Radiation treatment was given daily, five fractions per week. There was no significant difference in median survival (13 vs. 18 months) or locoregional failure and persistence of disease (56% vs. 52%) between the high-dose and standard-dose arms, respectively. There was also a significantly higher rate of deaths in the high-dose arm (nine vs. two events). However, of the nine treatment-related deaths in the high-dose arm, six occurred before reaching 5040 cGy, suggesting that the deaths were not related to the higher radiation dose. Nevertheless, higher radiation dose does not appear to provide a survival or local control benefit, so 5040 cGy remains the standard of care in the United States at this time.

Patients of advanced age or those with significant comorbidities may not be good candidates for surgical management of esophageal cancer. Older patients have higher rates of postoperative morbidity and mortality after esophagectomy.⁸ Studies suggest that definitive chemoradiation may be well tolerated in elderly patients, with outcomes comparable to those in younger patients.^8,9

As definitive chemoradiation in contemporary series have shown near equivalent long-term survival as surgical series, the question becomes whether patients undergoing chemoradiation for esophageal cancer benefit from surgery. The results of the 8501 study demonstrated a locoregional recurrence rate of over 45%. Although there is no clear survival benefit, studies suggest that resection leads to improved locoregional control, predominantly in squamous cell carcinoma.

In a French trial, patients with operable T2N0–1 thoracic esophageal cancer received induction chemoradiation with either protracted (46 Gy in 4.5 weeks) or split course (15 Gy, days 1–5 and 22–26) radiation concurrent with 5-FU/cisplatin. Patients with at least a partial response (n = 259) were randomly assigned to surgery or continued chemoradiation. At a median follow-up of 47 months, survival was not significantly different between the two groups (median survival 18 vs. 19 months), but the surgical arm had lower locoregional recurrence (34% vs. 43%, p = 0.001) and need for palliative stents.¹⁰ Quality of life was similar between the arms at 2 years.¹¹ A German trial randomized 172 patients with locally advanced esophageal squamous cell carcinoma to induction chemotherapy and chemoradiation (40 Gy) followed by surgery or the same induction regimen followed by chemoradiation (65 Gy or greater) without surgery.¹² At 5 years, overall survival was equivalent between the arms, but local progression-free survival was better in the surgery group (64% vs. 53%, p = 0.003). However, treatment-related mortality was also higher in the surgery group (12.8% vs. 3.5%, p = 0.03).

Two studies have compared definitive chemoradiation (with surgical salvage) and surgery alone for resectable thoracic esophageal squamous cell carcinoma. In a Chinese prospective randomized trial where 80 such patients received 50 to 60 Gy with concurrent 5-FU/cisplatin or esophagectomy alone, there was no difference in survival.¹³ In a similar but nonrandomized study comparing those receiving 60 Gy concurrent with 5-FU/cisplatin or surgery alone, survival rates were similar, but the surgery alone arm had a higher rate of metastases.¹⁴

The rationale for preoperative (neoadjuvant) therapy for esophageal cancer is three-fold: to decrease micrometastases and control systemic disease, to improve resectability by tumor shrinkage, and to assess in vivo tumor response to therapy.

Preoperative Chemoradiation

The use of chemotherapy concurrent with radiation preoperatively is based on the notion of radiosensitization, namely, that the chemotherapy makes radiation more effective. This treatment paradigm is also called trimodality treatment: chemotherapy and radiation preoperatively, followed by surgical resection. There are seven major randomized trials of surgery with or without preoperative chemoradiation for patients with potentially resectable esophageal cancer, three of which showed a statistically significant survival benefit for chemoradiation.

In the Irish trial by Walsh and colleagues, 113 patients with esophageal or gastroesophageal (GE) junction adenocarcinoma were randomized to surgery alone or cisplatin/5-FU–based chemoradiation before surgery.¹⁵ Chemoradiation improved survival (32% vs. 6% at 3 years), although the surgery alone arm had a lower than expected survival. CALGB 9781 randomized 56 patients with squamous cell carcinoma or adenocarcinoma of the thoracic esophagus to trimodality therapy or surgery alone. With median follow-up of 6 years, median survival was significantly improved with preoperative chemoradiation (4.5 vs. 1.8 years, p = 0.002).¹⁶ In the Dutch CROSS trial, 366 patients with potentially resectable esophageal or GE junction cancer (mostly adenocarcinoma) were randomized to preoperative chemoradiation (chemotherapy was paclitaxel and carboplatin, radiation was given to 41.4 Gy over 5 weeks) versus surgery alone. The study showed a significant overall survival benefit with preoperative chemoradiation (median survival 49 months vs. 24 months), and this 29% of patients in this arm achieved a pathologic complete response.¹⁷ There was also a higher complete (R0) resection rate in the chemoradiation-surgery group (92% vs. 69%, p <0.001).

Three other randomized trials of surgery alone versus preoperative chemoradiation followed by surgery did not show a statistically significant survival benefit to trimodality treatment.^18-20 Adequate statistical power may have been an issue. As well, three studies comparing sequentially delivered chemotherapy and radiation preoperatively to surgery alone also failed to show any survival benefit to trimodality therapy.^21-23

Preoperative Chemotherapy

Multiple randomized phase III trials have evaluated the benefit of preoperative chemotherapy followed by surgery compared with surgery alone (no radiation in either arm). Results are mixed, with four negative trials^24-27 and three showing a survival benefit to preoperative chemotherapy.^28-30

Comparing Preoperative Regimens

A German study published in 2009 randomized 119 patients with locally advanced adenocarcinoma of the lower esophagus or gastric cardia to (A) induction chemotherapy (5-FU, leucovorin, cisplatin, 15 weeks) followed by surgery or (B) chemotherapy (12 weeks) followed by chemoradiation (30 Gy in 15 fractions over 3 weeks with concurrent cisplatin/etoposide) followed by surgery.³¹ The study closed early on account of low accrual. At a median follow-up of 46 months, the chemoradiation arm had a higher rate of pathologic complete response (16% vs. 2%, p = 0.03) and node negative disease at surgery (64% vs. 38%, p = 0.01). Three-year overall survival was 47% in the chemoradiation arm compared with 28% in the chemotherapy arm, but this did not reach statistical significance (p = 0.07). This may have been due to insufficient power.

Meta-Analyses

With conflicting results from different randomized trials, we can look to meta-analyses with their increased statistical power to determine whether preoperative treatment provides a survival benefit in esophageal cancer. Sjoquist et al.³² looked at 12 trials of neoadjuvant chemoradiation versus surgery alone (n = 1854), 9 trials of neoadjuvant chemotherapy versus surgery alone (n = 1981), and 2 trials comparing neoadjuvant chemoradiation and neoadjuvant chemotherapy (n = 194), all in patients with resectable esophageal cancer. Overall, neoadjuvant chemoradiation improved survival by 22% over surgery alone. This survival benefit was seen in both adenocarcinoma (17% survival benefit) and squamous cell carcinoma (20% survival benefit). Neoadjuvant chemotherapy without radiation, compared with surgery alone, provided a 13% survival improvement. This benefit was seen in adenocarcinoma patients (17% survival benefit), but there was no significant survival benefit of neoadjuvant chemotherapy over surgery alone for squamous cell carcinoma. When chemoradiation and chemotherapy were indirectly compared, there was weak evidence favoring chemoradiation, although this was not statistically significant.

Another recently published meta-analysis found similar results.³³ The authors report that neoadjuvant chemoradiation provides a 19% overall survival benefit over surgery alone, and neoadjuvant chemotherapy a 7% survival benefit. They did not find a significant increase in morbidity rates after neoadjuvant chemoradiation, and reported that the likelihood of a complete (R0) resection was 15% higher after neoadjuvant treatment. They found no significant survival difference between definitive chemoradiation and neoadjuvant treatment followed by surgery or surgery alone, although treatment-related mortality rates were over seven-fold lower.

These meta-analyses provide strong evidence for a survival benefit of neoadjuvant chemoradiation or chemotherapy over surgery alone in esophageal cancer. Although there appears to be no clear advantage of neoadjuvant chemoradiation over neoadjuvant chemotherapy, the data do suggest that chemoradiation may have a marginally greater survival benefit and higher pathologic response rate.

Measuring Response to Neoadjuvant Therapy: Implications for Individualized Treatment

Although neoadjuvant therapy followed by surgery has emerged as the standard of care for locally advanced esophageal cancer, there is variability in response to neoadjuvant treatment. Some patients may have a complete response but will go on to resection and be exposed to the risks of surgery. Similarly, for patients who do not respond to neoadjuvant chemotherapy, their prognosis after this treatment might be worse than that of a primary surgical approach, especially if there was local disease progression.

A German phase II trial (MUNICON) used FDG-positron emission tomography (PET) imaging to identify early responders to neoadjuvant chemotherapy.³⁴ Patients (n = 119) with locally advanced adenocarcinoma of the distal esophagus or gastric cardia had a baseline PET scan, then went on to receive 2 weeks of 5-FU/platinum-based induction chemotherapy. Those with a decrease of 35% or more in tumor glucose standard uptake value (SUV) on repeat PET were defined as metabolic responders and went on to receive 12 more weeks of chemotherapy and then surgery. Patients who did not have a sufficient response on PET to the 2 weeks of induction chemotherapy were considered nonresponders and went straight to surgical resection. Median event-free survival was higher among PET responders (30 months vs. 15 months, p = 0.002), and these patients were also more likely to have an R0 resection and complete pathologic response at surgery.

A follow-up study, MUNICON II, treated nonresponders to induction chemotherapy with salvage neoadjuvant chemoradiation before surgery.³⁵ With this salvage treatment, nonresponders had an improved histopathologic response rate when compared with the MUNICON I results. A recent retrospective study showed that PET complete response predicted improved survival after definitive chemoradiation, but not after trimodality therapy for esophageal cancer (neoadjuvant chemoradiation + surgery).³⁶ These preliminary data suggest that FDG-PET may help to identify patients in whom surgery might be avoided, but would need to be tested in a prospective study. The goal of identifying responders to neoadjuvant treatment is to potentially tailor therapies based on tumor biology and response.

In patients who undergo surgery for esophageal cancer, there is limited data on postoperative radiation. There are only three older randomized trials which address this issue. A trial from Hong Kong randomized 130 patients after curative or palliative esophageal resection to postoperative radiation (49 Gy in 14 fractions) or no additional treatment.³⁷ Among patients who had palliative resections, postoperative radiation decreased local recurrence (20% vs. 46%), but there was no significant effect on local control in patients who had curative resection. In fact, there was a detriment to overall median survival with postoperative radiation (9 months vs. 15 months) and a higher rate of complications (37% vs. 6%), with 17 cases of gastric ulceration and 5 deaths from bleeding. These data are from 1986 to 1989 using older radiation planning and delivery techniques. A French study, from 1979 to 1985, randomized 221 patients with esophageal squamous cell carcinoma to surgery alone or surgery followed by postoperative radiation (45–55 Gy).³⁸ Although local recurrence was lower in the radiation arm (15% vs. 30%), there was no difference in survival (median 18 months both arms). A Chinese study randomized 485 patients with esophageal squamous cell cancer from 1986 to 1997 to surgery versus surgery followed by radiation (50–60 Gy).³⁹ The addition of radiation in this group of patients did not improve overall survival.

To assess whether the addition of concurrent chemotherapy improves outcomes with adjuvant radiation in esophageal patients, we can extrapolate from the Intergroup 0116 trial.⁴⁰ In this study 556 patients with adenocarcinoma of the stomach or GE junction were randomized to resection alone or resection plus postoperative treatment. Approximately 20% of the patients had tumors located in the GE junction. The adjuvant treatment arm consisted of one cycle of postoperative chemotherapy (5-FU/leucovorin), then concurrent chemoradiation (45 Gy and 5-FU/leucovorin), followed by two more cycles of chemotherapy. Adjuvant chemoradiation improved overall survival (median 36 months vs. 27 months) and relapse-free survival (median 30 months vs. 19%). Based on these data, adjuvant chemoradiation is generally recommended for patients with GE junction cancers who undergo surgery as initial management.

We do not routinely use postoperative radiation therapy in esophageal cancer above the GE junction, since modern studies exhibit improved outcomes with neoadjuvant treatment. In addition, postoperative radiation fields are generally much larger to treat the anastomotic site (Fig. 27-1). Larger fields that extend high into the thorax and neck can involve normal tissues that are sensitive to the effects of radiation, particularly the lung and heart. Neoadjuvant chemoradiation is thus preferable to postoperative treatment for locally advanced esophageal cancer. Although the results of INT-0116 support the use of postoperative treatment for GE junction tumors, the rationale for neoadjuvant therapy applies to tumors of the cardia and GE junction, and neoadjuvant chemoradiation is therefore our preferred treatment strategy for locally advanced tumors of the esophagus, GE junction, and cardia.

Brachytherapy, which involves the temporary or permanent insertion of radioactive sources into the tumor or peritumoral tissues, is the most conformal radiation treatment modality available. It allows for focal dose escalation to the tumor, with sparing of surrounding normal tissue.⁴¹ RTOG 9207 was a prospective phase I/II study of 49 patients with inoperable esophageal cancer who received concurrent chemoradiation (50 Gy with 5-FU/cisplatin) followed by esophageal brachytherapy (15–20 Gy).⁴² Unfortunately, life-threatening treatment-related toxicity was 24% and treatment-related deaths were 10%. Six patients developed esophageal fistulas, resulting in three deaths, and median survival was only 11 months.

Intraluminal brachytherapy was used alone or as additional dose following external beam radiation therapy for patients with T1N0 esophageal cancer in a retrospective Japanese series.⁴³ The authors reported good overall outcomes (75% local control and 84% overall survival at 5 years) but 7% grade 3 to 4 toxicity including grade 4 and 5 esophageal ulcers. Another recent Japanese series reviewed 97 cases of stage I esophageal cancer treated with radiation (external beam with or without brachytherapy) and chemotherapy.⁴⁴ The addition of brachytherapy showed no survival or local control benefit, but the rate of serious adverse effects including lethal esophageal ulcers was significantly higher. Thus, for curative esophageal cancer, the addition of brachytherapy to external beam radiation has yet to demonstrate a local control or survival advantage. However, treatment-related toxicities, including esophageal ulcers and fistulas, can be significant.

The use of intraluminal brachytherapy for palliation of obstructive symptoms in locally advanced esophageal cancer is much more promising. A multicenter Danish phase III trial randomized 209 patients with dysphagia from inoperable esophageal or GE junction cancer to esophageal stent placement or single-dose brachytherapy (12 Gy).⁴⁵ Dysphagia improved more rapidly after stent placement, but long-term relief was more sustained after brachytherapy. Quality of life scores were better with brachytherapy. A Polish series of 91 patients with unresectable locally advanced esophageal cancer demonstrated a significant improvement in dysphagia with palliative high-dose-rate brachytherapy (22.5 Gy in 3 fractions per week).⁴⁶ Whether the addition of external beam radiation to brachytherapy can improve palliation has been addressed in two prospective phase III randomized trials. A Canadian multicenter trial randomized 60 patients with inoperable esophageal cancer to intraluminal brachytherapy (16 Gy in 2 fractions over 3 days) with or without additional external beam radiation (30 Gy in 10 fractions over 2 weeks).⁴⁷ At 6 months follow-up, the authors found no significant benefit of external beam radiation for symptom relief or overall survival. A study by the International Atomic Energy Agency randomized 219 patients with obstructive squamous cell esophageal cancer to high-dose–rate brachytherapy (2 fractions of 8 Gy each, within 1 week) with or without additional external beam radiation therapy (30 Gy in 10 fractions).⁴⁸ Combined modality treatment (brachytherapy + external beam) improved symptoms of dysphagia, odynophagia, regurgitation, and chest pain compared with brachytherapy alone, although there was no difference in survival outcomes.

Besides brachytherapy, other esophageal recanalization methods, with the aim of debulking the tumor, include laser therapy,⁴⁹ photodynamic therapy (PDT),⁵⁰ and argon plasma coagulation (APC).⁵¹ These methods were compared directly in a randomized trial of 93 patients with obstructive dysphagia from esophageal cancer: APC + brachytherapy, APC + PDT, or APC alone. Palliative combination treatment with APC and brachytherapy or PDT was more effective than APC alone. APC with brachytherapy was associated with the best quality of life and the least treatment-related complications.

Radiation treatment for esophageal cancer can be challenging because of the advanced presentation of esophageal tumors, and the large fields that are often needed to adequately cover gross disease and submucosal spread. These large fields can include normal tissues, such as lung, heart, and liver, and can result in treatment-related toxicity.

Pulmonary Complications

Patients receiving chemoradiation for esophageal cancer are at risk for pulmonary complications. Radiation pneumonitis is a subacute complication characterized by persistent cough or shortness or breath arising 6 weeks to 6 months after therapy. There is often also a radiologic correlation seen on CT within the photon beam path. Radiation oncologists try to minimize complications based on dose metrics to normal tissue. There is evidence that the risk of pneumonitis correlates with mean lung dose and data points on a dose–volume histogram (Fig. 27-2), although there is no definitive set of dose parameters that correlates perfectly with low pneumonitis risk.⁵²

Pulmonary complications that occur after esophagectomy can contribute substantially to postoperative morbidity and mortality. A study of 110 esophageal cancer patients treated with preoperative chemoradiation followed by esophagectomy found that dosimetric factors such as higher mean lung doses and worse lung sparing correlated with postoperative pulmonary complications.⁵³ However, a recent German study comparing patients who received preoperative chemoradiation followed by surgery or surgery alone found no difference in perioperative pulmonary toxicity.⁵⁴

Cardiac Toxicity

Long-term cardiac morbidity from esophageal radiation is generally unknown because of the small number of long-term survivors of locally advanced esophageal cancer and the fact that late cardiac toxicity from radiation does not occur until 10 to 20 years after treatment. Extrapolating from the breast cancer literature, however, shows that there are excess cardiac deaths in patients who received high doses of radiation to the left ventricle.⁵² Thus, as cure rates for esophageal cancer improve, cardiac late effects may become more important. Techniques such as three-dimensional (3D) CT planning, intensity-modulated radiation therapy, and high resolution imaging may help to improve dose distributions to the heart and lung.

Metastatic Disease

Unfortunately, 50% to 60% of patients with esophageal cancer present with unresectable locally advanced or metastatic disease. The most common sites of distant metastases in patients with esophageal cancer are the liver, lungs, bone, and adrenal glands. Metastases can progress quickly and cause significant pain and discomfort, impacting quality of life. Palliative external beam radiation can be very effective in improving pain from symptomatic bone metastases.⁵⁵ New advances in stereotactic body radiation therapy (SBRT) have been used successfully for palliation of spinal⁵⁶ and liver metastases.⁵⁷ Thus radiation therapy is frequently part of the overall palliative management of patients with metastatic esophageal cancer.

This chapter reviews the data supporting the use of radiation therapy (XRT) as an adjunct in treating patients with esophageal cancer. Whether as definitive therapy, used in conjunction with chemotherapy, or as a form of palliation (with external beam or brachytherapy), XRT has an important role and should be considered early in the decision-making process. One of the most important factors currently under evaluation is the ability to minimize toxicity from XRT, not only to adjacent organs such as the heart, lungs and spine, but also to minimize XRT dose to the future gastric conduit. Several techniques are available that should be pursued prospectively in these patients. First, a determination of the degree of lymph node involvement should be made. This can generally be done at the time of EUS with FNA. If these lymph nodes are negative or not evaluable, thoracoscopy/laparoscopy lymph node staging can be performed. These techniques will allow the radiooncologist to limit the dose, using 3D conformal and IMRT dose planning, to a smaller field if lymph nodes are not involved as opposed to a large field (conceivably more than 5 cm beyond the proximal and distant extent of the primary tumor). Another potential technique is to treat patients with a full stomach (using a liquid “meal”) at each treatment. This has been shown to act as a weight and pull the greater curvature of the stomach (the future conduit) laterally into the left upper quadrant. This allows full dose XRT to be delivered to the celiac axis without directly damaging the conduit or its vascular branches.