Neoadjuvant and adjuvant therapy for locally advanced resectable thoracic esophageal cancer

INTRODUCTION — 

Esophageal cancers are aggressive tumors with a generally poor prognosis. Locally advanced cancers of the thoracic esophagus (figure 1), which include both adenocarcinoma and squamous cell carcinoma (SCC) histologies, are typically managed using combined multimodality therapy with surgery, systemic therapy, and/or radiation therapy (RT).

This topic will provide an overview of neoadjuvant (ie, preoperative) and adjuvant (ie, postoperative) management for locally advanced resectable cancers involving the thoracic esophagus. The surgical management of esophageal and gastric cancer, neoadjuvant and adjuvant therapy for gastroesophageal junction (GEJ; which can also be called the esophagogastric junction [EGJ]) and gastric cancer, management of locally advanced unresectable esophageal cancer (including cervical esophageal cancer), and other related topics on the management of esophageal and gastric cancer are discussed separately.

●(See "Surgical management of resectable esophageal and esophagogastric junction cancers".)

●(See "Surgical management of invasive gastric cancer".)

●(See "Neoadjuvant and adjuvant therapy for locally advanced resectable esophagogastric junction and gastric cardia adenocarcinoma".)

●(See "Adjuvant and neoadjuvant treatment of gastric cancer".)

●(See "Management of locally advanced unresectable or inoperable esophageal cancer".)

●(See "Management of superficial esophageal cancer".)

●(See "Initial systemic therapy for metastatic esophageal and gastric cancer".)

SQUAMOUS CELL VERSUS ADENOCARCINOMA — 

There are two major histologies of esophageal cancer: squamous cell carcinoma (SCC) and adenocarcinoma. Cancers found in the thoracic esophagus can either be of SCC or adenocarcinoma histology [1].

There are increasing data that these histologies differ in terms of pathogenesis, epidemiology, tumor biology, and prognosis. In acknowledgment of these differences, the Tumor, Node, Metastasis (TNM) staging system provides separate stage groupings (but similar definitions for T, N, M, and grade [G] categories) for SCCs and adenocarcinomas of the esophagus and gastroesophageal junction (GEJ) (table 1). In addition, tumor location (for SCCs only) is incorporated into stage grouping. (See "Clinical manifestations, diagnosis, and staging of esophageal cancer".)

The impact of histology on the initial management of locally advanced esophageal cancer is evolving. Therefore, clinical studies on esophageal cancer are analyzing and reporting the results of various therapeutic strategies according to histology and tumor location, (figure 1) as these factors impact selection of therapy, prognosis, and patterns of recurrence [2-11]. For example, pathologic complete response (pCR) rates are higher in SCC following chemoradiation (CRT). For patients with esophageal SCC who are treated with initial CRT, subsequent nonoperative management is an option for those who achieve an endoscopic complete response or those who decline or are medically unfit for surgery. By contrast, limited data exist on the adequacy of nonoperative management for adenocarcinomas, and most patients undergo surgery after initial therapy with either CRT or systemic therapy. (See 'Necessity for surgery' below.)

Further details on the pathogenesis, epidemiology, and tumor biology of esophageal SCC and adenocarcinoma are discussed separately. (See "Epidemiology and risk factors for esophageal cancer" and "Barrett's esophagus: Pathogenesis and malignant transformation".)

THORACIC ESOPHAGUS TUMORS

Clinical T3-4 or node-positive disease

Overview of our suggested initial approach — Our suggested approach to patients with clinical (c)T3-4 or node-positive thoracic esophageal cancer that is resectable (figure 1 and table 1) is as follows:

●We recommend combined modality therapy rather than surgery alone, regardless of whether the histology is adenocarcinoma or squamous cell carcinoma (SCC) (table 1).

•Adenocarcinoma – For most patients with cT3-T4 or node-positive thoracic esophageal adenocarcinoma who are younger with good performance status and minimal comorbidities, we suggest perioperative chemotherapy with fluorouracil, leucovorin, oxaliplatin, and docetaxel (FLOT) rather than neoadjuvant chemoradiation (CRT). For older patients, those with multiple comorbidities, and/or those who are unlikely to tolerate the toxicities of FLOT (eg, neuropathy), neoadjuvant CRT is an appropriate alternative. Patients treated with neoadjuvant CRT followed by surgery should also be evaluated for adjuvant therapy (eg, adjuvant immunotherapy for those without a pathologic complete response [pCR]). (See "Neoadjuvant and adjuvant therapy for locally advanced resectable esophagogastric junction and gastric cardia adenocarcinoma", section on 'Perioperative chemotherapy versus neoadjuvant CRT'.)

•SCC – For most patients with cT3-T4 or node-positive thoracic esophageal SCC, we suggest initial therapy with CRT rather than chemotherapy. Patients who are treated with initial CRT may subsequently be evaluated for surgery versus nonoperative management, depending upon their treatment response; and patients who are treated with surgery should be subsequently evaluated for adjuvant therapy, if necessary. (See 'Chemoradiation responders' below and 'Role of postoperative (adjuvant) therapy' below.)

Compared with initial chemotherapy, initial CRT confers similar overall survival (OS) benefit, but is generally better tolerated, results in higher rates of pCRs and R0 resections, and offers the potential for nonoperative management in clinical responders. However, initial chemotherapy (either in the neoadjuvant or perioperative setting) may be the preferred approach in regions of the world where esophageal SCC is more prevalent, such as in Asia, and we acknowledge that clinical practice is variable. (See 'Chemotherapy versus chemoradiation' below.)

●For those who are not surgical candidates, regardless of histology, definitive CRT is an appropriate alternative. (See "Management of locally advanced unresectable or inoperable esophageal cancer".)

●For patients who are treated with CRT, we suggest concurrent radiosensitization with weekly carboplatin plus paclitaxel (table 2) rather than other systemic regimens due to better tolerance and ease of administration. FOLFOX is an alternative radiosensitizing agent. Selection of regimen and rationale is the same as for gastroesophageal junction (GEJ) and gastric cardia cancers and is discussed separately. (See "Neoadjuvant and adjuvant therapy for locally advanced resectable esophagogastric junction and gastric cardia adenocarcinoma", section on 'Selection of radiosensitizing regimen'.)

●Given the higher rate of locally persistent/recurrent disease after CRT alone and a lack of data on nonsurgical management of patients with adenocarcinomas, who have a low rate of pCR after CRT, we suggest inclusion of surgery rather than definitive CRT for clinically resectable adenocarcinomas. (See 'Necessity for surgery' below.)

We also suggest inclusion of surgery rather than definitive CRT for SCC. However, for those with SCC who have an endoscopic complete response, nonoperative management is an option, balancing the risks of surgical mortality versus improved locoregional control. Nonoperative management may be preferred for a patient with a higher burden of comorbidities who is less likely to tolerate surgery and/or has a tumor located in the proximal esophagus (eg, close to the larynx) where laryngectomy would be necessary in order to obtain optimal resection margins. In addition, data are equivocal for the OS benefits of neoadjuvant CRT followed by surgery versus definitive CRT. (See 'Chemoradiation responders' below.)

An important point is that the optimal way to define "complete CRT responders" and detect those with residual disease is not established. At some institutions, an endoscopic ultrasound (EUS)-guided fine needle aspiration biopsy is routinely undertaken [12,13], but others (including the National Comprehensive Cancer Network [NCCN] [14]) advise only upper endoscopy with biopsy if avoidance of surgery is being considered [15].

Until further data are available, we suggest against using postinduction therapy positron emission tomography (PET) scanning as an exclusive determinant in order to select patients for nonsurgical therapy. (See 'Utility of postinduction therapy PET scans' below.)

Surgery is an important component of treatment for those patients with less than a complete tumor response to neoadjuvant CRT. (See 'Nonresponders' below.)

●For patients who have residual disease at the time of surgery after neoadjuvant CRT, we suggest adjuvant nivolumab. (See 'After preoperative therapy' below.)

●Neoadjuvant management of GEJ and proximal gastric (ie, cardia, (figure 2)) adenocarcinoma is discussed separately. (See "Neoadjuvant and adjuvant therapy for locally advanced resectable esophagogastric junction and gastric cardia adenocarcinoma".)

Efficacy of individual strategies

Surgery alone — Although only 30 to 40 percent of patients have potentially resectable disease at presentation, surgery has been the standard treatment for early-stage thoracic esophageal cancer. Its utility as monotherapy has been challenged [16-22]. Observational surgical series report five-year survival rates of as high as 30 to 46 percent for surgery alone, with the most favorable outcomes in patients with node-negative SCC [23-26]. In one analysis of 4627 patients with esophageal cancer who were treated with surgery alone, without adjuvant or neoadjuvant therapy, five-year survival rates were 42 percent, but they were <50 percent for all disease stages except T1N0 cases (table 1), and they were 15 percent for any patient with node-positive disease [23].

This relatively poor long-term outcome has prompted an evaluation of neoadjuvant (preoperative), adjuvant (postoperative), and nonoperative strategies aimed at improving survival in patients with apparently localized disease. (See "Surgical management of resectable esophageal and esophagogastric junction cancers", section on 'Thoracic cancer resection'.)

Radiation therapy — Before the era of modern chemotherapy and CRT, radiation therapy (RT) alone was frequently used for local control of esophageal cancer. Modern radiation techniques (eg, three-dimensional conformal radiation therapy [3D-CRT], intensity-modulated radiation therapy [IMRT]) are associated with more favorable toxicity profiles than those associated with units which deliver lower energy radiation that were used in earlier years. The success of advanced radiation technology can be illustrated by a Chinese trial in which surgery was compared with RT alone in 269 patients with esophageal SCC [27]. RT was planned using 3D-CRT technique, and 69 Gy were delivered in 41 fractions over 45 days (45 Gy in 25 fractions over five weeks followed by 24 Gy in twice daily 1.5 Gy fractions for eight days using IMRT). Three- and five-year OS rates in the RT alone group (56 and 35 percent, respectively) were not significantly different from those in the surgery group (62 and 37 percent, respectively). These results cannot be extrapolated to patients with adenocarcinoma.

Although these data are encouraging, the strategy of treating with RT alone has been supplanted by combined CRT in the majority of patients because of significantly better outcomes [28,29], despite a higher rate of treatment-related toxicity.

Preoperative chemoradiation — Several trials and meta-analyses [29] have demonstrated improved survival with preoperative concurrent CRT compared with local therapy alone, and this approach is frequently used for potentially resectable stage T3 or 4, or node-positive localized cancer of the thoracic esophagus. However, the optimal regimen is not established.

The poor long-term survival associated with surgery alone and the radiosensitizing effect of concurrent chemotherapy provided the impetus to evaluate preoperative CRT. At least seven trials have directly compared surgery with or without preoperative CRT for patients with potentially resectable esophageal carcinoma [19,30-35]. Three studies demonstrated a significant survival benefit from combined modality therapy, all using a concurrent rather than sequential approach [34-36]. A survival benefit from preoperative CRT over surgery alone was confirmed in a network meta-analysis (hazard ratio [HR] for death for CRT over surgery alone 0.75, 95% CI 0.67-0.85) [29].

Concurrent chemoradiation — For patients who are treated with CRT, we suggest concurrent radiosensitization with weekly carboplatin plus paclitaxel (table 2) rather than other systemic regimens due to better tolerance and ease of administration. FOLFOX is an alternative radiosensitizing agent. Selection of regimen and rationale is the same as for GEJ and gastric cardia cancers and is discussed separately. (See "Neoadjuvant and adjuvant therapy for locally advanced resectable esophagogastric junction and gastric cardia adenocarcinoma", section on 'Selection of radiosensitizing regimen'.)

Concurrent CRT has been evaluated in patients with thoracic esophageal SCC. However, these studies mainly used concurrent cisplatin and fluorouracil (FU) with RT, which is no longer a preferred radiosensitization regimen due to toxicity and the availability of other effective and better-tolerated regimens.

●Versus RT alone – In randomized trials and a meta-analysis, the addition of cisplatin-based chemotherapy to RT significantly improved survival over RT alone [29,37-39]. However, the available data are almost exclusively in patients with SCC, and none of the trials have performed adequate pretreatment staging to reliably correlate outcome with locoregional tumor extent (ie, locally advanced unresectable versus potentially operable disease). The following sections will summarize the data for patients with disease confined to the primary and regional nodes based upon radiographic imaging.

•RTOG 85-01 – A landmark RTOG trial compared RT alone (64 Gy in 32 fractions over 6.5 weeks) versus concurrent CRT (two cycles of infusional FU [1000 mg/m² per day, days 1 to 4, weeks 1 and 5] plus cisplatin [75 mg/m² day 1 of weeks 1 and 5] and RT [50 Gy in 25 fractions over five weeks]) in patients with locoregional thoracic esophageal cancer [37]. Patients were required to have no evidence of spread beyond mediastinal and supraclavicular lymph nodes; 90 percent had SCC. The CRT group received two additional chemotherapy cycles, three weeks apart, after RT. Surgery was not part of the treatment schema.

The trial was closed prematurely with 121 patients, when an interim analysis showed a significant survival advantage for CRT (five-year OS 26 versus 0 percent) [38]. Analysis of failure patterns showed a significant reduction in both locoregional and distant failure for CRT. However, despite this benefit, 46 percent of patients in the CRT group had locally recurrent or persistent disease in the esophagus at 12 months. (See 'Necessity for surgery' below.)

As a result of this trial, definitive CRT became the standard of care for patients with inoperable disease. (See "Management of locally advanced unresectable or inoperable esophageal cancer", section on 'Benefits'.)

The issue of the unacceptably high locoregional failure rate was addressed in a follow-up trial, INT 0123.

•INT 0123 – In a randomized trial (US Intergroup 0123/RTOG 94-05), 236 patients with clinical stage T1 to 4, NX, M0 (table 1) esophageal cancer (SCC or adenocarcinoma) were randomly assigned to definitive CRT using either high-dose RT (64.8 Gy) or standard-dose RT (50.4 Gy), both with concurrent cisplatin plus infusional FU [40]. Two additional cycles of chemotherapy were repeated four weeks after the completion of RT. Surgical resectability was not an inclusion criterion. At a median follow-up of 16 months, relative to standard-dose RT, high-dose RT failed to improve OS (median OS 13 [95% CI 11-19] versus 18 [95% CI 15-23] months, two-year OS 31 versus 40 percent) with a similar incidence of locoregional persistent or recurrent disease (56 versus 52 percent). The rate of treatment-related deaths was increased with high-dose compared with standard-dose RT (10 versus 2 percent), although most of the deaths occurred before reaching the escalated RT dose. The trial was closed early due to the lack of benefit with higher-dose RT. (See "Management of locally advanced unresectable or inoperable esophageal cancer", section on 'Radiation therapy dosing'.)

•IMRT plus chemotherapy – Contemporary radiation techniques, such as IMRT, are associated with more favorable toxicity profiles. (See "Radiation therapy techniques in cancer treatment", section on 'Intensity-modulated radiation therapy'.)

Although few trials have been conducted, IMRT with concurrent chemotherapy is beginning to be studied for the treatment of esophageal cancer [41-43]. At least one Chinese trial of 170 patients with locally advanced esophageal cancer suggests that definitive CRT using the combination of IMRT plus concurrent cisplatin plus docetaxel improved local control and prolonged survival over IMRT alone, but side effects were more prominent [41]. No trial has compared IMRT plus concurrent chemotherapy with the same chemotherapy regimen plus standard fractionation 3D-CRT, and thus, the safety and efficacy of this approach compared with standard 3D-CRT remains undefined.

•ARTDECO trial – For patients with thoracic esophageal SCC or adenocarcinoma receiving definitive CRT, RT is delivered at a dose of 50.4 Gy. This approach is extrapolated from a randomized trial (ARTDECO) conducted in patients with medically inoperable or unresectable esophageal cancer that compared high-dose RT (61.6 Gy) with standard-dose RT (50.4 Gy), both concurrent with carboplatin plus paclitaxel [44]. Further details of this study are discussed separately. (See "Management of locally advanced unresectable or inoperable esophageal cancer", section on 'Radiation therapy dosing' and 'Technique for preoperative RT' below.)

●Versus surgery alone – Of the six completed randomized trials that compared preoperative concurrent CRT with surgery alone, three showed a statistically significant survival benefit for CRT [34,36,45], and three others did not, two of which were underpowered [30,35,46]. The three most important trials are the Dutch CROSS, CALGB 9781, and NEOCRTEC5010 trials:

•CROSS trial – Dutch investigators randomly assigned 363 patients with potentially resectable esophageal or gastroesophageal junction (GEJ) cancer (86 SCC, 273 adenocarcinoma, 4 other; majority distal esophageal, 11 percent GEJ) to preoperative CRT using weekly paclitaxel 50 mg/m² plus carboplatin (area under the curve of concentration X time [AUC] of 2) plus concurrent RT (41.4 Gy over five weeks) or surgery alone [45]. Preoperative CRT was well tolerated, with grade 3 or worse hematologic toxicity in 7 percent, and grade 3 or higher non-hematologic toxicity in <13 percent; there were also no differences in postoperative morbidity or mortality between the two groups. The microscopically complete (R0) resection rate was higher with CRT (92 versus 69 percent), and 29 percent of those treated with CRT had a pCR. At a median follow-up of 32 months, median OS was significantly better with preoperative CRT (HR for death 0.657, 95% CI 0.495-0.871, three-year survival rate 58 versus 44 percent).

The survival benefit persisted with longer follow-up (five-year survival 47 versus 33 percent, HR for death 0.67, 95% CI 0.51-0.87 [47]; 10-year survival 38 versus 25 percent, HR for death 0.60, 95% CI 0.46-0.80 [48]). In long-term analysis, although there was a clear effect of neoadjuvant CRT on reducing isolated locoregional and synchronous locoregional plus distant relapses, isolated distant relapse rates were not significantly different in the two groups [48].

•CALGB 9781 – CALGB 9781 was originally designed as a randomized Intergroup trial of trimodality therapy versus surgery in 500 patients with stages I-III esophageal or GEJ cancer, staged with esophagogastroduodenoscopy, barium esophagram, and computed tomography (CT). Staging EUS and thoracoscopy/laparoscopy were encouraged. Due to poor accrual, the study was closed prematurely with only 56 patients enrolled (42 adenocarcinomas, 14 SCC). A pCR was achieved in 10 of 25 assessable patients in the trimodality arm (40 percent), and neither perioperative morbidity nor mortality were increased compared with surgery alone [35]. Five-year survival was 39 versus 16 percent in favor of trimodality therapy, although the difference was not statistically significant.

•NEOCRTEC₅₀10 trial – In a Chinese trial, 451 patients with potentially resectable thoracic esophageal SCC were randomly assigned to neoadjuvant CRT (RT concurrent with vinorelbine plus cisplatin) or surgery alone [36,49,50]. At surgery, the pCR rate was 43 percent in those receiving CRT. Compared with surgery alone, patients receiving neoadjuvant CRT had a higher R0 resection rate (98 versus 91 percent), and a similar incidence of postoperative complications. In the latest analysis (median follow-up 54 months), neoadjuvant therapy was associated with better five-year overall (60 versus 49 percent, HR 0.74, 95% CI 0.57-0.97), and disease-free survival (64 versus 43 percent, HR 0.60, 95 CI 0.45-0.80) [49].

•Meta-analyses – Several meta-analyses have addressed the benefit of trimodality therapy over surgery alone for esophageal cancer [29,51-54]. As an example, one individual patient data network meta-analysis included 12 randomized trials comparing neoadjuvant CRT (either concurrent or sequential) versus surgery alone for thoracic esophageal or GEJ cancers [54]. In this study, neoadjuvant CRT increased OS compared with surgery alone (HR 0.77, 95% CI 0.68-0.87). In a subgroup analysis, this OS benefit was seen regardless of histology (SCC [HR 0.79] versus adenocarcinoma [(HR 0.73]) or tumor location (thoracic esophagus [HR 0.79] versus GEJ [HR 0.57]). Effect modification was not present by either histology or tumor location. (See "Neoadjuvant and adjuvant therapy for locally advanced resectable esophagogastric junction and gastric cardia adenocarcinoma", section on 'Neoadjuvant CRT'.)

●Older adults – Few data are available on the safety and efficacy of CRT in older adults. However, at least one report suggests that patients over age 70 tolerate cisplatin-based CRT without a major increase in adverse events and with outcomes that seem comparable with those achieved in younger individuals [55]. Patients with significant comorbidity (ie, Charlson score ≥1 (table 3)) did experience more severe toxicity and chemotherapy delays/dose reductions than those without comorbidity. (See "Comprehensive geriatric assessment for patients with cancer".)

Sequential chemoradiation — In contrast to the data on concurrent CRT, at least three trials comparing sequentially administered chemotherapy and RT followed by surgery with surgery alone have failed to show any survival advantage to combined modality therapy [19,31,32].

Technique for preoperative RT — The degree of response of a tumor and normal tissues/organs to radiation depends upon several radiotherapeutic factors [56-58]:

●Fraction size (standard fraction size, 1.8 Gy to 2 Gy) and interfractional intervals (standard interval, 24 hours)

●Total dose (standard preoperative dose in once daily schedule, 41.4 to 50.4 Gy)

●Duration of treatment (5 to 5.6 weeks for standard fractionation, without a rest during treatment)

●The arrangement of radiation portals in a manner that achieves the maximum dose differential between tumor and adjacent vital organs

Significant deviations from standard techniques should be avoided in a potentially curative setting. Fraction sizes that are larger than 2.5 Gy, treatment breaks of longer than one week, split-course fractionation schedules [59], and suboptimal radiation plans with a potential for increased risk of injury to the lung, heart, and spinal cord should be avoided.

●Conformal RT – Conformal therapy is a term that describes a strategy for matching ("conforming") the high-dose radiation region to the target volume while minimizing the radiation dose to normal tissues. This term is typically used when the target volumes are defined on a CT or other high-definition imaging study used during the treatment planning. (See "Radiation therapy techniques in cancer treatment", section on 'Conformal therapy'.)

Three-dimensional-CRT provides better coverage of the target volume while protecting the surrounding normal organs from excess RT dose. Conformal treatment planning also enables measurement of the dose of radiotherapy delivered to a volume of given organs, allowing more refined assessment of the interaction between radiotherapy dose and potential side effects.

3D-CRT uses four to eight beams if necessary to conform the distribution of radiation dose to the gross tumor volume (GTV), clinical tumor volume (CTV), and planning tumor volume (PTV), while the surrounding normal structures are protected from excessive radiation dose to the greatest extent possible. 3D-CRT plan provides a dose-volume histogram for GTV, CTV, and PTV as well as for normal organs at risk for complications. Thus, it is feasible to formulate a radiation dose schedule for the desired level of tumor control probability that is balanced with an acceptable level of toxicities. However, such risk assessment is not feasible with two-dimensional RT (2D-RT).

IMRT is an advanced radiation modality. Unlike 2D and even standard 3D-CRT, it uses inverse treatment planning to deliver highly conformal RT plans, enabling increased sparing of normal structures while still delivering the same dose of radiotherapy to the GTV, CTV, and PTV. As a result, the volume of normal tissue in the high radiation dose region (40 to 60 Gy) is smaller with IMRT than with 3D-CRT while the volume of normal tissue in the low-dose region (5 to 10 Gy) is larger with IMRT than with 3D-CRT. It may be expected (though not yet proven) that treatment with IMRT will result in fewer side effects than 3D-CRT [60-62].

●Target volume – The target volume consists of the GTV with a margin of clinically uninvolved tissue but potentially harboring microscopic tumor tissue CTV. The CTV should include 4 to 5 cm margins beyond the radiographic tumor extent in the cephalad-caudad direction and 2 to 2.5 cm beyond the radial border of GTV (defined by barium esophagogram or CT scan). For lesions of the lower third of the esophagus and GEJ, the caudal extension CTV beyond GTV includes a 3 to 4 cm margin of gastric cardia below the lower border of GTV. CTV for regional lymph nodes includes the celiac, gastric, and gastrohepatic lymph node groups for primary tumors at the GEJ. For primary tumors involving the upper two-thirds of the thoracic or the cervical esophagus, CTV includes both supraclavicular regions. It is necessary to add another 0.5 to 0.7 cm beyond CTV as the PTV in order to compensate for daily set-up error and respiratory tumor motion.

Esophageal cancers that are located at the GEJ can have a significant degree of tumor motion associated with respiration. We reported peak-to-peak motion of the primary tumor in 10 patients (9 near the GEJ with their involved lymph nodes at the celiac region) [58]. The peak-to-peak tumor motion in craniocaudal directions ranged from 0.6 cm to 4.8 cm for the primary tumor and from 1.2 cm to 4.4 cm for the involved lymph nodes. To avoid geographic miss in some of these patients when using 3D-CRT, four-dimensional CT (4D-CT) treatment planning and delivery of 3D-CRT has a significant advantage for optimum coverage of the target volume over 3D-CRT planned with the conventional helical CT.

●Optimal dose and fractionation schedules – Tumor size and radiation dose are important considerations for locoregional tumor control. The preferred radiation dose for CRT is a total dose of 50.4 Gy administered in daily 1.8 Gy fractions, five days per week. This dosing results in reasonable results with acceptable toxicity [44].

Other RT dosing and schedules have also been evaluated. Altered fractionation schedules such as accelerated schedules (45 Gy in 30 fractions over three weeks using twice daily 1.5 Gy fractions) or hybrid schedules using twice daily radiation during chemotherapy and once daily treatment between chemotherapy cycles (45 Gy in 25 fractions over five weeks to CTV, and 58.5 Gy in 34 fractions over five weeks to GTV, respectively) are tolerable, with encouraging tumor response, high pCR rates, and survival [30,63].

Patients judged inoperable because of either comorbidities or the presence of distant metastases can be treated by hypofractionated schedules to reduce overall treatment time. A total dose of 40 to 45 Gy at 2.5 Gy daily fractions five days a week is a reasonable schedule for patients who require palliation of esophageal obstruction.

Intensification of preoperative therapy — A consistent finding in many studies is that response to preoperative therapy, particularly the absence of residual disease in the surgical specimen, is an indicator of better disease-free and OS [30,64-75]. In a comprehensive literature review of 22 studies in which patients with esophageal or GEJ cancer underwent esophagectomy after neoadjuvant CRT, patients with a pCR were two- to threefold more likely to survive, as were those with residual disease in the esophagectomy specimen [74]. These benefits translate into a 33 to 36 percent mean absolute survival benefit when a pCR is achieved than when it is not.

These results provide the rationale for intensification of preoperative treatment through adding several cycles of induction chemotherapy prior to preoperative CRT, or increasing the number of cytotoxic agents administered concurrent with RT.

Several groups have reported their experience with sequential induction chemotherapy followed by CRT [76-81].

While no randomized trials have compared this sequential induction chemotherapy followed by CRT with standard CRT, only one phase III trial (POET), compared this approach with induction chemotherapy alone followed by surgery.

In the German POET trial, 126 patients with GEJ adenocarcinoma were randomly assigned to 16 weeks of chemotherapy alone (cisplatin plus leucovorin and short-term infusional FU) versus 12 weeks of the same chemotherapy regimen followed by low-dose RT concurrent with cisplatin and etoposide; both groups underwent subsequent surgical resection [79]. The pCR rate was significantly higher after induction chemotherapy followed by CRT, and there was a nonsignificant trend towards better median and three-year survival (47 versus 28 percent) in this group as well. Whether these results can be extrapolated to SCC of the thoracic esophagus is uncertain. This trial and the implication of its findings for treatment of GEJ cancers are discussed in detail separately. (See "Neoadjuvant and adjuvant therapy for locally advanced resectable esophagogastric junction and gastric cardia adenocarcinoma", section on 'Perioperative chemotherapy versus neoadjuvant CRT'.)

Although early results from many trials evaluating intensified chemotherapy during RT are encouraging [82-87], whether the added toxicity of any of these approaches is counterbalanced by substantial survival gains is unclear. Further randomized trials are needed to confirm the benefit of these more toxic approaches.

Necessity for surgery

Chemoradiation responders — For patients with clinically resectable adenocarcinoma who are treated with initial CRT, we suggest inclusion of surgery rather than definitive CRT alone.

We also suggest inclusion of surgery rather than definitive CRT for those with clinically resectable thoracic esophageal SCC who are treated with initial CRT. However, nonoperative management is an option for those who have an endoscopic complete response to initial CRT, balancing the risks of surgical mortality versus improved locoregional control.

Regardless of histology, definitive CRT is a reasonable approach for patients who are not surgical candidates. (See "Management of locally advanced unresectable or inoperable esophageal cancer".)

Approximately 50 percent of patients with SCC and 25 percent of those with adenocarcinoma have a pCR after neoadjuvant CRT [45]. The necessity of resection for these patients is controversial. Whether there is a survival benefit for trimodality therapy (which included post-RT or post-CRT esophagectomy) over definitive RT or CRT alone has been addressed in several meta-analyses, which came to opposite conclusions:

●One meta-analysis included data from eight studies (16,647 patients), only one of which was a randomized controlled trial [88]. OS was improved with neoadjuvant CRT as compared with definitive CRT in combined populations (HR for death 0.55, 95% CI 0.49-0.62), and the benefit was seen in both the adenocarcinoma and SCC subgroups. The authors did not address the impact of treatment response on outcomes.

●On the other hand, a year 2016 Cochrane review of eight randomized trials concluded that there was no difference in long-term mortality between definitive and neoadjuvant CRT (HR 0.88, 95% CI 0.76-1.03; 602 participants; four studies; low-quality evidence), and no difference in long-term recurrence between nonsurgical treatment and surgery (HR 0.96, 95% CI 0.80-1.16; 349 participants; two studies; low-quality evidence) [89]. The difference between nonsurgical and surgical treatment was imprecise for short-term mortality (risk ratio [RR] 0.39, 95% CI 0.11-1.35; 689 participants; five studies; very-low quality evidence) and the risk for local recurrence at maximal follow-up (RR 0.89, 95% CI 0.70-1.12; 449 participants; three studies; very-low quality evidence) [89].

An important point is that the optimal way to define "complete CRT responders" is not established. At some institutions, an EUS-guided fine needle aspiration biopsy is routinely undertaken [12,13], while others advise only upper endoscopy with biopsy if patients are being evaluated for nonoperative management (ie, no surgery following CRT). The role of fluorodeoxyglucose (FDG)-PET/CT to evaluate the local response to therapy as a means of selecting patients for whom surgery might be avoided is controversial and discussed below [15]. (See 'Utility of postinduction therapy PET scans' below.)

There are much more data on definitive CRT in SCC as compared with adenocarcinomas.

●Squamous cell carcinoma – In contemporary series, definitive CRT provides long-term survival in up to 27 percent of patients with SCC [37,40,59,90,91], a result that is not dissimilar to that achieved with preoperative CRT followed by surgery [16,30,34], neoadjuvant chemotherapy and surgery (see 'Neoadjuvant chemotherapy' below) [18,92], and surgery alone [18,19]. However, nearly all reports note a higher rate of locally persistent/recurrent disease when surgery is not a component of treatment [37,59,91].

At least two randomized trials directly comparing CRT alone with trimodality therapy (CRT followed by surgery) have failed to demonstrate better survival, although both show better locoregional control and a lesser need for palliative procedures when surgery is a component of multimodality treatment [93,94]. The patient populations in both were either exclusively or predominantly SCC. A Cochrane analysis of these two trials [95] came to the following conclusions:

•There was high-quality evidence that the addition of esophagectomy had no significant impact on survival (HR 0.99, 95% CI 0.79-1.24).

•There was moderate-quality evidence that the addition of esophagectomy improved freedom from locoregional relapse (HR 0.55, 95% CI 0.39-0.76), but low-quality evidence suggested that it increases the risk of treatment-related mortality (RR 5.11, 95% CI 1.74-15.02) [94,96,97].

•All other endpoints (quality of life, treatment-related toxicity, use of salvage procedures for dysphagia) were only reported in one trial, which found only very-low-quality evidence that surgery reduced the use of salvage procedures for dysphagia (HR 0.52, 95% CI 0.36-0.75).

Another meta-analysis by the American Society of Clinical Oncology (ASCO) came to similar conclusions, although they considered the evidence of a lack of survival benefit from resection to represent moderate quality evidence [53].

If definitive CRT is chosen, the choice of concurrent chemotherapy regimen is discussed separately. (See 'Concurrent chemoradiation' above.)

●Adenocarcinoma – In contrast to SCC, there are no randomized trials directly comparing trimodality versus bimodality (CRT alone) therapy in adenocarcinoma. Some data from retrospective analyses suggest inferior outcomes in this group with nonsurgical management [98-101]. As examples:

•Nonsurgical management of adenocarcinoma was evaluated in a retrospective analysis of 276 patients treated with definitive CRT at the University of Texas MD Anderson Cancer Center for esophageal cancer, 215 of whom had adenocarcinoma [99]. RT was planned and delivered using modern techniques (4D-CT and IMRT), and a moderate dose (50.4 Gy in 28 fractions) was combined with chemotherapy. The majority of patients had T3 (83 percent), N1 (69 percent), and M0 (87 percent) disease. Nearly all (98 percent) received concurrent chemotherapy, and 37 percent also received induction chemotherapy prior to CRT. At a median follow-up of 54 months, 140 (51 percent) had experienced a local recurrence, while 144 (52 percent) encountered distant failure with or without a locoregional recurrence, and 92 (33 percent) had no evidence of disease at last follow-up.

•In a second retrospective series of 154 patients with esophageal adenocarcinoma treated at a single center over an 11-year period, 60 were treated without surgery while the remainder received trimodality therapy [100]. Despite the fact that the surgically-treated patients had more advanced disease stage, survival was significantly better in this group (median survival 4.6 versus 1.9 years; five-year OS 44 versus 36 percent).

•The specificity of clinical complete response (cCR) for adenocarcinomas is too low to be used for clinical decision-making for delaying or avoiding surgery. In one series of 284 esophageal/GEJ cancers (92 percent adenocarcinomas), 218 (77 percent) achieved a cCR (as defined by endoscopic biopsy negative for cancer and FDG-PET showing only physiologic uptake) after CRT, but only 67 (31 percent) were true pCRs [102].

It is likely that some of these patients have a high risk for early disease recurrence within one year of completing treatment, and in this group, the benefit of surgery might not outweigh its potential side effects. Investigators at MD Anderson Cancer Center used data from 568 consecutive patients with potentially resectable esophageal adenocarcinomas who underwent CRT at MD Anderson Cancer Center over a 10-year period to construct a preoperative nomogram to risk stratify patients for the benefit of trimodality therapy in esophageal adenocarcinoma [101]. In the entire cohort, 373 underwent esophagectomy, while 195 did not; median follow-up was 62 months. Five-year OS in the trimodality and bimodality groups was 56.3 percent (95% CI 47.9-64.7) versus 36.9 percent (95% CI 31.4-42.4). However, when stratified according to risk factors (histologic grade, signet ring cell morphology, clinical nodal stage, intensity of FDG uptake) and balanced with propensity score matching, the survival benefit of trimodality therapy was much less pronounced in the high-risk cohort (five-year OS 32 versus 21 percent) compared with the low-risk group (five-year OS 66 versus 46 percent). Importantly, this nomogram did not identify a population for whom there was no benefit from resection. Furthermore, although propensity score matching was used to improve the comparability between the two treatment groups, unknown confounding factors may have influenced the outcomes. This model requires external validation.

Utility of postinduction therapy PET scans — Until further data are available, we do not use postinduction therapy PET scanning to select patients for nonsurgical therapy.

Postinduction therapy FDG-PET provides information about metabolic response in the primary tumor that may be clinically useful for selection of subsequent therapy. In particular, some retrospective data suggest that post-CRT FDG-PET scanning may serve to identify those patients for whom surgery might be avoided. One series included 105 patients with stage I to IVA esophageal cancer (75 percent adenocarcinoma) who were evaluable for a post-CRT PET response, 50 of whom received CRT alone [103]. In this cohort, those whose PET response was characterized by a post-treatment maximum standard uptake value (SUVmax) ≤3 in the tumor (n = 19, 38 percent) had an excellent outcome without resection (two-year OS 71 versus 11 percent for those with a post-treatment tumoral SUVmax ≥3.1; the corresponding two-year rates of freedom from local failure were 75 versus 28 percent). By contrast, those patients undergoing trimodality therapy (n = 55) showed no difference in outcome according to the post-CRT PET findings, probably because those patients who had residual disease underwent resection.

However, others have failed to find a correlation between the post-CRT SUV on PET and pathologic response at the time of resection. A systematic review of 13 studies, totaling 697 patients, concluded that the available data are too contradictory to conclude that interim PET-directed therapy for esophageal cancer can be considered a standard approach [104]. A major issue is that the best method to quantify FDG uptake for clinical use in esophageal cancer remains to be determined. A later systematic review and meta-analysis concluded that even when combined with endoscopic biopsies and EUS, FDG-PET/CT was insufficiently accurate for detecting residual disease after induction therapy for esophageal cancer [105].

The main use of PET in esophageal cancer may be to recognize which patients are not responding to induction chemotherapy prior to CRT, administered prior to resection. This subject is discussed separately. (See "Neoadjuvant and adjuvant therapy for locally advanced resectable esophagogastric junction and gastric cardia adenocarcinoma", section on 'PET-directed therapy'.)

Nonresponders — For patients whose disease does not respond to initial CRT, regardless of histology, surgery is an important component of therapy for those whose tumor remains operable.

Timing of surgery after chemoradiation — For most patients who will be treated with neoadjuvant CRT followed by immediate surgery, we prefer that surgery be performed within five to seven weeks of completing CRT. For those who need extra time to recuperate from CRT, surgery can be delayed until the patient is medically fit to undergo the operation. The risk of postoperative complications increases if surgery is performed too early (ie, sooner than four weeks of completing CRT). However, data are mixed for whether a delay in surgery is associated with worsened OS.

The typical interval between CRT and surgery, five to seven weeks, was initially chosen arbitrarily to allow resolution of acute inflammation and tumor regression while minimizing chronic fibrotic changes in the surgical field. In patients undergoing trimodality treatment for esophageal cancer, pCRs can predict decreased local and distant recurrence and improved survival [66,69,79]. Some initial studies suggest that delaying surgery beyond this timeframe resulted in higher pCR rates (due to more time for tumor regression) with no increase in surgical morbidity [106]. However, subsequent studies are mixed, with some suggesting that delayed surgery does not impact surgical outcomes or OS, whereas others suggest that delayed surgery is associated with worsened OS relative to earlier surgery [71,106-109]. As examples:

●In a randomized trial (NeoRes II) of 249 patients with locally advanced esophageal cancer treated with neoadjuvant CRT, compared with standard delay in surgery (four to six weeks after CRT), a prolonged delay in surgery (10 to 12 weeks after CRT) demonstrated no difference in pCR rates, tumor regression, resection margins, and number of resected lymph nodes [110]. Delayed surgery also resulted in worsened first-quartile OS (median 14 versus 26 months); however, for all patients, at a median follow-up of 36 months, the overall risk of death was not statistically significant (HR 1.35, 95% CI 0.94-1.95).

●By contrast, one observational study from the National Cancer Database suggested inferior 90-day mortality and long-term survival rates in patients who waited 65 days or longer between CRT and surgery [109].

Neoadjuvant chemotherapy — Multiple randomized trials have evaluated the benefit of chemotherapy administered prior to resection in patients with esophageal cancer limited to the primary and regional nodes by clinical assessment [18,65,92,111-114]. Four trials with a surgery-alone control arm are negative, including the United States Intergroup 0113 trial [18,65,111,115], while five others (including the MRC OE2 trial, the United Kingdom MAGIC trial, and the French FNLCC/FFCD trial), demonstrate a survival benefit compared with resection alone [92,112,113,116,117].

Meta-analyses — A survival benefit for neoadjuvant chemotherapy relative to surgery alone has been shown in several meta-analyses [51,53,54,118,119]. As an example, in one individual patient data meta-analysis included 12 randomized trials comparing neoadjuvant chemotherapy versus surgery alone for thoracic esophageal or GEJ cancers [54]. In this study, neoadjuvant chemotherapy followed by surgery increased OS compared with surgery alone (HR 0.86, 95% CI 0.75-0.99). In a subgroup analysis, tumor location was an effect modifier of this OS benefit, in favor of GEJ tumors (HR 0.68) relative to thoracic esophageal tumors (HR 0.88). Although effect modification by histologic subtype was not present, this OS benefit was less pronounced for SCC (HR 0.88) relative to adenocarcinoma (HR 0.78). (See "Neoadjuvant and adjuvant therapy for locally advanced resectable esophagogastric junction and gastric cardia adenocarcinoma", section on 'Selection of perioperative chemotherapy regimen'.)

Chemotherapy versus chemoradiation

●Thoracic esophageal adenocarcinoma – For patients with clinical T3-4 or node-positive resectable thoracic esophageal adenocarcinoma who are younger with good performance status and minimal comorbidities, we suggest perioperative chemotherapy with fluorouracil, leucovorin, oxaliplatin, and docetaxel (FLOT) rather than neoadjuvant CRT. For older patients, those with multiple comorbidities, and/or those who are unlikely to tolerate the toxicities of FLOT (eg, neuropathy), neoadjuvant CRT is an appropriate alternative. Patients treated with neoadjuvant CRT followed by surgery should be evaluated for adjuvant therapy (eg, adjuvant immunotherapy for those without a pCR).

In a phase III trial (ESOPEC) that included patients with thoracic esophageal and GEJ adenocarcinoma, perioperative FLOT improved OS relative to neoadjuvant CRT with good compliance rate and similar postoperative morbidity and mortality [120]. Further details on the ESOPEC trial and other studies comparing neoadjuvant chemotherapy and neoadjuvant CRT in esophageal and GEJ adenocarcinoma are discussed separately. (See "Neoadjuvant and adjuvant therapy for locally advanced resectable esophagogastric junction and gastric cardia adenocarcinoma", section on 'Perioperative chemotherapy versus neoadjuvant CRT'.)

●Thoracic esophageal SCC – For most patients with a cT3 to T4 or node-positive resectable esophageal SCC, we suggest initial therapy with CRT rather than chemotherapy. However, we recognize that clinical practice is variable, and initial chemotherapy (either in the neoadjuvant or perioperative setting) is an acceptable strategy that is used frequently in regions of the world where esophageal SCC is more prevalent, such as in Asia.

Patients who are treated with initial CRT may subsequently be evaluated for surgery versus nonoperative management, depending upon the treatment response; and patients who are treated with surgery should be subsequently evaluated for adjuvant therapy, if necessary. (See 'Chemoradiation responders' above and 'Role of postoperative (adjuvant) therapy' below.)

In patients with thoracic esophageal SCC, compared with initial chemotherapy, initial CRT is generally better tolerated, results in higher rates of pCRs and R0 resections, and offers the potential for subsequent nonoperative management in clinical responders. However, most studies suggest that neoadjuvant CRT and neoadjuvant chemotherapy result in similar OS [121-124]. As examples:

•In an open-label phase III trial (JCOG1109 NExT), 601 patients with previously untreated locally advanced thoracic esophageal SCC were randomly assigned to one of three neoadjuvant treatment arms, followed by esophagectomy [121]:

-Doublet chemotherapy with two cycles of cisplatin and FU every three weeks

-Triplet chemotherapy with three courses of FU, cisplatin, and docetaxel every three weeks

-Neoadjuvant CRT (Two cycles of cisplatin plus FU every four weeks concurrent with RT at 40.1 Gy)

•At a median follow-up of 51 months, results were as follows:

-In a post-hoc analysis, neoadjuvant triplet chemotherapy resulted in higher OS compared with neoadjuvant CRT, but the difference was not statistically significant (three-year OS 72 versus 68 percent, HR 0.80, 95% CI 0.59-1.10).

-Neoadjuvant triplet chemotherapy improved OS over doublet chemotherapy (three-year OS 72 versus 63 percent, HR 0.68, 95% CI 0.50-0.92).

-Neoadjuvant CRT resulted in higher OS compared with neoadjuvant doublet chemotherapy, but the difference was not statistically significant (three-year OS 68 versus 63 percent, HR 0.84, 95% CI 0.63-1.12).

-Rates of pCR were higher for neoadjuvant CRT (39 percent) than triplet chemotherapy (20 percent) and doublet chemotherapy (2 percent).

-Relative to neoadjuvant CRT, neoadjuvant triplet chemotherapy resulted in higher rates of grade 3 or 4 neutropenia (85 versus 45 percent), grade 3 or 4 febrile neutropenia (16 versus 5 percent), and treatment discontinuation (9 versus 6 percent). However, relative to triplet chemotherapy, neoadjuvant CRT resulted in higher rates of esophagitis (61 versus 5 percent).

•In another randomized clinical trial, 264 patients with locally advanced (cT3-4a, N0-N1, M0) esophageal SCC received either neoadjuvant CRT or neoadjuvant chemotherapy (both with cisplatin plus paclitaxel), followed by minimally invasive esophagectomy [122,125]. Compared with neoadjuvant chemotherapy, neoadjuvant CRT improved pCR rates (28 versus 3 percent) but failed to improve OS (64 versus 55 percent, HR 0.82, 95% CI 0.58-1.18) [122].

•Most meta-analyses of randomized clinical trials, but not all [29], demonstrate similar OS for neoadjuvant chemotherapy and neoadjuvant CRT in thoracic esophageal SCC [52-54,126]. As an example, in one network meta-analysis that included four randomized trials comparing neoadjuvant CRT (either concurrent or sequential) to neoadjuvant chemotherapy [31,123,124,127], OS was similar between the two treatment arms (HR 0.90, 95% CI 0.74-1.09) [54]. In a subgroup analysis, OS was also similar for neoadjuvant CRT and neoadjuvant chemotherapy regardless of histology (SCC versus adenocarcinoma) or tumor location (thoracic esophagus versus GEJ).

Neoadjuvant versus postoperative chemotherapy — The superiority of neoadjuvant as compared with postoperative adjuvant chemotherapy was shown in a Japanese trial (JCOG9907) in which 330 patients with clinical stage II or III SCC of the thoracic esophagus were randomly assigned to surgery preceded or followed by two 21-day courses of cisplatin (80 mg/m² on day 1) plus infusional FU (800 mg/m² daily for five days) [114]. Five-year OS was significantly higher in the group receiving preoperative chemotherapy (55 versus 43 percent).

Impact of preoperative treatment on local control — Although most trials focus upon survival as a primary endpoint, locoregional control is also important when selecting among treatment options. Local failures can be defined as recurrent local disease following a margin-negative esophagectomy or after a margin-negative or margin-positive resection. Using either definition, the frequency of local failure appears to be higher in trials in which patients were treated with surgery alone or definitive CRT alone as compared with those receiving CRT followed by surgery [18,19,30,59,128,129]. (See 'Patterns of failure' below.)

As examples:

●In a combined analysis of the Dutch CROSS trial (described above) in conjunction with data from the preceding phase II trial investigating the same preoperative regimen followed by surgery, there were 374 patients who underwent resection, 49 percent allocated to surgery and 51 percent to CRT followed by surgery [129]. At a minimum follow-up of 24 months, patients undergoing preoperative CRT had significantly lower rates of recurrent disease overall (35 versus 57 percent), locoregional recurrence (14 versus 34 percent), and isolated locoregional recurrence (3.3 versus 9.3 percent). (See 'Concurrent chemoradiation' above.)

●In the previously described meta-analysis of trials comparing preoperative CRT with surgery alone [128], the odds ratio (OR) for locoregional recurrences was significantly lower with preoperative therapy (OR for local recurrence 0.38, 95% CI 0.23-0.63). (See 'Concurrent chemoradiation' above.)

Whether local recurrence rates are lower in patients treated with neoadjuvant chemotherapy alone as compared with surgery is uncertain; the following data are available (see 'Neoadjuvant chemotherapy' above):

●In the MAGIC trial, local failure was confirmed before death in fewer patients in the chemotherapy group as compared with surgery alone (14 versus 20 percent), but the p value for the comparison was not reported [112].

●In the MRC trial, the local failure rates were nearly identical in the chemotherapy and surgery groups (12 and 11 percent, respectively) [92]. The meta-analysis of preoperative chemotherapy versus surgery alone did not address the issue of locoregional control [130]. (See 'Meta-analyses' above.)

Clinical T1N0 disease — Patients with clinical Tis or T1N0 thoracic esophageal adenocarcinoma or SCC (table 1) are treated with initial resection. Options include esophagectomy or endoscopic resection. Patients who are not surgical or endoscopic resection candidates can be treated with radiation with or without chemotherapy. Further details, including selecting between these management strategies, are discussed separately. (See "Management of locally advanced unresectable or inoperable esophageal cancer".)

Patients who are treated with surgery should also be evaluated for adjuvant therapy, if necessary, based on the postoperative pathology. (See 'Role of postoperative (adjuvant) therapy' below.)

Clinical T2N0 disease — For most patients with clinical T2N0 thoracic esophageal cancer that is resectable, regardless of histology, we suggest combined modality therapy rather than surgery alone. For those with T2N0 thoracic esophageal SCC who are low-risk (ie, well-differentiated, <2 cm in size, no lymphovascular invasion), upfront surgery is an appropriate alternative.

Patients who complete initial combined modality therapy should subsequently be evaluated for surgical resection; and patients who complete surgery should be evaluated for adjuvant therapy, if necessary, based on the postoperative pathology. (See 'Necessity for surgery' above and 'Role of postoperative (adjuvant) therapy' below.)

The optimal approach to clinical T2N0 disease is debated, and guidelines from expert groups differ. Some suggest initial CRT for SCC, and either neoadjuvant chemotherapy or CRT for adenocarcinomas of the distal esophagus or GEJ [131]. Others, including the NCCN [14], suggest initial resection for clinical T2N0 adenocarcinomas or SCCs as long as they are <3 cm and well differentiated, but initial CRT for others with either histology who have high-risk disease. Year 2020 ASCO guidelines also state that initial surgery may be more appropriate for patients with low-risk clinical T2N0 lesions (ie, well-differentiated, <2 cm) [53].

The contribution of neoadjuvant CRT to outcomes in clinical T2N0 thoracic esophageal cancer is uncertain. These patients were included in the three positive trials examining the benefit of preoperative CRT, although the actual representation is only known for CALGB 9781 (3 of the total 56 enrolled patients) [34,35,45]. None of the trials stratified outcomes according to histologic stage.

As noted above, the French FFCD 9901 trial failed to show a benefit for preoperative CRT compared with surgery alone in 195 patients with stage I or II esophageal or GEJ cancer [46]. Thirty-seven of the enrolled patients (19 percent) had clinical stage I disease. It is possible that the study was underpowered to show a significant survival benefit, if one was present [132]. (See 'Concurrent chemoradiation' above.)

Although several retrospective studies and a systematic review of several of these studies have examined whether surgical resection or neoadjuvant therapy provides better outcomes for these patients, they have come to disparate conclusions, and there is no consensus as to the best approach [133-138]. Guidelines from the NCCN suggest that patients with clinical T2N0 esophageal cancer be approached similarly to those with more advanced stage potentially resectable disease (ie, with initial CRT preferred) [14]. However, a major problem is the accuracy of preoperative staging. In one report of 499 patients with clinical T2N0 esophageal cancer, preoperative staging was accurate in only 14 percent, and patients were both understaged (50 percent) and overstaged (36 percent) [139]. There is a need for better clinical staging modalities. Increasingly, endoscopic resection is being pursued in this patient population to improve the selection of patients who need surgery [140]. (See "Endoscopic ultrasound for evaluating patients with esophageal cancer".)

Role of postoperative (adjuvant) therapy

After preoperative therapy — Patients with a pCR on the surgical specimen after initial CRT do not require further therapy and may proceed to post-treatment surveillance. (See 'Post-treatment cancer surveillance' below.)

For patients without a pCR (ie, with residual disease) on the surgical specimen after initial CRT, regardless of histology, we suggest adjuvant nivolumab for up to one year. For patients who are ineligible for nivolumab, options include post-treatment surveillance (for ypN0-1 tumors only) or adjuvant chemotherapy with different agents than those given preoperatively (eg, FOLFOX if the initial CRT regimen included concurrent carboplatin plus paclitaxel).

●Nivolumab – Benefit for nivolumab was shown in the CheckMate 577 trial, in which 794 patients who had received neoadjuvant CRT for esophageal or GEJ cancer (70 percent adenocarcinoma) and had residual pathologic disease at the time of surgery were randomly assigned to nivolumab (240 mg) or placebo every 2 weeks for 16 weeks followed by nivolumab 480 mg or placebo every 4 weeks; the maximum treatment duration was one year [141]. Enrollment was irrespective of programmed death receptor-1 ligand 1 (PD-L1) overexpression. Tumor site was esophagus in 60 percent and GEJ in 40 percent; histology was adenocarcinoma in 71 percent and SCC in 29 percent. At a median follow-up of 24.4 months, median disease-free survival, the primary endpoint, was twice as long with nivolumab (22.4 versus 11 months, HR for disease progression or death was 0.69, 95% CI 0.56-0.86), and the benefits were seen across all patient subgroups (histology, location, initial and post-treatment disease stage, PD-L1 overexpression or not). OS data were not mature. Although treatment-related adverse effects were frequent, most were grade 1 or 2 and only 9 percent of patients discontinued adjuvant nivolumab because of adverse effects. The benefits were gained without any significant decline in patient-reported health-related quality of life over the year of nivolumab treatment.

Based on these results, the US Food and Drug Administration has approved nivolumab as adjuvant therapy for patients who previously received neoadjuvant CRT following complete resection of esophageal or GEJ cancer with residual pathologic disease.

There is no established clinical role for the addition of immunotherapy to perioperative chemotherapy or postoperative (adjuvant) chemotherapy in esophageal cancer. Further details are discussed separately. (See "Neoadjuvant and adjuvant therapy for locally advanced resectable esophagogastric junction and gastric cardia adenocarcinoma", section on 'Is there a role for adding immunotherapy to perioperative chemotherapy?'.)

●Chemotherapy – In the postoperative setting, there are concerns as to tolerability of further chemotherapy [38,112,142], but a potential survival benefit for adjuvant chemotherapy in patients with and without residual nodal disease following CRT was suggested in two retrospective analyses derived from the National Cancer Database of patients with esophageal cancer who were treated with neoadjuvant CRT followed by resection [143,144], and a meta-analysis of 10 studies involving 6462 patients, nine of which were cohort studies; most utilized neoadjuvant CRT rather than induction chemotherapy [145]. When compared with neoadjuvant therapy and esophagectomy alone, adjuvant therapy groups had a significantly lower mortality at both one-year (RR 0.52, 95% CI 0.41-0.65) and five-year follow-up (RR 0.91, 95% CI 0.86-0.96).

Although the optimal regimen is not established, administering adjuvant chemotherapy with different agents than those given preoperatively is a reasonable approach. For example, patients who were treated with CRT with carboplatin plus paclitaxel could receive adjuvant chemotherapy with FOLFOX, and vice-versa.

No prior neoadjuvant therapy — For patients with completely resected node-positive or pathologic T3 or T4 node-negative thoracic esophageal adenocarcinoma who have not received neoadjuvant therapy, we suggest adjuvant chemotherapy in an attempt to improve outcomes. We also suggest adjuvant chemotherapy for those with high-risk pathologic T2N0 adenocarcinoma (ie, those that are poorly differentiated, have lymphovascular or perineural invasion, or arise in a patient under the age of 50). We do not offer postoperative CRT, as data are unclear on the benefits of this approach.

For patients with resected thoracic esophageal SCC who have not received neoadjuvant therapy, we only offer postoperative therapy if the margins are positive. It is difficult to come to any conclusions as to whether there are specific advantages for adjuvant CRT over chemotherapy alone, and either approach is reasonable. Further confirmatory trials, particularly randomized trials, are necessary before specific recommendations can be made.

Chemoradiation — For patients with a node-positive adenocarcinoma of the GEJ, postoperative CRT is a standard approach, at least in the United States, based upon results from the United States Intergroup trial. (See "Neoadjuvant and adjuvant therapy for locally advanced resectable esophagogastric junction and gastric cardia adenocarcinoma", section on 'Adjuvant CRT'.)

For other patients, particularly those with a thoracic esophageal SCC, the optimal approach is uncertain. Some uncontrolled trials and retrospective comparisons of patients treated with and without CRT suggest potential benefit for postoperative CRT. However, others do not, and there is only a single randomized trial demonstrating benefit of adjuvant therapy as compared with surgery alone:

●In a retrospective report, outcomes of 38 patients with node-positive disease after esophagectomy alone who received postoperative CRT (concurrent or sequential RT plus cisplatin and FU with or without epirubicin) were compared with 28 similar patients who did not receive further therapy [146]. Local recurrence rates were lower in the group receiving postoperative therapy (35 versus 13 percent), and the median OS was longer (48 versus 14 months).

●Mature data from a prospective phase II uncontrolled trial also support benefit for adjuvant CRT in this setting. In this trial, 50 patients with locally advanced esophageal cancer (90 percent T3, 81 percent node-positive, 13 percent with extraregional nodal metastases) received CRT (RT 50.4 to 59.4 Gy plus concurrent cisplatin and FU) after esophagectomy [147]. At a median follow-up of 47 months, four-year projected survival was 51 percent, a rate that is far higher than might be expected for this relatively poor prognosis cohort.

●Benefit was also suggested in a Chinese trial in which 172 patients with stage IIb-III esophageal SCC were randomly assigned to surgery alone (n = 54) or followed by postoperative RT alone (n = 54) or CRT (n = 64), concurrent chemotherapy was with paclitaxel plus cisplatin or nedaplatin [148]. When both postoperative groups were combined, postoperative therapy was associated with a significantly improved three-year disease-free (53.8 versus 36.7 percent) and OS (63.9 versus 48 percent).

Benefit for CRT over RT alone was suggested in a retrospective review of 304 patients with thoracic esophageal SCC who had undergone upfront esophagectomy with a three-field lymph node dissection and determined to have lymph node but not distant metastases [149]. Postoperative chemotherapy (cisplatin plus paclitaxel) plus RT (50 Gy) was administered to 164, while 140 underwent postoperative RT alone (50 Gy). Margin status was not addressed. Five-year OS rates were significantly better with CRT (47 versus 39 percent), and recurrence rates (including distant, combined regional and distant, and overall) were all significantly less with chemotherapy. Although early toxicity was significantly more common with CRT, there were no significant differences in late toxic effects between the two groups.

Chemotherapy alone — For patients with esophageal cancer who have not received preoperative chemotherapy or CRT, postoperative chemotherapy alone may be beneficial, although proof of a survival benefit from randomized trials with a surgery alone control group is scant [150-153]. The use of adjuvant chemotherapy in thoracic esophageal adenocarcinoma is mainly extrapolated from data in resected gastric adenocarcinoma. (See "Adjuvant and neoadjuvant treatment of gastric cancer", section on 'Adjuvant chemotherapy'.)

One randomized trial compared surgery alone versus surgery followed by adjuvant chemotherapy (two courses of cisplatin 80 mg/m² on day 1 and FU 800 mg/m² daily for five days) included 242 patients with esophageal SCC recruited from 17 Japanese institutions [153]. The five-year disease-free survival rate (the primary endpoint) was significantly better with chemotherapy (55 versus 45 percent), but OS was not significantly different (61 versus 52 percent). The short duration of adjuvant chemotherapy, and the fact that 25 percent of the patients in the chemotherapy group did not receive both full courses of therapy, may have compromised the ability to document a survival difference.

Post-treatment cancer surveillance

Patterns of failure — Upper or mid-thoracic esophageal SCCs (figure 1) tend to recur locoregionally first, while distal esophageal adenocarcinomas more commonly recur with distant dissemination. Most recurrences develop within one year, and recurrences tend to develop earlier in patients treated with neoadjuvant therapy as compared with surgery alone. This was illustrated in a series of 590 patients who underwent esophagectomy for adenocarcinoma [154]. The peak interval for recurrence after esophagectomy alone was six to nine months, and more than 90 percent of the disease recurrences occurred by three years. By contrast, among patients treated with neoadjuvant CRT (ie, trimodality therapy), the peak time frame for recurrence was the first three months, and >90 percent of recurrences were evident by 21 months. The pattern of recurrence was distant, locoregional, or both in 60, 30, and 10 percent of patients, respectively, and did not differ in patients treated with surgery alone. Of note, these results may be impacted by selection bias, as patients with more advanced tumors likely had an increased likelihood of receiving neoadjuvant therapy. A similar distribution of recurrences (distant, locoregional, or both in 55, 28, and 17 percent) have been reported by others following trimodality therapy [155].

On the other hand, isolated local recurrences are more frequent after definitive CRT, and salvage surgery may benefit a greater number of these patients. This was shown in a retrospective analysis of 276 patients with esophageal cancer (78 percent adenocarcinoma) who were treated with definitive CRT at MD Anderson Cancer Center over a nine-year period (2002 to 2011) [99]. The site of first failure was local only in 64 (23 percent); and 23 (36 percent, 8 percent of the entire cohort) of these were amenable to salvage surgery. At a median follow-up of 54 months for the entire cohort, the estimated three- and five-year OS rates for those undergoing salvage surgery were 61 and 45 percent, respectively. Ninety-one percent of the local recurrences developed within two years, suggesting that vigilant surveillance is more important in this time frame. (See 'Impact of preoperative treatment on local control' above and 'Necessity for surgery' above.)

Surveillance strategy — There are no randomized trials to guide the postoperative surveillance strategy and no data that demonstrate improvement in quality of life or longevity from earlier detection of asymptomatic recurrences. At our institutions, we perform history, physical examination, and targeted blood work (for a symptomatic patient, or if there was a serum tumor marker that was elevated preoperatively) every four months for the first three years and also perform restaging contrast-enhanced CT scans of the chest and abdomen at four-month intervals. We do not carry out surveillance endoscopy unless there was a preoperative history of Barrett's esophagus, a questionable margin at the time of surgery, or if the patient has a recalcitrant stricture that is worrisome for an occult local recurrence. More vigilant surveillance in the first two years after treatment may be warranted in patients who underwent definitive CRT.

The primary purpose of post-treatment surveillance is to implement a potentially beneficial salvage therapy in cases of locoregional failure. However, the incidence of locoregional failure is low, particularly after trimodality therapy, and the number of potentially curable recurrences that will be detected by intensive post-treatment surveillance is small.

This was demonstrated in a report of 518 patients with esophageal adenocarcinoma who were treated with preoperative CRT and followed for a median of 29.3 months [156]. The post-treatment surveillance strategy included CT or PET-CT every three months for the first year, every six months for two additional years, then annually for at least five years. Endoscopic examination was performed every six months for 18 months, then annually. Isolated locoregional failure developed in 27 (5 percent), only 11 of which were intraluminal. By contrast, distant metastases developed in 188 (36 percent). Twelve patients with locoregional failure had salvage CRT, and four underwent salvage surgery (three of whom later developed metastatic disease). Overall, only 10 of the 27 patients with a locoregional failure survived longer than 10 years. Thus, only 2 percent of the 518 patients benefited from post-treatment surveillance.

Consensus-based guidelines from the NCCN suggest the following [14]:

●History and physical examination every three to six months for years one and two, then every six to twelve months for years 3 through 5, then annually.

●Complete blood count (CBC) and chemistry profile, as clinically indicated

●Radiologic imaging and upper gastrointestinal (GI) endoscopy, as clinically indicated

●Dilation for anastomotic stenosis

●Nutritional counseling

By contrast, consensus-based guidelines from the European Society for Medical Oncology emphasize the lack of evidence that regular follow-up after initial therapy has an impact on survival outcomes, with the possible exception of patients who might be potential candidates for endoscopic reintervention or early "salvage surgery" after failing definitive CRT [131]. They advise that follow-up visits concentrate on symptoms, nutrition, and psychosocial support. In the case of a complete response to CRT and no surgery, a three-month follow-up based on endoscopy, biopsy, and CT scan may be recommended to detect early recurrence, leading to a discussion about salvage surgery. (See 'Patterns of failure' above.)

Long-term follow-up is needed to address treatment-related late complications associated with high-dose RT.

When planning the post-treatment surveillance strategy, care should be taken to limit the number of CT scans, particularly in younger individuals, given concerns about radiation exposure and the risk for second malignancies. (See "Radiation-related risks of imaging".)

CERVICAL ESOPHAGUS TUMORS — 

Squamous cell carcinoma (SCC) of the cervical or upper esophageal sphincter esophagus presents a unique management situation. If surgery is performed, it usually requires removal of portions of the pharynx, the larynx, the thyroid gland, and portions of the proximal esophagus. In addition, radical neck dissections are usually carried out; as such, the management is more closely related to SCC of the head and neck than for malignancies involving the more distal portions of the esophagus. In general, radiation therapy (RT) combined with chemotherapy is preferred over surgery for proximal esophageal cancers where laryngectomy would be necessary for a good cancer operation since survival appears to be comparable and major morbidity is avoided in most patients [157]. The management of cervical esophageal tumors is discussed separately. (See "Surgical management of resectable esophageal and esophagogastric junction cancers", section on 'Cervical esophageal cancer' and "Management of locally advanced unresectable or inoperable esophageal cancer", section on 'Cervical esophageal cancer'.)

SOCIETY GUIDELINE LINKS — 

Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Esophageal cancer".)

INFORMATION FOR PATIENTS — 

UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)

●Basics topic (see "Patient education: Esophageal cancer (The Basics)")

SUMMARY AND RECOMMENDATIONS

●General principles – Locally advanced cancers of the thoracic esophagus (figure 1), which include both adenocarcinoma and squamous cell carcinoma (SCC) histologies, are typically managed using combined multimodality therapy with surgery, systemic therapy, and/or radiation therapy (RT). (See 'Introduction' above.)

●Initial therapy for clinical Tis or T1N0 disease – Patients with clinical Tis or T1N0 thoracic esophageal adenocarcinoma or SCC are treated with initial resection. Options include surgery (esophagectomy) or endoscopic resection. Patients who are not surgical or endoscopic resection candidates can be treated with RT with or without chemotherapy. Further details, including selecting between these management strategies, are discussed separately. (See "Management of superficial esophageal cancer".)

Patients who are treated with surgery should also be evaluated for adjuvant therapy, if necessary, based on the postoperative pathology. (See 'Role of postoperative (adjuvant) therapy' above.)

●Initial therapy for clinical T2N0 disease – For most patients with clinical T2N0 thoracic esophageal cancer that is resectable, regardless of histology, we suggest combined modality therapy rather than surgery alone (Grade 2C). For those with T2N0 thoracic esophageal SCC who are low-risk (ie, well-differentiated, <2 cm in size, no lymphovascular invasion), upfront surgery is an appropriate alternative. (See 'Clinical T2N0 disease' above.)

Patients who complete initial combined modality therapy should subsequently be evaluated for surgical resection; and patients who complete surgery should be evaluated for adjuvant therapy, if necessary, based on the postoperative pathology. (See 'Necessity for surgery' above and 'Role of postoperative (adjuvant) therapy' above.)

●Initial therapy for clinical T3-T4 or node-positive disease – For patients with clinical T3-T4 or node-positive thoracic esophageal adenocarcinoma or SCC that is resectable (table 1), we recommend combined modality therapy rather than surgery alone (Grade 1B). (See 'Clinical T3-4 or node-positive disease' above and 'Neoadjuvant chemotherapy' above and 'Preoperative chemoradiation' above.)

The approach to combined modality therapy is histology dependent:

•Adenocarcinoma – For most patients with thoracic esophageal adenocarcinoma who are younger with good performance status and minimal comorbidities, we suggest perioperative chemotherapy with fluorouracil, leucovorin, oxaliplatin, and docetaxel (FLOT) rather than neoadjuvant chemoradiation (CRT) (Grade 2B). For older patients, those with multiple comorbidities, and/or those who are unlikely to tolerate the toxicities of FLOT (eg, neuropathy), neoadjuvant CRT is an appropriate alternative. Patients treated with neoadjuvant CRT followed by surgery should also be evaluated for adjuvant therapy (eg, adjuvant immunotherapy for those without a pathologic complete response [pCR]). Further details are discussed separately. (See "Neoadjuvant and adjuvant therapy for locally advanced resectable esophagogastric junction and gastric cardia adenocarcinoma", section on 'Perioperative chemotherapy versus neoadjuvant CRT'.)

•SCC – For patients with thoracic esophageal SCC, we suggest initial therapy with CRT rather than chemotherapy (Grade 2C). Compared with initial chemotherapy, initial CRT confers similar overall survival (OS) benefits, but is generally better tolerated, results in higher rates of pCRs and R0 resections, and offers the potential for subsequent nonoperative management in clinical responders. However, we recognize that practice is variable, and initial treatment with chemotherapy (either in the perioperative or neoadjuvant setting) is an acceptable strategy that is used frequently in regions of the world where esophageal SCC is more prevalent, such as Asia. (See 'Chemotherapy versus chemoradiation' above.)

Patients who are treated with initial CRT may be subsequently evaluated for surgery versus nonoperative management, depending upon the treatment response; and patients who are treated with surgery should be subsequently evaluated for adjuvant therapy, if necessary. (See 'Necessity for surgery' above and 'Role of postoperative (adjuvant) therapy' above.)

●Ineligible for surgery – Regardless of histology, for patients who are not surgical candidates or those in whom surgery is not technically feasible, definitive CRT is appropriate. (See "Management of locally advanced unresectable or inoperable esophageal cancer".)

●Selection of radiosensitizing regimen for CRT – For patients who are treated with CRT, we suggest concurrent radiosensitization with weekly carboplatin plus paclitaxel (table 2) rather than other systemic regimens due to better tolerance and ease of administration (Grade 2C). FOLFOX is an alternative radiosensitizing regimen. Selection of regimen and rationale is the same as for gastroesophageal junction (GEJ) and gastric cardia cancers and is discussed separately. (See "Neoadjuvant and adjuvant therapy for locally advanced resectable esophagogastric junction and gastric cardia adenocarcinoma", section on 'Selection of radiosensitizing regimen'.)

●Role of surgery following initial CRT

•Adenocarcinoma – For patients with clinically resectable thoracic esophageal adenocarcinoma who are treated with initial CRT, we suggest inclusion of surgery rather than definitive CRT alone (Grade 2C). (See 'Necessity for surgery' above.)

•SCC – For those with clinically resectable thoracic esophageal SCC who are treated with initial CRT, we also suggest inclusion of surgery rather than definitive CRT (Grade 2C). However, nonoperative management is an alternative option for those who have an endoscopic complete response to initial CRT. (See 'Chemoradiation responders' above.)

•Nonresponders to CRT – For patients whose disease does not respond to initial CRT, regardless of histology, surgery is an important component of therapy for those whose tumors remain operable. (See 'Nonresponders' above.)

•What is the role of post-CRT PET imaging? – We do not use post-CRT fluorodeoxyglucose (FDG)-positron emission tomography (PET) scanning to evaluate the local response to therapy as a means of selecting those patients for whom surgery might be avoided. (See 'Utility of postinduction therapy PET scans' above.)

●Postoperative (adjuvant) therapy

•No prior neoadjuvant therapy – For patients with completely resected node-positive or pathologic T3 or T4 node-negative thoracic esophageal adenocarcinoma who have not received neoadjuvant therapy, we suggest postoperative (adjuvant) chemotherapy (Grade 2C). We also suggest adjuvant chemotherapy for selected patients with high-risk (poorly differentiated histology, lymphovascular or perineural invasion, or patients younger than age 50), resected pathologic T2N0 adenocarcinoma (Grade 2C) We do not offer postoperative CRT, as data are unclear on the benefits of this approach. (See 'No prior neoadjuvant therapy' above.)

For patients with resected thoracic esophageal SCC who have not received neoadjuvant therapy, we only offer postoperative therapy if the margins are positive. It is difficult to come to any conclusions as to whether there are specific advantages for adjuvant CRT over chemotherapy alone, and either approach is reasonable. Further randomized trials are necessary.

•After initial CRT – Patients with a pathologic complete response (pCR) on the surgical specimen after initial CRT do not require further therapy and may proceed to post-treatment surveillance. (See 'Post-treatment cancer surveillance' above.)

For patients without a pCR (ie, residual disease) on the surgical specimen after preoperative CRT, regardless of histology, we suggest adjuvant nivolumab for up to one year (Grade 2B). For those who ineligible for nivolumab, options include post-treatment surveillance (for ypN0-1 tumors only) or adjuvant chemotherapy with different agents than those given preoperatively (eg, FOLFOX (table 4) if the initial CRT regimen included carboplatin plus paclitaxel, and vice-versa). (See 'After preoperative therapy' above.)

●Post-treatment surveillance

•There are no randomized trials to guide the postoperative surveillance strategy and no data that demonstrate improvement in quality of life or longevity from earlier detection of asymptomatic recurrences. We perform history, physical examination, targeted blood work, and contrast-enhanced CT of the chest and abdomen every four months for the first three years. (See 'Surveillance strategy' above.)

•We do not carry out surveillance endoscopy unless there was a preoperative history of Barrett's esophagus, a questionable margin at the time of surgery, or if the patient has a recalcitrant stricture that is worrisome for an occult local recurrence.

ACKNOWLEDGMENT — 

The UpToDate editorial staff acknowledges Noah C Choi, MD, who contributed to an earlier version of this topic review.