Multimodality Therapy in the Management of Locally Advanced Esophageal Cancer

Does Perioperative Chemotherapy Improve Survival in Patients With Esophageal Cancer and, If So, Does This Depend on Histology?

Table 38.1 lists high-quality randomized studies comparing survival outcomes in patients receiving perioperative chemotherapy compared with curative intent surgery alone. A number of early studies failed to demonstrate a clear survival advantage in patients receiving preoperative chemotherapy compared with curative intent surgery alone. However, despite these overarching findings, important information regarding the utility of chemotherapy in the management of EC can be inferred. For example, in the negative studies conducted by Roth et al., Schlag, Law et al., Ancona et al., and Kelsen et al., subgroup analysis demonstrates improved survival outcomes in patients who demonstrate a response to the preoperative regimen.

Roth et al. demonstrated improved survival in squamous cell carcinoma (SCC) patients who demonstrated a major (47%) or complete (5%) response to vinblastine-, cisplatin-, and bleomycin-based chemotherapy (median survival, 20 months vs. 6 months; P = .008). Patients who responded to preoperative therapy also fared better than patients who received surgery alone. Schlag similarly demonstrated significant increases in survival for patients with SCC who demonstrated a response (minor 12%, major 32%, complete 6%) to 3 cycles of cisplatin and 5-fluorouracil (5FU)-based chemotherapy. Schlag et al. randomized SCC patients to receive 2 cycles of cisplatin and 5FU-based chemotherapy in the preoperative setting or to curative intent surgery alone. The authors demonstrated reduced recurrence rates, predominantly as a result of improved locoregional control in patients who received chemotherapy compared with surgery alone. This finding likely relates to higher R0 resection rates in chemotherapy-treated patients (67% vs. 35%; P = .003). This did not translate into an overall survival (OS) advantage (median and 2-year survival in chemotherapy vs. surgery alone, 16.8 months vs. 13 months, and 44% vs. 31%, respectively; P = .17). However, in patients who demonstrated a response, as in the previous studies, median and 2-year survival times were improved (chemotherapy vs. surgery, 42.2 months vs. 13.8 months; P = .008 and 59% vs. 33%, respectively). Ancona et al. similarly randomized patients with SCC to two cycles of preoperative cisplatin and 5FU or curative intent surgery alone. Combined complete and major response rates of resected tumors in the neoadjuvant arm were 40%, with a complete response rate of 12.8%. No difference in median survival was noted on an intention to treat basis (24 and 25 months for surgery alone vs. neoadjuvant chemotherapy, respectively). In the 40% of patients who demonstrated a major response to chemotherapy, a significant improvement with respect to median (53 months) and 3-year (74%) and 5-year (60%) survival was observed compared with patients undergoing surgery alone (28 months, 46%, 26%; P = .01) and nonresponders to chemotherapy (19 months, 38%, 19%; P < .05). In keeping with this theme, the survival benefit was most pronounced in the 12.8% of patients who demonstrated a complete response, leading the authors to conclude that pathologic response is a significant determinant of long-term outcome in addition to R0 resection.

Kelsen et al. randomized patients with both SCC and esophageal adenocarcinoma (EAC) in roughly equal proportions to three cycles of preoperative cisplatin and 5FU or curative intent surgery alone. Complete response rate in patients receiving chemotherapy was 2.5%, with a major objective response observed in 19%. In keeping with previous studies, survival was only improved in patients who demonstrated a major response (response vs. no response, hazard ratio [HR], 2.83; 95% confidence interval [CI], 1.84 to 4.35; P < .001).

Collectively, these data suggest that effective chemotherapy, as measured by an objective regression response, is protective and confers a significant survival advantage, in addition to R0 resection in patients with both SCC and EAC compared with surgery alone. Furthermore, effective chemotherapy may improve R0 resection rates and thus contribute to improved survival through improved locoregional control.

More recent studies, including the MRC/OE2, Boonstra, Ychou, and Cunningham (MAGIC) trials, were positive on an intent to treat basis, supporting the use of neoadjuvant chemotherapy over surgery alone in patients with locally advanced esophageal cancer. In these studies, histologies were commonly mixed. Therapeutic regimens included cisplatin and 5FU-based doublets, with the exception of the Cunningham (MAGIC) trial, which administered triplet therapy encompassing an anthracycline in addition to cisplatin and 5FU. All studies administered chemotherapy in the preoperative and postoperative periods, with the exception of the MRC study, where chemotherapy was administered in the preoperative setting only.

In the MRC study and its subsequent follow-up (OE2), preoperative cisplatin and 5FU was associated with a significant reduction in primary tumor size and regional lymph node positivity, compared with specimens from untreated patients. This was associated with improved R0 resection rates in patients subject to neoadjuvant chemotherapy (60% vs. 54%, P < .001). Similarly, OS was improved in patients subject to neoadjuvant treatment (median and 2-year survival, neoadjuvant chemotherapy followed by surgery versus surgery alone: 16.8 months vs. 13.3 months and 43% vs. 34%, respectively; HR, 0.79; 95% CI, 0.67 to 0.93; P = .004). In long-term follow-up (OE2), patients subject to chemotherapy demonstrated reduced T and N stage compared with patients receiving surgery alone following pathologic analysis. Overall 5-year survival was 23% in neoadjuvant group versus 17.1% in patients randomized to surgery alone ( P < .001).

Boonstra et al. conducted a clinical trial comparing perioperative chemotherapy using a regimen of cisplatin and etoposide in patients with SCC. Following chemotherapy, a partial response rate of 40% and complete response rate of 7% were observed. No difference in R0 resection (71% vs. 57%) rate or lymph node positivity between the two treatment arms was noted; however, significantly more patients in the surgery alone arm demonstrated unresectable tumors or underwent an R2 resection. Accordingly, a significant survival advantage in patients receiving chemotherapy was noted (median, 2- and 5-year survival CS vs. surgery alone, 16 months vs. 12 months, 42% vs. 30%, and 26% vs. 17%, respectively; P = .03).

The trials by Ychou and Cunningham et al. included patients with adenocarcinoma of the esophagus and stomach. Ychou et al. randomized patients, of whom 75% harbored lower esophageal or gastroesophageal (GE) junction tumors to receive perioperative cisplatin and 5FU versus surgery alone. Patients randomized to preoperative therapy demonstrated improved outcome with respect to R0 resection rate (87% vs. 74%; P = .04) and OS (5-year survival chemotherapy vs. surgery alone 38% vs. 24%; P < .05). In the study by Cunningham et al., patients with gastric and lower esophageal/GE junction tumors (25%) were randomized to receive perioperative chemotherapy with 5FU, cisplatin, and epirubicin. In patients receiving perioperative therapy, an improvement in R0 (CS vs. surgery alone, 79.3% vs. 70.3%; P = .03) resection rate and a tendency for smaller tumors and less advanced nodal disease was noted. In keeping with these findings, improved OS was noted in the chemotherapy arm compared with surgery alone (HR, 0.75; 95% CI, 0.6 to 0.93; P = .009; 5-year survival, 36.3% vs. 23%).

Collectively, the data demonstrate that when given in the perioperative setting, chemotherapy can effectively reduce tumor burden, facilitate curative resection, and impart a significant survival benefit in patients with locally advanced SCC or EAC. With respect to whether efficacy differs according to tumor histology, overall the data suggest that chemotherapy provides a benefit in both SCC and EAC. The MRC/OE2 study specifically assessed treatment effects according to histology and found no evidence of a difference in the two, with a significant reduction in both T and N stage overall. Furthermore, the benefits of chemotherapy with respect to survival did not differ according to histology, with absolute 5-year survival rates in chemotherapy versus surgery alone of 22.6% versus 17.6% in EAC patients and 25.5% versus 17% in SCC patients, respectively. Thus effective chemotherapy may be considered in esophageal cancer patients in addition to surgery, regardless of histology.

Do Triplet Regimens Confer an Advantage Over Doublets?

The studies mentioned thus far have predominantly employed chemotherapy regimens composed of a doublet. These are typically composed of a platinum agent and 5FU. The rationale for this regimen stems from the observation of significant response rates in conjunction with a good toxicity profile. The MAGIC trial suggests that treatment with a chemotherapy triplet via the addition of an anthracycline to the cisplatin and 5FU doublet is also highly effective. Proponents of triplet therapy highlight the potential to improve response rates with acceptable toxicity. A recent meta-analysis examined 21 randomized studies comparing a doublet to triplet regimen in patients with distal esophageal, esophagogastric, and gastric adenocarcinoma. Overall, outcomes in 3475 patients were assessed. The authors demonstrated improved OS (HR, 0.9; 95% CI, 0.83 to 0.97) and progression-free survival (PFS) (HR, 0.8; 95% CI, 0.69 to 0.93), favoring taxane-based triplet therapies. Fluoropyrimidine-based triplets were associated with an objective response rate (ORR) compared with doublets alone (RR, 1.25; 95% CI, 1.09 to 1.44). This benefit, however, was associated with increased incidence of grade 3 to 4 thrombocytopenia, infection, and mucositis by a factor of approximately 2.

With respect to SCC, doublet therapies composed of cisplatin and 5FU have demonstrated effectiveness, as demonstrated by the studies noted previously. Similarly, triplet therapies comprising an additional taxane have also demonstrated efficacy. However, to date no extensive randomized data exist specifically comparing doublet to triplet regimens in patients with squamous histology.

Recently there has been much interest in taxane-based triplets in the neoadjuvant setting, with excellent oncologic results from phase II studies employing docetaxel in both major histologic subtypes. We have previously shown excellent tolerance and response (pathologic complete response [pCR] of 10%) in a phase II trial in patients with locally advanced (cT3 and/or N1) adenocarcinoma of the esophagus, GE junction, and stomach using a regimen of docetaxel, cisplatin, and 5FU (DCF). Long-term results demonstrated an impressive 5-year survival of over 55%, despite a high burden of residual disease (median positive pathologic lymph nodes = 5). These results have been replicated in several German series, including a recent study employing a docetaxel-based regimen, but replacing cisplatin with oxaliplatin (FLOT: 5-Flourouracil, Leucovorin, Oxaliplatin, and Taxotere [Docetaxel]) for esophagogastric adenocarcinoma with impressive pathologic response rates—20% pCR and another 20%, with near pCR (<10% residual tumor). Another German study employing a similar regimen demonstrated a pCR rate of 15% and a median OS of just over 4 years. These strong data set the foundation for a large German multicenter phase three trial investigating FLOT versus the regimen established by the MAGIC trial ( ClinicalTrials.gov identifier, NCT01216644), ECF (epirubicin, cisplatin, and 5-fluorouracil regimen) in patients with locally advanced esophagogastric adenocarcinoma. This study has completed accrual. Several Japanese studies have supported the use of taxane-based triplets, most notably docetaxel, in esophageal SCC. Indeed, impressive results from these studies have led to inclusion of DCF as one of three arms in the ongoing large Japanese multiinstitutional NExT (JCOG1109) trial investigating neoadjuvant chemotherapy (cisplatin/5FU or docetaxel/cisplatin/5FU) versus cisplatin/5FU with concurrent radiotherapy. This study is well underway, with preliminary results due in 2017.

Is Chemotherapy More Effective When Administered in the Pre- or Postoperative Setting?

Current data are conflicted regarding the use of postoperative chemotherapy alone in the adjuvant setting only. The 1997 study by Ando et al. randomized patients with esophageal SCC to receive either curative intent surgery alone or adjuvant chemotherapy with cisplatin and vindesine, delivered in two cycles. Patients predominantly harbored locally advanced disease; however, in this study the exact rate of curative R0 resection was not clearly indicated. No survival benefit was observed with the addition of postoperative chemotherapy, although the efficacy of the regimen and uncertainty regarding the extent of resection diminish the generalizability of the study.

In 2003, Ando and colleagues published a Japanese Clinical Oncology Group multiinstitutional phase III trial investigating postoperative cisplatin and 5FU in patients with SCC versus surgery alone. In patients who were ultimately found to harbor node positive disease, adjuvant chemotherapy was associated with a significant improvement in 5-year disease-free survival (DFS) compared with surgery alone (52% vs. 38%; P = .041). Conversely, there was a nonsignificant trend toward worse DFS outcomes in node negative patients receiving adjuvant chemotherapy compared with surgery alone (70% vs. 76%; P = .433).

These seemingly contradictory results led to a follow-up study comparing the same regimen given either pre- or postoperatively. In this study, the authors directly compared preoperative versus postoperative chemotherapy in 330 patients with locally advanced esophageal SCC. Patients received two cycles of cisplatin and 5FU, either preceding or following curative intent surgery. OS was significantly improved in patients who received preoperative compared with postoperative therapy, with 5-year survival of 55% versus 43% ( P = .04). Overall response rates to chemotherapy were 38%, which translated into fewer preoperative-treated patients being found to harbor T4 or N+ tumors. Furthermore, significantly more patients who received neoadjuvant chemotherapy underwent curative (R0) resection (96% vs. 91%; P = .04). Patients in the adjuvant arm experienced increased toxicity and reduced completion (75% vs. 85%; P = .04) of therapy compared with patients in the neoadjuvant arm. Overall, the results of the studies to date support the use of preoperative chemotherapy with respect to compliance, assessment of tumor response, and survival.