- Email: [email protected]
Neoadjuvant Therapy for Locally Advanced Esophageal Cancer Should Be Targeted to Tumor Histology
Neoadjuvant Therapy for Locally Advanced Esophageal Cancer Should Be Targeted to Tumor Histology Brendon M. Stiles, MD, Mohamed K. Kamel, MD, Sebron W. Harrison, MD, Mohamed Rahouma, MD, Benjamin Lee, MD, Abu Nasar, MS, Jeffrey L. Port, MD, and Nasser K. Altorki, MD Division of Thoracic Surgery, Department of Cardiothoracic Surgery, Weill Cornell Medicine, New York-Presbyterian Hospital, New York, New York
Background. Controversy exists over the optimal neoadjuvant therapy in patients with locally advanced esophageal cancer (EC). Although most groups favor neoadjuvant chemoradiation (nCRT), some prefer preoperative chemotherapy (nCT) without radiation. The objective of this study was to compare outcomes in EC patients undergoing either regimen, followed by surgery. Methods. We reviewed a prospectively collected database of EC patients undergoing esophagectomy after nCT or nCRT from 1989 to 2016. Choice of therapy was at the discretion of the multidisciplinary team. Diseasefree survival (DFS) and cancer-specific survival (CSS) were compared by the Kaplan-Meier log-rank test. Independent predictors of CSS were estimated by Cox regression analysis. Results. Among 700 EC patients 338 patients were treated with nCRT (n [ 112) or nCT (n [ 226) followed by surgery. Patients were well matched for age, gender, and clinical stage, although patients with squamous cell carcinoma were more likely to receive nCRT (49% vs 26%, p < 0.001). At surgery 90% and 91% of nCRT and nCT patients, respectively, underwent transthoracic esophagectomy. nCRT, in comparison with nCT, was associated with similar rates of Calvien-Dindo grade III/IV complications (34% vs 33%, p [ 0.423) but with a trend toward higher perioperative mortality (5% vs 1%, p [ 0.064). Among adenocarcinoma patients (n [ 239) the use of nCRT was associated with higher rates of complete clinical response (18% vs 7.4%), pathologically
negative lymph nodes (52% vs 30%, p [ 0.001), and complete pathologic response (21% vs 5.1%, p < 0.001). However, there was no difference between nCRT and nCT for 5-year DFS (28% vs 31%, p [ 0.636) or CSS (51% vs 52%, p [ 0.824) among adenocarcinoma patients. For patients with squamous cell carcinoma (n [ 98), nCRT and nCT had similar rates of complete clinical response (31% vs 26%, p [ 0.205), but the rates of negative nodes (65% vs 46%, p [ 0.064) and of complete pathologic response (42% vs 12%, p < 0.05) were higher with nCRT. For these patients nCRT was associated with no statistical difference in 5-year DFS (57% vs 40%, p [ 0.595) but with improved 5-year CSS (87% vs 68%, p [ 0.019) compared with nCT. On multivariable analysis for CSS, nCRT predicted improved survival for patients with squamous cell carcinoma (hazard ratio, 0.242; 95% confidence interval, 0.071–0.830) but not for those with adenocarcinoma (univariate hazard ratio, 0.940; 95% confidence interval, 0.544–1.623). Conclusions. For adenocarcinoma patients undergoing surgery for EC, nCRT leads to increased local tumor response compared with nCT alone but with no difference in survival. For squamous carcinoma patients nCRT appears to improve CSS compared with nCT. For patients with locally advanced EC targeted neoadjuvant regimens should be used depending on tumor histology.
C
cancer (EC). Three previous randomized trials have been performed with conflicting results. The POET trial [1] showed a trend toward improved 3-year survival with nCRT, whereas a trial performed by Burmeister and colleagues [2] and more recently the NeoRes trial [3] showed no difference between the two neoadjuvant regimens with regard to 3-year overall survival. How tumor histology impacts neoadjuvant treatment is unclear, and only one of the randomized trials enrolled patients with squamous cell carcinoma. However it is likely that adenocarcinoma and squamous cell carcinoma are biologically distinct and respond differently to systemic
onflicting data exist regarding the choice of neoadjuvant chemoradiation (nCRT) versus neoadjuvant chemotherapy (nCT) alone before esophagectomy for patients with locally advanced esophageal Accepted for publication July 30, 2018. Presented at the Fifty-fourth Annual Meeting of The Society of Thoracic Surgeons, Fort Lauderdale, FL, Jan 27–31, 2018. Address correspondence to Dr Stiles, Division of Thoracic Surgery, Department of Cardiothoracic Surgery, Ste M404, Weill Cornell Medicine of Cornell University, 525 East 68th St, New York, NY 10065; email: [email protected].
Ó 2018 by The Society of Thoracic Surgeons Published by Elsevier Inc.
(Ann Thorac Surg 2018;-:-–-) Ó 2018 by The Society of Thoracic Surgeons
0003-4975/$36.00 https://doi.org/10.1016/j.athoracsur.2018.07.089
2
STILES ET AL INDUCTION THERAPY FOR ESOPHAGEAL CANCER
therapy and radiation. Indeed, in subgroup analyses of those trials it has been suggested that patients with squamous cell carcinoma derived the most benefit from the addition of radiation therapy. Similarly, in the CROSS (ChemoRadiotherapy for Oesophageal cancer followed by Surgery Study) trial it was shown that patients with squamous cell carcinoma were more likely to respond to nCRT than patients with adenocarcinoma. In a multiinstitutional study from our group, MD Anderson Cancer Center, and McGill University we previously demonstrated that for EC patients with adenocarcinoma undergoing en bloc esophagectomy, the type of neoadjuvant therapy was not predictive of survival [4]. In the current study we sought to examine the effect of neoadjuvant regimen on survival depending on tumor histology.
Ann Thorac Surg 2018;-:-–-
gastroesophageal junction cancer after nCT or nCRT between April 1989 and September 2016. Variables of interest included demographics, clinical, operative approach, and pathologic variables. Clinical and pathologic staging were based on the American Joint Committee on Cancer seventh edition TNM classification. Choice of therapy was at the discretion of the multidisciplinary team. Generally, we used nCT before publication of the CROSS trial and nCRT after, although in some instances exceptions for each period generally depended on referring oncologists. After neoadjuvant treatment patients were restaged using a combination of computed tomography and endoscopy and more recently using positron emission tomography. Assessment of clinical response to induction therapy was assessed by computed tomography, endoscopy, and more recently positron emission tomography.
Follow-Up, Recurrence, and Survival
Patients and Methods Study Design and Patient Selection The Weill Cornell Medicine-New York Presbyterian Hospital Institutional review board approved the study, and patient consent was waived. We reviewed a prospectively assembled EC database for patients undergoing esophagectomy for EC (upper, middle, or lower third) and
All patients were seen in follow-up every 3 months in the first year, then every 6 months for 2 additional years, then yearly thereafter. Computed tomography was performed every 6 months for 3 years and subsequently at yearly intervals. Positron emission tomography and endoscopy were only performed in response to specific clinical indications.
Table 1. Clinical Characteristics Chemotherapy Only (n ¼ 226)
Patients Demographics and Clinical Characteristics Median age, year (IQR) Male gender Median Charlson comorbidity index (IQR) Performance status 0 1 Histology Adenocarcinoma Squamous cell carcinoma Others Median preinduction tumor SUVmax (n ¼ 273), g/dL (IQR) Median preinduction tumor SUVmax (n ¼ 223), g/dl (IQR) Median reduction in tumor SUVmax (n ¼ 181), % (IQR) Clinical T stage T1/2 T3/4 Clinical N stage cN –ve cN þve Clinical stage I/II III/IV Clinical response Complete response Partial response Minimal response to stable/progressive disease
Chemotherapy–Radiation Therapy (n ¼ 112)
63 (55–68) 183 (81) 0 (0–1)
64 (57–69) 86 (77) 0 (0–2)
0.264 0.369 0.059
156 (69) 70 (31)
89 (79.5) 23 (20.5)
0.043
176 (78) 50 (22) 0 9.2 (5.9–13.3) 3.6 (2.2–5.6) 64 (37–88)
N/A ¼ not applicable;
63 (57) 48 (43) 1 10.5 (6.6–14.9) 4.0 (2.5–5.4) 68 (49–87)
<0.001 N/A 0.103 0.774 0.383
38 (17) 188 (83)
17 (15) 95 (85)
0.701
60 (26.5) 166 (73.5)
24 (21) 88 (79)
0.305
68 (30) 158 (70)
34 (30) 78 (70)
0.960
26 (11.5) 134 (59) 66 (29.2)
26 (23.2) 78 (69.6) 8 (7.1)
Values are n (%) unless otherwise defined. IQR ¼ interquartile range;
p Value
SUVmax ¼ maximum standardized uptake value.
<0.001
Ann Thorac Surg 2018;-:-–-
STILES ET AL INDUCTION THERAPY FOR ESOPHAGEAL CANCER
3
Overall survival was defined as the time from surgery until death from any cause. Disease-free survival (DFS) was defined as the time from surgery until recurrence or death from any cause. Cancer-specific survival (CSS) was defined as the time from surgery until death from EC.
Statistical Analysis Continuous variables were expressed as median (interquartile range) and were compared between the study groups using the Mann-Whitney U test. Categorical variables were expressed as frequencies (percentages) and were compared between the study groups by using the two-tailed Pearson’s c2 test. Cox proportional hazards regression analysis was done to determine the independent effects of clinical predictors on CSS. Variables included in the model were age, gender, performance status (PS), tumor location, clinical T stage, clinical N stage, clinical stage, type of induction treatment, clinical response to induction treatment, preinduction treatment tumor maximum standardized uptake value, surgical approach, extent of resection, and extent of lymph node dissection. Univariate predictors with probability values less than 0.20 were included in the multivariable analysis model. Survival probabilities were estimated using the Kaplan-Meier method, and differences in survival were compared using the log-rank test. Data analysis was performed using IBM SPSS software (version 24.0; SPSS, Inc, Chicago, IL).
Results Patients Among our cohort of patients we identified 338 treated with nCRT (n ¼ 112) or nCT (n ¼ 226) before esophagectomy (Table 1). Age and gender were similar, as were clinical stages of the patients receiving each type of therapy. Patients receiving nCT were more likely to have PS 1 (p ¼ 0.04). Finally, patients with squamous cell carcinoma were more likely to receive nCRT (49% vs 26%, p < 0.001) (Fig 1). The median dose of radiation therapy given was 45 Gy (interquartile range, 41.4–50.4). Chemotherapy was predominantly platinum-based (98.5%), with 64% of patients also receiving a taxane.
Clinical Response to Treatment Patients receiving nCRT were more likely to have a complete or partial clinical response (92.8% vs 70.5%, p < 0.001). This was particularly true for patients with adenocarcinoma (95.3% vs 69.3%, p < 0.001).
Surgical Treatment and Morbidity With evolution in our surgical approach more recently patients were treated by minimally invasive esophagectomy (Table 2). That combined with the recent results of the CROSS trial, patients receiving nCRT were more likely to undergo minimally invasive esophagectomy than patients treated with nCT (43% vs 14%, p < 0.001).
Fig 1. Proportion of esophageal cancer patients with (A) adenocarcinoma and (B) squamous cell carcinoma receiving neoadjuvant chemoradiation (nCRT) versus neoadjuvant chemotherapy (nCT).
Similarly, these patients were more likely to undergo Ivor Lewis esophagectomy (37% vs 9%, p < 0.001) because we had adopted that approach predominantly with minimally invasive esophagectomy. Despite these different strategies there were no differences in surgical complications after nCRT or nCT. The rate of Clavien-Dindo grade III/IV complications was 34% and 33%, respectively (p ¼ 0.423). No differences were apparent with regard to pulmonary or cardiovascular complications or for anastomotic leaks. There was a trend toward increased perioperative mortality with nCRT compared with nCT, but it did not reach statistical significance (5% vs 1%, p ¼ 0.064). Radial resection margins were greater after nCRT (0.5 vs 0.3, p < 0.001), but R0 resection rates were similar compared with nCT (90% vs 95%, p ¼ 0.136).
Pathologic Downstaging After nCRT for both adenocarcinoma and squamous cell carcinoma, patients were more likely to have a complete pathologic response (CPR) and to have negative lymph nodes (Table 3). Among adenocarcinoma patients (n ¼ 239) the rate of CPR after nCRT and nCT were 20.6% and 5.1%, respectively (p < 0.001), with rates of pathologic N0 classification of 52.4% and 29.5% (p ¼ 0.001), respectively. For squamous cell carcinoma patients (n ¼ 98) CPR rates for nCRT and nCT were 41.7% and 12% (p ¼ 0.001), whereas pN0 rates were 64.6% and 46% (p ¼ 0.064).
Survival We next analyzed survival in the different cohorts of EC patients receiving nCRT or nCT (median follow-up, 44 months for alive patients). Overall survival (Fig 2A) was 49% for nCRT patients versus 45% for nCT patients at 5 years (p ¼ 0.624). Regarding DFS (Fig 2B) no statistically significant differences were seen at 5 years between nCRT and nCT for all patients (40% vs 33%, p ¼ 0.768), for adenocarcinoma patients (28% vs 31%, p ¼ 0.636), or
4
STILES ET AL INDUCTION THERAPY FOR ESOPHAGEAL CANCER
Ann Thorac Surg 2018;-:-–-
Table 2. Surgical Approaches and Postoperative Outcomes Operative Characteristics and Postoperative Outcomes Tumor site Upper/middle third Lower third/gastroesophageal junction Surgical approach Open Minimally invasive esophagectomy/hybrid Transhiatal Ivor-Lewis Mckeown Anastomosis site Thorax Neck En-bloc resection Lymphadenectomy Two fields lymph node dissection Three fields lymph node dissection Complications Clavien-Dindo 0 (no complications) I/II III/IV Pulmonary Cardiovascular Anastomotic leak Perioperative mortality
Chemotherapy only (n ¼ 226)
Chemotherapy–Radiation Therapy (n ¼ 112)
36 (16) 190 (84)
13 (12) 99 (88)
195 31 21 21 184
64 48 11 42 59
p Value
0.288
(57) (43) (10) (37) (53)
<0.001
21 (9) 205 (91) 189 (84)
42 (37.5) 70 (63) 75 (67)
<0.001
100 (44) 126 (56)
86 (77) 26 (23)
<0.001
120 29 77 38 41 31 3
54 20 36 15 18 21 6
(86) (14) (9) (9) (82)
<0.001
<0.001
0.423 (53) (13) (34) (17) (18) (14) (1)
(49) (18) (33) (13) (16) (19) (5)
0.416 0.637 0.227 0.064
Values are n (%).
for squamous cell carcinoma patients (57% vs 40%, p ¼ 0.595). Because the survival curves appeared to separate for squamous cell carcinoma patients we evaluated CSS (Fig 3). For the whole cohort 5-year CSS was 65% versus 56% (p ¼ 0.077) for nCRT and nCT, respectively. There was no difference for adenocarcinoma patients (51% vs 52%, p ¼ 0.824), whereas for squamous cell carcinoma patients there was a survival advantage to nCRT over nCT (87% vs 68%, p ¼ 0.019). On multivariable analysis of the whole study cohort independent clinical predictors of cancer-specific mortality were clinical stage (cStage III/IV: hazard ratio [HR], 1.82; 95% confidence interval [CI], 1.15–2.88; p ¼ 0.011), PS 1/2 (HR, 1.59; 95% CI, 1.06–2.38; p ¼ 0.024), and response to induction therapy (nonresponsive: HR, 1.55; 95% CI, 1.01– 2.38; p ¼ 0.045). In patients with adenocarcinoma tumors PS was the only independent clinical predictor of CSS approaching significance (PS 1/2: HR, 1.64; 95% CI, 0.99– 2.74; p ¼ 0.056). In patients with squamous cell tumors response to induction therapy (nonresponsive: HR, 2.62; 95% CI, 0.98–7.00; p ¼ 0.055) was the only independent predictor of CSS approaching significance. For the whole cohort independent predictors of allcause mortality were cStage III/IV (HR, 1.61; 95% CI, 1.07–2.43; p ¼ 0.024), PS 1/2 (HR, 1.71; 95% CI, 1.18–2.49; p ¼ 0.005), response to induction therapy (nonresponsive:
HR, 1.61; 95% CI, 1.08–2.39; p ¼ 0.020), preinduction tumor maximum standardized uptake value (HR, 1.03; 95% CI, 1.01–1.05; p ¼ 0.045), and surgery year (1990–1999: HR, 3.57; 95% CI, 1.01–12.5; p ¼ 0.050) (Table 4).
Comment The optimal multimodality treatment for locally advanced EC is unknown. Most centers in the United States prefer nCRT, as evidenced by data from the National Cancer Database (2006–2012) in which only 12.5% of patients with induction therapy for EC received nCT [5]. However, conflicting data exist from previous randomized trials regarding the two strategies. The NeoRes (Neoadjuvant Chemotherapy Versus Radiochemotherapy for Cancer of the Esophagus or Cardia) trial recently published by Klevebro and colleagues [3] randomized 181 patients at multiple institutions in Sweden and Norway with esophageal squamous cell carcinoma or adenocarcinoma to either three cycles of chemotherapy (cisplatin and fluorouracil) followed by surgery or to the same chemotherapy followed by concomitant radiation (40 Gy) and then surgery. The primary endpoint of this study was CPR rate, and therefore it might not have been powered to detect survival differences. Not surprisingly, nCRT patients showed improved pathologic response to
Ann Thorac Surg 2018;-:-–-
5
STILES ET AL INDUCTION THERAPY FOR ESOPHAGEAL CANCER
Table 3. Pathologic Characteristics Pathologic Characteristics Resection margin Proximal, cm Distal, cm Radial, cm <0.1 0.1 R0 R1 Positive radial (n ¼ 332) Positive proximal/distal (n ¼ 332) Change from cStage to pStage No change Downstaging Upstaging Pathologic complete response pN stage N0 N1 N2 N3 No. of lymph nodes resected No. of pathologic þve lymph nodes pStage 0 I II III IV
Chemotherapy Only (n ¼ 226)
Chemotherapy–Radiation Therapy (n ¼ 112)
9 4.9 0.3 30 98 203 23 10 12
(5.1–12) (3–6.5) (0.1–0.8) (23) (77) (90) (10) (4.5) (5)
105 89 32 15
(46.5) (39.4) (14.2) (6.6)
37 62 13 33
(33) (55.4) (11.6) (29.5)
(33.2) (26.5) (21.7) (18.6) (21–38) (0–5)
64 23 19 6 21 0
(57.1) (20.5) (17) (5.4) (14–32) (0–2)
(6.6) (11.9) (25.2) (54) (2.2)
33 13 22 38 6
(29.5) (11.6) (19.6) (33.9) (5.4)
75 60 49 42 28 1 15 27 57 122 5
5.1 6.1 0.5 5 62 106 6 2 1
(3.2–9) (4.3–9.6) (0.2–0.9) (7.5) (92.5) (95) (5) (2) (1)
p Value <0.001 <0.001 0.011 0.006 0.136 0.350 0.068 0.020
<0.001 <0.001
<0.001 <0.001 <0.001
Values are n (%) or median (IQR).
Fig 2. (A) Overall survival (OS) and (B) disease-free survival (DFS) for patients with locally advanced esophageal cancer receiving neoadjuvant chemoradiation (CRT) versus neoadjuvant chemotherapy (CT).
6
STILES ET AL INDUCTION THERAPY FOR ESOPHAGEAL CANCER
Ann Thorac Surg 2018;-:-–-
trial of 75 EC patients published by Burmeister and colleagues [2] that demonstrated a 3-year overall survival of 49% after nCT compared with 52% after nCRT (p ¼ 0.97). However, a third randomized trial by Stahl and colleagues [1], the POET trial (Preoperative Chemotherapy or Radiochemotherapy in Esophagogastric Adenocarcinoma Trial), which included 126 patients, suggested a strong trend toward a benefit of nCRT compared with nCT, with 3-year survivals of 47% with nCRT compared with 28% with nCT (p ¼ 0.07). Both the Burmeister and Stahl trials showed higher rates of CPR and of nodal downstaging with nCRT, similar to the NeoRes results. However differences in surgical quality may explain the distinctions between the Stahl, Klevebro, and Burmeister trials in which only the nCT arm in the Stahl trial is a survival outlier. In the POET trial only 47% of patients receiving nCT had transthoracic esophagectomy, compared with 90% and 100% in the Klevebro and Burmeister trials, respectively. In subgroup analysis of randomized trial data [6] patients with moderate nodal metastatic burden treated by transthoracic esophagectomy had superior survival compared with those treated by transhiatal esophagectomy (5-year DFS of 64% vs 23%, p ¼ 0.02). We expect that with the infrequent use of transthoracic resections, such as in the Stahl trial, radiation may indeed improve local control and survival. Our institutional study presented here mirrors the data from the Klevebro and Burmeister trials. When grouping all adenocarcinoma and squamous cell EC patients together we found no difference in overall survival, DFS,
Fig 3. Cancer-specific survival of esophageal cancer patients receiving neoadjuvant chemoradiation (CRT) versus neoadjuvant chemotherapy (CTH), by histologic subtype adenocarcinoma or squamous cell carcinoma.
neoadjuvant therapy, with a 28% rate of CPR and a 65% rate of negative lymph nodes, comparing favorably with the 9% CPR rate and 38% rate of negative nodes after nCT. However this did not translate to improved survival, with 3-year overall survival of 49% in the nCT arm and 47% in the nCRT arm (p ¼ 0.77). Progression-free 3-year survival was 44% in both treatment arms. This was similar to what was previously reported in a randomized
Table 4. Multivariable Analysis Predictors of Overall (All-Cause) Mortality Multivariable Analysisa
Univariate Analysis 95% CI
95% CI
All Patients (Events 177/338)
HR
Lower
Upper
p Value
Age Male gender Performance status 1/2 Tumor site (upper/middle thirds) Tumor histology (squamous cell cancer) cT stage (III/IV) cNstage (Nþve) cStage (III/IV) Induction type (CRT) Clinical response (nonresponsive) Preinduction tumor SUVmax Approach (minimally invasive esophagectomy/hybrid) Extent of resection (en bloc) Lymphadenectomy (2 fields) Surgery year 2010–2016 2000–2009 1990–1999
1.01 1.48 1.97 1.58 0.56 1.78 1.35 1.67 0.98 1.50 1.02 0.71
0.99 0.99 1.45 0.99 0.39 1.14 0.95 1.18 0.69 1.08 0.99 0.44
1.03 2.22 2.66 2.52 0.82 2.78 1.92 2.35 1.40 2.08 1.04 1.15
0.216 0.058 <0.001 0.053 0.002 0.012 0.099 0.004 0.908 0.017 0.152 0.161
0.74 0.88
0.51 0.64
1.08 1.19
1.63 1.52
0.94 1.02
Reference 2.83 2.27
a
HR
Lower
Upper
p Value
1.08 1.71 0.92 0.70
0.65 1.18 0.43 0.42
1.79 2.49 1.96 1.15
0.777 0.005 0.824 0.156
1.61
1.07
2.43
0.024
1.61 1.03 1.25
1.08 1.01 0.58
2.39 1.05 2.69
0.020 0.045 0.564
0.114 0.401
0.86
0.53
1.40
0.543
0.081 0.039
1.14 3.57
0.60 1.01
2.17 12.5
0.697 0.050
Univariate variables with p < 0.1 were included in the multivariable model.
CI ¼ confidence interval;
CRT ¼ chemoradiation;
HR ¼ hazard ratio;
SUVmax ¼ maximum standardized uptake value.
Ann Thorac Surg 2018;-:-–-
or CSS after nCRT versus nCT. This was despite higher rates of pathologic response after nCRT, as demonstrated by higher rates of CPR and pN0 classification. Given our institutional high rate of transthoracic resection (90%), our preference for en bloc resection (78%), and our low rate of R1 resections (8.6%) we believe that any benefits of adding a second local control modality (radiation) are limited, particularly with regard to patients with adenocarcinoma. However it appears there may be some benefit to preoperative radiation therapy in patients with squamous cell carcinoma. In our series patients with squamous cell cancer undergoing nCRT had a suggestion of improved 5-year overall survival (73% vs 55%, p ¼ 0.157) and 5-year DFS (57% vs 40%, p ¼ 0.595) and had a statistically significant improvement in 5-year CSS (87% vs 68%, p ¼ 0.019) compared with those undergoing nCT. Although not powered to detect differences in outcome based on tumor histology, the Klevebro trial also suggested that any potential benefit of nCRT is driven by patients with squamous cell carcinoma. In patients with adenocarcinoma the survival curves actually favored nCT, although they were not statistically significant. Similarly, in the CROSS trial [7] it is apparent that the benefit of nCRT is much more clinically important in patients with squamous cell carcinoma than in those with adenocarcinoma (HR, 0.46, p ¼ 0.004, vs HR, 0.75, p ¼ 0.059). The most recent meta-analysis of the two strategies also suggested only modest differences in HRs comparing nCT versus nCRT to surgery alone (0.83 and 0.75, respectively) in adenocarcinoma patients [8]. Our study is limited by its retrospective nature and by our inability to control for biases in treatment allocation. Additionally, there were modifications to our surgical approach during the time period studied. In summary, although radiation therapy may improve outcomes in patients with locally advanced squamous cell EC, there appears to be no benefit to adding radiation therapy to the induction regimen for patients with esophageal adenocarcinoma. Improved pathologic response in these patients does not translate to improvements in survival when an appropriate esophagectomy is performed. The addition of radiation therapy increases cost, hypothetically could lead to dose reductions of chemotherapy, and potentially increases noncancer-related deaths in the first year after treatment. The 20% CPR rate and excellent survival reported with the use of neoadjvuant FLOT (5-fluorouracil, leucovorin, oxaliplatin, and docetaxel) add strength to the argument for chemotherapy alone [9]. The ESOPEC (Perioperative Chemotherapy Compared to Neoadjuvant Chemoradiation in Patients with Adenocarcinoma of the Esophagus) trial meant to compare FLOT with nCRT using the CROSS regimen is currently accruing patients and should more clearly define the role of nCT versus nCRT in patients with adenocarcinoma [10].
STILES ET AL INDUCTION THERAPY FOR ESOPHAGEAL CANCER
7
We believe that our institutional data are consistent with a recent study from the National Cancer Database [5] and with data from randomized trials. We therefore believe that a tailored approach to neoadjuvant therapy for patients with locally advanced EC is appropriate, depending on tumor histology. For patients with esophageal squamous cell carcinoma we recommend adding radiation treatment to induction therapy. However, for adenocarcinoma induction chemotherapy alone appears to be a sound treatment regimen.
References 1. Stahl M, Walz MK, Riera-knorrenschild J, et al. Preoperative chemotherapy versus chemoradiotherapy in locally advanced adenocarcinomas of the oesophagogastric junction (POET): long-term results of a controlled randomised trial. Eur J Cancer 2017;81:183–90. 2. Burmeister BH, Thomas JM, Burmeister EA, et al. Is concurrent radiation therapy required in patients receiving preoperative chemotherapy for adenocarcinoma of the oesophagus? A randomised phase II trial. Eur J Cancer 2011;47:354–60. 3. Klevebro F, Alexandersson von D€ obeln G, Wang N, et al. A randomized clinical trial of neoadjuvant chemotherapy versus neoadjuvant chemoradiotherapy for cancer of the oesophagus or gastro-oesophageal junction. Ann Oncol 2016;27:660–7. 4. Spicer JD, Stiles BM, Sudarshan M, et al. Preoperative chemoradiation therapy versus chemotherapy in patients undergoing modified en bloc esophagectomy for locally advanced esophageal adenocarcinoma: is radiotherapy beneficial? Ann Thorac Surg 2016;101:126. 5. Samson P, Robinson C, Bradley J, et al. Neoadjuvant chemotherapy versus chemoradiation prior to esophagectomy: impact on rate of complete pathologic response and survival in esophageal cancer patients. J Thorac Oncol 2016;11:2227–37. 6. Omloo JM, Lagarde SM, Hulscher JB, et al. Extended transthoracic resection compared with limited transhiatal resection for adenocarcinoma of the mid/distal esophagus: five-year survival of a randomized clinical trial. Ann Surg 2007;246:992–1000; discussion 1000–1. 7. Shapiro J, van Lanschot JJ, Hulshof MC, et al. Neoadjuvant chemoradiotherapy plus surgery versus surgery alone for oesophageal or junctional cancer (CROSS): long-term results of a randomised controlled trial. Lancet Oncol 2015;16:1090–8. 8. Sjoquist KM, Burmeister BH, Smithers BM, et al. Survival after neoadjuvant chemotherapy or chemoradiotherapy for resectable oesophageal carcinoma: an updated metaanalysis. Lancet Oncol 2011;12:681–92. 9. Schulz C, Kullmann F, Kunzmann V, et al. NeoFLOT: multicenter phase II study of perioperative chemotherapy in resectable adenocarcinoma of the gastroesophageal junction or gastric adenocarcinoma—very good response predominantly in patients with intestinal type tumors. Int J Cancer 2015;137:678–85. 10. Hoeppner J, Lordick F, Brunner T, et al. ESOPEC: prospective randomized controlled multicenter phase III trial comparing perioperative chemotherapy (FLOT protocol) to neoadjuvant chemoradiation (CROSS protocol) in patients with adenocarcinoma of the esophagus (NCT02509286). BMC Cancer 2016;16:503.
Background. Controversy exists over the optimal neoadjuvant therapy in patients with locally advanced esophageal cancer (EC). Although most groups favor neoadjuvant chemoradiation (nCRT), some prefer preoperative chemotherapy (nCT) without radiation. The objective of this study was to compare outcomes in EC patients undergoing either regimen, followed by surgery. Methods. We reviewed a prospectively collected database of EC patients undergoing esophagectomy after nCT or nCRT from 1989 to 2016. Choice of therapy was at the discretion of the multidisciplinary team. Diseasefree survival (DFS) and cancer-specific survival (CSS) were compared by the Kaplan-Meier log-rank test. Independent predictors of CSS were estimated by Cox regression analysis. Results. Among 700 EC patients 338 patients were treated with nCRT (n [ 112) or nCT (n [ 226) followed by surgery. Patients were well matched for age, gender, and clinical stage, although patients with squamous cell carcinoma were more likely to receive nCRT (49% vs 26%, p < 0.001). At surgery 90% and 91% of nCRT and nCT patients, respectively, underwent transthoracic esophagectomy. nCRT, in comparison with nCT, was associated with similar rates of Calvien-Dindo grade III/IV complications (34% vs 33%, p [ 0.423) but with a trend toward higher perioperative mortality (5% vs 1%, p [ 0.064). Among adenocarcinoma patients (n [ 239) the use of nCRT was associated with higher rates of complete clinical response (18% vs 7.4%), pathologically
negative lymph nodes (52% vs 30%, p [ 0.001), and complete pathologic response (21% vs 5.1%, p < 0.001). However, there was no difference between nCRT and nCT for 5-year DFS (28% vs 31%, p [ 0.636) or CSS (51% vs 52%, p [ 0.824) among adenocarcinoma patients. For patients with squamous cell carcinoma (n [ 98), nCRT and nCT had similar rates of complete clinical response (31% vs 26%, p [ 0.205), but the rates of negative nodes (65% vs 46%, p [ 0.064) and of complete pathologic response (42% vs 12%, p < 0.05) were higher with nCRT. For these patients nCRT was associated with no statistical difference in 5-year DFS (57% vs 40%, p [ 0.595) but with improved 5-year CSS (87% vs 68%, p [ 0.019) compared with nCT. On multivariable analysis for CSS, nCRT predicted improved survival for patients with squamous cell carcinoma (hazard ratio, 0.242; 95% confidence interval, 0.071–0.830) but not for those with adenocarcinoma (univariate hazard ratio, 0.940; 95% confidence interval, 0.544–1.623). Conclusions. For adenocarcinoma patients undergoing surgery for EC, nCRT leads to increased local tumor response compared with nCT alone but with no difference in survival. For squamous carcinoma patients nCRT appears to improve CSS compared with nCT. For patients with locally advanced EC targeted neoadjuvant regimens should be used depending on tumor histology.
C
cancer (EC). Three previous randomized trials have been performed with conflicting results. The POET trial [1] showed a trend toward improved 3-year survival with nCRT, whereas a trial performed by Burmeister and colleagues [2] and more recently the NeoRes trial [3] showed no difference between the two neoadjuvant regimens with regard to 3-year overall survival. How tumor histology impacts neoadjuvant treatment is unclear, and only one of the randomized trials enrolled patients with squamous cell carcinoma. However it is likely that adenocarcinoma and squamous cell carcinoma are biologically distinct and respond differently to systemic
onflicting data exist regarding the choice of neoadjuvant chemoradiation (nCRT) versus neoadjuvant chemotherapy (nCT) alone before esophagectomy for patients with locally advanced esophageal Accepted for publication July 30, 2018. Presented at the Fifty-fourth Annual Meeting of The Society of Thoracic Surgeons, Fort Lauderdale, FL, Jan 27–31, 2018. Address correspondence to Dr Stiles, Division of Thoracic Surgery, Department of Cardiothoracic Surgery, Ste M404, Weill Cornell Medicine of Cornell University, 525 East 68th St, New York, NY 10065; email: [email protected].
Ó 2018 by The Society of Thoracic Surgeons Published by Elsevier Inc.
(Ann Thorac Surg 2018;-:-–-) Ó 2018 by The Society of Thoracic Surgeons
0003-4975/$36.00 https://doi.org/10.1016/j.athoracsur.2018.07.089
2
STILES ET AL INDUCTION THERAPY FOR ESOPHAGEAL CANCER
therapy and radiation. Indeed, in subgroup analyses of those trials it has been suggested that patients with squamous cell carcinoma derived the most benefit from the addition of radiation therapy. Similarly, in the CROSS (ChemoRadiotherapy for Oesophageal cancer followed by Surgery Study) trial it was shown that patients with squamous cell carcinoma were more likely to respond to nCRT than patients with adenocarcinoma. In a multiinstitutional study from our group, MD Anderson Cancer Center, and McGill University we previously demonstrated that for EC patients with adenocarcinoma undergoing en bloc esophagectomy, the type of neoadjuvant therapy was not predictive of survival [4]. In the current study we sought to examine the effect of neoadjuvant regimen on survival depending on tumor histology.
Ann Thorac Surg 2018;-:-–-
gastroesophageal junction cancer after nCT or nCRT between April 1989 and September 2016. Variables of interest included demographics, clinical, operative approach, and pathologic variables. Clinical and pathologic staging were based on the American Joint Committee on Cancer seventh edition TNM classification. Choice of therapy was at the discretion of the multidisciplinary team. Generally, we used nCT before publication of the CROSS trial and nCRT after, although in some instances exceptions for each period generally depended on referring oncologists. After neoadjuvant treatment patients were restaged using a combination of computed tomography and endoscopy and more recently using positron emission tomography. Assessment of clinical response to induction therapy was assessed by computed tomography, endoscopy, and more recently positron emission tomography.
Follow-Up, Recurrence, and Survival
Patients and Methods Study Design and Patient Selection The Weill Cornell Medicine-New York Presbyterian Hospital Institutional review board approved the study, and patient consent was waived. We reviewed a prospectively assembled EC database for patients undergoing esophagectomy for EC (upper, middle, or lower third) and
All patients were seen in follow-up every 3 months in the first year, then every 6 months for 2 additional years, then yearly thereafter. Computed tomography was performed every 6 months for 3 years and subsequently at yearly intervals. Positron emission tomography and endoscopy were only performed in response to specific clinical indications.
Table 1. Clinical Characteristics Chemotherapy Only (n ¼ 226)
Patients Demographics and Clinical Characteristics Median age, year (IQR) Male gender Median Charlson comorbidity index (IQR) Performance status 0 1 Histology Adenocarcinoma Squamous cell carcinoma Others Median preinduction tumor SUVmax (n ¼ 273), g/dL (IQR) Median preinduction tumor SUVmax (n ¼ 223), g/dl (IQR) Median reduction in tumor SUVmax (n ¼ 181), % (IQR) Clinical T stage T1/2 T3/4 Clinical N stage cN –ve cN þve Clinical stage I/II III/IV Clinical response Complete response Partial response Minimal response to stable/progressive disease
Chemotherapy–Radiation Therapy (n ¼ 112)
63 (55–68) 183 (81) 0 (0–1)
64 (57–69) 86 (77) 0 (0–2)
0.264 0.369 0.059
156 (69) 70 (31)
89 (79.5) 23 (20.5)
0.043
176 (78) 50 (22) 0 9.2 (5.9–13.3) 3.6 (2.2–5.6) 64 (37–88)
N/A ¼ not applicable;
63 (57) 48 (43) 1 10.5 (6.6–14.9) 4.0 (2.5–5.4) 68 (49–87)
<0.001 N/A 0.103 0.774 0.383
38 (17) 188 (83)
17 (15) 95 (85)
0.701
60 (26.5) 166 (73.5)
24 (21) 88 (79)
0.305
68 (30) 158 (70)
34 (30) 78 (70)
0.960
26 (11.5) 134 (59) 66 (29.2)
26 (23.2) 78 (69.6) 8 (7.1)
Values are n (%) unless otherwise defined. IQR ¼ interquartile range;
p Value
SUVmax ¼ maximum standardized uptake value.
<0.001
Ann Thorac Surg 2018;-:-–-
STILES ET AL INDUCTION THERAPY FOR ESOPHAGEAL CANCER
3
Overall survival was defined as the time from surgery until death from any cause. Disease-free survival (DFS) was defined as the time from surgery until recurrence or death from any cause. Cancer-specific survival (CSS) was defined as the time from surgery until death from EC.
Statistical Analysis Continuous variables were expressed as median (interquartile range) and were compared between the study groups using the Mann-Whitney U test. Categorical variables were expressed as frequencies (percentages) and were compared between the study groups by using the two-tailed Pearson’s c2 test. Cox proportional hazards regression analysis was done to determine the independent effects of clinical predictors on CSS. Variables included in the model were age, gender, performance status (PS), tumor location, clinical T stage, clinical N stage, clinical stage, type of induction treatment, clinical response to induction treatment, preinduction treatment tumor maximum standardized uptake value, surgical approach, extent of resection, and extent of lymph node dissection. Univariate predictors with probability values less than 0.20 were included in the multivariable analysis model. Survival probabilities were estimated using the Kaplan-Meier method, and differences in survival were compared using the log-rank test. Data analysis was performed using IBM SPSS software (version 24.0; SPSS, Inc, Chicago, IL).
Results Patients Among our cohort of patients we identified 338 treated with nCRT (n ¼ 112) or nCT (n ¼ 226) before esophagectomy (Table 1). Age and gender were similar, as were clinical stages of the patients receiving each type of therapy. Patients receiving nCT were more likely to have PS 1 (p ¼ 0.04). Finally, patients with squamous cell carcinoma were more likely to receive nCRT (49% vs 26%, p < 0.001) (Fig 1). The median dose of radiation therapy given was 45 Gy (interquartile range, 41.4–50.4). Chemotherapy was predominantly platinum-based (98.5%), with 64% of patients also receiving a taxane.
Clinical Response to Treatment Patients receiving nCRT were more likely to have a complete or partial clinical response (92.8% vs 70.5%, p < 0.001). This was particularly true for patients with adenocarcinoma (95.3% vs 69.3%, p < 0.001).
Surgical Treatment and Morbidity With evolution in our surgical approach more recently patients were treated by minimally invasive esophagectomy (Table 2). That combined with the recent results of the CROSS trial, patients receiving nCRT were more likely to undergo minimally invasive esophagectomy than patients treated with nCT (43% vs 14%, p < 0.001).
Fig 1. Proportion of esophageal cancer patients with (A) adenocarcinoma and (B) squamous cell carcinoma receiving neoadjuvant chemoradiation (nCRT) versus neoadjuvant chemotherapy (nCT).
Similarly, these patients were more likely to undergo Ivor Lewis esophagectomy (37% vs 9%, p < 0.001) because we had adopted that approach predominantly with minimally invasive esophagectomy. Despite these different strategies there were no differences in surgical complications after nCRT or nCT. The rate of Clavien-Dindo grade III/IV complications was 34% and 33%, respectively (p ¼ 0.423). No differences were apparent with regard to pulmonary or cardiovascular complications or for anastomotic leaks. There was a trend toward increased perioperative mortality with nCRT compared with nCT, but it did not reach statistical significance (5% vs 1%, p ¼ 0.064). Radial resection margins were greater after nCRT (0.5 vs 0.3, p < 0.001), but R0 resection rates were similar compared with nCT (90% vs 95%, p ¼ 0.136).
Pathologic Downstaging After nCRT for both adenocarcinoma and squamous cell carcinoma, patients were more likely to have a complete pathologic response (CPR) and to have negative lymph nodes (Table 3). Among adenocarcinoma patients (n ¼ 239) the rate of CPR after nCRT and nCT were 20.6% and 5.1%, respectively (p < 0.001), with rates of pathologic N0 classification of 52.4% and 29.5% (p ¼ 0.001), respectively. For squamous cell carcinoma patients (n ¼ 98) CPR rates for nCRT and nCT were 41.7% and 12% (p ¼ 0.001), whereas pN0 rates were 64.6% and 46% (p ¼ 0.064).
Survival We next analyzed survival in the different cohorts of EC patients receiving nCRT or nCT (median follow-up, 44 months for alive patients). Overall survival (Fig 2A) was 49% for nCRT patients versus 45% for nCT patients at 5 years (p ¼ 0.624). Regarding DFS (Fig 2B) no statistically significant differences were seen at 5 years between nCRT and nCT for all patients (40% vs 33%, p ¼ 0.768), for adenocarcinoma patients (28% vs 31%, p ¼ 0.636), or
4
STILES ET AL INDUCTION THERAPY FOR ESOPHAGEAL CANCER
Ann Thorac Surg 2018;-:-–-
Table 2. Surgical Approaches and Postoperative Outcomes Operative Characteristics and Postoperative Outcomes Tumor site Upper/middle third Lower third/gastroesophageal junction Surgical approach Open Minimally invasive esophagectomy/hybrid Transhiatal Ivor-Lewis Mckeown Anastomosis site Thorax Neck En-bloc resection Lymphadenectomy Two fields lymph node dissection Three fields lymph node dissection Complications Clavien-Dindo 0 (no complications) I/II III/IV Pulmonary Cardiovascular Anastomotic leak Perioperative mortality
Chemotherapy only (n ¼ 226)
Chemotherapy–Radiation Therapy (n ¼ 112)
36 (16) 190 (84)
13 (12) 99 (88)
195 31 21 21 184
64 48 11 42 59
p Value
0.288
(57) (43) (10) (37) (53)
<0.001
21 (9) 205 (91) 189 (84)
42 (37.5) 70 (63) 75 (67)
<0.001
100 (44) 126 (56)
86 (77) 26 (23)
<0.001
120 29 77 38 41 31 3
54 20 36 15 18 21 6
(86) (14) (9) (9) (82)
<0.001
<0.001
0.423 (53) (13) (34) (17) (18) (14) (1)
(49) (18) (33) (13) (16) (19) (5)
0.416 0.637 0.227 0.064
Values are n (%).
for squamous cell carcinoma patients (57% vs 40%, p ¼ 0.595). Because the survival curves appeared to separate for squamous cell carcinoma patients we evaluated CSS (Fig 3). For the whole cohort 5-year CSS was 65% versus 56% (p ¼ 0.077) for nCRT and nCT, respectively. There was no difference for adenocarcinoma patients (51% vs 52%, p ¼ 0.824), whereas for squamous cell carcinoma patients there was a survival advantage to nCRT over nCT (87% vs 68%, p ¼ 0.019). On multivariable analysis of the whole study cohort independent clinical predictors of cancer-specific mortality were clinical stage (cStage III/IV: hazard ratio [HR], 1.82; 95% confidence interval [CI], 1.15–2.88; p ¼ 0.011), PS 1/2 (HR, 1.59; 95% CI, 1.06–2.38; p ¼ 0.024), and response to induction therapy (nonresponsive: HR, 1.55; 95% CI, 1.01– 2.38; p ¼ 0.045). In patients with adenocarcinoma tumors PS was the only independent clinical predictor of CSS approaching significance (PS 1/2: HR, 1.64; 95% CI, 0.99– 2.74; p ¼ 0.056). In patients with squamous cell tumors response to induction therapy (nonresponsive: HR, 2.62; 95% CI, 0.98–7.00; p ¼ 0.055) was the only independent predictor of CSS approaching significance. For the whole cohort independent predictors of allcause mortality were cStage III/IV (HR, 1.61; 95% CI, 1.07–2.43; p ¼ 0.024), PS 1/2 (HR, 1.71; 95% CI, 1.18–2.49; p ¼ 0.005), response to induction therapy (nonresponsive:
HR, 1.61; 95% CI, 1.08–2.39; p ¼ 0.020), preinduction tumor maximum standardized uptake value (HR, 1.03; 95% CI, 1.01–1.05; p ¼ 0.045), and surgery year (1990–1999: HR, 3.57; 95% CI, 1.01–12.5; p ¼ 0.050) (Table 4).
Comment The optimal multimodality treatment for locally advanced EC is unknown. Most centers in the United States prefer nCRT, as evidenced by data from the National Cancer Database (2006–2012) in which only 12.5% of patients with induction therapy for EC received nCT [5]. However, conflicting data exist from previous randomized trials regarding the two strategies. The NeoRes (Neoadjuvant Chemotherapy Versus Radiochemotherapy for Cancer of the Esophagus or Cardia) trial recently published by Klevebro and colleagues [3] randomized 181 patients at multiple institutions in Sweden and Norway with esophageal squamous cell carcinoma or adenocarcinoma to either three cycles of chemotherapy (cisplatin and fluorouracil) followed by surgery or to the same chemotherapy followed by concomitant radiation (40 Gy) and then surgery. The primary endpoint of this study was CPR rate, and therefore it might not have been powered to detect survival differences. Not surprisingly, nCRT patients showed improved pathologic response to
Ann Thorac Surg 2018;-:-–-
5
STILES ET AL INDUCTION THERAPY FOR ESOPHAGEAL CANCER
Table 3. Pathologic Characteristics Pathologic Characteristics Resection margin Proximal, cm Distal, cm Radial, cm <0.1 0.1 R0 R1 Positive radial (n ¼ 332) Positive proximal/distal (n ¼ 332) Change from cStage to pStage No change Downstaging Upstaging Pathologic complete response pN stage N0 N1 N2 N3 No. of lymph nodes resected No. of pathologic þve lymph nodes pStage 0 I II III IV
Chemotherapy Only (n ¼ 226)
Chemotherapy–Radiation Therapy (n ¼ 112)
9 4.9 0.3 30 98 203 23 10 12
(5.1–12) (3–6.5) (0.1–0.8) (23) (77) (90) (10) (4.5) (5)
105 89 32 15
(46.5) (39.4) (14.2) (6.6)
37 62 13 33
(33) (55.4) (11.6) (29.5)
(33.2) (26.5) (21.7) (18.6) (21–38) (0–5)
64 23 19 6 21 0
(57.1) (20.5) (17) (5.4) (14–32) (0–2)
(6.6) (11.9) (25.2) (54) (2.2)
33 13 22 38 6
(29.5) (11.6) (19.6) (33.9) (5.4)
75 60 49 42 28 1 15 27 57 122 5
5.1 6.1 0.5 5 62 106 6 2 1
(3.2–9) (4.3–9.6) (0.2–0.9) (7.5) (92.5) (95) (5) (2) (1)
p Value <0.001 <0.001 0.011 0.006 0.136 0.350 0.068 0.020
<0.001 <0.001
<0.001 <0.001 <0.001
Values are n (%) or median (IQR).
Fig 2. (A) Overall survival (OS) and (B) disease-free survival (DFS) for patients with locally advanced esophageal cancer receiving neoadjuvant chemoradiation (CRT) versus neoadjuvant chemotherapy (CT).
6
STILES ET AL INDUCTION THERAPY FOR ESOPHAGEAL CANCER
Ann Thorac Surg 2018;-:-–-
trial of 75 EC patients published by Burmeister and colleagues [2] that demonstrated a 3-year overall survival of 49% after nCT compared with 52% after nCRT (p ¼ 0.97). However, a third randomized trial by Stahl and colleagues [1], the POET trial (Preoperative Chemotherapy or Radiochemotherapy in Esophagogastric Adenocarcinoma Trial), which included 126 patients, suggested a strong trend toward a benefit of nCRT compared with nCT, with 3-year survivals of 47% with nCRT compared with 28% with nCT (p ¼ 0.07). Both the Burmeister and Stahl trials showed higher rates of CPR and of nodal downstaging with nCRT, similar to the NeoRes results. However differences in surgical quality may explain the distinctions between the Stahl, Klevebro, and Burmeister trials in which only the nCT arm in the Stahl trial is a survival outlier. In the POET trial only 47% of patients receiving nCT had transthoracic esophagectomy, compared with 90% and 100% in the Klevebro and Burmeister trials, respectively. In subgroup analysis of randomized trial data [6] patients with moderate nodal metastatic burden treated by transthoracic esophagectomy had superior survival compared with those treated by transhiatal esophagectomy (5-year DFS of 64% vs 23%, p ¼ 0.02). We expect that with the infrequent use of transthoracic resections, such as in the Stahl trial, radiation may indeed improve local control and survival. Our institutional study presented here mirrors the data from the Klevebro and Burmeister trials. When grouping all adenocarcinoma and squamous cell EC patients together we found no difference in overall survival, DFS,
Fig 3. Cancer-specific survival of esophageal cancer patients receiving neoadjuvant chemoradiation (CRT) versus neoadjuvant chemotherapy (CTH), by histologic subtype adenocarcinoma or squamous cell carcinoma.
neoadjuvant therapy, with a 28% rate of CPR and a 65% rate of negative lymph nodes, comparing favorably with the 9% CPR rate and 38% rate of negative nodes after nCT. However this did not translate to improved survival, with 3-year overall survival of 49% in the nCT arm and 47% in the nCRT arm (p ¼ 0.77). Progression-free 3-year survival was 44% in both treatment arms. This was similar to what was previously reported in a randomized
Table 4. Multivariable Analysis Predictors of Overall (All-Cause) Mortality Multivariable Analysisa
Univariate Analysis 95% CI
95% CI
All Patients (Events 177/338)
HR
Lower
Upper
p Value
Age Male gender Performance status 1/2 Tumor site (upper/middle thirds) Tumor histology (squamous cell cancer) cT stage (III/IV) cNstage (Nþve) cStage (III/IV) Induction type (CRT) Clinical response (nonresponsive) Preinduction tumor SUVmax Approach (minimally invasive esophagectomy/hybrid) Extent of resection (en bloc) Lymphadenectomy (2 fields) Surgery year 2010–2016 2000–2009 1990–1999
1.01 1.48 1.97 1.58 0.56 1.78 1.35 1.67 0.98 1.50 1.02 0.71
0.99 0.99 1.45 0.99 0.39 1.14 0.95 1.18 0.69 1.08 0.99 0.44
1.03 2.22 2.66 2.52 0.82 2.78 1.92 2.35 1.40 2.08 1.04 1.15
0.216 0.058 <0.001 0.053 0.002 0.012 0.099 0.004 0.908 0.017 0.152 0.161
0.74 0.88
0.51 0.64
1.08 1.19
1.63 1.52
0.94 1.02
Reference 2.83 2.27
a
HR
Lower
Upper
p Value
1.08 1.71 0.92 0.70
0.65 1.18 0.43 0.42
1.79 2.49 1.96 1.15
0.777 0.005 0.824 0.156
1.61
1.07
2.43
0.024
1.61 1.03 1.25
1.08 1.01 0.58
2.39 1.05 2.69
0.020 0.045 0.564
0.114 0.401
0.86
0.53
1.40
0.543
0.081 0.039
1.14 3.57
0.60 1.01
2.17 12.5
0.697 0.050
Univariate variables with p < 0.1 were included in the multivariable model.
CI ¼ confidence interval;
CRT ¼ chemoradiation;
HR ¼ hazard ratio;
SUVmax ¼ maximum standardized uptake value.
Ann Thorac Surg 2018;-:-–-
or CSS after nCRT versus nCT. This was despite higher rates of pathologic response after nCRT, as demonstrated by higher rates of CPR and pN0 classification. Given our institutional high rate of transthoracic resection (90%), our preference for en bloc resection (78%), and our low rate of R1 resections (8.6%) we believe that any benefits of adding a second local control modality (radiation) are limited, particularly with regard to patients with adenocarcinoma. However it appears there may be some benefit to preoperative radiation therapy in patients with squamous cell carcinoma. In our series patients with squamous cell cancer undergoing nCRT had a suggestion of improved 5-year overall survival (73% vs 55%, p ¼ 0.157) and 5-year DFS (57% vs 40%, p ¼ 0.595) and had a statistically significant improvement in 5-year CSS (87% vs 68%, p ¼ 0.019) compared with those undergoing nCT. Although not powered to detect differences in outcome based on tumor histology, the Klevebro trial also suggested that any potential benefit of nCRT is driven by patients with squamous cell carcinoma. In patients with adenocarcinoma the survival curves actually favored nCT, although they were not statistically significant. Similarly, in the CROSS trial [7] it is apparent that the benefit of nCRT is much more clinically important in patients with squamous cell carcinoma than in those with adenocarcinoma (HR, 0.46, p ¼ 0.004, vs HR, 0.75, p ¼ 0.059). The most recent meta-analysis of the two strategies also suggested only modest differences in HRs comparing nCT versus nCRT to surgery alone (0.83 and 0.75, respectively) in adenocarcinoma patients [8]. Our study is limited by its retrospective nature and by our inability to control for biases in treatment allocation. Additionally, there were modifications to our surgical approach during the time period studied. In summary, although radiation therapy may improve outcomes in patients with locally advanced squamous cell EC, there appears to be no benefit to adding radiation therapy to the induction regimen for patients with esophageal adenocarcinoma. Improved pathologic response in these patients does not translate to improvements in survival when an appropriate esophagectomy is performed. The addition of radiation therapy increases cost, hypothetically could lead to dose reductions of chemotherapy, and potentially increases noncancer-related deaths in the first year after treatment. The 20% CPR rate and excellent survival reported with the use of neoadjvuant FLOT (5-fluorouracil, leucovorin, oxaliplatin, and docetaxel) add strength to the argument for chemotherapy alone [9]. The ESOPEC (Perioperative Chemotherapy Compared to Neoadjuvant Chemoradiation in Patients with Adenocarcinoma of the Esophagus) trial meant to compare FLOT with nCRT using the CROSS regimen is currently accruing patients and should more clearly define the role of nCT versus nCRT in patients with adenocarcinoma [10].
STILES ET AL INDUCTION THERAPY FOR ESOPHAGEAL CANCER
7
We believe that our institutional data are consistent with a recent study from the National Cancer Database [5] and with data from randomized trials. We therefore believe that a tailored approach to neoadjuvant therapy for patients with locally advanced EC is appropriate, depending on tumor histology. For patients with esophageal squamous cell carcinoma we recommend adding radiation treatment to induction therapy. However, for adenocarcinoma induction chemotherapy alone appears to be a sound treatment regimen.
References 1. Stahl M, Walz MK, Riera-knorrenschild J, et al. Preoperative chemotherapy versus chemoradiotherapy in locally advanced adenocarcinomas of the oesophagogastric junction (POET): long-term results of a controlled randomised trial. Eur J Cancer 2017;81:183–90. 2. Burmeister BH, Thomas JM, Burmeister EA, et al. Is concurrent radiation therapy required in patients receiving preoperative chemotherapy for adenocarcinoma of the oesophagus? A randomised phase II trial. Eur J Cancer 2011;47:354–60. 3. Klevebro F, Alexandersson von D€ obeln G, Wang N, et al. A randomized clinical trial of neoadjuvant chemotherapy versus neoadjuvant chemoradiotherapy for cancer of the oesophagus or gastro-oesophageal junction. Ann Oncol 2016;27:660–7. 4. Spicer JD, Stiles BM, Sudarshan M, et al. Preoperative chemoradiation therapy versus chemotherapy in patients undergoing modified en bloc esophagectomy for locally advanced esophageal adenocarcinoma: is radiotherapy beneficial? Ann Thorac Surg 2016;101:126. 5. Samson P, Robinson C, Bradley J, et al. Neoadjuvant chemotherapy versus chemoradiation prior to esophagectomy: impact on rate of complete pathologic response and survival in esophageal cancer patients. J Thorac Oncol 2016;11:2227–37. 6. Omloo JM, Lagarde SM, Hulscher JB, et al. Extended transthoracic resection compared with limited transhiatal resection for adenocarcinoma of the mid/distal esophagus: five-year survival of a randomized clinical trial. Ann Surg 2007;246:992–1000; discussion 1000–1. 7. Shapiro J, van Lanschot JJ, Hulshof MC, et al. Neoadjuvant chemoradiotherapy plus surgery versus surgery alone for oesophageal or junctional cancer (CROSS): long-term results of a randomised controlled trial. Lancet Oncol 2015;16:1090–8. 8. Sjoquist KM, Burmeister BH, Smithers BM, et al. Survival after neoadjuvant chemotherapy or chemoradiotherapy for resectable oesophageal carcinoma: an updated metaanalysis. Lancet Oncol 2011;12:681–92. 9. Schulz C, Kullmann F, Kunzmann V, et al. NeoFLOT: multicenter phase II study of perioperative chemotherapy in resectable adenocarcinoma of the gastroesophageal junction or gastric adenocarcinoma—very good response predominantly in patients with intestinal type tumors. Int J Cancer 2015;137:678–85. 10. Hoeppner J, Lordick F, Brunner T, et al. ESOPEC: prospective randomized controlled multicenter phase III trial comparing perioperative chemotherapy (FLOT protocol) to neoadjuvant chemoradiation (CROSS protocol) in patients with adenocarcinoma of the esophagus (NCT02509286). BMC Cancer 2016;16:503.