Abstract
Objective:
To compare the efficacy and safety of induction FOLFOX followed by PET-directed neoadjuvant chemoradiation therapy (nCRT), induction carboplatin plus paclitaxel (CP) followed by PET-directed nCRT, and nCRT with CP alone in patients with esophageal adenocarcinoma (EAC).
Summary Background Data:
nCRT with CP is a standard treatment for locally advanced EAC. The results of Cancer and Leukemia Group B 80803 support the use of induction chemotherapy followed by PET-directed chemoradiation therapy.
Methods:
We retrospectively identified all patients with EAC who underwent the treatments above followed by esophagectomy. We assessed incidences of pathologic complete response (pCR), near-pCR (ypN0 with ≥90% response), and surgical complications between treatment groups using Fisher’s exact test and logistic regression; disease-free survival (DFS) and overall survival (OS) were estimated by the Kaplan-Meier method and evaluated using the log-rank test and extended Cox regression.
Results:
In total, 451 patients were included: 309 (69%) received induction chemotherapy before nCRT (FOLFOX, n=70; CP, n=239); 142 (31%) received nCRT with CP. Rates of pCR (33% vs 16%, P=0.004), near-pCR (57% vs 33%, P<0.001), and 2-year DFS (68% vs 50%, P=0.01) were higher in the induction FOLFOX group than in the induction CP group. Similarly, the rate of near-pCR (57% vs 42%, P=0.04) and 2-year DFS (68% vs 44%, P<0.001) were significantly higher in the FOLFOX group than in the no-induction group.
Conclusions:
Induction FOLFOX followed by PET-directed nCRT may result in better histopathologic response rates and DFS than either induction CP plus PET-directed nCRT or nCRT with CP alone.
Mini-Abstract
We sought to compare the efficacy and safety of induction FOLFOX followed by PET-directed neoadjuvant chemoradiation therapy (nCRT), induction carboplatin plus paclitaxel (CP) followed by PET-directed nCRT, and nCRT with CP alone in patients with esophageal adenocarcinoma (EAC). In these patients, induction FOLFOX followed by PET-directed nCRT may result in better histopathologic response rates and disease-free survival than either induction CP plus PET-directed nCRT or nCRT with CP alone.
INTRODUCTION
The CROSS trial established the effectiveness of neoadjuvant chemoradiation therapy (nCRT) followed by surgery for locally advanced esophageal adenocarcinoma (EAC).1 Histopathologic tumor regression after nCRT is the most important predictor of survival among patients with EAC.2–8 Patients with pathologic complete response (pCR) have the most favorable prognosis,4,9,10 whereas those with minimal pathologic response to preoperative chemoradiation have poor outcomes and no survival advantage, compared with patients with similar stage treated with upfront esophagectomy.11–13
Assessment of the metabolic tumor response to induction chemotherapy using 18F-fluorodeoxyglucose positron emission tomography (PET) with integrated computed tomography (CT) has been shown to accurately predict treatment response.14 The identification of metabolic nonresponse before nCRT may enable an individualized, response-guided treatment approach to tailor multimodality therapy.
We have shown that patients with a PET nonresponse who continue the same chemotherapy regimen during radiation therapy (RT) have worse survival outcomes than patients who switch regimens.15,16 On the basis of these data, Cancer and Leukemia Group B (CALGB) launched the 80803 study to evaluate whether switching to a different chemotherapy regimen during nCRT among patients with a PET nonresponse could improve rates of pCR and survival. The study met its primary endpoint of improving the pCR rate in PET nonresponders who changed chemotherapy regimens during RT. Additionally, PET responders to induction FOLFOX had higher rates of pCR than PET responders to carboplatin plus paclitaxel (CP).17 As a result, FOLFOX has replaced CP as the preferred induction chemotherapy regimen at our institution.
As CALGB 80803 was not a randomized study, it remains unclear whether FOLFOX and PET-directed nCRT has better efficacy and safety than standard CP and RT without induction chemotherapy. Our study aims to answer two questions: (1) If induction chemotherapy is used, does induction FOLFOX provide better efficacy (measured by rates of pCR, near-pCR, disease-free survival [DFS], and overall survival [OS]) and safety (measured by incidence of serious postoperative complications) than induction CP? (2) If induction FOLFOX has better efficacy than induction CP, is induction FOLFOX followed by PET-directed nCRT better than nCRT with CP alone?
METHODS
Patient Population
After approval from our institutional review board, we retrospectively queried our database to identify all patients with locally advanced EAC who were treated surgically from January 2010 to June 2019. All patients gave written informed consent. Patients included in the study received any of the following treatment regimens: (1) FOLFOX group: induction FOLFOX or, rarely, capecitabine plus oxaliplatin, followed by PET assessment and PET-directed nCRT (FOLFOX for responders and CP for nonresponders); (2) CP group: induction CP followed by PET assessment and PET-directed nCRT (CP for responders and FOLFOX for nonresponders); and (3) no-induction group: standard nCRT with CP alone (Supplementary Figure 1). The standard regimen at our center was induction CP from 2012 to 2017. After the initial results of CALGB 80803 were presented in January 2017, we changed our standard approach to induction FOLFOX. Some patients included in the analysis received CROSS-style nCRT with CP alone at an outside institution before coming to Memorial Sloan Kettering for surgery.
Patients treated with other chemotherapy regimens, salvage esophagectomy, and induction chemotherapy without follow-up PET imaging were excluded. Additionally, patients treated with induction chemotherapy who, in the absence of any significant clinical or PET response during nCRT, continued the same chemotherapy regimen were excluded (Supplementary Figure 2).
To answer our first question, we compared patients treated with induction FOLFOX and patients treated with induction CP; to answer our second question, we compared patients treated with induction FOLFOX and patients treated with nCRT with CP alone.
Pretreatment Evaluation and Treatment
Pretreatment clinical staging was performed using endoscopic ultrasound and/or cross-sectional imaging, including PET-CT. Outside records were requested if staging and nCRT occurred at an outside institution. All pathology reports were reviewed, and staging was performed using the 8th edition of the AJCC Staging Manual of the tumor-node-metastasis classification.18 Pathologic staging for patients who underwent resection before the institution of the 8th edition was modified to reflect this edition. Details of chemotherapy, radiation, and surgery are presented in Supplementary Figure 1.
PET Response
PET response was defined as a decrease in maximum standardized uptake value (SUVmax) after induction chemotherapy of ≥35%. This calculation was made by the treating medical oncologist on the basis of the reported SUVmax in the baseline and post–induction chemotherapy reports.14,15,19,20 Whereas standard practice was to switch chemotherapy regimens during nCRT in patients with a PET nonresponse, some patients with a PET nonresponse remained on the same chemotherapy during RT because of clear clinical benefit (e.g., substantial improvement in dysphagia) in the context of a borderline PET response (a decrease in SUVmax <35%). These patients were subsequently classified in subanalyses as having had a “treatment response.”
Statistical Analysis
Categorical variables were summarized as frequencies and percentages and compared between treatment groups using the Chi-square test or Fisher’s exact test; continuous variables were summarized as medians and ranges and compared using the Wilcoxon rank-sum test.
Primary endpoints were rates of pCR and near-pCR. In posttreatment specimens, pCR was defined as a complete absence of viable tumor grossly and microscopically in the entire surgical specimen, consisting of the resected esophagus and all harvested lymph nodes, and near-pCR was defined as N0 disease in which the pathologist identified ≥90% treatment effect and only scant viable tumor cells within the primary tumor specimen. Previous work from our institution revealed almost identical survival benefits associated with pCR and near-pCR.21 For this reason, near-pCR was chosen as the surrogate outcome measure, as opposed to pCR.
Univariable and multivariable logistic regression analyses were performed to quantify the relationship between induction treatment groups and near-PCR. In the multivariable models, all adjustment variables were chosen for inclusion a priori on the basis of known relationships with prognosis and histologic tumor response. Crude and adjusted odds ratios (ORs) and corresponding 95% confidence intervals (CIs) were reported.
Long-term survival endpoints (DFS and OS) were calculated from the date of surgery to the date of event (death for OS and recurrence or death for DFS) or were censored on the date of the last follow-up. Survival was estimated using the Kaplan-Meier method and compared between pretreatment groups using the max-combo weighted log-rank test, to address any potential violation of the proportional hazards assumption.22 The association of treatment groups with DFS was further assessed using multivariable Cox regression. The proportional hazards assumption was assessed graphically using the scaled Schoenfeld residuals and the score test for the independence of the residuals with time. In the presence of nonproportional hazards, the model-adjusted relative hazard of FOLFOX versus no induction was quantified by including an interaction term of log-transformed DFS and induction therapy in the Cox model.23,24 The time-varying effects were displayed graphically for up to 3 years from the time of surgery. For ease of presentation, hazard ratios (HRs) with 95% CIs were reported at select time points. Clinically important covariates, including age, sex, clinical T and N stages, grade, Zubrod score, and serious complications, were selected a priori for inclusion in the multivariable analysis.
Surgical complications, in-hospital mortality, 30-day mortality, 90-day mortality, rates of serious complications, and length of stay were compared between treatment groups using Fisher’s exact test or the Wilcoxon rank sum test. A serious postoperative complication was defined as any grade ≥III complication, in accordance with the Clavien-Dindo classification.25
The cumulative incidence of each disease recurrence type was estimated such that death without a recurrence was considered a competing risk. The association of each recurrence type with induction chemotherapy was assessed using Gray’s test.
All P values are two-tailed, and statistical significance was defined as two-sided P<0.05. Statistical analyses were performed using R (v3.6.1, R Foundation for Statistical Computing, Vienna, Austria), including the survival and coxphw packages.
RESULTS
Patient Characteristics
A total of 451 patients were included in the study: 309 (69%) received induction chemotherapy before nCRT (FOLFOX, n=70; CP, n=239), and 142 (31%) received nCRT with CP alone. Relevant demographic, perioperative, and pathologic characteristics are summarized in Table 1. Age, race, and baseline stage of disease were not statistically significantly different between groups.
Table 1.
Relevant demographic, clinical, and pathologic characteristics of all included patients
| Characteristic | All (N=451) | FOLFOX (n=70) | CP (n=239) | No Induction (n=142) | P |
|---|---|---|---|---|---|
| Age, years | 64 (29 to 84) | 65 (41 to 80) | 64 (29 to 84) | 65 (36 to 84) | 0.35 |
| Sex | 0.49 | ||||
| Male | 382 (85) | 56 (80) | 205 (86) | 121 (85) | |
| Female | 69 (15) | 14 (20) | 34 (14) | 21 (15) | |
| Race | 0.51 | ||||
| White | 409 (91) | 63 (90) | 215 (90) | 131 (92) | |
| Black | 3 (0.7) | 0 (0) | 3 (1) | 0 (0) | |
| Asian | 16 (4) | 4 (6) | 6 (3) | 6 (4) | |
| Other | 23 (5) | 3 (4) | 15 (6) | 5 (4) | |
| Smoking status | 0.58 | ||||
| Former/current | 301 (67) | 44 (63) | 158 (66) | 99 (70) | |
| Never | 150 (33) | 26 (37) | 81 (34) | 43 (30) | |
| Baseline grade | >0.9 | ||||
| Well/moderate | 249 (56) | 38 (57) | 132 (55) | 79 (56) | |
| Poor | 196 (44) | 29 (43) | 106 (45) | 61 (44) | |
| cT | 0.47 | ||||
| 1/2 | 55 (12) | 6 (9) | 26 (11) | 23 (16) | |
| 3 | 373 (83) | 59 (87) | 200 (84) | 114 (80) | |
| 4a | 21 (5) | 3 (4) | 13 (5) | 5 (4) | |
| cN | 0.51 | ||||
| N0 | 135 (30) | 22 (31) | 66 (28) | 47 (33) | |
| N+ | 316 (70) | 48 (69) | 173 (72) | 95 (67) | |
| ASA | 0.62 | ||||
| ASA II | 63 (14) | 9 (13) | 37 (15) | 17 (12) | |
| ASA III | 362 (80) | 58 (83) | 185 (77) | 119 (84) | |
| ASA IV | 26 (6) | 3 (4) | 17 (7) | 6 (4) | |
| Esophagectomy type | 0.049 | ||||
| Ivor Lewis | 420 (93) | 65 (93) | 217 (91) | 138 (97) | |
| Other | 31 (7) | 5 (7) | 22 (9) | 4 (3) | |
| LVI | 100 (22) | 7 (10) | 58 (24) | 35 (25) | 0.028 |
| Neural invasion | 105 (23) | 13 (19) | 57 (24) | 35 (25) | 0.59 |
| Pathologic grade | 0.020 | ||||
| Well/moderate | 192 (44) | 25 (39) | 106 (45) | 61 (44) | |
| Poor | 142 (32) | 14 (22) | 83 (35) | 45 (32) | |
| No residual cancer | 103 (24) | 25 (39) | 45 (19) | 33 (24) | |
| ypTNM stage | 0.59 | ||||
| I | 205 (45) | 36 (51) | 104 (44) | 65 (46) | |
| II | 61 (14) | 12 (17) | 30 (13) | 19 (13) | |
| III | 168 (37) | 20 (29) | 97 (41) | 51 (36) | |
| IV | 17 (4) | 2 (3) | 8 (3) | 7 (5) |
Data are no. (%) or median (range). Statistical analysis was performed using the Pearson χ2 test or Fisher’s exact test when any expected cell count was <5 for categorical variables and the Wilcoxon rank-sum test for continuous variables. ASA, American Society of Anesthesiology; CP, carboplatin plus paclitaxel; LVI, lymphovascular invasion; ypTNM stage, pathologic stage of disease after neoadjuvant chemoradiation.
The proportion of patients with PET response was slightly higher in the FOLFOX group (n=50 [71%]) than in the CP group (n=154 [64%], P=0.32). All patients with PET response remained on the same chemotherapy regimen during nCRT, and most patients with PET nonresponse were switched to a different regimen.
Pathologic Response
Ninety-three patients (21%) had pCR, and 179 had near-pCR (40%). Table 2 shows the differences in rates of pCR and near-pCR among treatment groups. The induction FOLFOX group had a statistically significantly higher rate of pCR (33% vs. 16%, P=0.004) and near-pCR (57% vs. 33%, P<0.001) than the induction CP group. The results of univariable and multivariable logistic regression analyses confirmed that, compared with CP, induction FOLFOX was independently associated with near-pCR (OR, 2.15 [95% CI, 1.17 to 4.00], P=0.014) (Table 3).
Table 2.
Rates of pCR and near-pCR between patients with and without PET response by induction chemotherapy group and between groups
| Characteristic | FOLFOX (n=70) | CP (n=239) | P (FOLFOX vs. CP) | No Induction (n=142) | P (FOLFOX vs. No Induction) |
|---|---|---|---|---|---|
| PET response | 50 (71) | 154 (64) | 0.32 | N/A | |
| pCR | |||||
| All patients | 23 (33) | 39 (16) | 0.004 | 31 (22) | 0.10 |
| PET response | 19/50 (38) | 31/154 (20) | 0.014 | 0.038 | |
| PET nonresponse | 4/20 (20) | 8/85 (9) | 0.24 | >0.99 | |
| P (within each induction group, PET response vs. nonresponse) | 0.17 | 0.043 | N/A | ||
| Near-pCR | |||||
| All patients | 40 (57) | 80 (33) | <0.001 | 59 (42) | 0.040 |
| PET response | 30/50 (60) | 60/154 (39) | 0.013 | 0.032 | |
| PET nonresponse | 10/20 (50) | 20/85 (24) | 0.027 | 0.48 | |
| P (within each induction group, PET response vs. nonresponse) | 0.59 | 0.022 | N/A |
Data are no. (%). Statistical analysis was performed using Fisher’s exact test. CP, carboplatin plus paclitaxel; N/A, not applicable; pCR, pathologic complete response; PET, positron emission tomography.
Table 3.
Univariable and multivariable logistic regression analysis to evaluate whether FOLFOX is independently associated with near-pCR
| Variable | Near-pCR | Univariable | Multivariable | |||
|---|---|---|---|---|---|---|
| No (n=189) | Yes (n=120) | OR (95% CI) | P | OR (95% CI) | P | |
| Induction group | ||||||
| CP | 159 (67) | 80 (33) | — | — | ||
| FOLFOX | 30 (43) | 40 (57) | 2.65 (1.54 to 4.60) | <0.001 | 2.15 (1.17 to 4.00) | 0.014 |
| Age, years | 65 (37 to 84) | 64 (29 to 80) | 1.0 (0.98 to 1.02) | 0.89 | 1.0 (0.97 to 1.03) | 0.98 |
| Sex | ||||||
| Male | 157 (60) | 104 (40) | — | — | ||
| Female | 32 (67) | 16 (33) | 0.75 (0.39 to 1.43) | 0.40 | 0.62 (0.29 to 1.32) | 0.22 |
| Race | ||||||
| White | 172 (62) | 106 (38) | — | — | ||
| Other | 17 (55) | 14 (45) | 1.34 (0.62 to 2.82) | 0.45 | 1.34 (0.57 to 3.19) | 0.50 |
| Smoking status | ||||||
| Former/current | 124 (61) | 78 (39) | — | |||
| Never | 65 (61) | 42 (39) | 1.03 (0.63 to 1.66) | 0.91 | 1.15 (0.65 to 2.04) | 0.63 |
| Clinical T stage | ||||||
| cT1–2 | 22 (69) | 10 (31) | — | — | ||
| cT3 | 158 (61) | 101 (39) | 1.41 (0.65 to 3.22) | 0.40 | 1.87 (0.81 to 4.56) | 0.15 |
| cT4 | 9 (56) | 7 (44) | 1.71 (0.49 to 5.98) | 0.40 | 1.66 (0.44 to 6.38) | 0.45 |
| Clinical N stage | ||||||
| cN0 | 53 (60) | 35 (40) | — | — | ||
| cN+ | 136 (62) | 85 (38) | 0.95 (0.57 to 1.58) | 0.83 | 0.72 (0.39 to 1.31) | 0.28 |
| Baseline grade | ||||||
| Poor | 80 (59) | 55 (41) | — | — | ||
| Well/moderate | 107 (63) | 63 (37) | 0.86 (0.54 to 1.36) | 0.51 | 0.76 (0.45 to 1.30) | 0.32 |
| LVI | 64 (98) | 1 (2) | 0.02 (0.00 to 0.08) | <0.001 | 0.02 (0.00 to 0.08) | <0.001 |
Data are no. (row percentages) or median (range). Analysis was performed using a logistic regression model. The multivariable logistic regression analyses were adjusted for age, sex, race, smoking status, pretreatment histologic grade, clinical T and N stage, and vascular invasion. CI, confidence interval; CP, carboplatin plus paclitaxel; OR, odds ratio; LVI, lymphovascular invasion; pCR, pathologic complete response.
Rates of pCR (33% vs. 22%, P=0.10) and near-pCR (57% vs. 42%, P=0.04) were also higher in the induction FOLFOX group than in the no-induction group, but this difference was statistically significant only for near-pCR. When patients in the FOLFOX group with a PET response were compared with all patients in the no-induction group, those in the FOLFOX group had statistically significantly higher rates of pCR (38% vs. 22%, P=0.04) and near-pCR (60% vs. 42%, P=0.03).
Survival
Median follow-up among survivors was 36 months (range, 1 to 120 months) overall, 18 months (range, 6 to 95 months) in the FOLFOX group, 46 months (range, 1 to 120 months) in the CP group, and 24 months (range, 1 to 98 months) in the no-induction group. Overall, 204 patients died during the observation period: 19, 115, and 70 in the FOLFOX, CP and no-induction groups respectively. Median OS was 46 months (95% CI, 41 to 59 months) in the overall cohort, not reached in the FOLFOX group, 49 months (95% CI, 41 to 84 months) in the CP group, and 39 months (95% CI, 31 to 60 months) in the no-induction group. OS was not statistically significantly different between the induction FOLFOX vs. induction CP group (P=0.43) or between the induction FOLFOX group and the no-induction CP group (P=0.26) (Supplementary Figure 3).
For the DFS endpoint, 22, 147, and 85 patients in the FOLFOX, CP, and no-induction groups, respectively, experienced recurrence or died. Two-year DFS was 68% (95% CI, 56% to 81%) in the induction FOLFOX group, 50% (95% CI, 44% to 57%) in the induction CP group, and 44% (95% CI, 36% to 54%) in the no-induction group. DFS was statistically significantly higher in the FOLFOX group than in the CP group (P=0.01) or the no-induction group (P<0.001) (Figure 1). The relative hazard of induction FOLFOX versus no induction varied over time (Supplemental Figure 4). In the first 6 months after surgery, the relative hazard was not statistically significant; thereafter, a statistically significant protective effect was observed for induction FOLFOX, compared with no induction. In the multivariable model that adjusted for age, sex, clinical T stage and N stage, grade, Zubrod score, and serious complications, patients treated with FOLFOX had better survival than those who did not receive induction therapy (HR [95% CI] at 6 months, 0.31 [0.09 to 1.08], P=0.067; HR [95% CI] at 24 months, 0.18 [0.04 to 0.75], P=0.019). Distant recurrences were the most common type of disease recurrence, with a 5-year cumulative incidence of 33% (95% CI, 28 to 38%), but they were not significantly affected by induction therapy (no induction vs. induction: 36% [95%CI, 27% to 45% vs. 31% [95% CI, 26% to 37%]; P=0.43).
Figure 1.
Disease-free survival between patients treated with induction FOLFOX or induction carboplatin plus paclitaxel (CP) and between patients treated with induction FOLFOX or neoadjuvant chemoradiation therapy with CP (no induction).
Postoperative Morbidity
In-hospital mortality, 30-day mortality, 90-day mortality, rates of serious complications, and length of hospital stay were not statistically significantly different between patients who received induction chemotherapy and those who did not (Table 4). Of note, because patients in the FOLFOX group were treated more recently than patients in the CP group, a statistically significantly higher proportion of patients in this group underwent minimally invasive esophagectomy (66% vs. 44%, P=0.002). As a result, median length of stay was statistically significantly shorter in the FOLFOX group (median [range], 9 [6 to 28] vs. 10 [5 to 144] days, P<0.001).
Table 4.
Observed rates of unadjusted intraoperative and postoperative outcomes
| Outcome | Induction FOLFOX vs. CP | Induction Chemotherapy vs. No Induction | ||||
|---|---|---|---|---|---|---|
| FOLFOX (n=70) | CP (n=239) | P | Induction (n=309) | No Induction (n=142) | P | |
| Length of stay, days | 9 (6 to 28) | 10 (5 to 144) | <0.001 | 10 (5 to 144) | 10 (1 to 71) | 0.86 |
| Minimally invasive | 46 (66) | 105 (44) | 0.002 | 151 (49) | 60 (42) | 0.22 |
| R1 resection | 1 (1) | 11 (5) | 0.31 | 12 (4) | 4 (3) | 0.79 |
| Serious complication | 12 (17) | 41 (17) | >0.99 | 53 (17) | 29 (20) | 0.43 |
| In-hospital mortality | 0 (0) | 7 (3) | 0.36 | 7 (2) | 2 (1) | 0.73 |
| 30-day mortality | 0 (0) | 5 (2) | 0.59 | 5 (2) | 2 (1) | >0.99 |
| 90-day mortality | 1 (1) | 13 (5) | 0.20 | 14 (5) | 4 (3) | 0.45 |
Data are no. (%) or median (range). Fisher’s exact test and the Wilcoxon rank-sum test were used. CP, carboplatin plus paclitaxel.
Treatment Response Versus PET Response
Of the 105 patients with PET nonresponse, 28 (27%; 3/20 in the FOLFOX group and 25/85 in the CP group) remained on the same chemotherapy regimen during RT because of a treatment response. Among these patients (n=28), the median decrease in SUVmax was 30.8% (interquartile range, 22.6% to 32.7%). In the analyses above, these patients were considered to have PET nonresponse on the basis of the strict definition of SUVmax change. We repeated the analyses of pathologic response and survival by including these patients with treatment response in the PET-responder group. These results are shown in Supplementary Table 1 and Supplementary Figure 5. The results seen in patients with PET response were similar to those seen in patients with treatment response. Of note, 2 of 3 patients with nonresponse to FOLFOX on PET but with treatment response had pCR at surgery (Supplementary Figure 6).
DISCUSSION
As the results of CALGB 80803 suggested that PET responders to induction FOLFOX had the highest rates of pCR and median OS, we sought to determine whether induction FOLFOX or CP before PET assessment was superior. Rates of pCR (33%) and near-pCR (57%) were significantly higher in the FOLFOX group than in the CP group (16% and 33%)—these rates are virtually identical to those in CALGB 80803: 31% in the FOLFOX group and 14% in the CP group. Two-year DFS was also statistically significantly higher in the FOLFOX group than in the CP group (68% vs. 50%, P=0.01). OS was not statistically significantly different between the two groups, although our analysis was likely underpowered to detect a benefit in OS, and the median follow-up for the FOLFOX group was shorter. In-hospital mortality, 30-day mortality, 90-day mortality, and rates of serious complications were also not statistically significantly different between the two groups. These findings suggest that, if induction chemotherapy is given, FOLFOX should be the regimen of choice.
We next attempted to determine whether induction FOLFOX is superior to nCRT with CP alone. The pCR rate was higher in the FOLFOX group (33%) than in the no-induction group (22%), although the difference was not statistically significant; however, the rate of near-pCR was significantly higher in the FOLFOX group (57% vs. 42%). The rate of pCR in the no-induction group (22%) was also virtually identical to that in the CROSS trial (23%). Two-year DFS was significantly higher in the induction FOLFOX group than in the no-induction group (68% vs 44%, P<0.001). OS, in-hospital mortality, 30-day mortality, 90-day mortality, and rates of serious complications were not statistically significantly different between the two groups. These findings suggest that induction FOLFOX followed by PET-directed nCRT may result in superior histopathologic tumor regression and longer DFS than nCRT with CP alone. Of note, although some patients underwent standard nCRT at an outside institution, surgery and pathologic review were performed at our center. Additionally, that the rates of pCR among these patients closely match those in the CROSS trial suggest that the quality of nCRT and the fitness of patients in our cohort are comparable to those of a trial-eligible population; our conclusions are therefore likely generalizable to other patients.
A unique aspect of our analysis is that it included patients who had PET nonresponse but experienced a clinical benefit from induction chemotherapy and who continued with the same chemotherapy regimen during nCRT. Two of 3 such patients had pCRs at surgery. Although based on few patients, these data provide a real-world perspective on the use of PET imaging and suggest there may be a subset of patients who do not meet the criteria for PET response who nevertheless experience substantial clinical benefit, including pCR at surgery, with continuation of the same chemotherapy regimen during nCRT.
Most patients underwent standard-of-care PET imaging. In many cases, at least 1 PET scan was performed outside of our center; however, in most instances, images were obtained for official review and comparison at our center. Despite the lack of rigorous standardization of PET imaging that was used in the CALGB 80803 study, our results recapitulate its results, suggesting that PET assessment is an accurate method to predict tumor response and can be performed as a standard procedure.
A major criticism of our study may be that it did not include a control group who received FOLFOX with RT without PET guidance. However, the rate of pCR in our FOLFOX group (33%) is numerically higher than that in SWOG 0356 (28%) or the randomized phase II study by Ajani and colleagues26 (13% in the no-induction group and 26% in the induction FOLFOX group). In the prospective MUNICON study by Ott and colleagues, patients with PET nonresponse who continued with 5-fluorouracil (5-FU) and cisplatin had worse rates of major pathologic response and inferior survival outcomes, compared with patients with PET response who continued 5-FU and cisplatin (44% vs. 5%, P=0.001). A subsequent effort to salvage the poor outcomes of patients with PET nonresponse by adding RT also yielded disappointing results.27 These data—taken together with our data showing poor outcomes when the same regimen is continued after PET nonresponse16—justify changing the chemotherapy regimen in patients with PET nonresponse to FOLFOX.
Our study represents a retrospective analysis performed at a single high-volume center; as such, results may not be generalizable to other practices. Nevertheless, all patients who received CP alone with RT were treated in smaller academic and community practices before surgery. The pCR rates among patients who received induction therapy in our study as well as those who received standard CP and RT are virtually identical to those in the CALGB 80803 and CROSS studies, respectively.
In conclusion, our results suggest (1) the use of induction FOLFOX followed by PET-directed nCRT can lead to higher rates of pathologic response and 2-year DFS than either induction CP followed by PET-directed nCRT or nCRT with CP alone and (2) this approach does not pose a risk of increased postoperative complications. Our results are consistent with the findings of CALGB 80803, and our study provides comparative data for a no-induction group. Our findings suggest that the use of induction FOLFOX followed by PET-directed nCRT may represent an optimization of the currently available preoperative therapy options. In the absence of a randomized study evaluating PET-directed FOLFOX versus nCRT with CP alone (the CROSS regimen), which may not be feasible, the results of the present study can help establish induction FOLFOX with PET assessment as the standard of care for patients with EAC.
Supplementary Material
Acknowledgments:
David B. Sewell of the Department of Surgery, Memorial Sloan Kettering Cancer Center, provided editorial assistance.
Financial Support:
This work was supported, in part, by the National Cancer Institute at the National Institutes of Health (grant number P30 CA008748).
Disclosures:
Matthew J. Bott is a consultant for AstraZeneca. James M. Isbell has stock ownership in LumaCyte and is a consultant/advisory board member for Roche Genentech. Yelena Y. Janjigian has financial relationships with Eli Lilly, ASCO, Michael J. Hennessy Associates, Paradigm Medical Communications, Zymeworks, AstraZeneca, Daiichi Sankyo, ONO Pharma, Merck, and Bristol Myers Squibb. Bernard J. Park has served as a proctor for Intuitive Surgical and a consultant for COTA. Valerie W. Rusch reports grant support (institutional) from Genelux and Genentech, travel support from Intuitive Surgical, and travel support and payments from National Institutes of Health Coordinating Center for Clinical Trials. Abraham J. Wu has received research grants (institutional) from CivaTech Oncology and serves as a consultant for Simphotek. David R. Jones serves as a consultant for AstraZeneca and on a clinical trial steering committee for Merck. Geoffrey Y. Ku has the following relationships: research funding from Arog, research funding/consulting from AstraZeneca, research funding/consulting from Bristol Myers Squibb, research funding from Daiichi, consulting from Eli Lilly, research funding/consulting from Merck, research funding/consulting from Pieris, and research funding from Zymeworks. Daniela Molena serves as a consultant for Johnson & Johnson, Urogen, and Boston Scientific. All other authors have no disclosures.
Footnotes
Presentation: This work was presented as a virtual poster at the 2020 Annual Meeting of the American Society of Clinical Oncology Virtual Scientific Program on May 29, 2020.
List of Supplemental Digital Content
Supplemental Figures and Tables.docx
Data availability:
Data are available from the corresponding author on request.
REFERENCES
- 1.van Hagen P, Hulshof MC, van Lanschot JJ, et al. Preoperative chemoradiotherapy for esophageal or junctional cancer. N Engl J Med. 2012;366(22):2074–2084. [DOI] [PubMed] [Google Scholar]
- 2.Langer R, Ott K, Feith M, Lordick F, Siewert JR, Becker K. Prognostic significance of histopathological tumor regression after neoadjuvant chemotherapy in esophageal adenocarcinomas. Mod Pathol. 2009;22(12):1555–1563. [DOI] [PubMed] [Google Scholar]
- 3.Lorenzen S, Thuss-Patience P, Al-Batran SE, et al. Impact of pathologic complete response on disease-free survival in patients with esophagogastric adenocarcinoma receiving preoperative docetaxel-based chemotherapy. Ann Oncol. 2013;24(8):2068–2073. [DOI] [PubMed] [Google Scholar]
- 4.Meredith KL, Weber JM, Turaga KK, et al. Pathologic response after neoadjuvant therapy is the major determinant of survival in patients with esophageal cancer. Ann Surg Oncol. 2010;17(4):1159–1167. [DOI] [PubMed] [Google Scholar]
- 5.Rohatgi PR, Swisher SG, Correa AM, et al. Failure patterns correlate with the proportion of residual carcinoma after preoperative chemoradiotherapy for carcinoma of the esophagus. Cancer. 2005;104(7):1349–1355. [DOI] [PubMed] [Google Scholar]
- 6.Schneider PM, Baldus SE, Metzger R, et al. Histomorphologic tumor regression and lymph node metastases determine prognosis following neoadjuvant radiochemotherapy for esophageal cancer: implications for response classification. Ann Surg. 2005;242(5):684–692. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Blum Murphy M, Xiao L, Patel VR, et al. Pathological complete response in patients with esophageal cancer after the trimodality approach: The association with baseline variables and survival-The University of Texas MD Anderson Cancer Center experience. Cancer. 2017;123(21):4106–4113. [DOI] [PubMed] [Google Scholar]
- 8.Chirieac LR, Swisher SG, Ajani JA, et al. Posttherapy pathologic stage predicts survival in patients with esophageal carcinoma receiving preoperative chemoradiation. Cancer. 2005;103(7):1347–1355. [DOI] [PubMed] [Google Scholar]
- 9.Donahue JM, Nichols FC, Li Z, et al. Complete pathologic response after neoadjuvant chemoradiotherapy for esophageal cancer is associated with enhanced survival. Ann Thorac Surg. 2009;87(2):392–398; discussion 398–399. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Berger AC, Farma J, Scott WJ, et al. Complete response to neoadjuvant chemoradiotherapy in esophageal carcinoma is associated with significantly improved survival. J Clin Oncol. 2005;23(19):4330–4337. [DOI] [PubMed] [Google Scholar]
- 11.Chevrollier GS, Giugliano DN, Palazzo F, et al. Patients with non-response to neoadjuvant chemoradiation for esophageal cancer have no survival advantage over patients undergoing primary esophagectomy. J Gastrointest Surg. 2020;24(2):288–298. [DOI] [PubMed] [Google Scholar]
- 12.Dittrick GW, Weber JM, Shridhar R, et al. Pathologic nonresponders after neoadjuvant chemoradiation for esophageal cancer demonstrate no survival benefit compared with patients treated with primary esophagectomy. Ann Surg Oncol. 2012;19(5):1678–1684. [DOI] [PubMed] [Google Scholar]
- 13.Rizk NP, Venkatraman E, Bains MS, et al. American Joint Committee on Cancer staging system does not accurately predict survival in patients receiving multimodality therapy for esophageal adenocarcinoma. J Clin Oncol. 2007;25(5):507–512. [DOI] [PubMed] [Google Scholar]
- 14.Lordick F, Ott K, Krause BJ, et al. PET to assess early metabolic response and to guide treatment of adenocarcinoma of the oesophagogastric junction: the MUNICON phase II trial. Lancet Oncol. 2007;8(9):797–805. [DOI] [PubMed] [Google Scholar]
- 15.Ilson DH, Minsky BD, Ku GY, et al. Phase 2 trial of induction and concurrent chemoradiotherapy with weekly irinotecan and cisplatin followed by surgery for esophageal cancer. Cancer. 2012;118(11):2820–2827. [DOI] [PubMed] [Google Scholar]
- 16.Ku GY, Kriplani A, Janjigian YY, et al. Change in chemotherapy during concurrent radiation followed by surgery after a suboptimal positron emission tomography response to induction chemotherapy improves outcomes for locally advanced esophageal adenocarcinoma. Cancer. 2016;122(13):2083–2090. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Goodman KA, Niedzwiecki D, Hall N, et al. Initial results of CALGB 80803 (Alliance): A randomized phase II trial of PET scan-directed combined modality therapy for esophageal cancer. Journal of Clinical Oncology. 2017;35(4_suppl):1–1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Amin MB, Edge SB. AJCC cancer staging manual. Eighth edition. ed. Switzerland: Switzerland: Springer; 2017. [Google Scholar]
- 19.Ott K, Weber WA, Lordick F, et al. Metabolic imaging predicts response, survival, and recurrence in adenocarcinomas of the esophagogastric junction. J Clin Oncol. 2006;24(29):4692–4698. [DOI] [PubMed] [Google Scholar]
- 20.Weber WA, Ott K, Becker K, et al. Prediction of response to preoperative chemotherapy in adenocarcinomas of the esophagogastric junction by metabolic imaging. J Clin Oncol. 2001;19(12):3058–3065. [DOI] [PubMed] [Google Scholar]
- 21.Sihag S, Nobel T, Hsu M, et al. Survival following trimodality therapy in patients with locally advanced esophagogastric adenocarcinoma: does only a complete pathologic response matter? Ann Surg. 2020. [Online ahead of print]. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Karrison TG. Versatile tests for comparing survival curves based on weighted log-rank statistics. Stata J. 2016;16(3):678–690. [Google Scholar]
- 23.Quantin C, Moreau T, Asselain B, Maccario J, Lellouch J. A regression survival model for testing the proportional hazards hypothesis. Biometrics. 1996;52(3):874–885. [PubMed] [Google Scholar]
- 24.Allison PD. Survival Analysis Using SAS: A Practical Guide. SAS Publishing; 1995. [Google Scholar]
- 25.Clavien PA, Barkun J, de Oliveira ML, et al. The Clavien-Dindo classification of surgical complications: five-year experience. Ann Surg. 2009;250(2):187–196. [DOI] [PubMed] [Google Scholar]
- 26.Ajani JA, Xiao L, Roth JA, et al. A phase II randomized trial of induction chemotherapy versus no induction chemotherapy followed by preoperative chemoradiation in patients with esophageal cancer. Ann Oncol. 2013;24(11):2844–2849. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.zum Büschenfelde CM, Herrmann K, Schuster T, et al. (18)F-FDG PET-guided salvage neoadjuvant radiochemotherapy of adenocarcinoma of the esophagogastric junction: the MUNICON II trial. J Nucl Med. 2011;52(8):1189–1196. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data are available from the corresponding author on request.