Esophageal cancer recurrence after multimodality therapy: analysis of patterns and implications on long-term survival Authors: Farah A Abdallah, Belisario A Ortiz, Robert A Vierkant, Carlos A Puig, K Robert Shen, Dennis A Wigle, Stephen D Cassivi, Janani S Reisenauer, Krishan R Jethwa, Cameron M Callaghan, Zhaohui Jin, Harry H Yoon, Christopher L Hallemeier, Luis F Tapias Published: 2026 ABSTRACT AbstractBackgroundMany patients with esophageal cancer experience recurrence after curative-intent multimodality therapy. This study aims to evaluate risk factors, patterns, and outcomes of esophageal cancer recurrence after neoadjuvant chemoradiotherapy and esophagectomy.MethodsSingle-institution retrospective review of patients who underwent multimodality therapy including esophagectomy for esophageal adenocarcinoma or squamous cell carcinoma (SCC) from 2011 to 2021. Detailed data on recurrence patterns was collected. Factors associated with time to recurrence, overall survival, and survival after recurrence were analyzed.ResultsThere were 665 patients included of which 627 (94%) had clinical stage ≥ II disease. Recurrence after esophagectomy occurred in 241 (36%) patients. Median time to recurrence was 9.9 months (interquartile range, 5.4–18.8 months). Recurrence was distant in 119 (49%) patients, regional in 72 (30%) patients, local in 12 (5%), locoregional in 2 (0.8%), and locoregional + distant in 35 (15%). Factors independently associated with a higher risk of recurrence included male sex, SCC histology, and pN stage. Downstaging after neoadjuvant therapy was associated with a lower risk of recurrence. Recurrence was associated with a higher risk of death (adjusted hazard ratio 5.95, 95% confidence interval: 4.73–7.50, P<0.001). Median overall survival after surgery was 12 months in patients with recurrence and 123 months in patients without recurrence. Recurrences in the pleura and bone were independently associated with decreased post-recurrence survival.ConclusionsThe incidence of recurrence, particularly distant, remains high despite neoadjuvant therapy and esophagectomy, consistent with observations from clinical trials. Better perioperative systemic therapies are needed for patients with locally advanced esophageal cancer.Keywords: Esophageal neoplasms, neoadjuvant therapy, esophagectomy Highlight box Key findings • Recurrence rates following curative intent multimodality therapy for esophageal cancer remain high and most recurrences are distant. • Recurrence dramatically reduces long-term survival, particularly recurrences in the pleura or bones. • Downstaging after neoadjuvant therapy was associated with a lower risk of recurrence. What is known and what is new? • Esophageal cancer is associated with a high risk of recurrence despite modern advancements in treatment. • Our study further investigates risk factors associated with recurrence and the impact of esophageal cancer recurrence on survival. What is the implication, and what should change now? • New treatment paradigms in esophageal cancer should focus on maximizing pathologic complete responses and reducing distant recurrences. Based on recent evidence, clinicians must quickly incorporate the use of neoadjuvant chemo-immunotherapy in the appropriate clinical settings. Introduction Esophageal cancer represents the 7th most common cause of cancer-related mortality in the world (1). Globally, squamous cell carcinoma (SCC) is the most common, however, in the United States, adenocarcinoma (ACA) has become the most common histology (2). In 2025, approximately 22,000 new cases and over 16,000 deaths from esophageal cancer are expected in the United States (3). In the last decade, multimodality therapy has been established as the standard of care in patients with resectable locally advanced esophageal and esophagogastric junction (EGJ) carcinomas, based on the results of the CROSS trial that demonstrated the superiority of neoadjuvant chemoradiotherapy (CRT) followed by surgery over surgery alone (4). Recently, data have emerged demonstrating that triplet chemotherapy regimens are superior to the CROSS CRT regimen in patients with esophageal or EGJ ACA and their combination with immune checkpoint inhibitors can results in improved outcomes (5-7). Unfortunately, recurrence of esophageal and EGJ cancer after multimodality therapy remains a major problem. In randomized clinical trials, recurrence rates after neoadjuvant therapy and surgery have ranged between 39-55%, with distant recurrences representing most cases (5,6,8). This underscores the need for better systemic control of disease. Despite recent developments in the multimodality management of patients with ACA, those with SCC will still be recommended to receive neoadjuvant CRT based on current evidence. Further, as the neoadjuvant treatment paradigm for patients with ACA shifts, it will be important to procure data that can serve as a benchmark to evaluate real-world results. Therefore, the objective of this study was to examine outcomes at a high-volume referral center with the treatment of esophageal and EGJ carcinoma during the era dominated by neoadjuvant CRT. Specifically, we aimed to examine rates and patterns of recurrence after multimodality therapy, explored factors associated with recurrence, and assessed the impact of recurrence and its characteristics on long-term survival. We present this article in accordance with the STROBE reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0506/rc). Methods Study population This was a single-institution retrospective cohort study that included patients with primary esophageal or EGJ carcinoma who underwent surgical treatment with a curative intent after neoadjuvant therapy at a high volume tertiary care center. The study population was derived from an institutional database of 810 patients who underwent surgery for esophageal or EGJ cancer from January 1, 2011, until December 31, 2021. Exclusion criteria included histology other than ACA or SCC, clinical M1 disease, not receiving neoadjuvant therapy, surgical resection other than esophagectomy with a gastric conduit reconstruction (e.g., total gastrectomy, colon or small bowel interposition), positive surgical margins, pathologic M1 disease, and lack of authorization for research participation (Figure S1). In total, 665 patients with primary esophageal/EGJ cancer who underwent multimodality therapy were included in the analysis. The database was reviewed for demographics, comorbidities, clinical characteristics and staging, operative procedures, pathologic characteristics and staging, 90-day post operative complications, cancer recurrence, and vital status during last known follow up. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the institutional review board at Mayo Clinic (No. 21-010529) and individual consent for this retrospective analysis was waived. All participants granted research authorization according to Minnesota law. Statistical analysis and study endpoints Clinical and pathological staging were according to the American Joint Committee on Cancer 8th edition (9). Tumor downstaging was defined as a pathologic T or N staging lower than the clinical T or N staging. Data were summarized using frequencies and proportions for categorical variables, and means with standard deviations (SDs), or medians with interquartile ranges (IQRs) for continuous variables. Missing data counts are presented but not included in analysis. Recurrence free survival was calculated from date of surgery to either date of first recurrence or date of last known recurrence-free follow-up. For time-to-event analyses, right-censoring was applied as the last known follow-up date. Associations of disease recurrence with demographic and clinical characteristics were visualized using Kaplan-Meier cumulative incidence curves. Formal associations were carried out using Cox proportional hazards regression models using a two-stage approach. First, univariable associations were assessed by fitting a series of Cox regression models using time to recurrence as the event and modeling each characteristic in turn as the lone exposure variable. Following this, we examined the independent associations of characteristics with recurrence using multivariable Cox regression, fitting time to recurrence as the outcome and simultaneously including all univariately significant variables (P<0.05) as exposure terms after removing redundant variables from the model (for instance, if global clinical stage was included in the multivariable model, then individual clinical T- and N-stage variables were not). Primary analyses fit any recurrence as the outcome. Separate secondary analyses examined in turn associations with local/regional recurrence and distant recurrence by right-censoring recurrence not of that type at date of recurrence. In addition, associations of demographic, clinical and recurrence-related attributes with time to death from any cause (overall survival) was assessed among patients who recurred, by following patients from date of recurrence to date of death or last known to be alive. Associations were again examined using Kaplan-Meier curves, and formal associations with risk of death were carried out using a two-stage Cox regression approach like that described above. We examined associations of demographic and clinical characteristics and recurrence status with time to death, among all patients, by following patients from date of initial procedure to date of death or last known to be alive. For these, recurrence status was treated as a time-dependent variable, such that patients who recurred migrated from the no recurrence group to the recurrence group at their date of recurrence. Recurrence was assessed as a simple binary outcome variable and by levels defined by number of foci and stage of recurrence. Associations again were visualized using Kaplan-Meier curves and were formally assessed using the two-stage Cox regression approach. All statistical tests were two-sided. Analyses were carried out using the SAS System (SAS Institute, Inc., Cary, NC, USA). Results Baseline characteristics and initial treatment approach There were 665 patients with primary esophageal/EGJ carcinoma who received multimodality therapy including neoadjuvant treatment and esophagectomy during the study period. Baseline characteristics and neoadjuvant therapy approaches are shown in Table 1. Most patients presented with locally advanced disease (cT2N0M0 or higher), with clinical stages III and IVA representing 425 (67%) and 122 (10%) of patients, respectively. ACA was the most frequent histology in 593 (89%) patients. Neoadjuvant therapy was administered to all patients, with CRT being the most frequently used method during the study period (99% of patients). Table 1Baseline characteristics of 665 patients with esophageal cancer that received surgical treatmentVariableResultsAge, years64.5 (9.7)Sex Female104 [16] Male561 [84]Smoking Never186 [28] Current105 [16] Former374 [56]Alcohol use Never150 [23] Current431 [66] Former73 [11] Missing11Comorbidities Diabetes145 [22] Coronary artery disease103 [16] Atrial fibrillation44 [7] Chronic kidney disease28 [4]Tumor length (cm)5.1 (3.5)Clinical T-stage T15 [1] T2127 [20] T3487 [77] T410 [22] Missing36Clinical N-stage N0181 [29] N1328 [52] N2111 [18] N315 [2] Missing30Clinical stage I7 [1] II80 [13] III425 [67] IVA122 [10] Missing31Histology Adenocarcinoma593 [89] Squamous cell carcinoma72 [11]Neoadjuvant therapy665 [100] Chemotherapy665 [100] Radiotherapy656 [99] Immunotherapy27 [4]Data are presented as n [%] or mean (SD). N, node; SD, standard deviation; T, tumor. Baseline characteristics of 665 patients with esophageal cancer that received surgical treatment Data are presented as n [%] or mean (SD). N, node; SD, standard deviation; T, tumor. Surgical treatment and pathologic staging Details on surgical treatment are shown in Table 2. Ivor-Lewis esophagectomy was the preferred technique used in 481 (72%) patients. An open surgical approach was reported in 508 (76%) patients while a minimally invasive or hybrid approach was used in 157 (24%) patients. Postoperative complications occurred in 434 (65%) patients, with anastomotic leak occurring in 54 (8%). Detailed postoperative complications are shown in Table S1. Postoperative mortality at 30 days and 90 days was observed in 16 (2%) and 36 (5%) patients, respectively. Table 2Surgical treatment, pathologic characteristics and adjuvant therapiesVariableResultsEsophagectomy type Ivor-Lewis481 [72] Transhiatal72 [11] McKeown111 [17] Missing1Surgical approach Minimally invasive102 [15] Hybrid55 [8] Open508 [76]Hospital length of stay (days)9 [7–13]Post-operative complications434 [65] Atrial fibrillation168 [25] Pneumonia77 [12] Anastomotic leak54 [8] Chylothorax52 [8] Sepsis46 [7]Mortality 30-day16 [2] 90-day36 [5]Number of lymph nodes18 [13–25] <1063 [10] 10–14146 [22] 15–19171 [26] 20–29197 [30] ≥3087 [13]Pathologic T-stage ypT0208 [32] ypT1106 [16] ypT2123 [19] ypT3222 [34] ypT42 [0.3] Missing4Pathologic N-stage ypN0426 [64] ypN1133 [20] ypN279 [12] ypN326 [4] Missing1Pathologic tumor stage I164 [25] II85 [13] IIIA64 [10 IIIB141 [22] IVA25 [4] pCR177 [27] Missing9Downstaging after neoadjuvant therapy T downstaging398 [60] N downstaging314 [47] Any (T or N) downstaging438 [66]Downstaging groups No downstaging183 [29] Downstaging, but residual disease266 [43] Pathologic complete response177 [28] Unable to classify39Adjuvant therapy75 [11] Chemotherapy33 [5] Radiotherapy3 [0.5] Immunotherapy47 [7]Data are presented as n [%] or median [IQR]. IQR, interquartile range; N, node; pCR, pathological complete response; T, tumor. Surgical treatment, pathologic characteristics and adjuvant therapies Data are presented as n [%] or median [IQR]. IQR, interquartile range; N, node; pCR, pathological complete response; T, tumor. Details of pathologic evaluation are also shown in Table 2. Downstaging was observed in most patients after neoadjuvant therapy. Of 626 patients with complete information to determine the degree of response to neoadjuvant therapy, a pathological complete response (pCR, ypT0N0) was achieved in 177 (28%) patients, while downstaging, but with evidence of residual disease was observed in 266 (42%). According to histology, pCR was achieved in 161 of 560 (29%) patients with ACA and 16 of 66 (24%) with SCC (P=0.262). Downstaging, but with evidence of residual disease was observed in 241 of 560 (43%) with ACA and 25 of 66 (38%) with SCC. Post-operative adjuvant therapy was administered to 75 (11%) patients, most commonly immunotherapy. Esophageal/EGJ cancer recurrence Median follow up time was 32.4 months (IQR, 13.2–62.4 months). Cancer recurrence was observed in 241 (36%) patients (Figure 1A). Details on patterns of recurrence are shown in Table 3. Most recurrences were distant, observed in 154 (23%) patients. Distant recurrence occurred in isolation in 119 (18%) patients. The most common sites of distant recurrence included liver in 47 (7%) patients, lungs in 42 (6%), and bone and pleura in 27 (4%) patients each. Among patients who experienced esophageal/EGJ cancer recurrence, median time to recurrence after surgery was 9.9 months (IQR, 5.4–18.8 months). Esophageal cancer recurrence in 665 patients after multimodality therapy (neoadjuvant therapy + esophagectomy). Recurrence in the entire cohort (A), and by sex (B), cancer histology (C), pathology T stage (D), pathologic N stage (E), and degree of downstaging after neoadjuvant therapy (F), including complete pathologic response (pcr). N, node; T, tumor. Table 3Recurrence of esophageal cancer after multimodality therapyVariableResults, n [%]Recurrence241 [36]Recurrence focality Single focus181 [75] Multiple foci58 [24] Missing4Recurrence type Local only12 [5] Regional only72 [30] Distant only119 [49] Local + regional2 [0.8] Local + distant3 [1] Regional + distant31 [13] Local + regional + distant1 [0.4] Missing4Recurrence type Any local 18 [7] Any regional 106 [44] Any distant154 [64] Missing4Recurrence site Liver47 [20] Lung42 [17] Thoracic lymph nodes29 [12] Bone27 [11] Pleura27 [11] Brain24 [10] Anastomosis18 [7] Mediastinum18 [7] Adrenal glands8 [3] Stomach3 [1] Other81 [34] Recurrence of esophageal cancer after multimodality therapy Table 4 shows the risk factor analysis for associations with time to recurrence. Cumulative incidence rates of recurrence a 1, 3, and 5 years were 26%, 43%, and 46% for the entire group, respectively (Figure 1A). Univariable analysis showed that male sex, advanced clinical T and N stages, increasing overall clinical stage, SCC histology, advanced pathologic T and N stages, and use of adjuvant therapy were associated with increased risk of recurrence (Figure 1B-1E). Conversely, pCR and tumor downstaging were significantly associated with a lower risk of recurrence (Figure 1F). Cumulative incidence rates of recurrence at 1, 3, and 5 years were 14%, 28%, and 29% in patients with a pCR, while they were 24%, 39%, and 43% in patients with downstaging but with residual disease, and 43%, 64%, and 69% in patients without downstaging on final pathology. On multivariable analysis, male sex, SCC and pathologic N stage were independent predictors of esophageal/EGJ cancer recurrence. Table 4Risk factors associated with time of recurrence in 665 patients receiving multimodality therapy for esophageal carcinomaAttributeUnivariableMultivariableHR (95% CI)P valueHR (95% CI)P valueAge at surgery0.14 <551.00 55–640.71 (0.50–1.02) 65–740.68 (0.48–0.96) ≥750.82 (0.53–1.29)Biological sex0.010.01 Female1.001.00 Male1.69 (1.13–2.54)1.72 (1.14–2.61)Race0.52 Caucasian1.00 Other/unknown0.79 (0.39–1.60)Diabetes0.64 No1.00 Yes0.93 (0.68–1.27)Chronic kidney disease0.67 No1.00 Yes0.86 (0.44–1.68)Smoking/tobacco status0.11 Never1.00 Current0.82 (0.55–1.21) Former0.74 (0.56–0.98)Alcohol use0.28 Never1.00 Current1.29 (0.93–1.81) Former1.36 (0.84–2.19)Clinical T-stage0.05 cT1/cT21.00 cT3/cT41.40 (1.00–1.96)Clinical N-stage0.02 cN01.00 cN11.07 (0.78–1.47) cN2/cN31.58 (1.10–2.28)Clinical stage<0.0010.11 I/II1.001.00 III1.94 (1.20–3.12)1.40 (0.83–2.36) IV2.69 (1.60–4.52)1.85 (1.01–3.40)Tumor length0.12 ≤3 cm1.00 4–5 cm0.94 (0.66–1.35) 5–6 cm1.48 (0.98–2.23) 8+ cm1.18 (0.76–1.84)Location of tumor epicenter0.62 <30 cm upper/middle esophagus1.00 30–39.5 cm lower esophagus1.34 (0.71–2.56) >40 cm EGJ1.24 (0.63–2.44)Cancer histology0.010.03 Adenocarcinoma1.001.00 Squamous cell carcinoma1.62 (1.11–2.36)1.55 (1.04–2.32)Resection type0.95 Ivor Lewis1.00 Transhiatal esophagectomy1.06 (0.71–1.60) McKeown esophagectomy1.03 (0.73–1.46)Surgery type0.21 Minimally invasive1.00 Hybrid0.62 (0.35–1.10) Open0.79 (0.57–1.11)Any post-surgical complications0.26 No1.00 Yes0.86 (0.66–1.12)Number of post-surgical complications0.71 01.00 10.89 (0.64–1.23) 20.85 (0.59–1.24) 3 or more0.84 (0.59–1.19)Post-surgical atrial fibrillation0.030.07 No1.001.00 Yes0.70 (0.51–0.97)0.73 (0.52–1.03)Post-surgical ileus0.44 No1.00 Yes0.83 (0.52–1.33)Post-surgical pneumonia0.29 No1.00 Yes0.78 (0.49–1.24)Post-surgical pulmonary embolism0.32 No1.00 Yes0.77 (0.47–1.28)Post-surgical intervention0.25 No1.00 Yes0.75 (0.47–1.22)Post-surgical respiratory failure0.07 No1.00 Yes0.54 (0.28–1.06)Pathologic T-stage<.00010.92 ypT01.001.00 ypT11.18 (0.75–1.85)0.98 (0.46–2.10) ypT21.88 (1.27–2.78)1.15 (0.57–2.30) ypT32.49 (1.78–3.47)1.14 (0.56–2.29)Pathologic N-stage<0.001<0.001 ypN01.001.00 ypN12.45 (1.80–3.34)1.96 (1.32–2.91) ypN24.22 (3.00–5.93)2.70 (1.63–4.46) ypN34.12 (2.39–7.10)2.66 (1.32–5.36)Pathologic complete response<0.001 No1.00 Yes0.47 (0.33–0.65)T-stage downstaging<0.001 No1.00 Yes0.50 (0.38–0.65)N-stage downstaging<0.001 No1.00 Yes0.58 (0.45–0.76)Downstaging<0.001 No1.00 Yes0.42 (0.32–0.54)Downstaging<0.0010.68 No downstaging1.001.00 Downstaging, but no complete response0.48 (0.36–0.65)0.81 (0.51–1.28) Complete response0.30 (0.21–0.44)0.73 (0.30–1.78)Number of lymph nodes harvested0.96 <101.01(0.62–1.67) 10–141.10 (0.77–1.58) 15–191.00 20–291.00 (0.71–1.41) ≥300.92 (0.58–1.45)Any adjuvant treatment<0.0010.24 No1.001.00 Yes2.00 (1.43–2.80)1.41 (0.79–2.53)Adjuvant immunotherapy0.0040.84 No1.001.00 Yes1.81 (1.21–2.71)0.93 (0.48–1.83)CI, confidence interval; EGJ, esophagogastric junction; HR, hazard ratio; N, node; T, tumor. Risk factors associated with time of recurrence in 665 patients receiving multimodality therapy for esophageal carcinoma CI, confidence interval; EGJ, esophagogastric junction; HR, hazard ratio; N, node; T, tumor. Table S2 shows the risk factor analysis for associations with time to recurrence according to the extent of recurrence. The risk of local/regional recurrence was higher with male sex, increasing clinical stage, and increasing pathologic T and N stage. It was lower in former smokers and with downstaging or pCR. The risk of distant recurrence was higher in clinical N2 or N3 disease, clinical stage IV, SCC histology, increasing pathologic T and N stage, and use of adjuvant therapy. Interaction tests showed that patients with diabetes and longer tumors were at increased risk of local/regional recurrence, but not distant recurrence. In addition, patients with clinical N2 or N3 disease were at increased risk of distant recurrence but not local/regional recurrence. Overall survival Median overall survival after neoadjuvant therapy and esophagectomy was 56.9 months for the entire cohort (Figure 2A). Overall survival for the entire cohort at 1 year was 82% [95% confidence interval (CI): 79–85%], at 3 years it was 59% (95% CI: 55–63%), and at 5 years it was 49% (95% CI: 45–53%). Table S3 shows the risk factor analysis for overall survival. Downstaging and pCR were associated with improved survival on univariable analysis (Figure S2). Multivariable analysis showed that age >75 years, male sex, SCC histology, occurrence of 3 or more postoperative complications, and pathologic N stage were independently associated with decreased overall survival (Figures S3-S5). Overall survival in 665 patients with esophageal cancer after multimodality therapy (neoadjuvant therapy + esophagectomy). Overall survival in the entire cohort (A). Overall survival by recurrence status (B). Overall survival among patients who experienced recurrence according to number of recurrence foci (C), type of recurrence (D), distant recurrence in the pleura (E), and distant recurrence in bone (F). Impact of esophageal cancer recurrence on overall survival Recurrence after multimodality therapy was associated with a significantly decreased overall survival (Table 5, Figure 2B). In patients who experienced recurrence, the number of recurrence sites and the type of recurrence were independently associated with decreased overall survival (Table 5, Figure 2C,2D). Among patients who died after recurrence, median time from recurrence to death was 8.7 months (IQR, 4.3–16.1 months). Table S4 shows analysis of factors associated with overall survival among patients who experienced recurrence. Recurrence in the pleura and in bone were independently associated with decreased overall survival after recurrence (Figure 2E,2F). Having had a pCR initially was associated with improved survival after recurrence on univariable analysis but this association was lost on multivariable analysis (adjusted HR 0.64, 95% CI: 0.41–1.01, P=0.05). Table 5Impact of esophageal cancer recurrence on overall survival after multimodality therapyAttributeUnivariableMultivariableHR (95% CI)P valueHR (95% CI)P valueRecurrence status<0.001<0.001 No1.001.00 Yes5.71 (4.56–7.15)5.92 (4.65–7.55)Foci of recurrence<0.001<0.001 No recurrence1.001.00 One focus5.07 (3.98–6.48)5.36 (4.14–6.94) 2+ foci8.34 (6.01–11.59)8.40 (5.91–11.93)Stage of recurrence<0.001<0.001 No recurrence1.001.00 Local5.31 (2.70–10.44)6.91 (3.47–13.76) Regional4.75 (3.44–6.55)4.77 (3.40–6.78) Distant6.29 (4.89–8.09)6.50 (4.98–8.48)CI, confidence interval; HR, hazard ratio. Impact of esophageal cancer recurrence on overall survival after multimodality therapy CI, confidence interval; HR, hazard ratio. Early recurrence or death after multimodality therapy Esophageal/EGJ cancer recurrence occurred within 6 months of surgery in 71 patients, representing 11% of the entire cohort and 30% of those who experienced recurrence. Death within 6 months of surgery was observed in 51 patients, representing 8% of the entire cohort and 15% of those who experienced death during follow up. In total, 109 (16%) patients experienced recurrence or death within 6 months after surgery. Discussion In this contemporary cohort, we showed that esophageal and EGJ cancer recurrence remains a major problem after trimodality therapy with an estimated 5-year recurrence rate of 46%. We found that most recurrences were distant and that male sex, SCC, and pathologic N stage were independent predictors of recurrence. Importantly, tumor downstaging was associated with a lower risk of recurrence. A pCR, achieved in 28% of patients after neoadjuvant therapy, was associated with the lowest risk of recurrence. We observed a pCR rate of 29% in ACA and 24% in SCC. This is similar to what was observed in the CROSS trial for ACA but much lower than the reported 49% for SCC in that study (4). This might account for the observed differences in outcomes by histology identified in our study. Our findings are in line with prior large retrospective studies that have reported recurrence rates of 31–56% after esophagectomy (10-12). As a referral center, we likely see patients with more advanced disease at presentation, as over 70% of patients had node-positive disease. Similarly, recurrence rates after neoadjuvant therapy and surgery in randomized clinical trials have ranged between 39–55% (5,6,8). In addition, we showed that recurrence has a dramatic negative impact on long-term survival. Further, distant recurrence and multiple sites of recurrence were associated with worse long-term survival. This is not a surprising finding, but it further highlights that, despite contemporary neoadjuvant CRT, systemic dissemination of disease remains a challenge and, that it is difficult to salvage patients that experience distant recurrence. Moreover, we found that recurrences in the bone or pleura were associated with reduced survival when compared to recurrences from other sites. This information can be helpful when counseling patients on their prognosis and treatment options and might inform clinicians to consider clinical trials in these patients. A recent report from the ESOPEC trial demonstrated that neoadjuvant FLOT chemotherapy results in significantly lower rates of distant recurrence in patients with ACA when compared to CROSS, with similar rates of local, regional, and locoregional recurrence (13). As shown here, prognosis is largely impacted by distant recurrence. Therefore, improved systemic control of disease is key in this patient population. Interestingly, we found that approximately 16% of patients experienced recurrence or death within 6 months after multimodality therapy including surgery. This finding raises the issue of futility in nearly 1 of every 6 patients, as it would be hard to argue that those patients experienced significant benefit from esophagectomy. It also underscores the limitations of current non-invasive clinical staging methods, where imaging studies certainly fall short in detecting the presence of more advanced disease upon presentation or after neoadjuvant therapy. The role of new biomarkers, such as circulating tumor DNA (ctDNA), in the initial staging and longitudinal assessment deserves careful consideration as it might help better select patients with esophageal or EGJ carcinoma that would benefit the most from an aggressive multimodal treatment approach. It remains to be seen how monitoring of molecular residual disease is incorporated into clinical practice. One of the strongest predictors of recurrence and long-term survival in patients with locally advanced esophageal/EGJ carcinoma was the response to neoadjuvant therapy. A pCR has been associated with a lower risk of recurrence and longer overall survival (10,14). Our findings provide further evidence to support this concept, with some benefit also observed in patients that had partial responses. On the other hand, our results are sobering as a significant proportion of patients with a pCR still experienced recurrence during follow-up (28% at 3 years). Prior studies have found that poor tumor differentiation grade is a strong predictor of recurrence in patients that had a pCR (15). Histology seems to impact the patterns of recurrence in this subgroup, with more distant recurrences seen in patients with ACA and more locoregional recurrences seen in SCC (15). Although an important achievement after neoadjuvant therapy, pCR does not necessarily translate into cure. These data should be accounted for when considering organ-preserving strategies with active surveillance as opposed to surgery as recently reported by the SANO trial (16). To note, patients that experience pCR currently are not candidates for adjuvant therapies after trimodality therapy. Therefore, there is an opportunity to study novel adjuvant therapies or the impact of surveillance of molecular residual disease with ctDNA, that might help detect recurrences earlier before they are evident on imaging studies. Recently, the ESOPEC trial—a phase 3, multicenter, unblinded, randomized, controlled trial of perioperative FLOT chemotherapy versus preoperative CRT in patients with resectable esophageal ACA—showed improved overall survival with FLOT over the CROSS trial CRT regimen (5). These results have led to a shift in clinical practice to neoadjuvant FLOT for good performance status patients with resectable esophageal/EGJ ACA. Still, progression-free survival in the FLOT arm was approximately 51%. Outside of a clinical trial, a recent retrospective study from two European centers reported recurrences in 48% of patients in spite of 84% of patients having received treatment with perioperative FLOT (12). Despite the improvements seen with FLOT, a giant leap is needed in the systemic treatment of patients with esophageal and GEJ carcinomas. Perioperative chemo-immunotherapy combinations are now emerging. The KEYNOTE-585 trial—a phase 3, multicenter, placebo-controlled, double-blinded, randomized trial in patient with EGJ and gastric cancer—included a cohort of patients comparing FLOT plus pembrolizumab versus FLOT plus placebo (17). This showed improved pCR rates with chemoimmunotherapy during an interim analysis, however, this did not translate into improved event-free survival or overall survival. The PALACE-2 trial—a phase 1/2, single-arm, multicenter trial—examined neoadjuvant concurrent chemoradiation plus pembrolizumab in patients with locally advanced resectable esophageal SCC and showed a pCR rate of 43.2% (18). Long-term outcomes from this study are pending. The MATTERHORN study—a phase 3, double-blind, randomized trial in resectable gastric and GEJ cancer—compared perioperative FLOT with or without durvalumab finding that chemoimmunotherapy resulted in significantly increased pCR rates and event-free survival (7,19). As both CRT and multi-agent systemic therapies have demonstrated improvements in progression-free and OS through distinct mechanisms, there could be consideration for total neoadjuvant strategies inclusive of sequential multi-agent systemic therapy and CRT for select patients. A subset analysis from the TOPGEAR trial suggested a trend towards improved OS with FLOT followed by CRT (HR 0.78, 95% CI: 0.55–1.12), compared with FLOT alone, although, this would need to be validated in the context of an appropriately designed prospective clinical trial (20). There may also be opportunities to incorporate CRT amongst a multi-agent systemic therapy backbone for “borderline resectable” characteristics such as extensive lymph node burden and/or in the subset of prognostically adverse patients with poor clinical response per interval imaging after 4 cycles of FLOT-based systemic therapy, although, each of these scenarios are also in need of prospective validation. In patients with esophageal SCC, a multicenter retrospective study showed that neoadjuvant chemoimmunotherapy was associated with improved overall survival, lower rates of recurrence, and lower rates of distant recurrence when compared to CRT (21). To note, this study included patients treated with multiple different immune checkpoint inhibitors. Moreover, treatment with immunotherapy of patients with deficient mismatch repair (MMR) tumors, appears promising (22,23). Our study has limitations. First, its retrospective nature limits the quality of the collected data. Second, we do not have comprehensive biomarker testing in this cohort that could allow for exploration of molecular factors associated with recurrence. Our results represent multidisciplinary care at a referral center for esophageal diseases; therefore, they cannot be extrapolated to other scenarios. Finally, the association found between adjuvant therapy and a higher risk of recurrence is likely confounded by indication, with an inability to fully adjust for it on multivariable analysis, as only patients with residual disease after surgery were candidates for adjuvant treatment based on the CheckMate 577 trial (24). Our study was not designed to assess if adjuvant immunotherapy effectively decreased the risk of recurrence in a real-word practice. Conclusions In conclusion, recurrence after neoadjuvant CRT followed by surgery remains a common problem in patients with locally advanced resectable esophageal and GEJ carcinoma, that negatively impacts survival. Particularly, patients who develop distant recurrence in the bone or pleura experience worse outcomes. We provide further evidence showing that the degree of pathologic response to neoadjuvant therapy is a strong predictor of recurrence and long-term survival. Our study provides benchmark outcomes data from a high-volume referral center and should serve for future comparisons of real-world results obtained with emerging neoadjuvant or perioperative treatment regimens. Supplementary The article’s supplementary files as Acknowledgments None. Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the institutional review board at Mayo Clinic (No. 21-010529) and individual consent for this retrospective analysis was waived. All participants granted research authorization according to Minnesota law. Footnotes Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0506/rc Funding: None. Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0506/coif). C.A.P. reports receiving honoraria for lectures from Curio Science. J.S.R. reports a research grant from Intuitive Surgical. K.R.J. reports receiving honoraria for lectures from RanOncQuestions.com, LLC. Z.J. reports consulting fees from Exelixis, Elevar Therapeutics, and Innovent Biologics; all payments were made to Mayo Clinic. H.H.Y. reports a research grant and support to attend a meeting from Beigene LTD; research grants from Amgen, and the National Cancer Institute; payment for expert testimony by the Food and Drug Administration (FDA); participation of advisory boards for Bayer Pharmaceuticals, Beigene LTD, Merck, Elevation Oncology, Novartis, Jazz Pharmaceuticals, Astellas, Amgen, and AstraZeneca; and leadership positions in Beigene LTD (global steering committee), LSK Biopharma (global steering committee), The Angiogenesis Foundation (advisor), Merck KN-585 (global steering committee), Merrimack (global steering committee), NCI Alliance, and the National Comprehensive Cancer Network (NCCN). L.F.T. reports consulting fees from AstraZeneca. The other authors have no conflicts of interest to declare. Data Sharing Statement Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0506/dss References Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0506/prf Supplementary Materials The article’s supplementary files as Data Availability Statement Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0506/dss