Key Points
Question
Is active surveillance cost-effective compared with standard surgery among patients with esophageal cancer who achieved a complete clinical response after neoadjuvant chemoradiotherapy?
Findings
In this prespecified cost-effectiveness secondary analysis of a cluster randomized clinical trial including 309 patients, active surveillance resulted in slightly higher quality-adjusted life-years and lower health care costs over 5 years compared with standard surgery, with 97% of bootstrap replications indicating cost-effectiveness.
Meaning
Active surveillance is a cost-effective alternative to standard surgery after neoadjuvant chemoradiotherapy and may reduce health care costs without compromising patient outcomes.
Abstract
Importance
Active surveillance is noninferior to standard surgery for 2-year survival and improves short-term health-related quality of life among patients with a complete clinical response (CCR) after neoadjuvant chemoradiotherapy (nCRT) for esophageal cancer. Although active surveillance reduces the upfront costs of surgery and hospital stay, it requires repeated diagnostic tests and, for some patients, delayed surgery and hospitalization during follow-up.
Objective
To assess the cost-effectiveness of active surveillance compared with standard surgery after nCRT.
Design, Setting, and Participants
This prespecified cost-effectiveness analysis from a health care perspective conducted at 12 hospitals in the Netherlands as a secondary analysis of the Surgery as Needed for Oesophageal Cancer (SANO) trial, a noninferiority, cluster randomized study, enrolled patients with esophageal cancer who achieved a CCR after nCRT between November 8, 2017, and January 17, 2021, with follow-up for up to 5 years. Data were analyzed on June 1, 2025.
Interventions
Active surveillance, consisting of repeated response evaluations at 6, 9, 12, 16, 20, 24, 30, 36, 48, and 60 months after nCRT, compared with standard surgery.
Main Outcome and Measures
Incremental cost-effectiveness of active surveillance vs standard surgery and quality-adjusted life-years (QALYs) with 95% CIs up to 5 years were derived with bootstrapping, with 80% of patients (247 of 309) having complete follow-up. Incremental net monetary benefit (iNMB) was calculated at varying willingness-to-pay thresholds. All analyses followed the modified intention-to-treat principle. Costs are given in Euros (currency exchange rate of €1 = US $1.16 as of June 11, 2026).
Results
Among 309 patients (198 in the active surveillance group; median age, 69 years [IQR, 63-74 years]; 156 men [79%]; and 111 in the standard surgery group; median age, 68 years [IQR, 61-73 years]; 86 men [77%]), those in the active surveillance group had a mean of 2.99 QALYs (95% CI, 2.73-3.26) at 5 years vs 2.88 QALYs (95% CI 2.69-3.06) in the standard surgery group. Mean health care costs per patient at 5 years were €36 733 (95% CI, €33 530-€40 009) in the active surveillance group vs €45 106 (95% CI, €39 449-€51 545) in the standard surgery group. The incremental QALY for active surveillance was 0.11 (95% CI, −0.10 to 0.33) and mean costs were €8374 lower (95% CI, €1792-€15 355) compared with standard surgery. At a willingness-to-pay threshold of €80 000 per QALY, the mean iNMB was €17 568 (95% CI, −€725 to €37 497), indicating that active surveillance is cost-effective. Bootstrap analysis showed that 97% of replications fell in the cost-effective region.
Conclusions and Relevance
In this secondary analysis of a randomized clinical trial of patients with esophageal cancer achieving a CCR after nCRT, active surveillance was cost-effective over a 5-year horizon compared with standard surgery. Broader implementation of this strategy among appropriately selected patients would most likely reduce health care costs without compromising health outcomes.
Trial Registration
The Dutch Trial Register: NTR 6803
This secondary analysis of a randomized clinical trial assesses the cost-effectiveness of active surveillance compared with standard surgery among patients with esophageal cancer who achieved a complete clinical response after neoadjuvant chemoradiotherapy.
Introduction
In a recent randomized clinical trial,1 active surveillance for patients with esophageal cancer and a complete clinical response (CCR) after neoadjuvant chemoradiotherapy (nCRT) was noninferior to standard surgery for 2-year survival and improved short-term health-related quality of life (HRQOL). As health care costs continue to increase globally and the incidence of esophageal cancer is increasing, assessment of the economic effect of an active surveillance strategy is important.2,3
Patients undergoing active surveillance undergo repeated clinical response evaluations (CREs) after nCRT, including fluorodeoxyglucose (FDG)–positron emission tomography–computed tomography (PET-CT) scans, esophagogastroduodenoscopy with biopsies, and endoscopic ultrasonography (EUS) with fine needle aspiration of suspicious lymph nodes. These CREs are continued for up to 5 years, or until locoregional regrowth or distant dissemination is diagnosed. Patients with locoregional regrowth in the absence of distant metastases are offered (postponed) esophagectomy. The Surgery as Needed for Oesophageal Cancer (SANO) trial showed that at a minimum follow-up of 2 years, one-third of patients continued to undergo active surveillance, died without recurrence, or refused further treatment, while around 10% developed distant metastases within 6 months after nCRT.1 Hence, surgery could potentially be avoided in more than 40% of patients, either because of futility or because of sustained CCR. On the other hand, almost half of patients underwent surgery for cancer regrowth, associated with incremental costs depending on the number of CREs performed prior to surgery.1 In addition, if active surveillance were to result in more distant dissemination compared with standard surgery, potential higher costs related to palliative care should be considered. Standard surgery involves esophagectomy in all patients after nCRT. The costs of surgery are substantial and vary depending on the complication rate and length of hospital stay. Moreover, surgery negatively affects HRQOL and may reduce the ability to return to work.4 In the Netherlands, during follow-up, no routine diagnostic tests such as radiologic imaging are performed.5 The aim of this study was to assess the cost-effectiveness of active surveillance compared with standard surgery among patients with CCR after nCRT for locally advanced esophageal cancer.
Methods
Study Design and Participants
We performed a prespecified cost-effectiveness analysis alongside the SANO trial, a phase 3, multicenter, stepped-wedge cluster randomized clinical trial conducted in 12 hospitals in the Netherlands, enrolling patients between November 8, 2017, and January 17, 2021 (trial protocol in Supplement 1). Patients with locally advanced esophageal or esophagogastric junction cancer who underwent nCRT according to the CROSS (Chemoradiotherapy for Oesophageal Cancer Followed by Surgery Study) regimen were included.6 Patients with CCR after nCRT underwent either active surveillance or standard esophagectomy, according to cluster randomization, after providing written informed consent. The study protocol was approved by the medical ethics committee of Erasmus MC.7 The SANO trial is registered in the Dutch trial register (NTR 6803). Details on the trial design and outcomes have been published previously.1,7,8 This economic evaluation is reported in accordance with the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) guideline9 and Dutch guideline for economic evaluation.10
Effectiveness
Quality-adjusted life-years (QALYs) were used as the measure of effectiveness. HRQOL was assessed using the EuroQoL 5-dimension 5-level questionnaire (EQ-5D-5L).11 Questionnaires were sent out at the time of achieving CCR, which was defined at baseline (3 months after completion of nCRT and before potential esophagectomy), and then at 3, 6, 9, 13, 17, 21, 27, 33, 45, and 57 months after CCR, concurrently with CREs in the active surveillance group. Questionnaire responses were grouped into predefined time windows if not completed at the exact scheduled time points. For example, responses intended for the 6-month follow-up were included if completed between 5 and 7 months. The acceptable time windows increased with longer follow-up intervals. Data collection and entry were conducted by members of either the POCOP (Prospective Observational Cohort Study of Oesophageal-Gastric Cancer Patients) project or the SANO study team.12 Quality of life weights were derived using the Dutch value set for the EQ-5D-5L.13 These weights were used to calculate QALYs over the 5-year period, with extrapolation, using the area under the curve method. Beyond the first year of follow-up, QALYs were discounted at an annual rate of 1.5%, in line with Dutch guidelines for economic evaluations, to account for the lower present value of future health benefits.10 Patients with CCR from the preSANO trial were included in the SANO trial, as described in the previously published protocol.7,8 These patients did not complete HRQOL questionnaires and their values were derived from estimated outcomes in these analyses.1,8,14
Costs
The cost analysis was conducted primarily from a health care perspective. This included all in-hospital costs related to esophageal cancer from the moment of achieving CCR. Costs were subdivided into 5 categories: (1) esophagectomy (including in-hospital stay and pathologic examination of the resection specimen), (2) diagnostic tests, (3) radiotherapy (some of which was given with curative intent) and palliative care (including stent placement, systemic antitumor therapy, day admissions) (4) outpatient clinic visits, and (5) (re)admissions and (re)interventions. Costs for hospital admissions for complications of palliative care were included in admissions. Costs incurred prior to CCR, including those related to nCRT and the first 2 CREs, were excluded, as these are identical between groups. Data were collected prospectively from electronic patient records. For patients referred to other hospitals, data were requested from those hospitals to minimize missing information. Total costs were calculated by multiplying resource use by the corresponding unit costs. Reference prices were obtained from the Dutch guideline for economic evaluation; if unavailable, benchmark Dutch Federation of University Medical Centers prices were used, and if these were also unavailable, internal cost calculations from the Erasmus University Medical Center were applied, all representing hospital-based cost estimates rather than reimbursement tariffs10,15 (eTable 1 in Supplement 2). Costs incurred after the first year of follow-up were discounted at an annual rate of 3%.16 All reference prices were inflation-adjusted to 2024 values using the Dutch consumer price index obtained from StatLine (Statistics Netherlands).17 Costs are given in Euros (currency exchange rate of €1 = US $1.16 as of June 11, 2026).
A societal perspective was adopted, including costs from productivity losses, to quantify the economic effect of lost productivity. These were measured using the Work Productivity and Activity Impairment questionnaire, distributed at the same time points as the EQ-5D-5L. Due to low response rates, these results were not included in the analysis (eAppendix 2 and eTables 2 and 3 in Supplement 2).
Statistical Analysis
Data were analyzed on June 1, 2025. All analyses followed the modified intention-to-treat principle, allowing one moment of crossover at time of reaching CCR, as described in the predefined statistical analysis plan.18 Mean differences in QALYs and costs between the active surveillance group and the standard surgery group were calculated at 5 years from CCR. Following the Dutch guideline for economic evaluation, a cost-effectiveness threshold of €80 000 per QALY was applied, and additional analyses were performed at €20 000 and €50 000 per QALY, as these thresholds are also commonly used in the Netherlands.10 By June 1, 2025, all patients had complete 4-year follow-up and 5-year follow-up was complete for 71% of patients (140 of 198) receiving active surveillance and 96% of patients (107 of 111) receiving standard surgery. Analyses were based on available 5-year EQ-5D-5L and cost data (eAppendix 2 in Supplement 2).
We used linear mixed models (LMMs) to adjust for potential differences in baseline covariates and to account for missing HRQOL data, within a nonparametric bootstrap framework (2000 replications), stratified by treatment group.19 In each bootstrap sample we fitted an LMM including fixed effects for age, sex, histologic subtype, histologic grading, clinical T category, clinical N category, World Health Organization (WHO) performance score, hospital of inclusion, and time (modeled as a factor), with a random effect for each individual patient. Baseline HRQOL scores were centered to ensure equality between groups before fitting the model.20 Missing values for WHO performance status, which were under 10%, were handled using single imputation via the MICE algorithm in R, version 4.3.2 (R Project for Statistical Computing), prior to the bootstrap procedure.21 We generated quality-of-life weights, conditional on being alive, from the LMM model. Quality-of-life weights were then set to zero at time points after death.
Cost data for years 1 to 4 were modeled within the same bootstrap samples using LMMs with similar fixed effects, but including hospital of inclusion as a random intercept to account for clustering by treatment center. Costs in year 5 were modeled within the same bootstrap using a linear regression with the same fixed effects, excluding hospital of inclusion, as costs were very low or absent for some hospitals in certain bootstrap samples. Estimations for costs were generated from these models. The total costs for years 1 to 4 and for year 5 were calculated for each bootstrap replication and then summed to obtain overall estimates.
Within each bootstrap sample, incremental costs and QALYs were estimated, and the incremental net monetary benefit (iNMB) was derived by multiplying the incremental QALYs by the willingness-to-pay threshold and subtracting the incremental costs. This approach combines costs and QALYs into a single value for a given willingness-to-pay threshold. Using the percentile method, 95% CIs were calculated. Cost-effectiveness planes and cost-effectiveness acceptability curves were generated to illustrate the proportion of bootstrap replications indicating cost-effectiveness across thresholds ranging from €0 to €120 000 per QALY. The iNMB and its confidence intervals were interpreted as measures of decision uncertainty rather than formal probabilities or hypothesis tests.
We performed a sensitivity analysis to account for the costs of the first 2 CREs across all patients, with these costs distributed to the patients who ultimately achieved a CCR, as these procedures are not part of standard practice. Furthermore, a sensitivity analysis was performed based on previous research indicating that if EUS were selectively performed only in cases with suspected lymph nodes on PET-CT scan, 97.8% of EUS procedures could be omitted without jeopardizing the accuracy of CREs.22 Accordingly, we excluded 97.8% of EUS procedures in the active surveillance group and recalculated the costs. In addition, we performed a sensitivity analysis using Dutch Pharmacotherapeutic Compass23 prices for chemotherapy and immunotherapy, as hospital-specific prices could not be published. These prices are higher than the actual costs for Dutch hospitals, but the analysis was performed and added to eAppendix 3 and eTable 4 in Supplement 2 for transparency. Last, a simplified budget effect analysis was included in eAppendix 4 in Supplement 2. All statistical analyses were conducted in R, version 4.3.2 (R Project for Statistical Computing).
Results
The modified intention-to-treat population comprised 309 patients included in the cost-effectiveness analysis: 198 in the active surveillance group (median age, 69 years [IQR, 63-74 years]; 156 men [79%] and 42 women [21%]) and 111 in the standard surgery group (median age, 68 years [IQR, 61-73 years]; 86 men [77%] and 25 women [23%]) (Figure 1; eTable 5 in Supplement 2). Five-year follow-up was complete for 71% of patients in the active surveillance group (140 of 198) and 96% of patients in the standard surgery group (107 of 111), with complete health care cost data except for 1 patient in the active surveillance group who emigrated.
Figure 1. Patient Flowchart.
The bottom 2 boxes comprise the modified intention-to-treat population. CCR indicates complete clinical response; CRE, clinical response evaluation; CROSS, Chemoradiotherapy for Oesophageal Cancer Followed by Surgery Study; CT, computed tomography; EGD, esophagogastroduodenoscopy; EUS, endoscopic ultrasonography; FDG, fluorodeoxyglucose; nCRT, neoadjuvant chemoradiotherapy; PET, positron emission tomography; and SANO, Surgery as Needed for Oesophageal Cancer trial.
A total of 175 patients (88%) in the active surveillance group and 72 patients (65%) in the standard surgery group completed at least 1 questionnaire; these patients were included in the QALY modeling to generate estimated outcomes for all patients. The median (IQR) number of responses per patient was 9 (6-11) in the active surveillance group and 8 (4-10) in the standard surgery group. Response rates ranged from 42% (47 of 111) to 87% (94 of 108) and were consistently lower in the standard surgery group than in the active surveillance group across all measurement points (eTable 6 in Supplement 2). Characteristics of respondents did not differ significantly between groups at any of the measured time points with respect to cT category, cN category, tumor location, histologic characteristics, tumor differentiation grade, WHO performance status, age, or sex. Similarly, no significant differences were observed between patients who completed questionnaires and those who did not.
Effectiveness Outcome
The active surveillance group had an estimated mean of 2.99 QALYs (95% CI, 2.73-3.26) per patient, compared with 2.88 QALYs (95% CI, 2.69-3.06) in the standard surgery group at 5-year follow-up. The incremental QALY for active surveillance was 0.11 QALYs (95% CI, −0.10 to 0.33) (Table 1).
Table 1. Mean Quality-of-Life Scores and Estimates Per Time Point.
| Time, moa | Score, mean (SD) | |||
|---|---|---|---|---|
| Standard surgery arm | Active surveillance arm | Standard surgery arm, bootstrap estimateb | Active surveillance arm, bootstrap estimate b | |
| Baseline | 0.87 (0.13) | 0.89 (0.11) | 0.89 (0.02) | 0.89 (0.02) |
| 3 | 0.81 (0.12) | 0.88 (0.13) | 0.79 (0.02) | 0.86 (0.02) |
| 6 | 0.83 (0.18) | 0.87 (0.13) | 0.79 (0.02) | 0.83 (0.02) |
| 9 | 0.87 (0.12) | 0.87 (0.13) | 0.79 (0.02) | 0.80 (0.02) |
| 13 | 0.88 (0.12) | 0.86 (0.13) | 0.76 (0.02) | 0.76 (0.02) |
| 17 | 0.85 (0.18) | 0.85 (0.15) | 0.69 (0.03) | 0.71 (0.03) |
| 21 | 0.84 (0.18) | 0.85 (0.17) | 0.64 (0.03) | 0.66 (0.03) |
| 27 | 0.87 (0.11) | 0.84 (0.16) | 0.59 (0.03) | 0.60 (0.03) |
| 33 | 0.85 (0.15) | 0.85 (0.17) | 0.54 (0.03) | 0.55 (0.03) |
| 45 | 0.83 (0.20) | 0.86 (0.14) | 0.46 (0.03) | 0.48 (0.03) |
| 57 | 0.81 (0.19) | 0.87 (0.16) | 0.42 (0.03) | 0.46 (0.03) |
Time points are in months after achieving complete clinical response (baseline).
Estimates from bootstrap are corrected for baseline differences and death. Within the bootstrap, scores were extrapolated to 5 years.
Cost Outcomes
The estimated mean costs per patient from a health care perspective were €36 733 (95% CI, €33 530-€40 009) in the active surveillance group and €45 106 (95% CI, €39 449-€51 545) in the standard surgery group (Table 2). This corresponds to mean costs that were €8374 lower (95% CI, €1792-€15 355) in the active surveillance group at 5-year follow-up. The cost difference was primarily driven by lower mean costs for surgery, which averaged €13 733 (95% CI, €11 142-€16 693) in the active surveillance group vs €30 583 (95% CI, €26 376-€35 777) in the standard surgery group. In addition, the mean costs of reinterventions and readmissions were higher in the standard surgery group than the active surveillance group (€7555 [95% CI, €5275-€10 241] vs €3605 [95% CI, €2768-€4524]). In contrast, the mean costs of diagnostic tests were higher in the active surveillance group than the standard surgery group (€14 295 [95% CI, €12 973-€15 598] vs €3527 [95% CI, €2690-€4419]). Similarly, radiotherapy and palliative care costs were higher in the active surveillance group than the standard surgery group (€3062 [95% CI, €2038-€4275] vs €1451 [95% CI, €557-€2608]) (Table 2; eTable 7 in Supplement 2).
Table 2. Health Care Costs Estimates Per Patient Per Group and Subtype Derived With Bootstrap.
| Cost derived with bootstrap | Cost, € (95% CI)a | ||
|---|---|---|---|
| Standard surgery arm | Active surveillance arm | Incremental difference | |
| Total | 45 106 (39 449 to 51 545) | 36 733 (33 530 to 40 009) | −8374 (−15 355 to −1792) |
| Esophagectomy | 30 583 (26 376 to 35 777) | 13 733 (11 142 to 16 693) | −16 850 (−22 849 to −11 652) |
| Diagnostic tests | 3527 (2690 to 4419) | 14 295 (12 973 to 15 598) | 10 768 (9170 to 12 427) |
| Radiotherapy and palliative care | 1451 (557 to 2608) | 3062 (2038 to 4275) | 1611 (98 to 3193) |
| Outpatient clinic visits | 2081 (1824 to 2342) | 2163 (1944 to 2386) | 82 (−252 to 423) |
| Readmissions and reinterventionsb | 7556 (5275 to 10 241) | 3605 (2768 to 4524) | −3951 (−6728 to −1599) |
Costs are given in Euros (currency exchange rate of €1 = US $1.16 as of June 11, 2026).
Including all admissions during follow-up.
Cost-Effectiveness
At 5 years, an average patient in the active surveillance group had 0.11 additional QALYs (95% CI, −0.10 to 0.33 QALYs) and simultaneously saved €8374 (95% CI, €1792-€15 355) in costs compared with standard surgery. At a willingness-to-pay threshold of €80 000 per QALY, the mean iNMB was €17 568 (95% CI, −€725 to €37 497), indicating that active surveillance is cost effective at this threshold. Bootstrap analysis showed that 97% of replications fell in the cost-effective region, and this proportion remained consistently high across thresholds from €0 to €120 000 per QALY, with reduced uncertainty at lower values (Figure 2 and Figure 3).
Figure 2. Line Graph Showing Cost-Effectiveness Plane With €80 000 Per Quality-Adjusted Life-Year (QALY) Threshold.
The standard surgery group was coded as 0. A total of 97% of bootstrap samples favor active surveillance. Q indicates quarter. Costs are given in Euros (currency exchange rate of €1 = US $1.16 as of June 11, 2026).
Figure 3. Line Graph Showing Cost-Effectiveness Acceptability Curve Across Increasing Willingness-to-Pay Thresholds.
The inset graph is an enlargement of the section where there is a decrease. Costs are given in Euros (currency exchange rate of €1 = US $1.16 as of June 11, 2026).
Sensitivity Analyses
In sensitivity analyses, allocating the additional cost required to identify 1 patient with a CCR (€5843; eFigures 1 and 2 in Supplement 2) to all active surveillance patients increased mean costs in the active surveillance group to €42 576 (95% CI, €39 373-€45 852), reducing the cost difference vs surgery to €2531 (95% CI, −€4051 to €9512). The iNMB at a threshold of €80 000 per QALY was €11 725 (95% CI, −€6568 to €31 654). In a second sensitivity analysis, excluding 97.8% of EUS procedures reduced mean costs in the active surveillance group to €32 413 (95% CI, €29 208-€35 739) and increased the cost difference vs surgery to €12 692 (95% CI, €5995-€19 590) with a mean iNMB of €21 887k (95% CI, €3537-€41 882) at the same threshold. When combining these sensitivity analyses for a future clinical simulation, mean costs in the active surveillance group were €38 256 (95% CI, €35 050-€41 582), with a cost difference vs surgery of €6850 (95% CI, €152-€13 747). The iNMB was €16 044 (95% CI, −€2306 to €36 039) at a €80 000 per QALY threshold. All sensitivity analyses indicated cost effectiveness of active surveillance, and across lower willingness-to-pay thresholds uncertainty decreased (eTable 8 in Supplement 2).
Discussion
We found that for patients with esophageal cancer achieving CCR after nCRT, active surveillance is cost effective compared with standard surgery from a health care perspective, offering similar QALYs at lower costs over 5 years of follow-up. The relatively modest differences in QALYs between groups were smaller than initially expected based on the known effect of esophagectomy on quality of life, likely reflecting substantial recovery over time and the proportion of patients in the active surveillance group ultimately undergoing surgery during follow-up. Across willingness-to-pay thresholds, the iNMB consistently favored active surveillance. Although uncertainty increased with additional costs for determination of CCR, the results remained consistent across scenarios, supporting its cost effectiveness vs standard surgery.
This evaluation of the cost effectiveness of active surveillance for esophageal cancer was directly on source data from a randomized clinical trial. Watch-and-wait strategies in rectal cancer have consistently shown cost effectiveness, but those analyses were largely simulation based.24,25 Moreover, sustained CCR is less frequent in esophageal cancer, with only one-fourth of SANO patients maintaining CCR at 3 years, compared with roughly two-thirds in rectal cancer.26,27 This lower rate of sustained CCR implies higher costs from CREs and delayed surgery in a substantial proportion of patients.
In the SANO trial, almost two-thirds of patients discontinued active surveillance within the first year. Most of these patients had locoregional regrowth necessitating surgery, while a smaller proportion developed distant metastases. We consider that all patients with disease progression within 6 months after nCRT (at CRE-3) were spared an unbeneficial esophagectomy with associated costs. The observed cost difference was primarily driven by lower surgical costs. This is reflected in the distribution of costs over time, with high costs during the first year. As expected, costs of diagnostic tests were higher among the active surveillance group; however, these costs did not outweigh the higher surgical costs in the standard surgery group. Similarly, palliative care costs were higher in the active surveillance group, likely due to the higher percentage of patients with distant metastases, but these also did not outweigh the savings from fewer esophagectomies.
With recent insights, costs in the active surveillance protocol can be further reduced to improve cost effectiveness. EUS with fine needle aspiration provides minimal added value during active surveillance, with only 0.1% of patients with a negative PET-CT scan having tumor-positive lymph nodes.22 Sensitivity analysis indicates that performing EUS only when FDG-avid lymph nodes are detected on PET-CT would save an additional mean of €4300 per patient. Furthermore, as time progresses, the a priori probability of detecting residual disease decreases. Standard diagnostic tests beyond 3 years after nCRT could likely be safely omitted if no recurrences are detected after this period in long-term follow-up. According to Dutch guidelines, routine follow-up scans are not recommended, whereas clinicians in other countries may perform more intensive imaging, which would increase the costs of standard surgery and, to a lesser extent (for roughly half of patients), the costs of active surveillance, further enhancing the cost advantage of active surveillance.5
The CheckMate 577 trial showed a disease-free survival benefit with adjuvant nivolumab in esophageal cancer, but no overall survival benefit.28 Median treatment duration was 10 months, although the number of cycles was not reported. In the Netherlands Cancer Registry, a subgroup of patients receiving adjuvant nivolumab had a minimum follow-up of 3 years and a median of 12 cycles (IQR, 6-13 cycles).29 Estimated costs were €58 320 per patient, excluding day admissions, blood tests, outpatient consultations, or toxicity-related admissions. Incorporating nivolumab into the treatment strategy according to European Society for Medical Oncology guidelines could further increase cost savings for active surveillance, as potentially fewer patients would require surgery followed by adjuvant therapy. However, results of the SANO-3 trial are needed to determine whether nivolumab adds benefit during active surveillance to allow a fair comparison.30
Limitations
This study has some limitations. The 5-year follow-up of the SANO trial was not yet fully complete; however, missing data were well within the acceptable range for modeling, and more than 70% of patients had complete 5-year cost and HRQOL data. In addition, the standard surgery group does not fully reflect clinical practice, as 2 CREs were necessary for selection, which may reduce the representativeness of the cost-effectiveness analysis; however, sensitivity analyses and the budget effect estimate corrected for this, and the conclusions remained unchanged. Societal cost data were limited because outpatient medical consumption was not captured and response rates for productivity questionnaires were low. Furthermore, the active surveillance and treatment protocols were implemented in specialized high-volume centers with extensive endoscopic and surgical expertise, which may limit generalizability to settings with different resources or care structures. Informed consent and privacy regulations prevented collection of data unrelated to esophagectomy. In-hospital costs related to esophageal cancer were collected in detail, and relevant data were requested when patients were referred to other hospitals. One could argue that all in-hospital costs should be included, as longer survival may increase costs from comorbidities. Capturing these costs would risk missing region-specific data, as patients receive care across multiple facilities.
Conclusions
In this secondary analysis of a randomized clinical trial of patients with esophageal cancer achieving a CCR after nCRT, active surveillance was cost-effective over a 5-year horizon compared with standard surgery. Given noninferior survival, active surveillance could be considered as a treatment option in shared decision making.1 Broader implementation of this strategy in appropriately selected patients would most likely reduce health care costs without compromising health outcomes.
Trial Protocol and Statistical Analysis Plan
eFigure 1. Flowchart of Costs During Clinical Response Evaluations (CRE) 1 and CRE 2 (Excluding Costs of PET-CT and Outpatient Clinic Visit During CRE 2) Divided by Patients With CCR to Determine Distributed Costs for One Patient With Complete Clinical Response (CCR)
eFigure 2. Budget Impact Estimate of Patients Eligible for Active Surveillance in 2024 Based on the Netherlands Cancer Registry and Potential Saved Costs if Active Surveillance Was Started
eAppendix 1. Productivity Loss
eAppendix 2. Supplementary Statistical Methods
eAppendix 3. Sensitivity Analyses: Systemic Therapy Prices
eAppendix 4. Simplified Budget Impact Analysis
eTable 1. Detailed Cost Prices per Group
eTable 2. Working Status per Group at Baseline and Productivity Loss During Follow-Up
eTable 3. Response Rates for Productivity Questionnaires
eTable 4. Measurements for Calculation of Chemotherapy and Immunotherapy
eTable 5. Baseline Characteristics of Patients Included in the SANO-Trial
eTable 6. Response Rates for EQ5D-5L Questionnaires
eTable 7. Healthcare Costs Per Group, Subtype and Year Unadjusted
eTable 8. Incremental Net Monetary Benefit of Active Surveillance Versus Surgery Across Willingness-to-Pay Thresholds and Sensitivity Analyses
eTable 9. Members of the SANO Study Group, Besides the Authors Who Are Indicated on the Title Page
eReferences.
Nonathor Collaborators. The Surgery as Needed for Oesophageal Cancer (SANO) Study Group Members
Data Sharing Statement
References
- 1.van der Wilk BJ, Eyck BM, Wijnhoven BPL, et al. ; SANO Study Group . Neoadjuvant chemoradiotherapy followed by active surveillance versus standard surgery for oesophageal cancer (SANO trial): a multicentre, stepped-wedge, cluster-randomised, non-inferiority, phase 3 trial. Lancet Oncol. 2025;26(4):425-436. doi: 10.1016/S1470-2045(25)00027-0 [DOI] [PubMed] [Google Scholar]
- 2.Muka T, Imo D, Jaspers L, et al. The global impact of non-communicable diseases on healthcare spending and national income: a systematic review. Eur J Epidemiol. 2015;30(4):251-277. doi: 10.1007/s10654-014-9984-2 [DOI] [PubMed] [Google Scholar]
- 3.Arnold M, Laversanne M, Brown LM, Devesa SS, Bray F. Predicting the future burden of esophageal cancer by histological subtype: international trends in incidence up to 2030. Am J Gastroenterol. 2017;112(8):1247-1255. doi: 10.1038/ajg.2017.155 [DOI] [PubMed] [Google Scholar]
- 4.Markar SR, Zaninotto G, Castoro C, et al. Lasting Symptoms After Esophageal Resection (LASER): European multicenter cross-sectional study. Ann Surg. 2022;275(2):e392-e400. doi: 10.1097/SLA.0000000000003917 [DOI] [PubMed] [Google Scholar]
- 5.Oesofaguscarcinoom. Federatie Medisch Specialisten. Richtlijnen. 2024. Accessed June 1, 2025. https://richtlijnendatabase.nl/richtlijn/oesofaguscarcinoom/startpagina_-_oesofaguscarcinoom.html
- 6.van Hagen P, Hulshof MCCM, van Lanschot JJ, et al. ; CROSS Group . Preoperative chemoradiotherapy for esophageal or junctional cancer. N Engl J Med. 2012;366(22):2074-2084. doi: 10.1056/NEJMoa1112088 [DOI] [PubMed] [Google Scholar]
- 7.Noordman BJ, Wijnhoven BPL, Lagarde SM, et al. ; SANO-study group . Neoadjuvant chemoradiotherapy plus surgery versus active surveillance for oesophageal cancer: a stepped-wedge cluster randomised trial. BMC Cancer. 2018;18(1):142. doi: 10.1186/s12885-018-4034-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Eyck BM, van der Wilk BJ, Noordman BJ, et al. ; SANO-study group . Updated protocol of the SANO trial: a stepped-wedge cluster randomised trial comparing surgery with active surveillance after neoadjuvant chemoradiotherapy for oesophageal cancer. Trials. 2021;22(1):345. doi: 10.1186/s13063-021-05274-w [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Husereau D, Drummond M, Augustovski F, et al. ; CHEERS 2022 ISPOR Good Research Practices Task Force . Consolidated Health Economic Evaluation Reporting Standards 2022 (CHEERS 2022) statement: updated reporting guidance for health economic evaluations. Value Health. 2022;25(1):3-9. doi: 10.1016/j.jval.2021.11.1351 [DOI] [PubMed] [Google Scholar]
- 10.Zorginstituut Nederland. Guideline for Economic Evaluations in Healthcare. Zorginstituut Nederland; 2024. [Google Scholar]
- 11.Herdman M, Gudex C, Lloyd A, et al. Development and preliminary testing of the new five-level version of EQ-5D (EQ-5D-5L). Qual Life Res. 2011;20(10):1727-1736. doi: 10.1007/s11136-011-9903-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Coebergh van den Braak RRJ, van Rijssen LB, van Kleef JJ, et al. ; Dutch Pancreatic Cancer Group, Dutch Upper GI Cancer Group and PLCRC Working Group. . Nationwide comprehensive gastro-intestinal cancer cohorts: the 3P Initiative. Acta Oncol. 2018;57(2):195-202. doi: 10.1080/0284186X.2017.1346381 [DOI] [PubMed] [Google Scholar]
- 13.Versteegh MM, Vermeulen KM, Evers SMAA, de Wit GA, Prenger R, Stolk EA. Dutch tariff for the five-level version of EQ-5D. Value Health. 2016;19(4):343-352. doi: 10.1016/j.jval.2016.01.003 [DOI] [PubMed] [Google Scholar]
- 14.Noordman BJ, Spaander MCW, Valkema R, et al. ; SANO study group . Detection of residual disease after neoadjuvant chemoradiotherapy for oesophageal cancer (preSANO): a prospective multicentre, diagnostic cohort study. Lancet Oncol. 2018;19(7):965-974. doi: 10.1016/S1470-2045(18)30201-8 [DOI] [PubMed] [Google Scholar]
- 15.NFU verlengt samenwerking met Performation voor landelijke kostprijzenbenchmark UMC’s. Performation. 2025. Accessed June 1, 2025. https://performation.com/blog/nfu-verlengt-samenwerking-met-performation-voor-landelijke-kostprijzenbenchmark-umcs/
- 16.Neumann PJ, Ganiats TG, Russell LB, Sanders GD, Siegel JE, eds. Cost-Effectiveness in Health and Medicine. Oxford University Press; 2016. doi: 10.1093/acprof:oso/9780190492939.001.0001 [DOI] [Google Scholar]
- 17.CBS StatLine - Consumentenprijzen; prijsindex. Centraal Bureau voor de Statistiek (CBS). 2024. Accessed June 1, 2025. https://opendata.cbs.nl/statline/#/CBS/nl/
- 18.Statistical analysis plan for the SANO-trial: neoadjuvant chemoradiotherapy followed by surgery versus active surveillance for oesophageal cancer (SANO). ClinicalTrials.gov. Updated July 11, 2023. Accessed January 1, 2025. https://clinicaltrials.gov/study/NCT05953181
- 19.Gabrio A, Plumpton C, Banerjee S, Leurent B. Linear mixed models to handle missing at random data in trial-based economic evaluations. Health Econ. 2022;31(6):1276-1287. doi: 10.1002/hec.4510 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Manca A, Hawkins N, Sculpher MJ. Estimating mean QALYs in trial-based cost-effectiveness analysis: the importance of controlling for baseline utility. Health Econ. 2005;14(5):487-496. doi: 10.1002/hec.944 [DOI] [PubMed] [Google Scholar]
- 21.van Buuren S, Groothuis-Oudshoorn K. mice: Multivariate imputation by chained equations in R. J Stat Softw. 2011;45(3):1-67. doi: 10.18637/jss.v045.i03 [DOI] [Google Scholar]
- 22.Gangaram Panday SSG, Pittacolo M, Lagarde S, et al. ; SANO Study Group. The value of routine endoscopic ultrasound in patients with esophageal cancer after neoadjuvant chemoradiotherapy undergoing active surveillance. Endoscopy. 2026;58(6):577-587. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Farmacotherapeutisch Kompas. Zorginstituut Nederland . Accessed June 1, 2025. https://www.farmacotherapeutischkompas.nl/
- 24.Cui CL, Luo WY, Cosman BC, et al. Cost effectiveness of watch and wait versus resection in rectal cancer patients with complete clinical response to neoadjuvant chemoradiation. Ann Surg Oncol. 2022;29(3):1894-1907. doi: 10.1245/s10434-021-10576-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Miller JA, Wang H, Chang DT, Pollom EL. Cost-effectiveness and quality-adjusted survival of watch and wait after complete response to chemoradiotherapy for rectal cancer. J Natl Cancer Inst. 2020;112(8):792-801. doi: 10.1093/jnci/djaa003 [DOI] [PubMed] [Google Scholar]
- 26.van der Valk MJM, Hilling DE, Bastiaannet E, et al. ; IWWD Consortium . Long-term outcomes of clinical complete responders after neoadjuvant treatment for rectal cancer in the International Watch & Wait Database (IWWD): an international multicentre registry study. Lancet. 2018;391(10139):2537-2545. doi: 10.1016/S0140-6736(18)31078-X [DOI] [PubMed] [Google Scholar]
- 27.Gangaram Panday SSG, van Klaveren D, Lagarde SM, et al. ; SANO Study Group. Accuracy of predicting residual disease and disease progression during active surveillance for esophageal cancer. Ann Surg Oncol. 2026;33(2):946-954. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Kelly RJ, Ajani JA, Kuzdzal J, et al. ; CheckMate 577 Investigators . Adjuvant nivolumab in resected esophageal or gastroesophageal junction cancer. N Engl J Med. 2021;384(13):1191-1203. doi: 10.1056/NEJMoa2032125 [DOI] [PubMed] [Google Scholar]
- 29.Netherlands Cancer Registry (NCR). Netherlands Comprehensive Cancer Organisation (IKNL). 2025. Accessed June 1, 2025. https://www.iknl.nl/en/ncr/ncr-data-figures
- 30.Obermannová R, Alsina M, Cervantes A, et al. ; ESMO Guidelines Committee . Oesophageal cancer: ESMO Clinical Practice Guideline for diagnosis, treatment and follow-up. Ann Oncol. 2022;33(10):992-1004. doi: 10.1016/j.annonc.2022.07.003 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Trial Protocol and Statistical Analysis Plan
eFigure 1. Flowchart of Costs During Clinical Response Evaluations (CRE) 1 and CRE 2 (Excluding Costs of PET-CT and Outpatient Clinic Visit During CRE 2) Divided by Patients With CCR to Determine Distributed Costs for One Patient With Complete Clinical Response (CCR)
eFigure 2. Budget Impact Estimate of Patients Eligible for Active Surveillance in 2024 Based on the Netherlands Cancer Registry and Potential Saved Costs if Active Surveillance Was Started
eAppendix 1. Productivity Loss
eAppendix 2. Supplementary Statistical Methods
eAppendix 3. Sensitivity Analyses: Systemic Therapy Prices
eAppendix 4. Simplified Budget Impact Analysis
eTable 1. Detailed Cost Prices per Group
eTable 2. Working Status per Group at Baseline and Productivity Loss During Follow-Up
eTable 3. Response Rates for Productivity Questionnaires
eTable 4. Measurements for Calculation of Chemotherapy and Immunotherapy
eTable 5. Baseline Characteristics of Patients Included in the SANO-Trial
eTable 6. Response Rates for EQ5D-5L Questionnaires
eTable 7. Healthcare Costs Per Group, Subtype and Year Unadjusted
eTable 8. Incremental Net Monetary Benefit of Active Surveillance Versus Surgery Across Willingness-to-Pay Thresholds and Sensitivity Analyses
eTable 9. Members of the SANO Study Group, Besides the Authors Who Are Indicated on the Title Page
eReferences.
Nonathor Collaborators. The Surgery as Needed for Oesophageal Cancer (SANO) Study Group Members
Data Sharing Statement
