Routine CT During Adjuvant Nivolumab Caught Most Recurrences in Resected Esophageal Cancer
Adjuvant nivolumab has changed the postoperative conversation in esophageal and gastroesophageal junction cancer, but it has not eliminated the problem that drives follow-up care: recurrence still emerges in a substantial minority of patients after surgery and neoadjuvant chemoradiotherapy. In this multicenter retrospective cohort from three hospitals in the Netherlands, investigators examined whether scheduled imaging during treatment was detecting relapse often enough to justify the effort and whether the timing of scans should differ from routine postoperative practice.
The answer was not subtle. Recurrences surfaced during treatment, and most were found on planned CT rather than because symptoms forced a scan.
What the Dutch Cohort Showed Across the First Year
The cohort began with 353 patients who underwent resection between September 2021 and October 2023; 151 received at least one dose of adjuvant nivolumab. The treated group was predominantly male, with 82% men, a median age of 66 years, and 86% adenocarcinoma. Those baseline features matter because they reflect the real-world postoperative population now being considered for immune checkpoint therapy after trimodality treatment, not a narrowly selected trial subgroup.
Among the 151 patients treated with nivolumab, 27 developed recurrent disease on treatment, which translates to 18% of the cohort. Distant spread dominated the relapse pattern: 89% of recurrences included distant metastasis. The biology is familiar to thoracic oncologists. Once systemic escape occurs after local therapy, recurrence often travels beyond the surgical bed, and imaging becomes the main practical tool for recognizing it before the disease presents with pain, weight loss, neurologic symptoms, or organ dysfunction.
Routine CT was responsible for finding 67% of these recurrences. In plain terms, scheduled imaging did most of the work.
Baseline Imaging Added Little, but Surveillance CT Had Clear Yield
Baseline CT did not identify recurrences in this study. Diagnostic yield started at 0% at baseline and rose to 9% by 4 months. That pattern suggests that the immediate postoperative scan, while reassuring in principle, contributed little to recurrence detection in this cohort when used as a routine baseline step after surgery and before or around the start of nivolumab.
The more informative scans were the ones performed during treatment. The study compared follow-up intervals of every 3 months versus every 4 months, reflecting how surveillance is often handled in practice when clinicians are balancing patient burden, resource use, and the risk of missing early metastatic spread. The recurrence distribution did not cluster into a single abrupt wave; instead, it unfolded over time. Disease-free survival was 89% at 4 months, 84% at 8 months, and 75% at 12 months, showing a gradual decline rather than a sharp early cliff.
That tempo argues for continued surveillance rather than a single early scan followed by long gaps. It also explains why the authors focused on interval imaging rather than trying to justify one isolated postoperative checkpoint.
Why Recurrence Detection During Immunotherapy Matters Clinically
For patients receiving adjuvant nivolumab, imaging does more than document relapse. It determines whether therapy should continue, whether a patient should be redirected to systemic salvage treatment, and whether metastatic disease can be recognized while performance status is still preserved. In esophageal cancer, delayed recognition of distant recurrence can mean the difference between a patient who remains eligible for second-line therapy and one who presents with symptomatic, more advanced disease.
The predominance of distant metastasis in this cohort reinforces that point. Local or locoregional recurrence may sometimes be approached with procedural or radiation strategies, but disseminated disease changes the entire treatment path. Scheduled CT, then, functions as a gatekeeper for treatment adaptation during the narrow postoperative window when relapse may still be clinically silent.
The study does not claim that one surveillance interval is superior to the other. It does, however, show that both 3-month and 4-month imaging rhythms were sufficient to capture recurrent disease in a real-world treatment population, and that recurrence was not front-loaded enough to make baseline imaging alone a useful substitute for continued surveillance.
A Practical Signal for Postoperative Follow-up
For clinicians managing resected esophageal or gastroesophageal junction cancer during adjuvant nivolumab, the message is straightforward: routine imaging during treatment appears to have real diagnostic value, while baseline CT contributed little to recurrence detection in this cohort. The gradual fall in disease-free survival over 12 months supports a surveillance model that keeps imaging in rotation rather than treating the start of immunotherapy as a single checkpoint.
This is not a randomized test of imaging frequency, and it does not settle whether every 3 months is better than every 4 months. It does show that recurrence detection was materially dependent on scheduled CT, with most relapses identified before they became clinically obvious. In a disease where distant metastasis was the dominant pattern of failure, that finding has immediate bedside relevance.
The study leaves one central question unresolved: which imaging interval achieves the best balance between early detection, patient burden, and resource use. For now, the data support routine surveillance during adjuvant nivolumab rather than a minimal or symptom-triggered approach alone.