SANO
ForLocally advanced oesophageal cancer, cCR after nCRT
TL;DR5y OS 50% vs 54%, HR 1.17 (0.77-1.77) active surveillance vs surgery after cCR to nCRT; distant mets 46% vs 29%, P<.01.
The RT read is that nCRT alone was the definitive treatment for 43% of the surveillance arm, but distant mets rose to 46% vs 29% without immediate surgery. That makes cCR after nCRT a reasonable organ-preservation endpoint for selected pts. Whether early oesophagectomy suppresses distant spread, or the gap comes from how surveillance detects disease, is the open question.
In locally advanced oesophageal cancer with cCR on biopsies, EUS and PET-CT after nCRT, this supports active surveillance with surgery as needed as a discussable option; it does not address pts without cCR or nCRT regimens other than those used in SANO.
nCRT alone was definitive treatment for a large part of the surveillance arm, which makes cCR on biopsies, EUS and PET-CT the decision gate. The nCRT regimen is not stated in source, so dose transfer cannot be judged, and the distant-mets excess (46% vs 29%) matters when counselling responders.
Surgery as needed spared 85/198 (43%) a non-beneficial oesophagectomy with no significant 5y OS difference (HR 1.17). The distant-mets excess under surveillance (P<.01) raises the question of whether delaying resection gives up systemic control, and that 43% includes pts who progressed within 6 mo.
9 details 1 trial watching
Multicentre stepped-wedge cluster-randomised non-inferiority trial, 2017-2021, Netherlands Trial Register NTR-6803. This report covers 5y outcomes at median f/u 76 mo (IQR 67-86).
Locally advanced oesophageal cancer with a complete clinical response after nCRT, defined as no tumour on endoscopic biopsies, endoscopic ultrasound, or PET-CT. N=309: 198 active surveillance, 111 standard surgery.
nCRT was the common backbone. The regimen, dose and fractionation are not stated in source, so whether the result transfers to a given chemoRT protocol cannot be judged from this abstract.
The earlier report showed non-inferiority on two-year OS with improved QoL under surveillance. This analysis compares 5y OS and DFS by Cox regression and estimates distant metastasis incidence with competing-risk analysis.
OS and DFS were not significantly different, but distant mets were higher under surveillance (P<.01). 85/198 (43%) were spared non-beneficial oesophagectomy.
| Endpoint | Active surveillance (n=198) | Standard surgery (n=111) | Effect |
|---|---|---|---|
| 5y OS | 50% (95%CI 34-57) | 54% (95%CI 46-64) | HR 1.17 (95%CI 0.77-1.77) |
| DFS | not reported per arm | not reported per arm | HR 1.39 (95% CI 0.94-2.06) |
| 5y cumulative incidence distant mets | 46% (95%CI 39-53) | 29% (95%CI 21-37) | P<.01 |
This extends SANO's own 2y non-inferiority result and follows the preSANO cCR-assessment framework. It parallels the watch-and-wait paradigm in rectal cancer (OPRA). The randomised definitive-CRT question is being tested separately in NEEDS.
The 43% spared figure counts pts with distant mets within 6 mo of nCRT, which inflates the organ-preservation benefit. At 5y, no NI margin is reported, and an OS HR upper bound of 1.77 cannot exclude clinically relevant harm.
Survival parity holds at 5y, but the distant-mets excess is the signal to explain. It could reflect micrometastatic seeding during delayed resection, or disease surfacing on surveillance imaging that resection would have found later. The abstract cannot separate the two.
Long-term f/u of a cluster-randomised NI trial supporting emerging organ preservation; wide OS CI and a significant distant-mets excess stop short of practice-changing.
- Why are distant metastases higher under active surveillance?
- Does systemic therapy during surveillance offset the distant-mets excess? recruiting An Organ Preservation Strategies After Chemoradiotherapy Combined With Immunotherapy for Esophageal Cancer (PALACE3). Phase 3n=356 · primary completion 2028-03 · ph3 nCRT+IO organ preservation vs nCRT+surgery, 3y OS
- Which nCRT regimen maximizes durable cCR?
📚 Sources · 📄 2 papers
Abstract
Abstract
The longer read
The five-year SANO data shift the debate on organ preservation in oesophageal cancer. The question is no longer whether active surveillance is safe in the short term. It is what kind of failure surveillance trades for avoiding surgery. Overall survival stays close between strategies (HR 1.17, 95% CI 0.77-1.77) after a median 76 months, and DFS points the same way without reaching significance (HR 1.39, 0.94-2.06). Taken alone, those numbers suggest the 2-year non-inferiority result held. The finding that complicates that reading is the cumulative incidence of distant metastases: 46% under surveillance vs 29% after standard surgery, P<.01.
There are two broad explanations, and they lead to different practice. The first is biological: residual microscopic disease that biopsies, EUS and PET-CT missed keeps seeding while the primary is watched, and an early oesophagectomy removes that source. If so, the survival parity may wear down with longer follow-up, or may be held up only by effective salvage of locoregional regrowth. The second is about how failure is counted. Under surveillance, distant disease may be picked up and recorded where, after resection, it would have shown up at a later restaging. The fact that OS has not diverged argues somewhat for the second explanation, but the abstract cannot settle it. A reader should treat the metastasis signal as unresolved, not as noise.
The design deserves attention. A stepped-wedge cluster-randomised trial allocates centres to the timing of a strategy switch, not individual patients to an arm. That explains the unequal groups (198 vs 111) and makes the comparison more exposed to calendar-time and centre effects than an individually randomised trial. The non-inferiority framing also does less work at five years than at two. The source reports no 5-year margin, and an upper confidence bound of 1.77 is wide enough to include a harm most clinicians would reject. "Not statistically significantly different" is the right description. Equivalence is not.
The 43% figure (85/198 spared non-beneficial oesophagectomy) needs careful reading. It includes patients whose distant metastases appeared within six months of nCRT, for whom avoiding surgery reflects early systemic failure rather than successful organ preservation. Those patients gain something real, because a futile oesophagectomy is avoided. But the number of patients with a durable, sustained cCR is smaller than 43%, and the abstract does not give it on its own.
For a radiation oncologist, SANO makes nCRT a potentially definitive treatment for responders, so cCR assessment becomes the gate that decides the management path. It raises two questions the abstract cannot answer. First, would a different nCRT dose or schedule raise the durable cCR fraction, given that the regimen is not stated here? Second, should the distant-mets excess prompt systemic intensification during surveillance? Ongoing randomised work on definitive chemoradiation with salvage surgery, such as NEEDS, is better placed to answer the dose question than SANO is.