Endoscopic response still falls short as a stand-alone gatekeeper for nonoperative care in RESET-C
A brief correspondence in Journal of Clinical Oncology pushes back against an overly direct reading of endoscopic response as a decision rule for nonoperative management in RESET-C. The message is narrow but clinically important: response seen on endoscopy should not yet be treated as sufficient evidence to steer patients away from surgery or other standard management pathways.
In oncology, the temptation to convert an early surrogate into a treatment selector is familiar. Endoscopic improvement can suggest biologic activity, local regression, and a more favorable immediate appearance, but visual response alone rarely captures the full behavior of residual disease. Microscopic persistence, submucosal spread, multifocality, and later regrowth can remain invisible when the decision is based only on what the endoscope shows at a single moment.
That is the core tension in this correspondence. Endoscopic response may be informative, but it is not yet validated as a sufficient discriminator for nonoperative management.
Why a visible response can still mislead bedside decisions
Endoscopy excels at documenting what can be directly inspected. It is far less reliable as a measure of what remains beyond the mucosal surface or in adjacent tissue planes. In cancers where the therapeutic question is whether surgery can be safely deferred, the clinical standard cannot rest on appearance alone. A lesion that looks improved can still harbor viable disease, and a short-term aesthetic response can outpace biologic certainty.
This matters because nonoperative strategies are only defensible when the probability of occult residual disease and later failure is low enough to justify observation. Without that confidence, a visually favorable lesion may become a false reassurance. The stakes are straightforward: an optimistic interpretation of endoscopic response can shift a patient from definitive local control into a surveillance pathway before the evidence base is ready for that move.
In practice, surrogate endpoints become dangerous when they are granted more authority than they have earned. Endoscopic findings may belong in a broader assessment, but not as a lone criterion for de-escalation.
RESET-C enters the discussion with an unresolved clinical threshold
The correspondence is framed as a reply to an accompanying article titled “Endoscopic Response Should Not Yet Select Nonoperative Management in RESET-C.” Even in the absence of trial details in the correspondence itself, the title signals the central issue plainly: the field has not reached the point where endoscopic response can be used to choose between operative and nonoperative care with confidence.
That restraint is clinically important because treatment de-escalation demands a higher evidentiary bar than treatment escalation. If clinicians accept a surrogate too early, the consequence is not simply an imperfect biomarker classification. It can mean delaying curative surgery, misclassifying residual disease risk, and exposing patients to an observation strategy that may prove inadequate for their tumor biology.
The correspondence therefore functions less as a technical critique than as a guardrail. It reminds readers that a favorable local appearance should be interpreted within the full oncologic context, not elevated into a stand-alone decision point.
Safety, selection, and the burden of proof before de-escalation
Nonoperative management has obvious appeal when a tumor appears to respond well. Avoiding surgery can reduce immediate morbidity and preserve anatomy and function. Yet every de-escalation strategy carries a hidden requirement: the selection tool must be strong enough to identify patients whose disease biology truly supports observation. Endoscopy, by itself, may not meet that threshold.
That limitation is especially relevant in a setting where clinicians may feel pressure to act quickly on an early response signal. Visual response is seductive because it is immediate, tangible, and easy to communicate. Biology is less accommodating. Tumors often leave behind viable cells even when the mucosal surface looks deceptively quiet. If the goal is durable oncologic control, the threshold for declaring a patient suitable for nonoperative care must remain high.
The correspondence does not argue against studying response-adapted strategies. It argues against premature adoption. Those are different positions, and the distinction matters at the bedside.
Conflicts disclosed and context for the discussion
The correspondence lists limited disclosures. Michael Chieng reported research funding from Reponex Pharmaceuticals A/S. Ismail Gögenur reported stock and other ownership interests in Nordic AI Medical, honoraria from Pharmacosmos, research funding from repoceuticals and reponex, and a patent related to AdhesionScore, described as an in-vitro assay for detecting perioperative changes in plasma proteins developed with Roskilde University. No other potential conflicts were reported.
Those disclosures matter in the standard way for interpreting correspondence and commentaries, but they do not change the central clinical point raised by the piece. The bar for using endoscopic response as a decision rule remains unmet on the basis of this communication alone.
The practical question now is whether forthcoming data can define a response framework that combines endoscopy with more durable markers of oncologic safety, rather than relying on appearance alone. Until then, endoscopic improvement should be treated as a signal for further evaluation, not as a green light for nonoperative management.