A unified toxicity framework for ADCs and checkpoint inhibitors enters the clinic
As antibody–drug conjugates and immune checkpoint inhibitors move deeper into routine oncology, the toxicity problem is no longer confined to a single drug class or a single organ. Pulmonary, hepatic, gastrointestinal and cutaneous adverse events now appear in patients receiving either approach, and the overlap becomes especially difficult to interpret when the two are combined in earlier lines of therapy or in curative-intent settings.
The practical challenge is attribution. A cough, transaminase rise or rash may reflect immune activation, payload injury, or an ADC-specific effect shaped by target biology, linker chemistry or host susceptibility. In the absence of a shared language, clinicians have been left to manage the same presentation through different mental models, which makes dose holds, steroid initiation and rechallenge decisions inconsistent across centers.
Three biological layers explain why toxicity looks the way it does
The review organizes ADC- and ICI-related adverse events around three interacting layers: immune activation, payload cytotoxicity and payload-independent effects. Each layer can produce a different phenotype, but the bedside picture often converges on the same organs. Lung injury, for example, may arise from immune-mediated pneumonitis or from a payload-linked interstitial process, yet both can present with hypoxemia, imaging changes and a short window for intervention.
A mechanistic framework matters because ADC toxicities are not uniform across compounds. The clinical signatures described across major payload classes follow recognizable patterns: interstitial lung disease and pneumonitis with deruxtecan, diarrhoea and neutropenia with SN-38, cumulative peripheral neuropathy with monomethyl auristatin E, keratopathy with monomethyl auristatin F, and sinusoidal obstruction syndrome with calicheamicin-containing conjugates. Those signatures are not academic labels. They point to the organ systems that deserve the closest surveillance before irreversible injury develops.
Overlap organs demand immune-style management until proven otherwise
In patients receiving ICI–ADC combinations, the lungs, liver, colon and skin carry the greatest diagnostic ambiguity. The framework recommends treating toxicities in these overlap organs as immune-mediated unless a clear cytotoxic pattern is predominant or demonstrated. That approach reflects the danger of missing a treatable immune event while assuming the injury is purely payload related.
Grade matters as well. If grade 2 or higher toxicity occurs, both agents should be held. This is a conservative but clinically logical step, because continuing one component while the true driver remains uncertain risks amplifying injury before the phenotype declares itself. Once the event resolves, the question shifts from emergency control to safe resumption, and the review proposes severity-based algorithms for work-up, steroid escalation, biologic rescue and rechallenge.
One sentence captures the bedside tension: delayed attribution can turn a reversible event into a treatment-ending one.
Why payload class changes the organ signature
The review’s value lies in linking anatomy to pharmacology. Deruxtecan-associated lung toxicity is a reminder that payload distribution and tissue vulnerability can outweigh the elegance of targeted delivery. SN-38 pushes the gut and marrow toward a classic cytotoxic pattern, while auristatin-based constructs accumulate cumulative nerve injury because axonal damage builds across exposure. MMAF-associated keratopathy reflects ocular surface exposure and local tissue sensitivity, and calicheamicin-containing agents carry a sinusoidal obstruction phenotype that fits vascular-endothelial injury in the liver.
This is where an apparently simple question becomes operationally difficult: is the patient experiencing a class effect, a target-driven toxicity or a host-specific event? The answer changes which organ systems are monitored, how aggressively clinicians suppress inflammation, and whether treatment can be resumed after recovery. In a combination regimen, the wrong assumption can lead to either unnecessary discontinuation or unsafe rechallenge.
Rechallenge should be auditable, not improvised
The framework also pushes oncology teams toward more explicit treatment resumption decisions. Rechallenge after toxicity resolution remains variable in practice, especially when the initial event occurred during combination therapy. A structured algorithm creates a record of what was held, what resolved, what steroid or biologic support was needed, and whether symptoms recur after restart.
That level of documentation is not bureaucratic excess. It is the only way to distinguish a one-time inflammatory flare from a reproducible toxicity signature. It also matters for trials, because the growing use of these agents earlier in disease means the safety threshold is being tested in patients with longer life expectancy and less tolerance for chronic organ injury.
Minimum reporting standards could make future trials comparable
The review closes with a call for minimum reporting standards that would capture baseline risk assessment, surveillance strategy and outcomes after treatment resumption. Without those elements, cross-trial comparisons remain difficult, and biomarker discovery stays fragmented. Two studies can describe the same adverse event using different grading practices, different imaging thresholds and different restart rules, leaving investigators unable to tell whether one regimen is genuinely safer or simply documented differently.
Standardized reporting would also help identify host-related determinants of toxicity. The current framework explicitly includes the host alongside target and platform effects, a useful reminder that age, organ reserve, prior treatment exposure and comorbidity burden likely shape vulnerability even when the same ADC or ICI is used.
Resource-stratified care is part of the proposal
The review does not assume every cancer center has the same infrastructure. Healthcare settings with limited access to specialized toxicity pathways can adapt the framework through resource-stratified workflows for surveillance and rescue. That makes the model more than a theoretical exercise. It becomes a practical guide for centers that may not have immediate access to subspecialty consultation, advanced imaging or biologic rescue protocols, yet still need to make rapid decisions when a patient deteriorates.
For oncologists, the central message is straightforward: ADCs and ICIs are converging in the same treatment space, but their toxicities do not respect the old boundaries between targeted therapy and immunotherapy. A unified framework gives clinicians a way to interpret overlapping lung, liver, bowel and skin events before the pattern becomes irreversible. The remaining gap is operational consistency across institutions, especially around rechallenge criteria, surveillance intervals and the minimum data needed to compare one regimen with another.