Pathologic response is becoming the anchor point in resectable NSCLC
Neoadjuvant chemoimmunotherapy has already changed the front line for patients with early-stage, resectable non-small-cell lung cancer without an oncogenic driver. The regimen now in routine use pairs an anti-PD-(L)1 antibody with platinum-doublet chemotherapy, and in perioperative strategies it extends into the adjuvant phase with continued immunotherapy after surgery. The clinical gain is no longer in doubt: pathological responses have improved, and long-term outcomes have moved in the right direction, including overall survival.
The harder question sits after resection. Uniform treatment delivered to biologically diverse tumors creates an obvious mismatch between intensity and need. Some patients appear to have already been “cured” of their dominant disease burden by preoperative treatment, while others still harbor resistant residual cancer that will relapse unless therapy is intensified. A response-adaptive framework is emerging to separate those two groups rather than treating them as if they were biologically identical.
Pathological complete response has become the most useful fixed point in that discussion.
Why pCR carries more weight than a simple response label
Pathologic complete response, assessed with standardized methods, has matured into a robust surrogate for durable benefit after neoadjuvant chemoimmunotherapy in resectable non-oncogene-driven NSCLC. In practice, that matters because pCR is not just a descriptive endpoint from the operating room specimen. It functions as a biologic readout of whether the combination of cytotoxic therapy and immune activation has actually extinguished the dominant malignant clone before it can seed recurrence.
The appeal is biological as much as statistical. A tumor that disappears histologically after preoperative treatment has likely shown deep dependence on the therapy-induced immune and cytotoxic pressure, while a tumor that persists has already advertised resistance. For surgeons, oncologists, and pathologists, the key issue is whether that post-treatment specimen can serve as an anchor for decisions about escalation or de-escalation in the adjuvant setting.
Standardization is essential here. Pathologic response assessment only becomes clinically actionable when it is reproducible across centers and interpreted consistently. The review places pCR at the center of management because it is currently the clearest available measure linking the surgical specimen to future disease control.
Biomarkers before surgery are informative, but not decisive
Baseline immune markers, especially high tumor PD-L1 expression, increase the likelihood of achieving pCR. Even so, PD-L1 by itself is not enough to guide adjuvant treatment decisions. It is a static baseline measurement, and static measurements often fail when the treatment effect depends on dynamic tumor-host interaction over time.
That limitation is clinically important. Tumors evolve during therapy, immune pressure changes the cellular composition of the lesion, and the immune microenvironment can shift in ways a pretreatment sample cannot capture. A biomarker that improves the odds of response is not necessarily a biomarker that can tell a clinician who should stop treatment and who should intensify it.
Residual disease after surgery is a different biological state from a tumor that was never sensitive to treatment in the first place. The review argues for dynamic markers that track response as it happens, rather than relying on one-time baseline profiling.
ctDNA adds a dynamic layer, but not yet enough to stand alone
Circulating tumor DNA has emerged as one of the most promising tools in the perioperative setting. Its strongest value lies in prognostication and in capturing early treatment response through clearance patterns. When ctDNA disappears, the biology is moving in the right direction; when it persists, minimal residual disease is still present even if imaging or pathology seem reassuring.
Yet ctDNA is not ready to replace pCR as a de-escalation tool. The review makes a careful distinction between information that is prognostic and information that is sufficiently robust to direct treatment. ctDNA currently appears most useful in selected contexts, especially when there is discordance between biomarkers or when postoperative minimal residual disease remains detectable.
A persistent ctDNA signal after surgery is the bedside equivalent of unfinished disease control. It suggests that microscopic relapse risk remains high, even when the resection itself looked complete.
Radiomics and metabolic imaging may narrow the gap between scans and biology
The review also points to radiomics and metabolic imaging as complementary approaches. Their promise lies in extracting biologic meaning from imaging patterns that conventional radiology often flattens into size change alone. Heterogeneous uptake, texture features, and evolving radiographic signatures may help stratify risk earlier and with more nuance than anatomic imaging can provide.
These approaches are unlikely to work in isolation. Their future value probably comes from combination with ctDNA and immune markers, creating a multi-layered model that better reflects how a tumor is behaving under treatment pressure. In a disease where surgical timing, pathologic response, and microscopic residual disease all matter, one biomarker is rarely enough.
That is why integrative models are drawing attention: not because they are fashionable, but because NSCLC is biologically messy and single-axis decision-making is too blunt for the task.
Intensification strategies will need biomarker enrichment
For biologically high-risk tumors, the review outlines intensified perioperative strategies meant to overcome primary resistance. Emerging options include antibody–drug conjugates and new immunotherapy combinations. These approaches are conceptually attractive because they attack resistant disease through different mechanisms: stronger cytotoxic delivery, broader immune engagement, or both.
Their clinical niche, however, will depend on careful trial design. Patients with an insufficient pathological response and those with postoperative minimal residual disease are the most obvious candidates for escalation. The inverse group, patients with pCR and clean dynamic biomarker profiles, may be better served by avoiding unnecessary adjuvant toxicity.
The treatment goal is straightforward even if the biology is not: maximize the chance of pCR where it is still possible, then use postoperative biology to decide who needs more therapy and who does not.
Response-adaptive perioperative care is the next practical model
The most useful framework emerging from this review is a response-adaptive perioperative strategy that merges pathology with dynamic tumor-host monitoring. In that model, surgery is not the end of biologic assessment. It is one checkpoint in a longer sequence that includes pretreatment biomarkers, pathologic response, ctDNA dynamics, and potentially imaging-derived features.
For clinicians, the implication is concrete. A patient with pCR and no molecular signal of residual disease may not need the same adjuvant pressure as a patient with persistent viable tumor or detectable ctDNA. For trialists, the message is equally direct: future studies should stop assuming that one perioperative pathway fits all resectable NSCLC.
The field is moving away from uniformity and toward biologic triage. In resectable NSCLC, the next challenge is not only to improve average outcomes, but to assign the right intensity of therapy to the right patient after the scalpel has done its part.