Combining immune checkpoint inhibitors (ICIs) with chemotherapy has reshaped cancer care by pairing immune reinvigoration with tumour debulking and microenvironmental reprogramming. Cytotoxics can heighten antigen release, expose damage-associated molecular patterns, prime dendritic cells (DCs), and transiently curb immunosuppressive populations—creating a therapeutic window in which PD-(L)1 or CTLA-4 blockade is more effective (1, 2). Clinically, this biology has translated into higher response rates and survival gains across several solid tumours, exemplified in metastatic non-small-cell lung cancer where pembrolizumab plus platinum/pemetrexed became a first-line standard (3). Evidence also supports triplet strategies that add anti-angiogenic therapy to ICI–chemotherapy, improving progression-free and overall survival in nonsquamous disease (4). In curative-intent pathways, neoadjuvant ICI–chemotherapy deepens pathological regression and prolongs event-free survival in resectable NSCLC, anchoring a perioperative paradigm (Figure 1) (5). Beyond lung cancer, early triple-negative breast cancer demonstrates durable event-free survival benefits when pembrolizumab is integrated with neoadjuvant chemotherapy and continued post-operatively (6).

Amid these advances, key questions remain around deployment, selection, and safety: how to sequence therapy (induction vs concurrent), how many cycles are sufficient in the neoadjuvant setting, which composite or on-treatment biomarkers best guide patient selection, and how to preserve dose intensity while managing immune-related and cytotoxic adverse events. This Editorial distils the mechanistic rationale and current practice patterns of ICI–chemotherapy combinations and outlines a forward-looking agenda—standardized reporting of timing and dose intensity, integration of adaptive biomarkers for real-time triage, and rational escalation to triplets when biology supports it—to help clinicians design and deliver chemo–immunotherapy safely and effectively across tumour types.

The systematic integration of chemo–immunotherapy evidence across tumour types converges on three axes—sequencing, patient selection, and safety/feasibility—offering a practical scaffold for clinical use.

Taken together, the contributions in this Topic argue that chemo–immunotherapy is most effective when it is deliberately timed, sequence-aware, and risk-adapted rather than used reflexively. Across lung and gastrointestinal cancers—and even in edge populations—the evidence supports three pragmatic shifts: (i) optimise how we combine (limited neoadjuvant cycles, explicit induction vs. concurrent choices, and biologically justified maintenance); (ii) upgrade what we measure (adopt MPR as a practical perioperative surrogate, paired with pre-operative risk tools); and (iii) strengthen guardrails (plan for overlapping toxicities and report sequence, cycle number, dose intensity/density uniformly). This synthesis provides a transferable scaffold for trial design and day-to-day decision-making across tumour types.

Looking ahead, four priorities can make chemo–immunotherapy consistently durable and mechanism-anchored. First, build sequence-aware perioperative regimens that align chemotherapy-driven antigen release with checkpoint reinvigoration and predefine the role/duration of adjuvant therapy (informed by KEYNOTE-671, AEGEAN, CheckMate 77T) (7–9). Second, upgrade endpoints by embedding major pathological response within hierarchical/composite frameworks and quantifying its bridge to EFS/OS, so early signals translate into patient-centred benefit (7, 8). Third, steer escalation or de-escalation with multimodal biomarkers—functional T-cell states, histopathology, ctDNA kinetics, and imaging–pathology AI—to enable organ preservation where appropriate. Fourth, evaluate rational triplets judiciously when resistance biology is explicit, and anchor maintenance choices to adjuvant benchmarks (e.g., IMpower010) with uniform reporting of sequence, cycle number, dose intensity, and dose density for reproducibility (10).