Oncology’s Next Product Wave: Reading the Pipeline Beyond Checkpoint Inhibitors

A Decade Built on One Mechanism

It is easy to forget how recently the checkpoint inhibitor was the entire story in oncology drug development. Pembrolizumab’s first approval, for advanced melanoma in 2014, opened a decade in which PD-1 and PD-L1 blockade expanded into roughly forty separate cancer indications and, in Keytruda’s case alone, built the best-selling drug in pharmaceutical history — a franchise examined in detail elsewhere in this issue. That era is not over; pembrolizumab remains the most deeply integrated immunotherapy across current oncology treatment guidelines, and no single new mechanism has displaced checkpoint blockade as the backbone of most modern treatment regimens. But the backbone is no longer where the interesting new drug design is happening.

The genuinely new mechanistic ground in 2026 is being broken elsewhere, in three overlapping categories that are increasingly designed to work alongside checkpoint blockade rather than replace it. That distinction — alongside rather than instead of — is the single most important framing shift in oncology drug development over the past three years, and it explains why the next wave of value creation in this category looks structurally different from the checkpoint-inhibitor land grab that preceded it. Nobody is trying to build the next Keytruda. Everyone is trying to build the thing that makes Keytruda work in the patients it currently does not.

The Antibody-Drug Conjugate Renaissance

The first category is the antibody-drug conjugate, which has quietly become one of the most productive engineering platforms in the industry. Fourteen ADCs now carry FDA approval, with dozens more in clinical development and hundreds of candidate molecules moving through earlier evaluation. The design logic has become genuinely sophisticated: a tumor-targeting antibody, a stable linker, and a cytotoxic payload combine to deliver chemotherapy with something closer to a scalpel than a firehose, sparing healthy tissue that a systemically administered chemotherapy agent would otherwise damage indiscriminately.

The engineering history behind that sophistication is worth a moment, because it explains why ADCs took roughly two decades to go from concept to genuine clinical workhorse. The first-generation ADCs, following gemtuzumab ozogamicin’s original 2000 approval, were held back by linker instability — payloads detaching from the antibody before reaching the tumor, causing exactly the systemic toxicity the whole design was meant to avoid — and by suboptimal drug-to-antibody ratios that left too little cytotoxic payload attached to do the job. Brentuximab vedotin’s 2011 approval and trastuzumab emtansine’s in 2013 marked the field’s genuine turning point, demonstrating that refined linker chemistry and better-calibrated payload ratios could dramatically improve the therapeutic index. Every ADC approved since has been, in some sense, a further iteration on that same linker-and-payload optimization problem, and the newest molecules in the class carry drug-to-antibody ratios and linker stability profiles that would have been considered implausible a decade ago.

Datopotamab deruxtecan, a TROP2-directed ADC, improved progression-free survival from 4.9 to 6.9 months over chemotherapy in the Phase 3 TROPION-Breast01 trial for HR-positive, HER2-negative breast cancer — a result solid enough to earn US approval and open combination trials with checkpoint inhibitors. What is striking is not any single result but the trajectory: ADCs are moving from salvage therapy in heavily pretreated patients toward first-line and even curative-intent settings, which is a fundamentally different commercial and clinical proposition than the class occupied even five years ago, when an ADC was, almost by default, a third- or fourth-line option reached for only after better-established therapies had failed.

TROP2: A Case Study in a Crowded Target

TROP2 itself has become one of the most contested targets in solid-tumor oncology, and the competitive dynamics there are instructive of how this next wave of oncology drug development actually plays out commercially. Sacituzumab tirumotecan — known as sac-TMT, MK-2870, or SKB264 depending on which side of the Merck licensing agreement with China’s Kelun-Biotech you are standing on — posted a positive Phase 3 result in first-line triple-negative breast cancer in May 2026, in a patient population largely excluded from the existing pembrolizumab-plus-chemotherapy standard of care because of low PD-L1 expression, which under current guidelines excludes roughly 60 percent of metastatic triple-negative breast cancer patients from the checkpoint-inhibitor-based first-line regimen entirely.

Merck is simultaneously running the molecule in combination with pembrolizumab itself across lung, gastroesophageal, and further breast cancer trials — a genuinely large program, spanning studies designated MK-2870-007 through at least MK-2870-015 in the public trial registry, each targeting a different tumor type or line of therapy. That is the shape of things to come: the ADC is not competing with the checkpoint inhibitor for the same patient, it is extending the checkpoint inhibitor’s reach into populations it could never treat alone. Gilead’s Trodelvy, the first-generation TROP2 ADC, occupies an increasingly narrow niche as newer entrants with more stable linker chemistry and higher drug-to-antibody ratios demonstrate superior efficacy in head-to-head-adjacent trial comparisons — a reminder that being first to a target does not guarantee remaining the best-in-class molecule against it once the underlying chemistry keeps improving.

Bispecifics and the Direct Recruitment of T Cells

The second category is the bispecific and multispecific antibody, which takes a different approach to the same problem — recruiting the immune system directly rather than relying on pre-existing anti-tumor immunity the way checkpoint blockade does. That distinction matters clinically in a specific way: checkpoint inhibitors only work in patients whose immune system has already, on its own, mounted some detectable response against the tumor, which the drug then unleashes by blocking the tumor’s ability to suppress it. A meaningful share of cancers never generate that initial immune response in the first place, and those patients simply do not respond to checkpoint blockade regardless of dose or combination partner. Bispecific T-cell engagers sidestep that requirement entirely by physically bridging a T cell to a tumor cell, forcing an immune interaction that would otherwise never occur.

T-cell engagers such as epcoritamab and glofitamab are now competing directly with CAR-T therapies in relapsed lymphoma, offering a genuinely important practical advantage: they are off-the-shelf biologics administered on an outpatient basis, rather than requiring the weeks-long, highly individualized manufacturing process that autologous CAR-T demands — a manufacturing distinction covered in more depth elsewhere in this issue’s look at cell and gene therapy. Tarlatamab, which targets DLL3 in small-cell lung cancer, represents a genuinely distinct mechanism within this category: it physically redirects T cells toward tumor cells regardless of whether the patient’s immune system was already primed to recognize them, in a cancer type that has historically been one of oncology’s most immunologically “cold” and treatment-resistant.

Bispecific ADCs — molecules engineered to hit two targets simultaneously while still delivering a cytotoxic payload — are now in early clinical trials for solid tumors, aiming to delay the resistance that eventually limits single-target therapies. ACR335, a tetravalent molecule targeting both cMET and EGFR while delivering a dual-payload combination, entered Phase 1 trials in 2026 and illustrates just how architecturally ambitious this next generation of molecules has become relative to the comparatively simple single-target, single-payload ADCs of a decade ago.

Cell Therapy’s Harder Problem

The third thread, still earlier-stage but worth tracking closely, is cell therapy’s slow migration from blood cancers into solid tumors — a much harder engineering problem, given the physical and immunological barriers a solid tumor presents that a circulating lymphoma cell does not. A blood cancer is, functionally, distributed throughout the bloodstream and accessible to an engineered immune cell almost by definition; a solid tumor is a physically dense mass with its own immunosuppressive microenvironment specifically evolved to exclude and deactivate exactly the kind of immune cells a CAR-T therapy introduces. Progress there is real but constrained less by biology than by manufacturing, a bottleneck significant enough that we have given it its own treatment elsewhere in this issue, and the solid-tumor cell-therapy programs that are advancing tend to pair a genuinely novel targeting mechanism with equally novel approaches to overcoming that microenvironment problem, rather than simply porting a blood-cancer CAR-T design into a new tumor type.

The Fourth Thread: Vaccines Re-enter the Conversation

There is a fourth thread that deserves mention on its own: the personalized cancer vaccine, which spent a decade as a research curiosity and has recently produced data hard to dismiss, treated at greater length elsewhere in this issue. Moderna and Merck’s individualized neoantigen therapy, given alongside pembrolizumab in resected high-risk melanoma, showed a sustained 49 percent reduction in the risk of recurrence or death at five years of follow-up — a result that held essentially unchanged from the three-year readout, evidence that the effect is durable rather than a transient immune bump. It is early, and it depends on a genuinely difficult manufacturing feat, but it is the clearest evidence yet that the checkpoint inhibitor’s next chapter is combination therapy with mechanisms nobody had fully validated five years ago.

The Sequencing Problem Nobody Has Fully Solved

None of this diminishes the safety and sequencing questions that come with moving increasingly potent therapies into earlier lines of treatment. On-target, off-tumor toxicity becomes a bigger clinical management problem as ADCs and bispecifics are used for longer durations in healthier patients who, unlike late-line patients with few remaining options, have genuine alternatives and lower tolerance for serious side effects. The field is still working out the right combination and sequencing strategies to manage it — which drug goes first, how long to wait before adding a second mechanism, how to detect resistance early enough to switch regimens before a patient progresses. That is exactly the kind of clinical-operational complexity that favors platform companies with genuine payload and linker chemistry expertise, and genuine translational and clinical-operations depth, over single-asset developers chasing whichever target is fashionable this cycle without the infrastructure to manage what happens once the drug actually reaches patients.

What This Means for Portfolio Construction

For a concentrated healthcare portfolio, the read is straightforward even if the science is not: oncology remains the single largest destination for biotech venture capital — roughly a third of total dollars deployed in 2025 and 2026 — not because the category is crowded with me-too checkpoint inhibitors, but because it has genuinely specialized into several distinct, investable sub-platforms. The companies worth backing are the ones building modular engines — proprietary linker chemistry, novel bispecific formats, neoantigen manufacturing pipelines — rather than the ones betting a single asset on a target that six other companies are also chasing. A single-target, single-asset oncology company in 2026 is, structurally, a considerably riskier bet than the same company would have been in 2016, precisely because the field has matured enough that a platform capable of iterating across multiple targets, payloads, and formats now has a genuine, demonstrated head start over a company that has to start from scratch with each new program.