Conference Intelligence AACR 2026: Seven Signals That Will Shape Oncology's Next Chapter

AI Earns Its Own Plenary — and This Time It's Biological

AACR gave AI a dedicated plenary for the first time this year. That's a structural signal — the conference committee decided AI warrants the same programmatic weight as clinical trials and discovery science. But the more important signal is in the abstracts themselves.

Consider two presentations from the poster and minisymposium programs. "OncoTwin: A multimodal digital twin framework for predicting treatment response and guiding trial design in ALK-rearranged non-small-cell lung cancer" isn't theoretical — it was validated against Phase 2/3 data and positions digital twins as a bridge between real-world evidence and clinical trial design. Separately, a deep-learning CT biomarker study demonstrated that fully automated imaging analysis provided earlier and more reliable efficacy readouts than best overall response or progression-free survival in simulated Phase 2 NSCLC trials.

The implication for pharma: AI at this meeting isn't a technology showcase. It's entering the workflow at decision-critical points — trial design, efficacy assessment, biomarker discovery. Companies still treating AI as an R&D experiment are falling behind companies embedding it in their clinical development pipeline.

Early-Onset Cancers Demand a Different Approach

AACR elevated early-onset cancers to a full plenary session this year. The data across 62 abstracts makes the reason clear: these aren't younger versions of the same disease.

The abstract "Spatial profiling reveals distinct immune microenvironments in early vs late onset colorectal cancer" found that EOCRC tumors carry a more immunosuppressive microenvironment — higher Treg infiltration, more M2 macrophage presence in immune-enriched regions. That finding has direct consequences for immunotherapy strategies: treatments validated in older populations may underperform in younger patients for microenvironmental reasons, not just genomic ones.

Gastric cancer data tells a parallel story. "Real-world clinicogenomic comparison of early- and average-onset gastric cancer" identified distinct genomic features and pathway dysregulation patterns in younger patients that may drive both tumorigenesis and therapy resistance. Meanwhile, epigenetic data from early-onset CRC models suggests that host age itself — not just tumor genetics — modifies methylation patterns in ways that affect tumor behavior.

For portfolio strategists, the practical question is whether existing clinical programs have adequate representation of early-onset populations. If the biology diverges this significantly, trials designed for average-onset disease may generate data that doesn't generalize to the fastest-growing patient segment. 

MRD in Solid Tumors Reaches a Clinical Inflection Point

The Discovery Science Plenary — chaired by Maximilian Diehn of Stanford — is titled "The Next Frontier in Minimal Residual Disease: Solid Tumors." That plenary title isn't accidental. MRD monitoring has already rewritten treatment algorithms in hematologic malignancies. The abstract data at AACR 2026 suggests solid tumors are approaching the same tipping point.

The evidence is stacking up across tumor types. "Prognostic utility of minimal residual disease detection using circulating tumor DNA in early-stage breast cancer: A systematic review and meta-analysis" found a nearly 10-fold increased recurrence risk in ctDNA-positive patients, with a consistent ~10-month lead time before clinical detection of relapse. That's not a research finding — that's a clinically actionable window.

In NSCLC, "Early ctDNA quantification by ctFE outperforms Max VAF for survival stratification across locally advanced and oligometastatic NSCLC treated with radiotherapy" introduced a new metric — circulating tumor fraction equivalent — that consistently beat traditional variant allele frequency using a tumor-naïve, off-the-shelf assay. The significance: scalable MRD monitoring that doesn't require tumor tissue for assay design.

Real-world breast cancer data adds another dimension. Published ctDNA-MRD studies now show 71.4% relapse sensitivity with a hazard ratio of 0.45 for intervention benefit — data that validates ctDNA not just as prognostic, but as a predictive biomarker that may guide treatment decisions.

For CI teams, the competitive map is shifting. ctDNA-based MRD is creating a new decision node in adjuvant treatment algorithms. Companies with MRD-guided strategies will have a differentiated clinical narrative. Those without one will need to explain why.

Precision Prevention Moves from Theory to Presidential Priority

The Presidential Select Symposium — typically the most agenda-setting session at AACR — is titled "Targeting Stage 0: Precision-Based Prevention." The framing matters. "Stage 0" implies intercepting cancer before it declares itself as a clinical entity.

This isn't isolated. Major symposia on "Aneuploidy and Mutations in Normal Tissues and Their Role in Cancer Initiation," "Genetic and Environmental Determinants of Cancer," and "Cancer Prevention and Screening: Exploring Policy Solutions" reinforce the theme across biology, genetics, and policy tracks simultaneously. The late-breaking minisymposium features "The genomic landscape of likely human precancers" — foundational mapping work that could define future screening paradigms.

AACR has historically been a treatment-first meeting. Elevating prevention to presidential-level programming signals a structural shift in how the oncology establishment frames its priorities. For the multi-cancer early detection category — and for companies building interception strategies — this institutional validation matters more than any individual data readout. 

Clinical Trials Plenary: Where the Pipeline Competition Sharpens

The Clinical Trials Plenary sessions concentrate the highest-impact data into four focused storylines. Each one reveals where the competitive pressure is building.

Next-generation KRAS inhibitors enter the clinic. The precision oncology plenary features two first-disclosure datasets: Elisrasib (D3S-001), a next-generation GDP-bound KRAS G12C inhibitor in advanced NSCLC, and zoldonrasib (RMC-9805), a RAS(ON) G12D-selective tri-complex inhibitor — the first oral G12D-targeting agent with clinical data. In the late-breaking minisymposium, daraxonrasib plus chemotherapy as first-line treatment for metastatic pancreatic adenocarcinoma represents the first combination RAS inhibitor data in a tumor type defined by KRAS mutations. These disclosures mark the transition from "can we drug KRAS" to "which KRAS strategy wins."

ADC combinations define the next wave. The dedicated ADC plenary features a first-in-human EGFR-targeted ADC (SYS6010) in nasopharyngeal carcinoma, a first-in-human Claudin 6-targeted ADC (QLS5132) in platinum-resistant ovarian cancer, and combination data from the ARTEMIS-009 study (risvutatug rezetecan + adebrelimab) in NSCLC. A T-DXd + olaparib dose-escalation study extends ADCs into the DNA damage response space. The strategic signal: the ADC field is shifting from single-target monotherapy to rational combination design.

Cell therapy expands beyond hematology. The CAR-PRISM trial brings ciltacabtagene autoleucel data in high-risk smoldering myeloma — an earlier-disease intervention. A first-in-human KIR-CAR study tests a novel receptor architecture in mesothelin-expressing solid tumors. And mRNA-4359 plus pembrolizumab in first-line advanced melanoma combines mRNA-based immune modulation with checkpoint inhibition, bridging vaccine and cell therapy concepts.

Novel mechanisms reach first-in-human. The WEE1 inhibitor zedoresertib combined with PKMYT1 inhibitor lunresertib is the first-in-class combination targeting parallel cell cycle checkpoints. CID-078, a novel Cyclin A/B-RxL inhibitor, represents an entirely new mechanism in solid tumors. Both present first data in the precision oncology plenary.

Spatial Omics Stops Being a Technology Story and Becomes a Clinical One

The shift in spatial omics at AACR 2026 isn't about adoption — it's about what the technology is now revealing. And for drug development, the findings are consequential.

"Spatial transcriptomics uncovers pharmacokinetic barriers and tumor-intrinsic determinants of resistance to trastuzumab deruxtecan in breast cancer" demonstrates that spatial analysis can identify why an ADC fails in specific tumor regions — not just whether it fails. Separately, spatial transcriptomics of sacituzumab govitecan-treated tumors identified distinct gene expression programs separating responders from resistant cases, with immune activation genes enriched in responders.

The implications extend beyond ADCs. "Spatial archetypes of conserved and cancer-specific immune-stromal niches" presents a large-scale framework for classifying microenvironmental architectures across tumor types — effectively a taxonomy of tissue organization patterns that predict therapy response. "Spatial 3D and multi-omics mapping of diffuse gastric cancer evolution from preinvasive to invasive lesions in CDH1 mutation carriers" tracks how cancers develop in three-dimensional tissue context, identifying candidate biomarkers of disease progression.

For biomarker and companion diagnostic strategy, the trajectory is clear. Tumor profiling is moving from "what mutations are present" to "how is the tissue spatially organized." Companies building next-generation patient selection approaches should be designing around spatial architecture, not just molecular markers.

The Microenvironment Gets Mechanistically Specific

TME research at AACR has historically been broad. This year, the program carves it into mechanistically distinct problems: necrotic microenvironments driving metastasis, neural-immune crosstalk, inflammaging, and stromal reprogramming each receive dedicated major symposia. That specificity reflects where the science has matured enough to generate actionable therapeutic hypotheses.

The abstract data supports this. "Spatial transcriptomics reveals distinct SPP1-CD44 signaling networks in neoadjuvant osimertinib-treated EGFR mutant non-small cell lung cancer" identifies a specific macrophage-tumor cell signaling axis (SPP1-CD44) that drives tolerance to EGFR-targeted therapy. This isn't a general observation about macrophages being present — it's a defined mechanism with a potential therapeutic target (the TM4SF4/SPP1/SERPINA3 axis).

In HER2-enriched breast cancer, "Tertiary lymphoid structure (TLS)-associated immune circuits define response to durvalumab, trastuzumab, and pertuzumab (DTP)" found that coordinated Tfh-B-cell and dendritic-B-cell interactions within TLS are the dominant determinants of DTP response — not bulk immune cell counts.

The strategic implication: the era of PD-L1 and TMB as primary immune biomarkers is ending. The next generation of IO patient stratification will require understanding spatial immune organization, specific cell-cell signaling networks, and microenvironmental architecture. Companies still relying on single-marker immune biomarkers for clinical development are building on a foundation that AACR 2026 data is actively undermining.

What This Means for CI Teams

AACR 2026 delivers 9,561 sessions across six days, with 393 late-breaking posters and 56 clinical trials plenary presentations running in parallel tracks. No team covers this comprehensively through selective attendance.

But coverage isn't the hard problem. Synthesis is. The insights that will shape competitive strategy over the next 12 months won't emerge from any single abstract. They'll come from connecting patterns: how spatial omics data reframes ADC resistance, how MRD monitoring creates new adjuvant treatment decision points, how KRAS combination strategies interact with the precision prevention agenda.

The teams that extract the most value from AACR 2026 will be the ones that move from session notes to strategic interpretation — quickly.