Oncology Formulation Is a $256 Billion Problem. AACR 2026 Shows Who's Solving It.

AACR 2026 and the Formulation Angle Nobody Covers

More than 30,000 cancer researchers are in San Diego this week for the American Association for Cancer Research Annual Meeting 2026. Over 7,400 abstracts. More than 100 scientific sessions. Presentations from Daiichi Sankyo, Pfizer, AstraZeneca, Merck, Revolution Medicines, Zymeworks, and over 100 Chinese biotech companies showcasing nearly 400 research achievements. The conference theme is "Precision, Partnership, Purpose: Advancing Cancer Science to Save Lives Globally." The cochairs are Paul S. Mischel of Stanford and Alice T. Shaw of Dana Farber.

The plenary talks will focus on targets, biomarkers, and clinical endpoints. The poster halls will be dominated by antibody drug conjugates, bispecific antibodies, and checkpoint inhibitor combinations. The financial press will cover stock moves: Greenwich LifeSciences gained 9.6% on the day its AACR presentation was announced.

What will be discussed far less is the discipline that determines whether any of these molecules actually reach patients. Formulation science. The pharmaceutical chemistry of linker stability, conjugation homogeneity, lyophilized powder reconstitution, high concentration protein viscosity, and cold chain integrity. Oncology accounts for 37% of the global drug pipeline, the single largest share of any therapeutic area. That pipeline is worth nothing if the drug product cannot be manufactured, shipped, stored, and delivered in a form that works.

This briefing covers what AACR 2026 reveals about the convergence of formulation innovation and clinical oncology, and why delivery science is increasingly the differentiator among pharmaceutical companies competing in oncology.

The ADC Revolution: Formulation as the Differentiator

An antibody drug conjugate is a three component therapeutic: a monoclonal antibody that binds a tumor specific antigen, a cytotoxic payload that kills cancer cells, and a chemical linker connecting them. The antibody delivers the payload to the tumor. The cell internalizes the conjugate. The linker cleaves. The payload releases. The cancer cell dies. In principle, this is straightforward. In pharmaceutical product development practice, every step involves formulation decisions that determine whether the drug works or fails.

As of early 2026, the FDA has approved 15 ADCs. AACR 2026 features 18 industry dual payload projects in a single session. Araris is presenting ARC 401, the first triple payload ADC. Bispecific ADCs targeting two tumor antigens simultaneously are appearing across multiple poster sessions. The ADC market grew from $1.6 billion in 2017 to $15.29 billion in 2025, and is projected to reach $32.66 billion by 2035 at a CAGR of 9.2%.

Linker Chemistry

Linker design is a critical and underappreciated element of ADC pharmaceutical chemistry. Twelve of 15 clinically approved ADCs use cleavable linkers, which release their payloads inside tumor cells through protease sensitivity, acid lability, or redox responsiveness. The linker categories tell the story of the field's evolution:

• Peptide based linkers (val cit, val ala): Cleaved by lysosomal cathepsin B. Used in Adcetris, Padcev, and Tivdak.
• Acid labile linkers (hydrazone): Cleaved at low pH in lysosomes. Used in earlier generation ADCs like Mylotarg and Besponsa.
• Disulfide based linkers: Cleaved by intracellular glutathione. Used in Elahere.
• Glucuronide based linkers: Cleaved by lysosomal beta glucuronidase. Emerging platform with improved hydrophilicity and stability.

Site specific conjugation platforms, including ThioMab, GlycoConnect, AJICAP, and AbClick, now enable homogeneous ADCs with controlled drug to antibody ratios, reducing batch to batch variability and improving safety profiles. This is a fundamental shift from the random conjugation methods used in earlier ADCs, where heterogeneous DAR distributions produced variable pharmacokinetics and unpredictable toxicity.

Payload Evolution

The payload field has shifted decisively from first generation auristatins and maytansinoids toward second generation topoisomerase I inhibitors. Over 150 ADCs using MMAE and MMAF payloads have been discontinued in clinical trials due to limited efficacy at tolerated doses. The new standard is DXd (deruxtecan, an exatecan derivative) and SN 38 (a camptothecin derivative), which power Enhertu and Trodelvy respectively. DXd disrupts DNA replication with greater potency and a wider therapeutic window than its predecessors.

At AACR 2026, dual payload ADCs integrating exatecan and triptolide, an RNA polymerase II inhibitor, are establishing a dual killing mechanism that overcomes resistance to single payload approaches. Degrader antibody conjugates, which promote targeted degradation of intracellular proteins rather than direct cytotoxicity, are another frontier. Fortitude Bio's FORT 101 is in discovery, and Pfizer is advancing DACs through its Seagen and Nurix Therapeutics partnership.

DAR Optimization

The drug to antibody ratio determines potency, toxicity, and pharmacokinetics. Most approved ADCs target a DAR of 3 to 4. Enhertu broke convention with a DAR of approximately 8, enabled by its hydrophilic DXd payload and stable tetrapeptide linker. The higher loading was initially considered risky. It proved clinically superior. Enhertu generated $4.98 billion in combined FY2025 sales, up from $3.75 billion in FY2024, and AstraZeneca projects it could reach $11.5 billion across all indications by 2030.

The lesson for pharmaceutical research: DAR is a formulation variable that interacts with linker hydrophilicity, payload charge, conjugation site, and the resulting biophysical properties of the conjugate. Getting it wrong produces aggregation, deconjugation, or off target toxicity. Getting it right produces the best selling ADC in history.

The Subcutaneous Reformulation Wave

While ADCs represent new molecular entities, a parallel trend is converting existing intravenous oncology biologics to subcutaneous delivery. Each conversion is a full pharmaceutical product development program that requires solving high concentration protein stability, viscosity management, injection volume constraints, and co formulation with permeation enhancers.

The engine driving most of these conversions is Halozyme Therapeutics' ENHANZE platform. ENHANZE uses recombinant human hyaluronidase PH20 to temporarily degrade hyaluronan in subcutaneous tissue, enabling large volume SC delivery. The effect is local and transient: hyaluronan restores within 24 to 48 hours. Halozyme reported record full year 2025 revenue of $1.4 billion, with royalty revenue from ENHANZE enabled products growing 52% to $868 million. The company projects royalty revenue exceeding $1 billion in 2026. ENHANZE is licensed to Roche, Takeda, Pfizer, Janssen, AbbVie, Eli Lilly, Bristol Myers Squibb, argenx, and others.

The commercial proof is Darzalex Faspro. Johnson & Johnson converted 85% of its U.S. Darzalex business to the SC formulation within four years of approval. Total Darzalex franchise sales reached $11.67 billion in 2024. Mizuho analysts estimate that successful IV to SC switches can limit biosimilar driven revenue erosion from 75% to approximately 30%.

The High Concentration Protein Problem

Achieving a therapeutic dose in less than 2 mL for SC injection requires protein concentrations of 100 mg/mL or higher. At these concentrations, proteins experience dramatically increased viscosity from self association driven by hydrophobic patches, charge interactions, and dipole forces. High viscosity affects filling accuracy during manufacturing, injection force for patients and caregivers, and aggregation propensity during storage.

The formulation strategies being deployed include excipient engineering with arginine, proline, and ionic strength optimization to mitigate viscosity; antibody engineering to reduce self association; spray dried or spray frozen protein suspensions for ultra high concentrations above 200 mg/mL; and tangential flow filtration for high concentration processing. Each of these is a pharmaceutical chemistry problem with no universal solution. Every antibody behaves differently at high concentration, and predicting which formulation approach will work for a given molecule remains largely empirical.

Padcev + Keytruda QLEX: Three Trends Converge

On December 4, 2025, more than four months ahead of its April 2026 PDUFA date, the FDA approved Padcev (enfortumab vedotin) in combination with Keytruda or Keytruda QLEX as perioperative treatment for cisplatin ineligible muscle invasive bladder cancer. This FDA drug approval is the first PD 1 inhibitor plus ADC combination with a subcutaneous reformulation option, a convergence of three major pharmaceutical biotechnology trends in a single regimen. It also illustrates how the drug development stages for oncology are compressing: an ADC, an immune checkpoint inhibitor, and a subcutaneous reformulation, each of which took years of independent pharmaceutical biotechnology work, now packaged together.

The EV 303 Data

The Phase III EV 303/KEYNOTE 905 trial evaluated perioperative Padcev plus pembrolizumab versus surgery alone in cisplatin ineligible MIBC patients. The results were definitive:

Approximately 83,000 new bladder cancer cases are diagnosed annually in the United States, with roughly 40% of MIBC patients ineligible for cisplatin based chemotherapy. The Padcev and Keytruda combination fills a critical unmet need for patients who previously had limited perioperative options. Grade 3 or higher adverse events occurred in 71.3% of the combination arm, including pruritus, alopecia, diarrhea, fatigue, and anemia.

What the QLEX Formulation Involves

Keytruda QLEX is a fixed combination drug product containing pembrolizumab co formulated with berahyaluronidase alfa, a variant of human hyaluronidase developed by Alteogen Inc. rather than Halozyme. This is a competitive entry in the hyaluronidase enabled SC delivery space.

The formulation is available in two strengths: 395 mg pembrolizumab with 4,800 units berahyaluronidase alfa per 2.4 mL for the every three week regimen, administered over approximately one minute; and 790 mg pembrolizumab with 9,600 units berahyaluronidase alfa per 4.8 mL for the every six week regimen, administered over approximately two minutes. Both contain 165 mg per 2,000 units per mL in single dose vials, administered subcutaneously into the thigh or abdomen.

Keytruda QLEX is the first and only SC immune checkpoint inhibitor that can be given in as little as one minute, compared to 30 minutes for IV Keytruda. Clinical validation came from a Phase III trial where 377 patients randomized 2:1 (SC versus IV) with platinum doublet chemotherapy showed confirmed objective response rates of 45% (SC) versus 42% (IV), with no notable differences in progression free survival or overall survival.

The Patent Cliff Strategy

The strategic significance of Keytruda QLEX extends well beyond clinical convenience. Keytruda generated $29.5 billion in 2024 revenue, making it the best selling drug in the world. Its U.S. patent expires in 2028. Without the SC reformulation, Merck faces estimated revenue erosion exceeding 75% from biosimilar competition. With QLEX, Merck CEO Rob Davis projects capturing 30 to 40% of Keytruda's U.S. patient base on the SC formulation, turning the patent cliff into what he calls "more of a hill" and limiting erosion to approximately 30%. This is a $300 billion question for the pharmaceutical companies facing oncology biologic patent expirations between 2026 and 2030.

Oncology Formulation Challenges: What the Drug Development Stages Actually Require

AACR 2026 abstracts describe biological hypotheses and clinical endpoints. What they do not describe, because it falls outside the scope of cancer research conferences, is the formulation engineering required to turn each of those hypotheses into a viable medicine. The challenges are specific, technical, and consequential at every stage of pharmaceutical product development.

Cytotoxic Drug Handling and Containment

ADC manufacturing sits at the intersection of biologics and small molecule cytotoxicity. Payloads like MMAE, DXd, and PBD dimers kill cells at picomolar concentrations. Manufacturing requires BSL 3 equivalent containment, specialized isolator technology, and dedicated suites. Any breach during conjugation, fill finish, or shipping creates both safety and regulatory exposure. The 18 dual payload ADC programs presented at AACR 2026 each introduce additional complexity: two payloads with different potency profiles, different stability requirements, and different handling protocols.

Targeted Release Engineering

The central formulation challenge in oncology is selective cytotoxicity. ADC linker chemistry, pH sensitive nanoparticles, enzyme activated prodrugs, and antibody masking (Probody technology) all represent different pharmaceutical chemistry solutions to the same problem: killing tumor cells while sparing healthy tissue. The dual mechanism payloads appearing at AACR 2026, such as exatecan plus triptolide combinations, add another dimension because each payload may require different release kinetics within the same conjugate.

Lyophilization and Stability

Most approved ADCs are supplied as lyophilized powders for reconstitution, reflecting the inherent instability of conjugated biologics in aqueous solution. Lyophilization converts unstable solutions into stable powders, but the process itself introduces stresses: freezing damage, drying induced denaturation, and reconstitution related aggregation. Recent advances include AI driven cycle optimization, digital twin simulations for process control, and automated visual inspection. ADC stability challenges go further. Hydrophobic cytotoxic payloads create surface patches that attract neighboring antibodies, initiating aggregation. Payload charge, not just hydrophobicity, is now recognized as a critical stability factor. Premature deconjugation in circulation, caused by plasma proteases acting on peptide linkers or retro Michael reactions on maleimide linkers, reduces efficacy and increases systemic toxicity.

Cold Chain and Companion Diagnostics

Biologics require storage at 2 to 8 degrees Celsius, with some ADCs demanding more stringent conditions. Cold chain management for a single biologic shipment costs 5 to 10 times more than ambient temperature products. Any temperature excursion can cause irreversible aggregation or deconjugation.

Separately, FDA drug approval for targeted oncology agents increasingly requires co approved companion diagnostics. PD L1 testing alone involves four different FDA cleared tests with different expression cutoffs and scoring algorithms depending on the drug and indication. Each companion diagnostic co development extends the clinical trial phases timeline and cost. If a diagnostic partner declines, the therapeutic cannot receive approval with that companion, regardless of clinical efficacy.

AI in Oncology Pharmaceutical Research

Over 173 AI discovered drug programs are in clinical development as of early 2026, with 15 to 20 expected to enter pivotal clinical trial phases this year. Oncology accounts for 72.8% of all AI driven drug discovery programs, dwarfing dermatology at 5.8% and neurology at 5.2%. The number of AI designed candidates entering clinical stages grew from 3 in 2016 to over 173 today, with IND filings reaching their highest single year jump in 2025. Key biotech companies driving this include Insilico Medicine, Recursion, BenevolentAI, Absci, and Generate Biomedicines.

AI Driven Biomarker Discovery

Machine learning models can identify complex biomarker signatures from heterogeneous data sources, including genomic, proteomic, imaging, and clinical datasets, that traditional statistical methods cannot detect. Research published in Cancer Cell in 2025 demonstrated that contrastive learning approaches achieve superior predictive performance for clinical trial outcomes by linking biomarkers with therapeutic response. AACR 2026 features dedicated sessions on this topic, including "AI Revolution in Cancer Research" on April 20 with speakers from Stanford, Harvard, Johns Hopkins, and TU Dresden, and "AI in Cancer Care: Practical Applications Transforming Oncology" from MD Anderson.

Mass Spectrometry Proteomics

Mass spectrometry proteomics has become a critical analytical platform in pharmaceutical research, enabling high throughput analysis of protein expression, interactions, and post translational modifications. In oncology, MS based proteomics serves as the analytical backbone for liquid biopsy biomarker development, measuring dysregulated proteins in blood, urine, or saliva as minimally invasive tools for cancer detection and monitoring. The integration of high resolution mass spectrometry with machine learning and network based analytics has accelerated identification of reliable biomarker panels for predicting treatment response and monitoring disease progression. For ADC development specifically, mass spectrometry proteomics provides the quantitative characterization needed to verify conjugation site occupancy, DAR distribution, and payload release kinetics, all of which feed directly into formulation optimization. As drug development stages grow more complex with multi payload and bispecific conjugates, mass spectrometry proteomics becomes indispensable for characterizing these increasingly sophisticated drug products at the molecular level.

Computational ADC Design

Machine learning is increasingly applied across the ADC development pipeline. Computational models can now predict linker behavior and stability before synthesis, model payload antibody interactions to optimize DAR, simulate ADC pharmacokinetics to predict therapeutic windows, and predict viscosity for high concentration antibody solutions using atomistic simulations. The field is moving toward integration of computational tools from target selection through formulation, though no single platform spans the full chain from molecular design to pharmaceutical product development today.

AI Guided Combination Therapy

AI platforms analyze datasets to identify optimal drug combinations, predict synergies, and anticipate resistance mechanisms. In 2025, research institutions including NCATS, MIT, and UNC built machine learning models using Random Forest, XGBoost, Deep Neural Networks, and Graph Convolutional Networks, with the MIT model achieving an 83% success rate in laboratory validation for combination therapy prediction. As oncology moves toward multi agent regimens like Padcev plus Keytruda, computational combination screening becomes directly relevant to clinical development strategy.

What This Means for Formulation Science

The pattern at AACR 2026 is consistent across every major category: ADCs, SC reformulations, combination regimens, and AI designed therapeutics. The biological hypotheses are advancing faster than the formulation infrastructure can support them.

Eighteen dual payload ADCs at a single conference session, each requiring novel linker chemistry, payload compatibility engineering, and manufacturing containment protocols. Seven approved IV to SC oncology conversions, each requiring high concentration protein formulation, viscosity management, hyaluronidase co formulation, and injection site characterization. A $29.5 billion drug franchise betting its patent cliff survival on a reformulation that delivers pembrolizumab subcutaneously in one minute. Over 173 AI discovered programs heading toward the same formulation bottleneck that has always existed: dissolving, stabilizing, scaling, and delivering the drug product.

The oncology clinical trials market alone is valued at $14.95 billion in 2025, projected to reach $25.61 billion by 2035. Phase III trials account for 48.89% of that market. Each of those trials depends on a drug product whose formulation was locked months or years earlier. A formulation error at that stage does not delay a program. It kills it.

• ADC linker instability in circulation reduces the amount of payload reaching the tumor, limiting efficacy regardless of target biology.
• High viscosity in a subcutaneous biologic formulation can make a syringe unusable for the patient population that needs it.
• Aggregation during lyophilization or cold chain excursion can invalidate a manufacturing batch worth millions of dollars.
• Companion diagnostic co development failures can block FDA drug approval for a clinically effective therapeutic.
• Every AI designed molecule entering the clinic faces the same formulation constraints: solubility, stability, bioavailability, manufacturability.

DeepC is built for this problem. The platform provides AI driven formulation intelligence, operating in the space between molecular discovery and clinical manufacturing where drugs succeed or fail. Its tools address the specific pharmaceutical product development challenges visible across AACR 2026: excipient selection and compatibility assessment for novel drug products, critical quality attribute imputation for incomplete CQA datasets, LC MS/MS spectral prediction from molecular structures for analytical method development, and structured data extraction from regulatory filings, batch records, stability reports, and CMC documents. Every recommendation is traceable to its regulatory source data, built to meet the transparency and data lineage requirements of the FDA/EMA joint AI principles and EMA Annex 22.

The pharmaceutical companies presenting at AACR 2026 are building the next generation of oncology therapeutics. The formulation science required to turn those molecules into medicines is the constraint they all share. The tools for addressing that constraint at the speed and scale the pipeline demands have not existed until now.