Transforming Cancer Therapy: Insights from 2025 H1 FDA Approvals and Preclinical Drug Discovery

Oncology drug approvals in H1 2025

In the first half of 2025, the FDA’s Center for Drug Evaluation and Research (CDER) approved a total of 16 novel drugs, with half of these drugs related to the treatment of cancer. Both the novel and supplemental oncology approvals reflected recent developmental innovations, including an increased focus on targeted, immunologically driven, and personalized oncology therapies, targeting a broad range of cancers.

New antibody-drug conjugates (ADCs) for solid tumors, small molecule targeted therapies, and biomarker-guided approaches emerged, representing significant progress in precision medicine. Several therapeutics addressing rare cancers also gained approval, including the first treatment for KRAS-mutated ovarian cancer and a non-surgical treatment option for patients with neurofibromatosis type 1.

Although 2025’s H1 approvals tracked slightly behind those in 2024, there is reason to be optimistic as innovation remained strong. Overall, these H1 approvals demonstrated continued progress with developing treatments that offer improved benefits and wider treatment options for oncology patients.

List of FDA novel cancer approvals H1 2025:

While these approvals are a reason for optimism, political and economic uncertainty remain ongoing challenges to cancer drug research. With attrition rates for novel drug discovery persistently high (approximately 95%), maintaining an effective drug discovery pipeline remains key if innovation is to continue to accelerate. This pipeline relies on patient-relevant, clinically predictive models within the preclinical screening process to drive improved decision-making and success rates.

With this in mind, let’s examine the role of different advanced preclinical screening models and how these can support effective biomarker strategies in drug development and approval.

Advanced Preclinical Screening Models

Cell lines

In many instances, cell lines represent the first step in drug discovery. They serve as an initial high-throughput method to evaluate drug candidates against multiple cancer types and diverse genetic backgrounds and discover predictive biomarkers for drug responses.

Additional cell line applications include:

• Drug efficacy testing
• High-throughput cytotoxicity screening
• In vitro drug combination studies
• Adhesion, migration, and invasion assays
• Colony-forming cell (CFC) assays
• Soft agar colony formation

As cell lines are a customary preclinical model, diverse and established collections are available to research teams. This includes, Crown Bioscience’s cell line and CDX database , which is the world’s largest commercial database of well-characterized cell lines.

This collection includes a range of different cell line panels:

• More than 500 genomically diverse cancer cells lines for drug response screening
  
  
• A diverse oncogenic in vitro screening panel of over 150 well-validated cell lines with associated cell line derived xenograft models for in vivo efficacy studies
  
  
• More than 170 genomically characterized cell lines to correlate mutation status, copy number variation, and expression levels in drug response
  
  
• Cancer cell lines with primary cells originating from genetically defined patient-derived xenograft (PDX) models
  
  
• A 3D ex vivo panel of freshly isolated and frozen cells from PDX models

Cell lines provide an essential starting point for drug discovery, offering reproducible and standardized testing conditions. As they are a versatile, quick and relatively low-cost platform, they are suitable for a variety of applications, including single agent assays and more complex combination studies.

However, there are inherent limitations associated with 2D models. For example, they have a limited ability to represent tumor heterogeneity and they don’t reflect tumor microenvironments (TME).

Organoids

Organoids have been described by Nature as “invaluable tools in oncology research” that are revolutionizing drug discovery workflows. These 3D models are grown from patient tumor samples, so they faithfully recapitulate the phenotypic and genetic features of the original tumor.

In April 2025, the FDA announced that its animal testing-requirement for monoclonal antibodies and other drugs will be reduced, refined or potentially replaced entirely with a range of advances approaches, including organoids.

And just days ago, the NIH announced it will no longer fund standalone animal model development projects and is shifting toward supporting alternative platforms like organ-on-a-chip technologies.

This is set to make organoids an even more central tool in the oncology drug pipeline, helping to “get safer treatments to patients faster and more reliably, while also reducing R&D costs and drug prices”.

Organoid biobank databases, like this one, allow researchers to find the right models to progress their oncology drug development, with various applications:

• Investigate drug responses
• Evaluate immunotherapies
• Explore genetic disease drivers
• Safety and toxicity studies
• Personalized medicines
• Disease modeling
• Predictive biomarker identification

Organoids can also be used to assess mechanisms of resistance and model the development of tumors. They support high-throughput screening of therapeutic candidates, more effectively predict tumor responses to treatments than cell lines, and are more cost-effective than animal models. As a result, their impact on drug development is already apparent.

For example, they have been used to identify MTAP as a new target in pancreatic cancer and SIRT1 as a new target in bladder cancer.

Although organoids can be seen as a bridge between in vitro and in vivo, there is no one perfect model. Organoids are more complex and time-consuming to create than cell lines, and they cannot fully represent a complete TME.

PDX models

Created by implanting patient tumor tissue into immunodeficient mice, PDX models preserve key genetic and phenotypic characteristics of patient tumors. They are the most clinically relevant preclinical models, which is why they have been called the gold standard of preclinical research.

PDX collections, such as Crown Bioscience’s PDX database (the world’s largest collection of clinically relevant PDX models) allow researchers to search models by indication, drug responses, patient history or multiomics data, providing a powerful tool for drug development and clinical translation.

PDX model applications:

• Biomarker discovery and validation
• Clinical stratification
• Exploring new indications
• Targeted research
• Drug combination strategies

PDX models enable more personalized treatment strategies and allow researchers to evaluate therapies across diverse tumor types and gain a better understanding of mechanisms of action. As these models preserve the architecture of the original tumor and include components of the TME, they can more closely mirror tumor responses to more accurately predict clinical outcomes and improve clinical trial success rates.

However, they are not without their drawbacks. Compared to other models, they are more expensive, resource-intensive and time-consuming to produce. Additionally, they cannot support the high-throughput testing made possible by 2D cell lines and organoids, and the ethics of animal testing must also be considered.

An integrated approach

As each model has associated pros and cons, adopting a holistic, multi-stage approach is essential. This allows research teams to leverage the inherent advantages of each model. PDX-derived cell lines offer an effective starting point. Organoids allow researchers to build on their understanding and develop their research. While PDX models represent the final preclinical stage before human trials. Each model offers crucial insights into how drugs may perform in patients, building a robust drug discovery pipeline to reduce attrition rates and improve the chances of success.

An integrated approach in action: biomarker hypothesis generation and validation

The early identification and validation of biomarkers is crucial to drug development as this allows researchers to identify patients with biological features that drugs target, track if drugs are working, and identify early indicators of drug effectiveness. By taking an integrated, holistic approach, biomarker discovery can be structured and optimized:

1. Using PDX-derived cell lines, researchers can identify potential correlations between genetic mutations and drug responses. This large-scale and targeted screening allows researchers to generate sensitivity or resistance biomarker hypotheses.
2. During organoid testing, biomarker hypotheses can be refined and validated using these more complex 3D tumor models. Multiomics, including genomics, transcriptomics and proteomics, can help identify more robust biomarker signatures.
3. PDX models can be used to validate these biomarker hypotheses before clinical trials. The unique attributes of PDX models give researchers a deeper understanding of biomarker distribution within a heterogeneous tumor environment.