To increase the success rate of drug discovery

In Japan, it takes a very long time for a single drug to be developed, and the success rate of drug development is declining year by year. In recent years, the success rate from compound screening (discovery phase) to approval has fallen to about "1 in 22,000 to 1 in 31,000."

This article summarizes the success rates and high drop-out rates by process in new drug development, and the key to drug discovery success rates.

Success rates by development process

Here is the success rate for each development process. The table below shows the average next phase transition rate.

*The overall success rate from the start of clinical trials to approval is 13%

Reasons for the high drop rate in Phase II

From the table above, it can be seen that the drop-out rate in Phase II is high. Regarding this part, reasons such as "the drug efficacy confirmed in animal experiments (non-clinical studies) was not reproduced in the human pathological environment (=translational gap)" have been cited for Phase II (a study involving a small number of patients). Thus, there are various reasons for drop-outs in each process.

Causes of drops in each process

From 1991 to 2000 (and even up to the present), challenges in pharmacokinetics (PK) have been overcome with advanced technologies such as high-precision bioanalysis. However, the barrier of "efficacy" still remains and is a cause for dropouts.

Lack of efficacy

"Lack of efficacy" refers to a state where the expected therapeutic effect (significant difference) is not observed in patients. The dropout rate due to this reason has not improved even after 10 years since 1991. This makes it the biggest reason for dropouts as of 2000.

Safety concerns in the clinical stage

"Safety concerns during clinical trials" refer to cases where a drug was discontinued due to concerns that unacceptable adverse events would occur if administered to humans (healthy adults or patients). The discontinuation rate was approximately 10.1% in 1991 and slightly over 10.1% in 2000.

Toxicity issues

"Toxicity issues" are primarily identified during the nonclinical phase (animal testing). For example, organ damage, genotoxicity, and carcinogenicity—which pose serious risks to survival and health—are common causes of program termination. The termination rate due to these causes was approximately 10.1% in 1991, but had increased to about 20.1% by 2000.

Pharmacokinetics / Bioavailability

"Pharmacokinetics/Bioavailability" refers to situations where a drug is not sufficiently absorbed into the body or does not reach the target tissue after administration, or where its effects do not last because it is metabolized or excreted too quickly. As of 1991, the dropout rate was approximately 40.1%, making it the primary cause of dropout at that time. By 2000, this rate had fallen to less than 10.1%, representing a significant improvement.

Formulation and prescription issues

Even if a substance (active pharmaceutical ingredient) is excellent in its own right, there are cases where it is dropped due to “formulation and dosage form issues.” This refers to situations where it is difficult to process and store the substance stably in the form of a pharmaceutical product, such as an injection or tablet. While the drop rate was nearly 0% in 1991, it had risen to approximately 51% by 2000.

Rising manufacturing costs (cost of sales)

"Soaring manufacturing costs (cost of goods sold)" refers to a situation where the costs associated with manufacturing a drug—such as raw material expenses and the complexity of the manufacturing process—are deemed too high, making it impossible to secure a profit once the drug is released to the market. The drop rate was nearly 0.1% in 1991, but by 2000, it had risen to just under 10.1%.

Decline in commercial value (loss of marketability)

This refers to cases where competitors launch superior new drugs before a company can bring its own new drug to market, or where management decides that it cannot recoup massive investments because the market size is smaller than anticipated. As for the causes of this, the dropout rate, which was approximately 51% in 1991, has increased significantly, reaching about 20% by 2000.

Overcoming the "Translational Gap" Key to Drug Discovery Success

It can be said that obtaining data with a low "translational gap" and a high "efficacy" wall improves the success rate of drug discovery, leading to reduced time and avoidance of losses.

One of the causes of the translational gap is said to be the "limitations of animal models." Even if dramatic effects are observed in preclinical studies, there are four definitive biological and anatomical gaps that explain why these effects are not reproduced in humans. The four gaps are explained below.

[Anatomy/Target Factor] Dissociation of Receptor and Signal Structure Due to Species Difference

First, there's the issue of "dissociation of receptor and signal structures due to species differences." For example, an inhibitor that was highly effective in animal studies might not fit well with human receptors, leading to insufficient clinical efficacy.

This is because even minor differences in the molecular structure of target receptors or signaling molecules between humans and animals can result in a critical difference in a drug’s binding affinity—that is, its “effectiveness.”

【Metabolism/Physiology】 Presence or absence of active metabolites due to interspecies differences in pharmacokinetic pathways (ADME)

The second gap is "the presence or absence of active metabolites due to species differences in pharmacokinetic pathways (ADME)." For example, if the "metabolite" produced in an animal's body is primarily responsible for the drug's efficacy, and that metabolite is hardly generated in the human body, then the expected efficacy will not manifest at all.

Such gaps occur because the activity of drug-metabolizing enzymes (such as CYPs) and the metabolic pathways themselves differ significantly between animals and humans.

[Pathophysiology] Artificially created "acute, simple models" and
Dissociation in Human Chronic and Complex Diseases

Many animal models are simple models where pathology is artificially induced over a short period through specific feeders or drug administration, or genetic manipulation. Therefore, a gap arises between "artificially created acute, simple models" and "human chronic, complex diseases."

For example, human MASH is a complex chronic disease that progresses over many decades through stages of "fat deposition," "inflammation," and "fibrosis," often against a background of lifestyle-related diseases like diabetes and dyslipidemia. Simple mouse models that merely accumulate fat over a short period cannot accurately predict drug accessibility to fibrotic tissue in advanced human cases or the inflammatory signals that arise over long durations.

[Immunology] Immunological immaturity due to the rearing environment (SPF environment)

The experimental animals used in tests are raised in an extremely clean SPF (Specific Pathogen-Free) environment in a near-sterile condition, which results in their immune systems being very immature and uniform.

Compared to humans (actual patients), who are exposed to various environmental factors and commensal bacteria and exhibit significant individual differences, animal models can be said to have an overly simplified immune response. This difference in "immune system reality" is the biggest reason for the lack of efficacy (dropout) in Phase II, especially in diseases like MASH where immune cells cause chronic inflammation leading to fibrosis.

Bridge the translational gap
Approaches to obtaining "human-like data" in non-clinical studies

Specifically, to avoid dropouts in Phase II, the key is to "acquire data that is as close as possible to the human physiological environment" in the early preclinical stages. To increase the success rate of drug discovery, it is important to verify "three approaches" in the preclinical stage when conducting preclinical studies, and it can be said that CRO selection needs to be done strategically.

Here are three checkpoints for your reference.

1. [Verification of Efficacy] Is it an "advanced disease model" with a proven correlation to human pathological conditions?

Confirming "advanced disease models" is an approach to bridge the "discrepancy between artificial models and human chronic diseases" that causes the translational gap. It is essential to select advanced models that can faithfully reproduce the human disease progression timeline and tissue structure, rather than catalog models that simply involve fat accumulation or temporary inflammation.

Specific checkpoints include the following:

• Is a parallel evaluation with clinical data possible?
  For MASH, the key points are whether scientific evidence of correlation has been established with clinical diagnostic criteria such as "NAS (NAFLD Activity Score)" and "fibrosis stage (Stage 0-4)," as well as gene expression patterns.
• Check the accumulated amount of background data.
  Focus on whether the CRO has a historical data foundation that can ensure stable reproducibility by reducing data variability due to feed composition and animal age.

2. Extrapolation to Humans: Using Human Cells
Are you combining 3D tissue models (organoids/MPS)?

Confirming the integration of "3D tissue models (organoids/MPS)" using human cells is an approach to completely eliminate the gap caused by "species differences." This technology, which was not possible with conventional 2D cell cultures, is utilized to reproduce "three-dimensional interactions between cells and the microenvironment."

Specific checkpoints include the following:

• Are biomimetic systems (MPS) being introduced?
  To confirm whether a screening system using "3D organoids" created from human iPS cells or patient-derived cells, and "Organ-on-a-chip" devices that can mimic blood flow and physical stimuli on a chip, is available.

3. Visualize Internal Dynamics: Capture "Intra-organizational PK/PD"
Is there high-precision bioanalysis technology?

This is an approach to visualize drug accessibility to target organizations and cells. Simply measuring blood drug concentrations (conventional PK) cannot determine whether a drug has penetrated deeply into diseased tissues that are highly degenerated or organized. Therefore, there is a need for technology that can proactively eliminate "uneven efficacy" in clinical practice.

Specific checkpoints are as follows:

• Microscopic imaging technology
  Regarding drug analysis in fine compartments within organs, such as "Mass Spectrometry Imaging (MSI)" and "3D imaging using tissue clearing technology," are there any methods that can be used for quantification?
• PK/PD analysis with a clinical perspective
  Verify that we have the analytical power to rigorously profile the correlation between tissue-level behavior and drug efficacy (in-tissue PK/PD), without solely relying on blood data.

Conclusion: Improving drug discovery success rates depends on selecting "strategic partners."

Currently, the success rate in drug development has fallen to between 1 in 20,000 and 1 in 30,000. In modern times, where "lack of efficacy" accounts for more than half of the reasons for dropout, the era of choosing a CRO solely based on criteria such as "low price" and "conducts studies as instructed" is over.

What's important is to choose a partner with whom you can discuss advanced approaches tailored to the modality of your company's candidate compounds and the characteristics of the target diseases, such as "chronic," "acute," or "immune-mediated," and thereby optimize the trial design.