 Home > Active Ingredient News > Drugs Articles > BeiGene Xiang Guo: Digital innovation starts from practicality and reshapes the new drug R&D ecosystem Encounter

BeiGene Xiang Guo: Digital innovation starts from practicality and reshapes the new drug R&D ecosystem Encounter

 Last Update: 2022-11-15
 Source: Internet
 Author: User

Tags

non potable water

chinese food new berlin

Search more information of high quality chemicals, good prices and reliable suppliers, visit www.echemi.com

PrevUnder Dr.
Zhirong Shen's introduction, we learned first-hand about the creation and growth of BeiGene's Translational Research and Translational Medicine Research Division (see BeiGene Shen Zhirong: Making translational medicine research a scout for new drug development).

If translational medicine is a pioneer in BeiGene's new drug development, then data science is the driving force behind BeiGene's future new R&D landscape
.
To that end, we will interview Dr.
Xiang Guo, Senior Vice President and Global Head of Statistics and Data Science at BeiGene, in the second part of this series.

Xiang Guo has extensive academic achievements and practical experience in the field of statistics and data science, and currently serves as the vice chairman of the Biostatistics Expert Committee of the Chinese Society of Clinical Oncology (CSCO), a member of the Global Statistics Community Committee of the Drug Information Association (DIA), a member of the Clinical Trial Professional Committee of the China Association for the Promotion of Pharmaceutical Innovation, and SCI Journal of Biopharmaceutical Statistics and Pharmaceutical Associate Editor
of Statistics.
Prior to joining BeiGene, Xiang Guo was the Asia Pacific Head of
Biostatistics at Merck Laboratories.

In the interview, Guo Xiang mentioned that the value of digital construction for drug research and development is becoming increasingly prominent
.
He is actively driving digital transformation in the field of new drug R&D, continuously introducing visualization, automation, intelligence and other technologies and working models into BeiGene's R&D department, and exploring the deep integration
of data science and drug innovation.

He proposed that digital innovation should not add too much additional burden of data collection to users, digital innovation should be closely integrated with specific businesses and solve business pain points, and new entry points should be explored for the realization of artificial intelligence in the pharmaceutical field, etc.
, these ideas are gradually exploring a feasible path
for the digitalization of new drug research and development through practice.

Dr.
Xiang Guo

Senior Vice President, BeiGene

Global Head of Statistics and Data Science

Since the 80s, regulatory agencies such as the US Food and Drug Administration (FDA) have used clinical research data as a key basis for
approval decisions.
This approach improves the scientific nature of regulatory decision-making, and also leads to a significant increase in the cost of new drug research and development, of which subject-related costs, research manpower and time investment are the main factors of
the increase in costs.
According to the financial report data disclosed by some listed pharmaceutical companies, in the international multicenter clinical study (MRCT), the average cost of enrolling a subject to complete the study is about 100,000 to 200,000 US dollars, and the labor cost accounts for more than 50% of
the total research and development expenses.

The introduction of digitalization is conducive to improving the efficiency of new drug research and development as a whole, shortening the research and development cycle, and controlling costs, thereby reducing drug prices and improving the accessibility and affordability of
drugs for patients.
"Digital innovation is an underlying technological revolution, which will reconstruct the new drug R&D ecology and profoundly affect the development and competitive landscape
of the pharmaceutical field.
" Dr.
Xiang Guo, Senior Vice President and Global Head of Statistics and Data Science at BeiGene, said
.

Guo always remembers that when he joined BeiGene in 2017, John V.
Johnson, Co-Founder, Chairman and CEO of BeiGene.
Oyler had a conversation with him: "As has happened in many industries in the last two decades, there will definitely be a top ten or even top five pharmaceutical companies in the world, not necessarily BeiGene, but if we try, the opportunity is there
.
" Guo Xiang said that one of the important reasons for him to leave Merck to join BeiGene is that company executives such as Ou Lei Qiang and Dr.
Wang Lai, the head of global research and development, highly recognize
the role of data science in new drug research and development.

Guo Xiang believes that the timing and volume of BeiGene's participation in R&D digital innovation is just right, and it is a rare opportunity
to provide the necessary resources without heavy historical baggage.
"We are creating a digital technology platform
that combines hardware and software.
In the future, data science will help new drug clinical research and more accurately evaluate the efficacy and safety of drugs.
At the same time, it realizes seamless connection between data and decision-makers, assists decision-makers in quickly screening and processing information, and improves the quality of decision-making; In addition, automated systems and integrated digital R&D platforms will significantly improve operational efficiency and lay a critical foundation for machine learning and artificial intelligence through data accumulation and user feedback, facilitating more modern
drug discovery.
”

Statistical innovation controls clinical research risks

At the end of clinical trials, in the hundreds of pages of reports that pharmaceutical companies submit to regulators, it is the estimates of the main efficacy indicators and the P values
of the corresponding hypothesis tests that are decisive for success or failure.
Therefore, in the context of digital innovation, BeiGene's Statistics and Data Science Division first applied statistical method innovation to assist clinical research to better grasp the P value and efficacy estimate.

Recently, BeiGene's zebrutinib achieved exciting results in the global Phase III ALPINE trial: "head-to-head" comparison of ibrutinib in patients with relapsed/refractory chronic lymphocytic leukemia/small lymphocytic lymphoma achieved a progression-free survival (PFS) efficacy result
.

“ 'Head-to-head' clinical trials require data science teams to have greater control over trial outcomes
.
This type of trial requires a reasonable setting of the test endpoint and test hypothesis
before initiation.
Guo Xiang mentioned that on the one hand, it is assumed that the difference between the two sets of data cannot be too small, and the numbers should have clinical significance; However, it should not be too large, if the difference in actual efficacy between the experimental group and the control group is 20%, and the trial assumes a difference of 30%, the final trial P value may be greater than 0.
05, and the statistical efficacy result
cannot be achieved.

Dr.
Guo Xiang participated in the internal event
to obtain approval for new indications for BeiGene's products.

At present, the competition in new drug research and development is becoming more and more fierce, and in order to seize market opportunities, save R&D time and reduce R&D costs, clinical research protocol design is still constantly innovating, but new designs often increase the uncertainty of research and increase the difficulty
of randomization and bias control.
Guo Xiang said: "This puts forward higher requirements for the sponsor's technical means and risk control capabilities, and requires data science innovation to serve as the gatekeeper of clinical research risk control
.
" ”

He gave some examples, such as in a number of completed and ongoing clinical trials, in order to speed up research and development, phase III trials
were launched in the absence of phase II trial results or before the phase II trial ended.
"For new target products, in addition to internal research data, there are not many external data that can provide reference for the hypothesis and endpoint setting of phase III studies, and clinical trials face greater risks
of uncertainty.
"

To this end, BeiGene's statistical and data science teams and clinical teams have adopted adaptive design
in multiple trials.

"We added medium-term invalidity testing, which can be terminated in time to protect the interests of participants once there is reliable evidence that the drug has not achieved the desired therapeutic effect; In addition, the content of preset protocol adjustment can be added to the experimental design, which can be released according to the internal phase II test results or external data during the trial, and the test hypothesis can be changed or the sample size can be adjusted
.
Guo Xiang said
.

In addition, sponsors must demonstrate to regulators that the new trial design does not increase the likelihood of observing effective results on ineffective drugs when adopting the new trial design that may increase the chance of
false positives or introduce more bias in trial results.

In 2021, BeiGene established a methodology research team to apply statistical methods such as simulation research and theoretical derivation to verify the rationality
of the new trial design in false positive control and bias control.

In 2021, the statistics department established a methodology research team to apply statistical methods such as simulation research and theoretical derivation to verify the rationality of the new experimental design in false positive control and bias control

For example, the team used simulation studies to deduce an adaptive design clinical trial, and the probability of achieving positive results would be less than the 5% threshold
set by the drug regulatory agency after repeated repeated simulations under the assumption that the drug would not be effective.
"Such studies can give regulators more confidence in approval, but the simulation study design process is very complex and requires a large number of influencing factors to consider," Guo said, "Sometimes the simulation study submitted to the regulatory authorities is even as
long as a small clinical study report.
" ”

In recent years, BeiGene's Department of Statistics and Data Science has brought in a large number of domestic and foreign data science talents
with rich experience in drug discovery and regulatory to support the development of new types of research.
For example, Dr.
Hong, Head of Statistics, was the Head of Oncology Early R&D Statistics at Johnson & Johnson before joining BeiGene; Dr.
Jingjing Ye, a leader in data science, was the review project leader of the FDA's biostatistics department, and led the statistical review of a number of new drugs, including zebrutinib.
Dr.
Lu Kaifeng, the leader of the statistical methodology team, has worked for Merck, AbbVie and other companies, and his ability in statistical theory research is excellent
.

Tian Hong

Ye Jingjing

Lu Kaifeng

"Their extensive experience can provide insight into risk and bias control for new study designs
.
Some problems are very challenging to solve in practice and require a lot of experience and technical means to accumulate
.
Guo Xiang said that the addition of top data science talents with international perspectives gives BeiGene enough courage to adopt new research designs to accelerate product clinical development
.

Connect data seamlessly with decision makers

As the statistics team matures, Guo has begun to devote more energy to BeiGene's data warehouse (Data) in the past two years Warehouse) and digital construction
.
The data science team under the department is also responsible for this work, and this number of personnel has been increasing
rapidly in the past year.

The data warehouse project will integrate BeiGene's cross-functional data, including clinical trial data, clinical operations data, clinical biomarker data, post-marketing safety and real-world data, and will further integrate preclinical research and other related data
in the future.
In Guo Xiang's plan, the data warehouse is expected to become a key infrastructure construction
to realize the digitalization, visualization, automation and artificial intelligence of innovative drug research and development in the future.

Data warehouses are expected to become a key infrastructure for the digitalization, visualization, automation and artificial intelligence of innovative drug R&D in the future

Dr.
Guo Xiang spoke
at the 2021 DIA China Drug Discovery Quantitative Science Forum.

"As the foundation of an integrated digital platform, the data warehouse is like a library library management system, where the user can quickly understand what data is in the warehouse, where and what kind of attributes
the data he needs.
" Guo Xiang said, "We will create data visualization dashboards according to the business needs of different departments, and use more intuitive methods such as a variety of chart displays, so that decision makers can obtain the required information
more completely, faster and with high quality.
" ”

Drug research and development has entered the era of information explosion, and it is difficult for decision-makers to grasp the information and various data
of dozens or even hundreds of similar competitive projects under research around the world with their own memory ability, as they did ten or twenty years ago.
He mentioned that in the future, data warehouses can integrate crawlers, search, and natural language processing technologies to instantly capture research design, data updates, and changes in market information for key competitive projects and push them to decision makers
.

"Based on this data and key information, the clinical R&D team can formulate an R&D strategy, and also get possible information updates in the execution of the strategy to quickly adjust the R&D plan
.
" Guo Xiang said
.

In the process of integrating data, Guo Xiang emphasized that the data science department should constantly connect with the business department to help the business team use digital innovation to improve decision-making quality and operational efficiency
.
"This is the key to
increasing the enthusiasm of all departments to participate in digital construction.
Digital innovation is not a castle in the air and cannot evolve into a burden for users, but strives to solve real problems
for different departments.
”

Digital innovation is not a castle in the air and cannot evolve into a burden for users, but strives to solve real problems for different departments

First, digital decision-making can save operational costs
for clinical research.
Clinical studies require the procurement of experimental drugs, but the previous method of estimating drug dosage was rough, and only calculated based on the expected number of patients enrolled in each research center
.
In the actual study, the dosage and time of each patient are different, and the patient enrollment time is earlier or later, and some patients will withdraw from the study
early.
Therefore, the roughly estimated amount of medication often differs greatly from the actual amount, resulting in problems such as
drug waste.

Through data integration, data science teams use the collected data to develop clinical drug procurement management tools
that dynamically predict drug usage in clinical research trials.

"We will dynamically measure the procurement demand
of drugs based on data such as patient enrollment speed, medication status, adverse events or withdrawal from clinical trials.
It is like a cistern that monitors the flow of water in and out at all times to keep it in dynamic balance
.
In the future, the tools under development can also trigger changes in drug procurement needs in real time based on protocol changes and test hypothesis changes
on the integrated platform.
"The procurement of experimental drugs in clinical research is a major expense, and 'head-to-head' study control drugs are especially expensive, and digitalization can improve the accuracy of drug dosage prediction and save research expenses
.
" ”

Second, digital decision-making can improve clinical research operational efficiency
.
Data warehouses contain some real-world data that can be analyzed and used to accelerate research project enrollment
.
"Through data comparison, we can understand in advance which research centers in past clinical trials have enrolled more patients in the same indication and population range, and the quality of execution is high, so that we can give priority to these centers for research and make trial patient enrollment more efficient
.
" Guo Xiang said, "Based on the data we provide, the clinical execution team can work
more targeted.
”

Similar digitalization can be used to improve the efficiency of
clinical research monitoring.
The system can monitor the data in real time, timely discover the abnormal values of the data, and provide early warning
for the obvious deviation of the data of a certain center from the normal value by comparing the data differences between different centers.
"Every data collected during the research process has an impact on the final P-value, so it is necessary
to find data anomalies in time and feed them back to the operations team.
" Guo Xiang mentioned that the drug regulatory department is also promoting risk-based trial monitoring (RBM, Risk Based Monitoring), but this can only be better achieved
if it is based on a digital decision-making system.

Automation lays the foundation for smart medicine

Another important part of BeiGene's digital innovation is to automate clinical operations, freeing researchers from tedious work while leveraging user feedback to lay the foundation
for machine learning and artificial intelligence.

According to Guo, BeiGene's business team is always pursuing process innovation to continuously improve work efficiency
.
For example, a phase III clinical trial, from database lock to completion of clinical trial report (CSR) charting, used to take about 4~6 weeks, but the company team has now reduced the time to 2~3 weeks
.
"This goal is achieved with the support of process improvements and new technologies, rather than relying on an unlimited increase in human input, where automation plays a huge role
.
"

Dr.
Xiang Guo participated in internal events
at BeiGene.

Similar automated processes will be used in the integrated digital R&D platform under construction at BeiGene for research
protocol writing and clinical trial electronic data acquisition system (EDC) design.
The machine can automatically design the first version
of the research plan according to the input keywords.
For example, by entering the keyword "neoadjuvant therapy for non-small cell lung cancer", the system can automatically parse the information stored in the data warehouse and similar research protocols scraped from various publicly available sources and generate the first version of the protocol
.
The company's clinical research protocol designers only need to make detailed modifications
to the version provided by the machine.

The automation system also receives continuous feedback – every modification made by the user in the research protocol and the reason for the modification are recorded as data to help improve the work of the automation system and make it more and more accurate
.
This is expected to further improve BeiGene's clinical operational efficiency and decision-making quality, as well as help explore and discover new AI rules
applicable to new drug discovery.

Guo Xiang believes that the realization process of artificial intelligence for the development of new drugs will not be achieved overnight, but requires continuous innovation and consolidation of underlying technology and data accumulation, so that artificial intelligence can be further refined in the pharmaceutical industry and may be directly applied to new drug discovery
in the future.

"The application of artificial intelligence in areas with clear rules and limited dimensions, such as playing Go, is very mature
.
However, human beings themselves have limited understanding of the rules for developing new drugs, and we cannot fully tell the machine the rules for screening new drugs, and new drug research and development does not allow artificial intelligence to use patient trial and error
.
Shopping sites can recommend diapers and beer packages to consumers based solely on the correlation of purchase behavior, and verify the causality
in subsequent consumer purchases.
However, it is impossible for new drug R&D companies to randomly recommend random combinations of two different target oncology drugs to patients, trying to optimize the recommendation system
through patient treatment feedback.
Therefore, Guo Xiang believes that if artificial intelligence such as machine learning is to play a role in the research and development of new drugs, we may need to achieve greater breakthroughs
in theory and technology.

"Artificial intelligence needs to find new entry points
into the field of new drug research and development.
At BeiGene, automating clinical operations and providing better data support to decision makers can be potential entry points
.
For pharmaceuticals, such an entry point is also in line with the needs of the industry, shortening the research and development time, accelerating the enrollment of subjects, and reducing the investment of research and development personnel, which are conducive to improving the efficiency of
new drug research and development.
Guo said that the faster a product enters the commercialization stage, the sooner the company can achieve revenue growth
.
By reducing R&D costs, new drugs have the opportunity to reach more patients in developing countries at better prices, which is itself a global goal
for BeiGene.

At BeiGene, automating clinical operations and providing better data support to decision makers can be potential entry points

"I expect BeiGene to truly integrate the genes of digital technology and drug innovation, so that data can be put in front of decision-makers anytime, anywhere, while using their feedback to drive digital thinking innovation and technological breakthroughs
in new drug development.
" With its exploration in statistics and data science, BeiGene is poised to become a pioneer and leader in the digitalization of new drug R&D globally
.
For the future, Guo Xiang is ambitious
.

This article is an English version of an article which is originally in the Chinese language on echemi.com and is provided for information purposes only. This website makes no representation or warranty of any kind, either expressed or implied, as to the accuracy, completeness ownership or reliability of the article or any translations thereof. If you have any concerns or complaints relating to the article, please send an email, providing a detailed description of the concern or complaint, to service@echemi.com. A staff member will contact you within 5 working days. Once verified, infringing content will be removed immediately.