Artificial intelligence (AI) is shaking up the healthcare industry. With applications in drug discovery, medical imaging, disease modeling, and clinical trial conduct, it promises to revolutionize the way we conduct research, treat disease, and work with patients.
In drug discovery, we’ve seen some awareness behind the hype and early demonstrations of AI-enabled target recognition and pipeline development. AI can also support diagnostic decisions in medical imaging, read scans with extreme speed and accuracy, and detect abnormalities that are invisible to the human eye.
At the same time, AI-based disease modeling can provide greater insight into the etiology, transmission, and progression of diseases such as motor neuron disease, cancer, and HIV. However, one of the most promising frontiers in the field is conducting clinical trials and improving the likelihood of regulatory or technical success.
Increasing the chances of a successful clinical trial requires careful tuning of several different factors, and clinical trial sponsors look for solutions that minimize time and maximize outcomes. The various operational and scientific decisions that need to be made during a clinical trial—from site selection to endpoint selection—can help reduce trial risk and lead to more successful outcomes. AI is increasingly being used to help research teams address some of the challenges they face—whether operational, scientific or ethical.
Artificial Intelligence Generates Actionable Operational Insights
From an operational perspective, trial sites may perform differently, particularly with regard to the speed and variety of patient enrollment. Through AI analytics, sponsors and contract research organizations (CROs) can leverage historical trial data or real-world data to gain a better understanding of site performance, allowing them to make more informed decisions about time and resource allocation.
This knowledge and oversight can shorten development time and ultimately benefit patients. This use of AI is especially important in the face of Covid-19, where AI has proven to be invaluable in rare disease and oncology trials, as it can help sponsors predict and base trials in real-time The backlog of insights makes a quick turnaround because of the influx of Covid patients. While still in its early stages, AI is being used to assess data on patient availability and diversity, enabling sponsors and CROs to de-risk their decisions in a competitive environment.
Scientific hypotheses can be stress-tested with artificial intelligence
The recipe for a successful trial requires a deep understanding of the disease in question, the patient population it affects, and potential treatments. Historically, this has been accomplished by reviewing the scientific literature and previous clinical studies.
AI is now being used to augment the intelligence supporting the trials. It allows us to improve protocols and accurately predict trial success by analyzing multiple sets of inputs, including historical trial design, drug biology, sponsor characteristics, and clinical trial results across development programs.
In particular, combining real-world data with clinical trial data can provide greater insight into patient outcomes and improve risk monitoring. It can also support decisions around endpoint selection and better equip sponsors and CROs to target the best and most clinically relevant endpoints. AI is also being used to flag real-time trends emerging in trials that might otherwise be apparent when all the data is analyzed at the end of the study.
AI supports more diverse trials
Another challenge that has long plagued clinical trials is the lack of diversity among trial participants. Addressing the underrepresentation of certain populations in trials is critical from a scientific and ethical standpoint. Studies that fail to target different races, ages, genders, and lifestyles will not result in effective treatments that are representative of patient populations.
AI can play a role in bridging this gap by identifying which trial sites are best suited to serve underrepresented communities. By simulating a patient model, certain conclusions and assumptions can be drawn about the proportion of patients in a subgroup who will respond to a particular treatment. This can inform how clinical trial teams consider diversity in recruiting and recruiting. However, those involved in developing and using AI systems need to pay close attention to dismantling rather than reproducing biases in the collection and use of data. This includes building models that can be translated into epidemiologically representative broad populations. As ever, regulation plays a role in shaping the approach to risk management, data sources and enforced transparency.
Synthetic control arm as a powerful data support tool
Synthetic Control Arm (SCA), also known as External Control Arm, is another innovative tool powered by big data, powerful computing and advanced analytics. While AI is used to simulate real life, SCA uses actual patient-level data and biostatistics to replicate the control arm, eliminating the need for a placebo group.
Similar to artificial intelligence, these advanced statistical methods and analyses require large amounts of data to accurately simulate real life. While well-established biostatistical methods may fall outside the definition of “artificial intelligence”, it is important to note that traditional methods combined with high-quality data have shown great promise and success in regulatory environments.
In addition to diversity, patient recruitment faces other challenges, notably the time pressure to recruit as quickly as possible, and the ethical implications of recruiting trial control groups in situations where effective treatments may not be available, such as in many rare diseases. Synthetic control arms create proxies for real clinical trial patient-level data and can provide representative datasets that provide valuable information about a disease, indication, or treatment.
Additionally, the model can be run iteratively, which means that dynamic datasets can be run through various analyses to model several different outcomes.A handful of synthetic control arm submissions have been Approved by the U.S. Food and Drug Administration, including a mixed design in a phase III trial for recurrent glioblastoma, a disease with few treatment options and unmet need. SCA is just one of many advanced analytical tools and statistical methods that have great potential in the clinical phase of drug development.
The untapped potential of artificial intelligence in clinical research
By harnessing the power of artificial intelligence, we gain greater insights into diseases, patient populations and potential treatments. Technology is changing the way we conduct clinical trials: it is improving elements of trial design, including target population selection, comparison arms, and clinical endpoints. It also improves patient safety and patient enrollment, and provides pharmaceutical companies with critical insights and analysis of how their drugs work. But we’ve only scratched the surface of what we can actually achieve. The potential is huge, and artificial intelligence will surely become an important part of clinical research and drug development in the future.
Photo: Blue Planet Studios, Getty Images
