AI-based Digitalized Clinical Trials – The In-Silico style




For the healthcare sector, clinical trials cost billions of euros and often proceed for years without any guarantee of success after the study. It is a method to assess a drug or medical device's effectiveness in a clinical setting but is very expensive due to complicated clinical protocol and strict record-keeping. Artificial intelligence techniques are increasingly being tested to remove some of these heavy organizational liftings and have resulted in many In-silico experiments.


The term In-silico, which in Latin means 'in silicon,' was coined in 1987 as an expression denoting biological experiments conducted on a computer or through computer simulation and was first used in 1989. In-silico tests and virtual pharmacological therapy take advantage of human-based modeling and simulation technologies. This methodology has been used for modeling and simulation in both the pre-clinical trials and clinical evaluation of a future medical device since its inception, taking this broad spectrum of applicability into account.


Virtual Patients - Will patients be replaced by computers?


Many startups are trying to digitize medical research using big data and the internet of things. AI healthcare startups are providing platforms to apply machine-learning algorithms to automate clinical trials. Boston based TriNetX is one such company that offers services around clinical trial design and patient feasibility. Its software collects and analyzes patient health records to identify ideal candidates for clinical trials. TriNetX enables researchers to analyze the patient population and perform 'what-if' analyses in real-time, thereby developing virtual patient cohorts that can be re-identified for potential recruitment into clinical trials. And all this can be done in minutes!


The clinical analytics platform Saama from Silicon Valley uses its Life Science Analytics Cloud to seamlessly integrate, curate, and animate unlimited sources of structured, unstructured, and real-world data to deliver more actionable insights. Saama aligns key stakeholders who are either part of, or involved with, clinical research teams, thereby providing its sponsors and CROs services such as data aggregation for faster decision-making and more effective collaboration.


Irish startup Teckro simplifies clinical trials by using machine learning to digitize paperwork, enabling researchers to collaborate on smartphones from anywhere. This company provides a mobile platform that allows clinical researchers to quickly retrieve information and control various clinical protocols using the portal.


Another startup is Owkin that helps companies design better clinical trials using AI. Owkin's platform, Socrates, uses machine learning to integrate biomedical images, genomics, and clinical data, among other sources, to discover biomarkers and mechanisms associated with diseases and treatment outcomes. A recent publication in Nature Medicine by Owkin reports a deep-learning program trained on 3,000 patients with an aggressive form of lung cancer that enabled the company to develop a predictive model for lung cancer.


Good news for the animal lovers!


San Francisco based startup VeriSIM Life uses AI to do away with testing new drugs on animals. This biotech is developing AI-powered biosimulation models that can quickly predict how a drug will interact with an animal's biological system. This means pharma companies could accelerate the pre-clinical phase and move on to the human clinical trials.


Columbus, an Ohio-based startup, owns an AI platform, Virtual Imaging for Pathology Education and Research (VIPER), that is leveraging all of the cloud-based data and its network of partner institutions to flag eligible patients at the time of their diagnosis to fast track enrollment in clinical trials of cancer treatment. Texas-based Litmus Health is involved in taking data from wearables, sensors, and other smart devices and turning them into medically relevant insights. Its platform uses machine learning to detect patterns in the data based on participant behavior and responses. There's also a dashboard for real-time monitoring of the clinical study.


Virtual patients


Novadiscovery aims to reduce the enormous cost of clinical trials by running the trials first in-silico. This saves costs by optimizing the design, dosage, timing, and patient selection before running the real trial. They help clinical trials directly from phase I to phase III of clinical trials and reduce the size of the phase III trial by focusing on optimal responder profiles. This can save up to one year a few million dollars. The company's technology Jinkō uses real data collected from scientific studies regarding disease pathobiology and then uses drug data from existing studies to model its effect on patients. This lets users predict the clinical trial outcome.


We are just at the beginning of the growth of in silico trials. The FDA is already using simulations in its regulatory strategy to accelerate the progress of drugs to market, such as in the safety of vaccines for children. For patient-specific computer models, this is currently an area of intense regulatory science research. Computer models of disease progression and treatment response can represent each physical individual (digital twin) or a hypothetical individual whose key characteristics (described by the inputs of the model) are sampled from the joint distribution of a representative population (digital trials).


Digitalizing clinical trials can be used to predict how an individual patient will respond to certa