Maximizing Efficiency in Clinical Trials for GERD Through AI

Efficiency in GERD clinical trials
Efficiency in GERD clinical trials

Gastroesophageal reflux disease (GERD) is a common condition that affects millions of people worldwide. Clinical trials are an essential component in developing new treatments and therapies for GERD. However, clinical trials can be time-consuming and expensive, with many challenges that can delay progress. An analysis of gastroenterology clinical trials found that over 20,000 trials were registered between 2007 and 2019, with over 7 million participants. The number of trials increased by 32.6% from 2007 – 2013 to 2014 – 2019. GERD is a complex digestive tract disease with a high worldwide prevalence estimated at 8 – 33%. Zhang et al. (2022) evaluated the latest global data on GERD (204 countries) to assess the overall burden of GERD – prevalent cases increased by 77.5% from 441.57 million in 1990 to 783.95 million in 2019. GERD has a spectrum of symptoms that overlap with other conditions making clinical distinction difficult.

In this article, we will explore strategies using AI for maximizing efficiency in clinical trials for GERD. By implementing strategies and modern technology for more efficiency, researchers can accelerate the pace of clinical trials and bring new treatments to market faster.

AI in Gastroenterology Practice and Research

To increase the efficiency of gastroenterology clinical trials and GERD trials in particular, artificial intelligence (AI) has been applied to different aspects, including diagnosis, classification of GERD, and prediction of GERD post-intervention. Machine learning (ML) is a field of AI, and deep learning (DL) is a subset of ML. In the last decade, the diversification of ML and DL has been sustained thanks to the advent of new devices, parallel computing, and multi-core graphics processing units. With advances in AI application in gastroenterology, researchers are increasingly adopting AI tools in clinical trials to increase efficiency in data reading and to realize cost savings.

AI in Digestive Endoscopy

AI has been adopted in digestive endoscopy, e.g., the application of AI algorithms significantly improved the adenoma detection rate (ADR) of polyps. In endoscopy, AI includes computer-assisted detection (CADe) and computer-assisted diagnosis (CADx). CADe supports colonoscopists as a “second pair of eyes” and is especially useful for beginner colonoscopists as it eliminates the need for immediate endoscopic reassessment, thereby reducing healthcare costs. AI in inflammatory bowel disease (IBD), for example, can eliminate the subjectivity in diagnosis according to the endoscopist and assess the disease severity for treatment decisions or clinical research purposes.

AI in GERD and GI Diagnosis

AI has been applied to clinical questionnaires for GERD. These AI‐based questionnaires are the ideal tool to timely and accurately diagnose GERD without invasive procedures (upper GI endoscopy and pH‐impedance) and avoid treatment delays. Visaggi et al. (2022) found that DL is capable of recognizing complex non‐linear patterns in datatypes (previously challenging to process), such as endoscopic images and videos. Further, DL algorithms can help reduce the number of questionnaire variables needed for a definite diagnosis, allowing clinicians to use shorter questionnaires that are more acceptable to patients and increase efficiency.

For the diagnosis of upper gastrointestinal (GI) diseases, AI, ML, and DL are increasingly being used in computer-aided diagnosis systems to improve the recognition and characterization of pathology. AI has also been applied to numerical and categorical data on upper GI pathology to automate and optimize the assessment of diseases, including GERD. The authors caution against limitations associated with DL models, including that such models cannot apply reason throughout the decision process.

Predicting GERD

While sleeve gastrectomy (SG) has proven to be effective in treating severe obesity, it has been estimated to increase the risk of adverse events, GERD in particular, when compared to other bariatric procedures. To predict the onset of de novo GERD after SG, Emile et al. (2022) developed and validated a machine-learning AI model which had excellent accuracy with an AUC (area under the curve) of 0.93. The model identified the top five parameters that substantially impacted the likelihood of postop GERD. Of the five, the technical factors were the size of the orogastric tube and the distance of the first stapler firing from the pylorus. Insights into this could improve the efficiency of medical device clinical trials concerned with postop GERD.

pH-Impedance Metrics for Diagnosis

pH impedance testing is the standard for confirming or excluding pathological GERD (not always conclusively). However, manual analysis of pH-impedance tracings is time-consuming. Impedance parameters (namely reflux episodes and post-reflux swallow-induced peristaltic wave (PSPW) index) augment the diagnosis of GERD. Wong et al. (2023) found that AI demonstrates high performance in measuring these impedance metrics and expects the saving of time, eliminating noise, and optimal diagnosis of GERD.

Classification of GERD

The Los Angeles classification system (LA-grading) helps assess the endoscopic severity of GERD. Ge et al. (2023) developed a five-category classification DL model based on the LA system. The authors found that the DL model could improve the accuracy of endoscopists in the LA-grading of GERD.

AI Limitations in Clinical Trials for GERD

On the role of AI in diagnosis, Visaggi et al. (2022) cautioned about the limitations of DL models where input data and output (diagnosis) are known. Still, the processes by which the diagnosis is arrived at are not. The authors highlight ongoing research to understand how DL models make decisions. Another study suggests that the high computational power of AI algorithms may lead to the risk of overfitting. Overfitting occurs when the model is too tightly fitted to the training data and does not generalize toward new data. They suggest following up with high-quality real-time studies to expand the early results found.

Gastroenterology CRO for Biotech, Powered by Technology

In conclusion, the use of artificial intelligence (AI) in clinical trials for GERD has the potential to greatly improve efficiency and accelerate the pace of developing new treatments for this prevalent disease. To further improve speed and efficiency, sponsors should work with CROs who have experience with GI indications such as GERD and who use modern technology and practices.

Vial was founded on a mission to reimagine clinical trials and deliver faster and more efficient trials. The Vial Technology Platform leverages connected systems and intuitive design to run global trials efficiently at scale. Vial’s eClinical Suite delivers high-quality data through connected data capture and review. Vial replaces paper source in trials — driving a significant change in trial workflows. Driven by process automatization, Vial’s Site Startup app enables lightning-fast onboarding. Sites can be activated in as few as 30 days. Vial’s Patient Recruitment Platform recruits patients across 15+ channels and converts them to randomizations — radically reducing enrollment periods. For more info on how Vial can help you with your next GI clinical trial, visit us at Vial Gastroenterology CRO or contact a team member today!

Contact Us

By submitting, you are agreeing to our terms and privacy policy
This field is for validation purposes and should be left unchanged.