The Future of Pharma: Using Artificial Intelligence to Optimize Clinical Trials – Part One

Precision Pathology and Clinical Trials

Anatomic pathology, as an objective measure, can provide valuable insights in clinical trials. However, interpretation of pathology results by human eyes can be limited, time consuming, and in some cases introduce significant variability and subjectivity. Pathology review is used for trial design, patient recruitment, patient stratification, analyzing results, and monitoring patient response to treatment. Digital pathology has made the process of pathology review more efficient by significantly reducing turnaround time from tissue acquisition to pathology results. A versatile cloud-based digital pathology image management platform such as imageDx™, which allows groups of pathologists from anywhere in the world to review images simultaneously, eliminates the need to ship physical pathology slides around the globe and facilitates conducting multi-sites large clinical trials.

Precision pathology is a growing subfield of digital pathology that uses artificial intelligence (AI) to identify, extract, and quantify visual and sub-visual features and patterns in whole slide images. The greater accuracy, reproducibility and standardization associated with precision pathology offers many opportunities in improving clinical trials.  Optimizing clinical trial designs, delivery, and execution can add billions of dollars of value. A recent study by Mckinsey showed that reducing the clinical-trial enrollment timeline by 13 percent can save 10 to 15 percent in total external costs. Aside from the financial benefits, improving clinical trials’ efficiency leads to a shorter drug development cycle which translates directly into impacting human lives. Here we discuss the application of precision pathology in patient selection, stratification, and outcome measure in clinical trials. 

Patient Selection and Enrollment

Fig. 1 imageDx IHC digital assay; Ki67-stained breast cancer sample.

Clinical trials recruit participants based on a specific set of data that would include demographics (e.g. gender, age, ethnicity) and presence or absence of certain health conditions. Eligibility for clinical trials can involve pathology, where the presence or absence of certain health factors are evaluated based on molecular test results and pathologic findings. By upgrading time-intensive manual pathology assays to automated AI-powered digital assays using tools such as imageDx™, pathologists can make more accurate decisions at a faster rate and streamline their workflows.

AI-powered digital assays can assist pathologists in scoring IHC-stained whole slide images in a fraction of the time and with higher accuracy. An example of the results of such assays is shown in Fig. 1, where the imageDx™ IHC digital assay was applied to a Ki67-stained breast cancer sample. Various IHC scores (0+:yellow, 1+:blue, 2+:green, and 3+:red) are visualized using cell-by-cell masks; imageDx also provided detailed quantitative results including the number of cells in each IHC score group that could be downloaded as spreadsheets. 

Fig. 2. imageDx Tumor Profiling; Ki67-stained breast cancer sample

Conventionally, to assign a slide-level IHC score in a sample, a pathologist would count and record the number of cells with various stain intensity levels to calculate an overall slide-level score.  This tedious process would conventionally be performed for small regions of interest in a sample by a pathologist.  An AI-powered digital assay (Fig. 2) can identify all tumor regions in IHC-stained whole slide images automatically.  All the nuclei present in the tumor region can then be scored according to their stain intensity and other criteria such as localization of the biomarker or abundance of the biomarker in the membrane.  The results shown above were produced by the imageDx™ tumor segmentation digital assay that is often used in combination with an imageDx IHC digital assay to generate a slide-level IHC score, with a comprehensive breakdown of the IHC scores for all the tumor cells.

In Part Two of our series, Dr. Misagh Naderi will discuss patient stratification and results interpretation.

Learn more about how you can use AI to optimize a clinical trial.

About the Author

Dr. Misagh Naderi is the manager of AI and digital assays at Reveal Biosciences, where he serves as a liaison between domain experts in pathology, artificial intelligence, and biopharma. Misagh’s scientific expertise spans across computational biology, biomedical sciences, and chemical engineering with publications in protein modeling, in-silico drug design, and oncolytic-virology.

Prior to joining Reveal Biosciences, Misagh served as scientific advisor at two prominent international law firms, focusing on intellectual property and innovations at the intersection of biology, engineering, and artificial intelligence. He has a passion for science communication and helping companies leverage AI in providing solutions that would impact human health and wellbeing.

Pathology Revealed

Sign up to receive updates on the latest AI-powered pathology breakthroughs, access digital pathology resources, and more.