Neuroimaging Based Clinical Trials are Failing. How Can You Fix Them?

There is a huge unmet need in treating patients with brain diseases; clinical trials fail and finding effective therapies is hard.

If your company is struggling or failing with its neuroimaging clinical development, you aren’t alone. One billion people worldwide suffer¹ from a heartbreaking array of neurological related disorders including cancers, Parkinson’s and Alzheimer’s Diseases, Multiple Sclerosis, stroke, epilepsy, psychiatric disorders, brain injury and more. Many of us have been touched directly or through the suffering of a loved one by this burden, and have discovered the sad truth that the therapy options are incredibly limited. This is because despite over 2000 neurological clinical trials starting each year in the US alone², only 5.9% of Phase I neurology clinical trial drugs are likely to make it cross the finish line to FDA approval.³ It’s an expensive business, with eighty percent of neurological drug development studies delayed by one month or more with up to $8 million per day in revenue loss.⁴ Some large pharma companies appear to be abandoning neurological research for more lucrative therapy areas, and may prefer to invest in biotechs working in the space.⁴ This all adds up to a shortage of effective treatment options for the patients who desperately need them, with very few new therapies entering the marketplace as compared to other disease areas.

Looking at the causes of this high failure rate reveals some surprising factors.

To begin with, a large proportion of biomedical research papers are not reproducible⁶, and few neurological biomarker algorithms developed by researchers are able to be applied to clinical trials or clinical practice. The lack of translation of breakthroughs from research labs to clinical trials to clinical practice contributes to a loss of therapeutic area progress.

But what are some of the factors at the core of the problem?

Scale, inefficiency and subjectivity make neurological clinical trials inherently difficult.

Images and data sourced for trials and studies come from multiple locations, different scanner modalities and vendors, and diverse IT systems and software. This results in slow, cumbersome and expensive manual processes. Clinicians visually examine images, which introduces unwanted variation, and can result in missed diagnostic details, misdiagnosis, and incorrectly quantified and contradictory conclusions.⁷ Human error in reading can add to clinical trial failure and fewer successful therapies reaching the market.

Neuroimaging is key to understanding and treating brain disease – but it’s extremely data intensive and complex.

Neuroimaging techniques such as Functional Magnetic Resonance Imaging (fMRI), structural Magnetic Resonance Imaging (MRI), Diffusion Tensor Imaging (DTI), Computed Tomography (CT), Dynamic Contrast Enhancement MRI (DCE-MRI), and Positron Emission Tomography (PET) are key tools to enable researchers and clinicians to “see” pathology inside the skull in real time and in rich detail. Brain images such as MRI raw images are comprised of hundreds of 2D black and white image “slices” of the brain. Details can be very difficult to detect and interpreted directly by human readers. Brain imaging analyses require a pipeline of complex pre- and post-processing steps to help transform the images in order to be interpreted by radiologists. The sheer volume of data is enormous, and requires large scale storage and high performance computing to digest the thousands of data points that exist in the raw images.⁸

In addition, there are myriad logistical and human factors that hinder certainty and performance in neurology clinical studies that can drive significant complexity and cost. These factors multiply with the number of sites, number of patients and the scans per patient.

• Massive volumes of neuroimaging data must be collected from individual trial sites and transferred to various storage solutions for archiving and analysis.  
• Highly variable and fragmented hardware and software solutions are often deployed across trial sites as well as other partners (e.g. CRO, sponsor, image readers), resulting in a significant and expensive IT integration and data management burden.
• Radiologists visually inspect a large number of images from different scanners, angles, resolutions, patient positions, and more; a time-consuming and subjective process that can undervalue the information in the imaging data as it is mostly used for detecting adverse events.
• The manually and highly intense approach increases risk of missed details, incorrect measurement, and results in a high level of radiologist fatigue and burn-out, all reducing diagnostic accuracy and consistency, and potentially leading to biased results and trial failure.

Clinical research leaders can take key steps now to transform their brain / neuro disease clinical trials.

Leaders can shift this paradigm by changing from the outdated methods of conducting neuroimaging related clinical trials to modern collaborative and intelligence driven platform approaches. This will result in the development of advanced biomarkers, more novel Artificial Intelligence algorithms for better, more timely interpretation of results, and ultimately improve trial outcomes for the benefit of all. 

Addressing the fragmentation and administrative challenges helps to increase speed and quality, reduce costs and improve trial outcomes.

Automatic collection, standardization and curation of data and seamless collaboration are some of the key attributes of an effective, 21st century approach to neuroimaging-based clinical trials.

• Deploying a fast, collaborative, infinitely scalable cloud-based solution that enables instantaneous transfer between partners. For example, deploying immediate and automatic quality checks for defects such as image artifacts that can be quickly addressed.
• Look for automated workflows with built-in privacy compliance, seamless data transfer, reduced number of scan readers required to greatly reduce costs.
• A solution with a simple, easy to use interface that changes per user profile, reducing human errors, automatic data harmonization, automated Quality Control and Assurance. 
• A reproducible and secure environment compliant with FDA 21 part 11, HIPAA, GDPR, ISO13485.

Should you digitally transform your clinical trials approach?

By conducting a simple root cause analysis you can determine the case for digitally transforming your neurology trials:

• Is there unneeded friction slowing the flow of data and information?
• Is there a burden of reading errors that could be reduced through AI tools and algorithms?
• Are you exposed to conformity or compliance risks that could be reduced through automation?
• Are you becoming an IT integration department by accident, including investing in non-core assets such as IT infrastructure, or expecting clinicians to become coders?
• Do you have the right partners that have real neurological expertise that you can work with to develop the right solutions for your disease area?

Examining these areas will enable you to quantify the benefits of digitally transforming, which when coupled with qualitative benefits should present a compelling case to drive the change needed.

Digital is transforming healthcare like never before. Now is the time for clinical trials, and particularly neurology imaging based trials, to reap the benefits of this transformation for the benefit of us all.

1. Neurological disorders affect millions globally: WHO report, https://www.who.int/mediacentre/news/releases/2007/pr04/en/
2. Evaluate Omnium, February 2021, Evaluate Ltd.; Total US Trials with 2020 Start Date: 2,352, Total US Trial with a 2021 Start Date: 563, Total Active US Trials in Neurology: 688; www.evaluate.com.
3. Clinical Development Success Rates and Contributing Factors 2011–2020, BIO.org report, https://www.bio.org/clinical-development-success-rates-and-contributing-factors-2011-2020
4. Is Pharma Running Out of Brainy Ideas? https://science.sciencemag.org/content/329/5991/502
5. As Big Pharma Abandons Neuroscience, Can Biotech Provide Answers? https://www.labiotech.eu/in-depth/neuroscience-big-pharma-biotech/
6. Reproducibility of science: Fraud, impact factors and carelessness https://www.sciencedirect.com/science/article/pii/S0022282817303334
7. The role of specialist neuroradiology second opinion reporting: is there added value? https://www.sciencedirect.com/science/article/abs/pii/S0009926007005144
8. Big Data and Neuroimaging https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5766007/