7 Imaging Data Management Strategies to Accelerate Medical Device Innovation

Medical device companies have huge amounts of data, but imaging data is often not configured to support R&D efforts. Here’s how to change that.

Jim Olson, Flying

Data is the lifeblood of medical device companies’ R&D groups. However, unlocking the full potential of your organization’s data assets is not easy, especially for complex data such as medical imaging.

Medical imaging assets are extremely valuable to R&D efforts, but they are often disorganized and lack consistent labeling. This means that before they can be used for analytics and/or for machine learning or AI applications, they need to be standardized and made accessible.

But maintaining complex medical imaging data poses a major challenge for many organizations. Even after data is organized, existing infrastructure and research processes can continue to hinder success and, ultimately, time to market.

To accelerate and scale medical device R&D, organizations must adopt scalable data practices and scalable data management solutions. To make the process easier, here are seven data management tips to help maximize the value of imagery data.

1. Extract data from silos

Information does not add value when it exists in isolation. A data silo occurs when data is not easily discoverable and cannot flow freely between departments, a common situation in life science organizations. In many cases, data may be held in different databases with different storage structures and conventions. A comprehensive data platform can mitigate these issues by centralizing data in a shared repository, with access controls and version history. A well-designed and maintained data platform gives organizations flexibility in how data is stored, accessed, and used, while providing users with greater visibility into available assets.

2. Clean up and standardize data upstream

The complex nature of medical imaging data means that researchers must leverage asset metadata to make it useful for large-scale projects. However, metadata conventions often differ between data sources, devices, and practitioners. Data scientists can standardize an organization’s conventions, both in the archive and as newly captured data arrives. Standardization may include, but is not limited to, standardized labeling for imaging modalities and/or body parts.

3. Ensure data hygiene practices can work at scale

The data normalization described above only makes sense if it can be applied in an automated way at the enterprise level. Manual curation, cataloging and organization of data, even if done by a trained team that adheres to agreed standards, is too time consuming and always carries the risk of inconsistency. Automating these processes as much as possible can avoid many challenges later.

4. Understand the data modalities and associated metrics used across the organization and ensure automation works for everyone

Your research teams can use image-based data such as DICOM and microscopy, time-series based data such as electroencephalography, and CSV (comma-separated variable) files or other self-contained text files. descriptive in their work. Even using the same modality, different analysis approaches and outcome measures may need to be tracked or captured. When designing a data management modernization approach, teams should consider every data modality, data type, and associated workflow they can integrate and enforce standardization for all.

5. Make sure you have an adequate volume of diverse data for AI training

The old adage “Garbage in, garbage out” is especially true for AI training. Models trained using an inadequate volume of data – or data that does not reflect the diversity of impact variables such as scan types or patient population – are likely to underperform. To avoid this, it’s important to leverage all available data within your company, but you can also seek to supplement your own data with publicly available datasets or licensed datasets from collaborators. In both cases, the need remains for consistent retention and use of data in validated workflows.

6. Take advantage of cloud-scale resources

Medical imaging data requires large amounts of storage and computing power. Relying exclusively on on-site resources can be both expensive and cumbersome. Leveraging resources at cloud scale, on the other hand, enables elastic compute infrastructure and more flexible storage. Organizations can create cloud instances as needed for unparalleled scalability.

7. Consider how to approach full provenance

Provenance (establishing a documented trail to the original data source and associated analytical processes and steps) is required for reproducibility and regulatory approval. Research teams should look for systems that can automate provenance, with the recording of access logs, versions, and processing actions. Automating this work not only removes the burden on researchers, but eliminates the risk of non-compliance and errors.

If your organization is struggling to leverage insights from medical imaging data, you are not alone. Fortunately, there are tools and resources available to automate and scale data capture, curation, and computation.

Embedding these data management tips requires an upfront investment of time and resources, but that investment can pay off with enriched data and accelerated innovation.

Jim Olson is the CEO of Flyinga biomedical research computing platform that harnesses the power of cloud-scale computing infrastructure to address the growing complexity of modern computational science and machine learning.

The opinions expressed in this blog post are those of the author and do not necessarily reflect those of Medical Design & Outsourcing or its employees.

How to Join the MDO Contributor Network