NCI Hosts FDA/NIH Workshop on Resource Requirements in Next Generation Sequencing and Radiomics

The NIH hosted the FDA-NIH Next Generation Sequencing and Radiomics Workshop: Resource Requirements for Acceleration of Clinical Applications Including AI on September 29-30, 2021 (agenda). The FDA-NIH Next Generation Sequencing and Radiomics Working Group organized this meeting for the scientific community to discuss existing resource gaps and their impact on NGS and radiomics as well as strategies to accelerate high-quality research, development, validation, and regulatory science.View the workshop recordings (Day 1; Day 2).

Background

Demonstrating analytical validity and concordance of results from different methodologies or between testing sites is critical to ensure reproducible science and generalizability to the clinical setting. Furthermore, well-characterized reference materials are essential for methods development and validation; in the NGS and radiomics fields, these resources are insufficient. This is especially true for NGS where highly characterized, broadly available, and sustainable reference materials are lacking. Annotated radiomics datasets of sufficient quality and associated clinical metadata to improve the application of AI tools to radiomics are also needed.

The Office of the Director, NIH, and the FDA jointly published two Requests for Information (RFIs) related to this workshop. The RFIs include details on the background, areas of interest, and where to view the responses (the comment form is now closed).

Workshop Goals

• Identify and characterize critical resource gaps in NGS and radiomics, especially relating to reference standards
• Explore the landscape of existing resources that might be leveraged to accelerate growth in these fields
• Identify opportunities to support NGS and radiomic tool development and validation, especially those using AI and ML

Workshop Highlights

The workshop consisted of a keynote presentation by Anthony Kerlavage, PhD, NCI, on the NCI Research Data Commons and two plenary sessions on genomics and radiomics.

Discussions at the workshop were largely focused on processes and methodologies and emphasized the need, both in NGS and radiomics, for reference materials, especially highly annotated datasets for AI/ML applications.

Genomics

• The need to scale and centralize expert knowledge of clinically relevant variants
• Supplementing physical reference materials with the generation of in silico datasets, with a special emphasis on the importance of copy number variants and other structural rearrangements

Radiomics

• The need for ML-ready annotated datasets that reflect real-world complexities, and standards to which these datasets should adhere
• Directing resources towards existing best-practices operations to reach a broader audience with standardized expertise and technologies

FDA-NIH Joint Leadership Council Next Generation Sequencing and Radiomics Working Group Members