PhUSE CSS March Poster Ensuring consistency of SEND Datasets
SEND is currently the preferred submission format for the US FDA and became a mandate on 18th December 2017 . In SEND, all findings (MI-Microscopic Findings, MA-Macroscopic Findings, CL-Clinical Observations) as reported by Pathologists/Toxicologist are needed, these original as collected observations are split appropriately and standardized to controlled terminology and then mapped to separate columns in SEND domains. These Result or Findings as Collected (ORRES) contain terms including base pathological process, severity, modifiers and at times specimen type.
The Global standard terminologies and ontologies for pathological findings are often limited. INHAND (International Harmonization of Nomenclature and Diagnostic Criteria) initiative provides a standardized nomenclature and differential diagnosis for classifying microscopic lesions observed in laboratory rats and mice in toxicity. It is often restricted to rats and mice and, as of August 2016, the total coverage was only 78% (1247 terms) of standard nomenclature for microscopic lesions, in addition to the fact that it is often restricted to only rats and mice and still needs to be developed for other species (currently it is not the scope of the project2 ). The published terms are still insufficient and INHAND itself still needs to be added and rectified. There is no global standard terminology for clinical observations and macroscopic findings, and it is not even in the pipeline of development.
Similarly, another challenge is SEND accepts a framework of study design in TE-Trial Elements, TA-Trial Arms and TX-Trial Sets domains. The trial design of a SEND dataset is designed to be machine-readable, and it is very granular. It takes into account all variations in the trial arm of each dose group leading to more trial sets of fewer and more narrowly similar subjects having a common set of experimental and sponsor defined parameters. Whereas study reports generally have study design defined at dose/treatment groups. SEND also requires standardized way of explicitly stating certain decision rules that may not be seamless in the study Report.
Hence increased granularity of SEND trial design domains compared to sponsor dose group assignment and the need of explicit rendition of base pathological process and modifiers introduce challenges when sponsors trying to create SEND datasets. This also leads consistency and quality issues between the SEND datasets and Study Report.