Fit-for-Purpose Real-World Imaging Data: The Difference Between More Data and Better Data
Reading time /
7 minutes
Data & AI
TL;DR
- Fit-for-purpose RWiD is built for a specific clinical question, not just scale.
- Five essentials: validated, relevant, reliable, standardized, and measurable.
- Imaging is the clinical reference not a supplemental data source.
- Essential for oncology, neurology, and cardiology research.
- Segmed delivers longitudinal, multimodal imaging data linked to clinical and molecular insights.
- Built for regulatory-grade real-world evidence generation.
Introduction
Real-world imaging data (RWiD) is reshaping drug development and clinical research. Data volume alone does not determine value in these sectors. A dataset with millions of scans can still be unfit for a specific regulatory or clinical purpose. The question is not how much data you have available. It is whether that data has been validated for your exact context of use (COU). This article explains what “fit-for-purpose” means in the context of real-world data. It outlines five essential criteria for RWiD validation. And it shows why imaging is irreplaceable in oncology, neurology, and cardiology and what happens when it is absent.
Related: Segmed White Paper - A Framework for Fit-for-Purpose Real-World Imaging Data
What Does “Fit for Purpose” Really Mean?
“Fit-for-purpose” real-world imaging data meets a validatedstandard for a defined context of use (COU).1A COU can be a regulatory submission, a clinical trial, or a real-worldevidence (RWE) study. “Fit-for-purpose” is not the same as “high-quality” ingeneral terms.1 Data must be valid,accurate, and relevant to the specific question being answered. While there isvalue for generic or prebuilt datasets, they often lack the clinicalspecificity required for specialized research. Therefore, “fit-for-purpose”datasets are built to answer specific questions. The concept is grounded inguidance from the FDA, the Duke-Margolis Center for Health Policy, and theSPIFD (Structured Process to Identify Fit-For-Purpose Data) framework.1,3
Five Criteria That Define Fit-for-Purpose RWiD
1. Context-Specific Validation
Imaging biomarkers must correlate with underlying biological processes and clinical outcomes.2 Validation must be demonstrated for a specific context, whether drug development or clinical diagnosis. Without it, imaging data cannot be trusted for clinical decision-making.
2. Relevance and Reliability
Relevance means the dataset includes key variables: exposures, outcomes, and covariates for a representative patient population.1,3 Reliability depends on accuracy, completeness, provenance, and traceability. Both are strengthened by linking imaging to structured EHR data.
3. Standardization and Harmonization
In multicenter studies, acquisition, processing, and annotation must follow consistent protocols.2 Variability in equipment and imaging settings can undermine pooled analyses. At the same time, real-world variability improves generalizability and AI model robustness.
4. Quality Dimensions
Fit-for-purpose data must be accurate, timely, and complete.4 Accuracy ensures data reflects the variable being measured. Completeness ensures sufficient coverage to support valid inferences across trials, RWE studies, or regulatory submissions.
5. Structured Frameworks
The FDA’s BEST (Biomarkers, Endpoints, and other Tools) initiative provides a structure for evaluating biomarker data quality. The SPIFD framework offers a stepwise method to assess validity, reliability, and operational readiness.3 These tools reduce uncertainty and support regulatory acceptance of RWiD.
Why One Data Modality Isn’t Enough
No single data source captures the full picture of patient health. Imaging reveals anatomy, pathology, and function. EHRs add diagnoses, medications, and outcomes. Molecular data provides a biological context. Combining these sources enables precise cohort definition and richer disease characterization. Without EHR linkage, even technically valid imaging data maybe insufficient.
Evidence Snapshot: Segmed NSCLC Case Study
Segmed sourced CT scans at two RECIST-aligned timepoints for a retrospective non-small cell lung cancer study. These were linked to PD-L1 expression, immunotherapy type (pembrolizumab or nivolumab), chemotherapy details, treatment response, and survival outcomes. This multimodal dataset enables radiomic-molecular biomarker discovery in real-world patient populations bridging imaging phenotypes with clinical outcomes at scale.
Imaging Requirements Across Therapeutic Areas
Fit-for-purpose RWiD looks different in each specialty. Imaging modality, longitudinal depth, and clinical linkage requirements vary significantly by indication.
Oncology: RECIST 1.1 standardized tumor response measurement across trials.5 iRECIST was later developed because standard criteria misclassified immunotherapy responses, mistaking pseudo progression for true disease advancement.6 Without standardized longitudinal imaging, oncology evidence is unreliable.
Neurology: Meta-analyses show that 10–30% of clinically diagnosed Alzheimer’s patients are amyloid-negative on PET imaging.7 The IDEAS study found that amyloid PET changed clinical management in 60% of over 16,000 real-world patients.8 Symptom-based diagnosis alone is insufficient for modern neurological research.
Cardiology: HFrEF and HFpEF are distinct types of heart failure, yet claims data routinely groups them together. Only echocardiography and cardiac MRI can reliably classify these phenotypes.9 Without imaging, cardiology datasets cannot support regulatory-grade evidence generation.
How Fit-for-Purpose RWiD Supports the Life Science Value Chain
Fit-for-purpose RWiD creates decision-relevant value at every stage of drug development.
Discovery: RWiD enables researchers to understand natural disease history in routine care. It supports biomarker identification using real clinical phenotypes that are not controlled by trial populations.
Translational and Preclinical Research: Molecular imaging characterizes target engagement and PK/PD relationships. Fit-for-purpose RWiD extends these insights to diverse, population-scale cohorts.
Clinical Development: RWiD informs trial design, supports feasibility assessment, and helps refine eligibility criteria. It can also support external control arms where randomization is not feasible or ethical.
Regulatory Submissions: FDA guidance emphasizes data provenance, completeness, and alignment with a defined COU.10 Appropriately curated multimodal RWiD can contribute supportive evidence across safety and effectiveness evaluations.
Post-Approval: Longitudinal imaging cohorts linked to outcomes support comparative effectiveness studies, long-term monitoring, and payer decision-making particularly in rare diseases and precision medicine.
Delivering Imaging Data That Fits the Question
Segmed curates longitudinal, multimodal imaging datasets linked to clinical, molecular, treatment, and outcomes data. This approach captures real-world heterogeneity across patient populations, imaging protocols, and care pathways.
Segmed’s database includes 150M+ de-identified radiology exams covering 31M+ patients across 2,800+ sites on five continents. All data is HIPAA-compliant and regulatory-grade.
"Real-world imaging and multimodal data only create value when it is fit for a defined decision. At Segmed, we curate imaging datasets around the exact clinical and research questions our partners need to answer, not around convenience or scale alone." — Christian Evans, VP of Business Development, Biopharma and Partnerships, Segmed
Ready to explore fit-for-purpose RWiD for your research program? Connect with us to discuss your data requirements or submit a project feasibility inquiry to evaluate modality and cohort availability for a specific program.
Frequently Asked Questions - F.A.Q.
What does “fit-for-purpose” mean in real-world imaging data?
Fit-for-purpose real-world imaging data (RWiD) is validate and curated for a specific context of use (COU).1 The data meets defined inclusion and exclusion criteria for a regulatory submission, clinical trial, or RWE study. General or prebuilt datasets often lack the clinical specificity required for a defined research question.
What is the difference between a prebuilt dataset and a fit-for-purpose dataset?
A prebuilt dataset is assembled for general use and is not optimized for a specific clinical question.1,3 A fit-for-purpose dataset is deliberately curated using defined criteria aligned with a research or regulatory objective. Fit-for-purpose datasets offer greater cohort precision, validity, and regulatory traceability.
Why is imaging data more valuable than claims or EHR data alone in specialty research?
Imaging captures information that claims, symptoms, and EHR data cannot represent. In oncology, imaging defines tumor burden and treatment response.5 In neurology, molecular imaging confirms pathological diagnosis.7,8 In cardiology, echocardiography and cardiac MRI classify disease phenotypes.9
Why did early Alzheimer’s drug trials fail to show efficacy?
Many early Alzheimer’s trials enrolled patients without confirmed amyloid pathology. Studies show 10–30% of clinically diagnosed AD patients are amyloid-negative on PET imaging.7 Including these patients diluted treatment effects and obscured signals from the drug’s true target population.
Can fit-for-purpose RWiD support FDA regulatory submissions?
FDA guidance on real-world evidence emphasizes that data should be fit for purpose, with documented provenance, quality, completeness, and reliability appropriate for the intended regulatory use.¹⁰ When curated to meet these standards, multimodal RWiD can contribute supportive evidence for safety and effectiveness evaluations.
What is multimodal imaging data and why is it essential?
Multimodal imaging data combines medical images CT, MRI, or PET with EHRs, pathology, molecular data, and outcomes.1,2 This integration enables precise cohort definition and richer disease characterization. Single-source imaging dataset slack the clinical context needed for meaningful analysis.
How does Segmed ensure datasets meet regulatory standards?
Segmed curates datasets with defined data provenance, documented quality measures, and alignment with a specific context of use.10 All datasets are reviewed by physicians and subject-matter experts. Segmed’s data is HIPAA-compliant, de-identified, and has supported regulatory submissions in drug development and medical AI validation.
References
1. Stafkey D, Lokhandwala T, Bakshi S. Fit-for-purpose real-world data: an integral component of evidence planning. Value Outcomes Spotlight [Internet]. 2025 Jul-Aug [cited 2026 Jul 3];11(4). Available from: https://www.ispor.org/publications/journals/value-outcomes-spotlight/vos-archives/issue/view/real-world-evidence-in-healthcare-decisions/fit-for-purpose-real-world-data--an-integral-component-of-evidence-planning
2. Eertink JJ, Bahce I, Waterton JC, Huisman MC, Boellaard R, Wunder A, et al. The development process of fit-for-purpose imaging biomarkers to characterize the tumor microenvironment. Front Med. 2024;11:1347267. doi:10.3389/fmed.2024.1347267.
3. Gatto NM, Campbell UB, Rubinstein E, Jaksa A, Mattox P, Mo J, et al. The structured process to identify fit-for-purpose data. Clin Pharmacol Ther. 2022;111(1):122-34. doi:10.1002/cpt.2466.
4. Lerro CC, Bradley MC, Forshee RA, Rivera DR. Evaluating fit-for-use oncology real-world data for regulatory decision making. JCO Clin Cancer Inform. 2024;8:e2300261. doi:10.1200/CCI.23.00261.
5. Eisenhauer EA, Therasse P, Bogaerts J, Schwartz LH, Sargent D, Ford R, et al. New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur J Cancer. 2009;45(2):228-47. doi:10.1016/j.ejca.2008.10.026.
6. Seymour L, Bogaerts J, Perrone A, Ford R, Schwartz LH, Mandrekar S, et al. iRECIST: guidelines for response criteria for use in trials testing immunotherapeutics. Lancet Oncol. 2017;18(3):e143-52. doi:10.1016/S1470-2045(17)30074-8.
7. Ossenkoppele R, Jansen WJ, Rabinovici GD, Knol DL, van der Flier WM, van Berckel BN, et al. Prevalence of amyloid PET positivity in dementia syndromes: a meta-analysis. JAMA. 2015;313(19):1939-49. doi:10.1001/jama.2015.4669.
8. Rabinovici GD, Gatsonis C, Apgar C, Chaudhary K, Gareen I, Hanna L, et al. Association of amyloid PET with subsequent change in clinical management among Medicare beneficiaries with mild cognitive impairment or dementia. JAMA. 2019;321(13):1286-94. doi:10.1001/jama.2019.2000.
9. Lang RM, Badano LP, Mor-Avi V, Afilalo J, Armstrong A, Ernande L, et al. Recommendations for cardiac chamber quantification by echocardiography in adults. J Am Soc Echocardiogr. 2015;28(1):1-39. doi:10.1016/j.echo.2014.10.003.
10. U.S. Food and Drug Administration. Considerations for the use of real-world data and real-world evidence to support regulatory decision-making for drug and biological products [Internet]. Silver Spring (MD): FDA; 2023 [cited 2026 Jul 3]. Available from: https://www.fda.gov/regulatory-information/search-fda-guidance-documents/considerations-use-real-world-data-and-real-world-evidence-support-regulatory-decision-making-drug
Related Resources
Segmed White Paper - A Framework for Fit-for-Purpose Real-World Imaging Data.
Medical imaging: an essential component of real-world data.
Multimodal data pipelines: the new gold standard in pharma research.
FDA expectations for real-world imaging data: what makes RWE regulatory-grade.