Products

/Products
Products2019-10-08T11:37:45-04:00

GLIMPS (Graticule Life Image Machine Parsed Set)

Deidentified patient level data from Life Image includes a broad set of geographically diverse U.S. providers and a rapidly growing cohort of 140,000+ patients, which are summarized using the contents of multiple DICOM fields and free text SR (structured report) data. 

Each patient is linked to one or more imaging studies in the Life Image data. Key terms in ICD9 and SNOMED have been tagged at the study level for each patient to allow for identification of cohorts or summaries of the data set. Multiple NLP approaches were used to map medical terms from free text including cTAKES, MetaMap, and Regex. The GLIMPS data represents 360K+ studies and over 92,707,498 associated independent images that can be licensed through LIRE. 

Intended uses: 

  • Longitudinal data to identify markers of disease prior to diagnosis and progression during treatment

  • Supply machine learning teams with ‘control’ data from Life Image vs. ‘study’ data from a clinical trial or internal data source

  • Research into identification of novel imaging biomarkers for a rare or newly identified disease subtype

  • Validate AI models on real world data that have been built on small or narrow data sets

  • Design a new AI algorithm intended for deployment to radiology teams  within the Life Image HIE installed network or other imaging archive such as a VNA, PACS, or EHR

  • Support RWE teams with a novel source of data to support new approaches or techniques

Data dictionary and statistical sample available upon request. 

Note: Full text radiology notes are not included in the GLIMPS data set. These can be licensed upon request by selecting cohorts of interest or working with the Graticule team to build an extract.

Contact Us

LIRE (Life Image Research Extract)

Extracts with deidentified images/videos, free text notes, and DICOM headers from the Life Image data archive. Detailed extracts for projects provide a data mart to clients matching agreed-upon inclusion and exclusion criteria, data fields and DICOM modalities. LIRE extracts can also contain curation of specific data of interest such as values embedded in free text notes or classification of features in images to accelerate analysis.

Intended use: 

  • Longitudinal data to identify markers of disease prior to diagnosis and progression during treatment

  • Supply machine learning teams with ‘control’ data from Life Image vs. ‘study’ data from a clinical trial or internal data source

  • Research into identification of novel imaging biomarkers for a rare or newly identified disease subtype

  • Validate AI models on real world data that have been built on small or narrow data sets

  • Design a new AI algorithm intended for deployment to radiology teams  within the Life Image HIE installed network or other imaging archive such as a VNA, PACS, or EHR

  • Support RWE teams with a novel source of data to support new approaches or techniques

Contact Us

GLEAN (Graticule-Life External Acquisition Network)

Graticule and Life Image work together with health system partners to execute specific on demand requests to acquire data sets to match requirements. This process is accelerated by multiple business and technology investments made by Graticule and Life Image. The GLIMPS and LIRE data resources provide a model set of data that allows us to frame questions against smaller sets of real world imaging data to estimate feasibility and obtain results from a small sample to test analysis methods. These are inputs into a scale-up where Graticule engages with health systems and 3rd party data suppliers to build strategic cohorts that extend the Life Image data set. These may include linking Life Image data with a claims data set or consumer data set to obtain better visibility into patient journeys. It may also include collaborations with two or more health systems to source a broader set of images and related data, e.g. CSF (Cerebral Spinal Fluid) on a deidentiifed or consented cohort of patients.

Contact Us