With Komprise, AWS S3 and PathAI, NewYork-Presbyterian is finding success with AI at the point of care in a bold
new initiative to improve treatment and outcomes.
Healthcare IT infrastructure team reduced cloud costs by 96% using the Komprise automated AI workflow that curates a small subset of files and then deletes cloud copies after 30 days. This approach pared down AWS storage from 1PB to a rolling 33TB.
Digital pathology has become one of the fastest growing applications of AI in healthcare, improving diagnostic accuracy and reducing turnaround time. However, digitizing medical images creates extreme data growth.
With Komprise, Everpure, AWS S3 and PathAI, NewYork-Presbyterian is finding success with AI at the point of care in a bold new initiative to improve treatment and outcomes. The solution has allowed IT to operationalize AI data ingestion with great economics, speed and accuracy for their constituents.
Read this case study to learn how Komprise curates and only sends the newest data to the cloud, at speeds 10x faster than previous data transfer methods in use at NewYork-Presbyterian.
More Komprise customer success stories
Learn more about Komprise for Healthcare Unstructured Data Management
Komprise AI Ingestion FAQs
How did NewYork-Presbyterian reduce cloud storage costs by 96% for its digital pathology AI program?
NewYork-Presbyterian was generating more than 2PB of new pathology imaging data every year, and the cloud data transfer tool in use was copying all source data to AWS S3, driving storage costs to nearly 1PB. By using Komprise to curate and ingest only the newest, relevant slides and automatically delete cloud copies after 30 days, the health system reduced its AWS S3 footprint from 1PB to a rolling 33TB. This surgical copy-and-delete workflow cut cloud storage costs by 96% and allowed the AI initiative to fund itself through savings, without being treated as a traditional infrastructure cost center.
How does Komprise support AI data ingestion for digital pathology without disrupting clinical workflows or exposing protected health information?
Komprise integrates with existing clinical storage environments including Everpure (Pure Storage) on-premises and Amazon S3 without requiring changes to clinical applications or PACS systems. For NewYork-Presbyterian, pathology slides are discovered, classified, tagged, and ingested to S3 for analysis by the PathAI application using automated Komprise Smart Data Workflows. The original data stays on-premises for long-term retention and regulatory compliance, reducing cloud security exposure significantly. The health system’s information security director noted the reduction of the cloud storage footprint from 1PB to 30TB as a major security improvement, since less data in the cloud means less PHI exposure.
What results did NewYork-Presbyterian achieve and what is next for AI data management with Komprise?
The results from the initial digital pathology AI workflow include 96% lower cloud storage costs, AI data ingestion speeds 10x faster than previous transfer methods, reduced PHI exposure in the cloud, and improved researcher access to clinical imaging files through metadata-based search of the Komprise Global Metadatabase. The team is now planning to enrich metadata for self-service search, integrate with Snowflake to merge unstructured data queries with structured data, and expand the AI ingestion workflow to new specialties including radiology. For healthcare IT teams evaluating AI data management for clinical imaging, the NewYork-Presbyterian case study is available here.
————–