Case Study: Pathware achieves faster, AI-powered biopsy quality assessments with Labelbox

How Pathware is accelerating pathology by delivering AI-powered expert analyses

Pathware, maker of the BIOPTIC™ cloud-based sample management platform, faced the challenge of speeding up routine visual pathology workflows and creating large, high-quality training datasets. Pathologists traditionally spend roughly one hour assessing slides and one in five needle biopsies fail to collect sufficient tissue, so Pathware needed to label thousands of tissue clusters for machine learning but lacked web developers and training-data infrastructure to do it in-house. They turned to Labelbox for an annotation and training-data management solution.

Using Labelbox’s SaaS training-data platform, Pathware uploaded thousands of 5×5–12×12 image tiles and annotated them with bounding boxes and masks, used Labelbox ontology management to streamline external labeling, and verified adequacy with third‑party pathologists. Labelbox saved the 15-person Pathware team months of R&D, enabled labels that took roughly 10 seconds each, centralized their training data, and let BIOPTIC™ deliver biopsy-quality assessments in minutes with greater than 90% certainty—dramatically reducing clinician assessment time and improving confidence and scalability for future use cases.

Pathware

Jaron Nix

Chief Technology Officer