Artera, a precision medicine company developing AI-powered cancer treatment planning, delivers prostate cancer prognostic results in 1-2 days through its AWS-based ArteraAI Prostate Test — down from 6 weeks for traditional genomic tests — after the U.S. FDA granted De Novo authorization making it the first and only AI-powered software authorized to prognosticate long-term outcomes for patients with nonmetastatic prostate cancer, now recognized as an FDA-regulated software as a medical device (SaMD) and part of the NCCN Clinical Practice Guidelines since 2024.
"Artera was founded with the belief that there were a lot of signals in the histopathology image data that were not being used, but if an AI algorithm could be specifically developed with this in mind, you could radically change cancer patient care," said Nathan Silberman, Chief Technology Officer of Artera. "We've heard from patients who have said that because of our test, they were able to avoid unnecessary treatments with a lot of side effects," he added, describing the company's mission as giving clinicians as many data-backed insights as possible.
The challenge was twofold. Clinically, patients with localized prostate cancer were often overtreated or undertreated because no AI-based tool existed to help physicians match therapy to the individual patient's biology — traditional chemical assays measured only a small set of genes, consumed the original tissue sample, and took 6 weeks. Technically, Artera had to manage high-resolution biopsy images sometimes reaching 8 GB each, breaking them into tens of thousands of patches for model training, while serving millions of image patches to AWS servers for foundation model training — all under HIPAA compliance and multi-country data residency requirements.
Artera built a comprehensive AWS architecture: medical professionals access the Artera Portal through AWS Global Accelerator and CloudFront; within a VPC, Amazon ECS hosts the web portal containers while an Amazon EKS cluster runs AI/ML inference workloads analyzing biopsy images with computer vision models; Amazon EFS provides shared file storage accessible by both ECS and EKS for biopsy images; Amazon RDS manages patient records and diagnostic results; Amazon ElastiCache provides in-memory caching; and IAM, KMS, and CloudWatch handle access control, encryption, and monitoring. The workflow: biopsy images upload securely to Amazon S3, the EKS cluster orchestrates preprocessing, ML models trained and deployed on EKS access preprocessed images from EFS and store metadata in RDS, and results return to providers through the secure portal. Region-specific S3 buckets and EKS clusters maintain data locality for jurisdictional compliance.
Over 3.5 million prostate cancer survivors live in the United States. The test processes tens of thousands of image files per biopsy slide through ML workflows, completing in hours instead of weeks, and is implemented at qualified pathology labs across multiple countries. Artera plans to expand with a breast cancer product, analyze additional biomarkers, integrate genomic data with imaging analysis, and partner with major healthcare systems to embed diagnostics directly into clinical workflows — building toward a pan-tumor foundation model capable of assessing patient risk and therapy benefit across any cancer sample.