TL;DR

Diagnostics firm Droplet Biosciences is using Nvidia’s GPU-accelerated Parabricks software to compress post-surgery cancer test results from four to six weeks down to 24 hours. The approach analyses lymphatic fluid collected during surgery rather than waiting for tumour remnants to appear in blood.

What Happened

Droplet Biosciences announced a collaboration with Nvidia to apply the chipmaker’s Parabricks platform — a GPU-accelerated software suite — to genomic data analysis for DNA sequencing after cancer surgery.

The company’s method detects residual disease by analysing lymphatic fluid collected post-surgery, rather than relying on traditional blood-based tests that typically require four to six weeks for tumour remnants to become detectable.

“By leveraging Nvidia Parabricks’ acceleration, we’ve been able to compress some of our most computationally intensive steps from more than a day down to just a few hours,” said Droplet’s chief scientific officer, Wendy Winckler.

Overall genomic analysis turnaround has dropped from 10 days to under five days, according to senior director of informatics Zhuosheng Gu.

Cost and Speed Benefits

Despite higher hourly costs for GPU compute, the dramatically reduced runtime results in lower overall cost per sample, the company said. Faster turnaround means patients can receive results while still in hospital, avoiding extra visits or long waits.

Droplet’s first clinical test targets HPV-negative head and neck cancer, validated under Clinical Laboratory Improvement Amendments standards. The company is a member of Nvidia Inception, an AI startup accelerator programme, and an Nvidia AI Enterprise customer.

Looking Forward

The collaboration shows how GPU-accelerated AI is moving from research settings into clinical workflows. If Droplet’s approach scales beyond head and neck cancer, it could set a new standard for post-surgical monitoring — turning what was previously a weeks-long waiting period into a same-day result.