Actionable intelligence for clinical pathology and diagnostics.
In a clinical dataset of approximately 80,000 samples spanning 20 cancer types, RNA sequencing identified 29.1% of actionable level 1 and 2 gene fusions that were not detected by DNA sequencing alone.
A single whole-slide image scanned at diagnostic magnification can consume between 1 GB and 5 GB of storage. Multiply that by a…
There is a call that comes across my desk with numbing regularity. A tumor-only NGS run finishes in the late afternoon, the…
Quality control in a non-waived clinical lab is one of those things everyone nods about in policy meetings and almost nobody…
A procurement committee reviewing companion-diagnostic platforms this quarter faces an inconvenient arithmetic. Across 20…
A positive Mycoplasma genitalium result is no longer the end of the diagnostic question. In many laboratories, it is the point at…
The central problem in pathology AI is no longer whether a model can recognize a pattern on a slide. Many models can do that. The…
A single aerosolized droplet, invisible to the unaided eye and measured in femtoliters, is sufficient to corrupt an entire batch…
A molecular test can be analytically sound, clinically useful, and fully operational in the laboratory—and still create…
I keep a sticky note on the side of my monitor that reads, simply: "Cycle 1, Day 1." It is a reminder, written for myself during…
Genotypic resistance assays are fast, focused, and very good at answering one narrow question: does this specimen contain a known…
The uncomfortable arithmetic of whole slide imaging begins with a two-digit number that laboratory directors have, for nearly a…
Two of the largest academic molecular pathology operations in the United States spent years and millions of dollars building…
I've been thinking about a particular bench scenario that keeps surfacing in conversations with colleagues, and it bothers me in…
A whole-slide image isn't a picture. It's a problem. One H&E slide from a routine prostate biopsy can hit 100,000 by 100,000…
I want to start with a scene that plays out in high-complexity labs more often than anyone wants to admit. A tech you've known…
The central mistake in oncology assay selection is treating genomic breadth as a virtue in itself. It is not. A broad…
A positive nucleic acid amplification test does not, by itself, establish active Clostridioides difficile infection. That…
A whole-slide imaging program can have excellent scanners, validated classifiers, and experienced pathologists—and still produce…
A randomized controlled trial published in JAMA Network Open in 2024 found that emergency department patients with suspected…
Between 60% and 70% of all laboratory errors originate in the preanalytical phase — the operational interval that begins with…
Turnaround time is a cost center. In advanced non-small-cell lung cancer, published clinical data demonstrate that plasma ctDNA…
A MALDI-TOF result can look definitive even when the organism in front of you is telling a different story.
There's a persistent fantasy in computational pathology that AI looks at a slide the way a pathologist does — holistically, all…
Actionable intelligence for clinical pathology and diagnostics.