The scariest thing about sepsis isn’t how it kills — it’s how long it hides. Its signs and symptoms are notoriously nonspecific, and there’s currently no reliable biomarker that clearly points to it. That blind spot makes sepsis one of the deadliest conditions to treat, and one of the most litigated. A new sepsis scan could change that. Failure-to-diagnose malpractice claims tied to sepsis are common in Washington, because the warning signs so often arrive late.
Researchers at Heidelberg University Hospital and the German Cancer Research Center just published a study in Science Advances aimed at that blind spot. Their tool skips vitals and labs. It reads sepsis risk straight off the skin, using nothing more invasive than a camera.
The device uses hyperspectral imaging — a camera tuned to light wavelengths the human eye can’t see. Aimed at a patient’s palm or fingers, it picks up subtle shifts in blood, oxygen, and water content in the skin, signs of the microcirculatory changes that occur early in sepsis, as small blood vessels become more permeable and leak fluid into surrounding tissue.
No lab draw. No wait. Just a scan and an answer in seconds.
The team scanned every adult admitted to the surgical ICU over roughly 14 months, 508 patients in all. Sepsis status was confirmed for 437 of them: 129, or 30%, had sepsis, most often tied to an abdominal infection. Among 483 patients with 30-day follow-up, 14% died. Mortality climbed sharply with severity, from 6% among patients without sepsis to 27% with sepsis and 49% with septic shock.
Why the Sepsis Scan Matters in Court
Sepsis cases rarely come down to whether a hospital knew treatment mattered. They come down to timing. Symptoms present as vague. Staff run the usual checks. The infection spreads while everyone waits on labs, or waits for the picture to get bad enough to act on. By then, organ damage has often started.
That timing gap is why delayed sepsis diagnosis appears so often in medical negligence claims here. Washington’s standard-of-care doctrine doesn’t require perfection — it asks whether a provider acted as a reasonably prudent practitioner would, given what was knowable at the time. A hyperspectral scan isn’t part of that calculation yet. But as tools like this prove out and become standard, what counts as “knowable” could shift with them. A fast, cheap sepsis scan could end up doing exactly that.
Washington ICUs, from Harborview to UW Medicine to MultiCare, face this challenge daily. The gap isn’t effort — it’s biology outpacing observation. Human pattern recognition is inconsistent, and confirmatory lab work takes time a septic patient may not have. A sepsis scan that surfaces risk in seconds, before instinct even kicks in, closes that gap.
No scan replaces clinical judgment, and none will end malpractice on its own. But it sharpens a question hospitals will eventually face: once a faster, cheaper way to catch a deadly condition exists, how long can waiting to adopt it stay defensible?
