Nucleo

Nucleo's core function is automatic segmentation of CT scans. It replaces the slice-by-slice tracing that manual analysis requires. Per Nucleo Research, this runs about 2,500x faster than the manual process. For a clinician, that's the difference between a task that shapes the day and one that runs in the background.

The platform detects and quantifies fat and muscle mass automatically, and it flags sarcopenia. Cancer teams care about this because muscle loss affects how patients tolerate treatment, which is exactly the kind of slow, hard-to-spot change that gets missed when nobody stops to measure it. The company reports state-of-the-art performance on sarcopenia detection in its own evaluation.

Nucleo measures tumor lesions in a CT scan and returns consistent numbers. Measurement consistency matters more than raw speed here. When sizing shifts between scans because two people traced it differently, it's hard to tell whether the tumor changed or the reading did. That's a real problem for treatment decisions.

Lesions get classified as target or non-target according to RECIST criteria. That's the standard oncology trials and many treatment decisions use. Automating that step means the classification follows a fixed rule set instead of a reader's judgment call on a busy day, and in a trial setting that consistency can matter as much as the measurement itself.

In validation studies cited by Nucleo Research, measurements from the platform agreed with medical experts 98% of the time. That number comes from the company. Treat it as a vendor figure, not an independent benchmark. It's still the claim most buyers will want to test.

Nucleo is built to slot into an existing oncology workflow rather than replace it. A doctor imports the scan and the software returns metrics. That's a deliberately small footprint, which lowers the change-management burden on a department already stretched thin. It also means adoption depends on how well it plays with the systems a hospital already owns.