One CT Scan, Several Readings: Checking an AI Lung-Nodule Result Against the CT Machine's Own Reconstruction Settings

An AI program reads a CT scan of the chest and reports lung nodules, with their size and a score. Today we see one answer from one way of building the pictures. This note describes a simple check that can sit on top of any such program. The patient is scanned once. From the raw data of that one scan, several sets of pictures are built by changing only the CT machine's own reconstruction settings, within the range the maker permits. The AI reads every set. If the answers stay together, the scan supports the answer firmly. If they spread out, the answer depends on a reconstruction setting and deserves a closer look. Other methods shake the picture with invented changes, such as flips or artificial noise. This method uses the settings the CT machine itself allows, so the uncertainty it measures is real, and it is read for each patient's own scan. One number, the spread value, tells which case it is. A computer-decided shunting action then moves the scan that needs care to the front of the radiologist's list, while all other scans carry on. The words used stay calm. It needs no second scan, no extra radiation and no new device, and the AI program is not changed. Patent pending. Keywords: CT; lung nodule; AI scan reading; reconstruction settings; uncertainty; spread value; shunting action