AI-Driven PET-CT Technology for Faster, Better Lung Health Assessments

A computer implemented method, system or programme for combining a three-dimensional PET scan and a CT scan of a body into a combined image for segmentation, comprising of:

• Co-registering each of the PET scan and the CT scan as an input with reference to each other. 
• Extracting metadata from the input scan and the reference scan relating to positional and orientation data. 
• Matching origin points of the input scans. 
• Matching the dimensions, orientation and axes of the input scans.
• Creating a co-registered PET scan and a co-registered CT scan as outputs.

Combining the co-registered PET scan and the co-registered CT scan into a single, resampled 3D image with a selected bias factor favouring the PET scan to preserve more PET features in the combined image. The combined image includes the organ of interest and was used for training a convolutional network model to perform lung segmentation. In the training process, FDG PET-CT scans from patients with pulmonary tuberculosis were predominantly used. As such, these included lesions with a vast spectrum of extent, and morphological and inflammatory characteristics, improving the robustness of the model.

With the training of the model completed, the method of combining the PET and CT scan has been included in a software pipeline in which the trained model performs segmentation of lungs on PET-CT images.

The pipeline produces as outputs accurate regions of interest for respectively the PET and the CT. This allows reproducible and informative downstream analysis through sub-segmentation, quantification, texture analysis, and radiomics.

Although initially designed for lung segmentation, this technology is highly adaptable and can be applied to other organ systems with minimal modifications, expanding its potential clinical and research applications.