Siemens Healthineers International Ag

Fernando Franco

Thomas Coradi

Benjamin M Haas

Tobias Gass

Angelo Genghi

Mário Joao Fartaria

Simon Mathias Fluckiger

Tomasz Morgas

Michael Waschbüesch

Anna Siroki-Galambos

A computer-implemented method of generating dose information for a region of patient anatomy includes determining a first set of dose values for a first three-dimensional (3D) image of the region, wherein each value in the first set of dose values is associated with a different voxel of the first 3D image, and wherein the first 3D image is associated with a specific application of dose to the region; determining a second set of dose values for a representative 3D image of the region, wherein each value in the second set of dose values is associated with a different voxel of the representative 3D image; determining a set of geometric error models for the representative 3D image of the region, wherein each geometric error model in the set of geometric error models indicates a geometric error between a voxel of the representative 3D image and one or more voxels of a treatment fraction 3D image of the region; and based on the second set of dose values and the set of geometric error models, determining a set of dose probability values for each voxel of the representative 3D image, wherein each set of dose probability values includes at least one dose value and a probability value that corresponds to the dose value.

Determining uncertainty in radiation therapy delivered dose accumulation

Embodiments described herein provide a method for generating training data for Al based atlas mapping slice localization, and a system and method for using the training data to train a deep learning network. Training data development maps each slice of an input medical image to a position in a full body reference atlas along the longitudinal body axis. The method constructs a landmarking table of 2D slices indicating known anatomic landmarks of a reference subject, and interpolated slices. A final step for obtaining training data uses regression analysis techniques to create a vector of longitudinal axis coordinates of all slices from the input image. The training data is used to train a deep learning model to create an AI-based atlas mapping slice localizer model. The trained AI-based atlas mapping slice localizer model can be applied to generate mapping inputs to autosegmentation models to improve efficiency and reliability of contouring.

AI-based atlas mapping slice localizer for deep learning autosegmentation

Growing community of inventors

Zurich, Switzerland

Fernando Franco