Medical Imaging Control Technique Based on Natural Language Request

ChangeBridge: Patent Apps - Health Informatics (G16H)

Detected March 26th, 2026

Summary

The USPTO has published a patent application (US20260088156A1) from Siemens Healthineers AG detailing a technique for controlling medical imaging using natural language requests processed by large language models (LLMs). This application outlines a method for generating control signals for medical imaging preparation, execution, or processing based on user input.

View original document View source feed page

What changed

This document is a published patent application from Siemens Healthineers AG (US20260088156A1) describing a novel technique for controlling medical imaging equipment. The core innovation involves using a large language model (LLM) to interpret natural language requests from users and translate them into control signals for imaging systems. This method aims to streamline the preparation, execution, and processing of medical imaging procedures by allowing for more intuitive, voice- or text-based commands.

While this is a patent application and not a regulation, it signals potential future technological advancements in the healthcare sector. Compliance officers in medical device manufacturing and healthcare IT should be aware of such innovations as they may influence future product development, data handling practices, and the integration of AI in clinical workflows. The application highlights the increasing role of AI, specifically LLMs, in medical device control and patient care processes.

Source document (simplified)

← USPTO Patent Applications

TECHNIQUE FOR MEDICAL IMAGING CONTROL BASED ON A REQUEST MESSAGE

Application US20260088156A1 Kind: A1 Mar 26, 2026

Assignee

Siemens Healthineers AG

Inventors

Christian TIETJEN, Sven KOHLE, Gerhard KOHL, Steffen WEICHERT, Christoph BRACHMANN, Diego BARRIOS ROMERO, Johannes HAUBOLD

Abstract

One or more example embodiments relates to a technique for outputting a control signal for at least one of preparing, executing or processing medical imaging of a patient based on a request in natural language. A computer-implemented method comprises receiving a digital request message comprising the request in connection with the medical imaging, wherein the request is formulated in natural language; processing the request, wherein the processing comprises applying at least one first large language model (LLM); and outputting the control signal to an imaging function facility for the at least one of the preparing the medical imaging, the executing the medical imaging or the processing the medical imaging, wherein the output control signal is based on a result of the processing of the request.