# US 12,109,434 B2 · Neural network calibration for radiotherapy

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 12,109,434 B2
- Document ID: US12109434B2
- Status: Granted
- Kind: B2
- Application number: 18/527,066
- Filed: 2023-12-01
- Published: 2024-04-11
- Granted: 2024-10-08
- Earliest claimed: 2020-12-16
- Assignees: Siemens Healthineers International AG
- Applicants: Siemens Healthineers International AG
- Inventors: Mikko Hakala; Esa Kuusela; Elena Czeizler; Shahab Basiri
- Web page: https://patentgrep.com/patent/US12109434B2

## CPC

- A61N 5/1064: {for adjusting radiation treatment in response to monitoring}
- G16H 20/40: relating to mechanical, radiation or invasive therapies, e.g. surgery, laser therapy, dialysis or acupuncture
- G06N 3/045: Combinations of networks
- G06N 3/08: Learning methods
- G06N 3/09: Supervised learning
- G06N 20/00: Machine learning
- A61N: ELECTROTHERAPY; MAGNETOTHERAPY; RADIATION THERAPY; ULTRASOUND THERAPY (measurement of bioelectric currents A61B; surgical instruments, devices or methods for transferring non-mechanical forms of energy to or from the body A61B18/00; anaesthetic apparatus in general A61M; incandescent lamps H01K; infrared radiators for heating H05B)
- G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

Disclosed herein are systems and methods for identifying radiation therapy treatment data for patients. A processor accesses a neural network trained based on a first set of data generated from characteristic values of a first set of patients that received treatment at one or more first radiotherapy machines. The processor executes the neural network using a second set of data comprising characteristic values of a second set of patients receiving treatment at one or more second radiotherapy machines. The processor executes a calibration model using an output of the neural network based on the second set of data to output a calibration value. The processor executes the neural network using a set of characteristics of a first patient to output a first confidence score associated with a first treatment attribute. The processor then adjusts the first confidence score according to the calibration value to predict the first treatment attribute.

## Claims

### Claim 1

A method comprising:
retrieving, by one or more processors via a radiation therapy treatment generation (RTTP) model, a set of treatment attribute predictions for a radiotherapy treatment of a patient;
executing, by the one or more processors, a calibration model using the set of treatment attribute predictions and a set of labels indicating expected treatment attribute predictions to predict a calibration value; and
revising, by the one or more processors, at least one configuration of the RTTP model using the calibration value.

### Claim 2 (depends on 1)

The method of claim 1, further comprising:
responsive to revising the at least one configuration of the RTTP model, executing, by the one or more processors, the calibration model to predict a second calibration value.

### Claim 3 (depends on 1)

The method of claim 1, wherein the RTTP model is initially partially trained and then calibrated using the calibration model.

### Claim 4 (depends on 1)

The method of claim 1, wherein the calibration value corresponds to a clinic, such that the RTTP model is trained for the clinic.

### Claim 5 (depends on 1)

The method of claim 1, further comprising:
adjusting, by the one or more processors, a configuration of a radiotherapy machine according to at least one treatment attribute prediction generated by the RTTP model.

### Claim 6 (depends on 1)

The method of claim 1, wherein the calibration model is one of a machine learning model or a mathematical optimization algorithm.

### Claim 7 (depends on 1)

The method of claim 1, further comprising:
generating, by the one or more processors, a confidence score for the set of treatment attribute predictions generated using the RTTP model.

### Claim 8 (depends on 1)

The method of claim 1, wherein at least one treatment attribute prediction within the set of treatment attribute predictions is an angle associated with a couch or a gantry of a radiotherapy machine.

### Claim 9

A system comprising:
at least one processor;
a computer readable medium comprising a set of instructions, that when executed, cause the at least one processor to:
retrieve, via a radiation therapy treatment generation (RTTP) model, a set of treatment attribute predictions for a radiotherapy treatment of a patient;
execute a calibration model using the set of treatment attribute predictions and a set of labels indicating expected treatment attribute predictions to predict a calibration value; and
revise at least one configuration of the RTTP model using the calibration value.

### Claim 10 (depends on 9)

The system of claim 9, wherein the set of instructions further cause the at least one processor to:
responsive to revising the at least one configuration of the RTTP model, execute the calibration model to predict a second calibration value.

### Claim 11 (depends on 9)

The system of claim 9, wherein the RTTP model is initially partially trained and then calibrated using the calibration model.

### Claim 12 (depends on 9)

The system of claim 9, wherein the calibration value corresponds to a clinic, such that the RTTP model is trained for the clinic.

### Claim 13 (depends on 9)

The system of claim 9, wherein the set of instructions further cause the at least one processor to:
adjust a configuration of a radiotherapy machine according to at least one treatment attribute prediction generated by the RTTP model.

### Claim 14 (depends on 9)

The system of claim 9, wherein the calibration model is one of a machine learning model or a mathematical optimization algorithm.

### Claim 15 (depends on 9)

The system of claim 9, wherein the set of instructions further cause the at least one processor to:
generate a confidence score for the set of treatment attribute predictions generated using the RTTP model.

### Claim 16 (depends on 9)

The system of claim 9, wherein at least one treatment attribute prediction within the set of treatment attribute predictions is an angle associated with a couch or a gantry of a radiotherapy machine.

### Claim 17

A system comprising:
a radiation therapy treatment generation (RTTP) model;
a calibration model;
a server in communication with the RTTP model and the calibration model, the server configured to:
retrieve, via the RTTP model, a set of treatment attribute predictions for a radiotherapy treatment of a patient;
execute the calibration model using the set of treatment attribute predictions and a set of labels indicating expected treatment attribute predictions to predict a calibration value; and
revise at least one configuration of the RTTP model using the calibration value.

### Claim 18 (depends on 17)

The system of claim 17, wherein the server is further configured to:
responsive to revising the at least one configuration of the RTTP model, execute the calibration model to predict a second calibration value.

### Claim 19 (depends on 17)

The system of claim 17, wherein the RTTP model is initially partially trained and then calibrated using the calibration model.

### Claim 20 (depends on 17)

The system of claim 17, wherein the calibration value corresponds to a clinic, such that the RTTP model is trained for the clinic.