IP Library › Granted Patent US 12,646,623

Granted Patent B2

US 12,646,623 · App. 18/155,176 · Granted Jun 2, 2026

Methods and systems for providing molecular data based on CT images

Inventors: Arnaud Arindra Adiyoso (Nuremberg, DE); Andre Aichert (Erlangen, DE); Marvin Teichmann (Erlangen, DE); Tobias Heimann (Erlangen, DE)

Assignee: SIEMENS HEALTHINEERS AG

G16H50/20 A61B6/03 G06T7/0012 G06T2207/20084

View Patent ↗

Loading inventors, assignments & file history…

Monitor This Case

Get email alerts when status or documents change.

Order Certified Copies

Most orders are placed with the USPTO same day — all within 24 business hours.

Order via The Patent Place →

Pre-filled with this patent's details

Quick Facts

Patent No.: US 12,646,623
App. No.: 18/155,176
Filed: Jan 17, 2023
Granted: Jun 2, 2026
Kind: B2
Examiner: YIP, KENT
Art Unit: 2681
USPC: 382/131

External Links

USPTO Patent Center↗ Full Patent PDF↗

Abstract

One or more example embodiments of the present invention is based on a computer-implemented method for providing molecular data. The method comprises receiving a computed tomography image of at least a part of a lung of a patient, wherein the computed tomography image depicts at least one lung nodule. The molecular data is determined by processing first input data with a first trained function, wherein the first input data is based on the computed tomography image, and wherein the molecular data relates to a biomarker within at least one of a genome of the patient, a transcriptome of the patient, a proteome of the patient or a metabolome of the patient. Furthermore, the molecular data is provided. Providing the molecular data can comprise at least one of displaying, transmitting or storing the molecular data.

Claims (41)

1 . A computer-implemented method for providing molecular data, the method comprising:

receiving a computed tomography image of at least a part of a lung of a patient, the computed tomography image depicting at least one lung nodule;

determining a nodule patch by processing the computed tomography image with a second trained function, the second trained function configured to localize lung nodules depicted in computed tomography images;

determining nodule-based data by processing the nodule patch with a third trained function, the nodule-based data relating to a biomarker within at least one of a genome of the patient, a transcriptome of the patient, a proteome of the patient, or a metabolome of the patient;

determining feature data by processing the computed tomography image with a fourth trained function, the feature data relating to image biomarkers not related to lung nodules;

determining the molecular data by processing first input data with a first trained function, the first input data including the nodule-based data and the feature data, and the molecular data relating to the biomarker within at least one of the genome of the patient, the transcriptome of the patient, the proteome of the patient, or the metabolome of the patient; and

providing the molecular data.

2 . The method of claim 1 , wherein the nodule-based data relates to at least one of an expression of programmed death-ligand 1 (PD-L1) within the proteome of the patient or a mutation of epidermal growth factor receptor (EGFR) within the genome of the patient.

3 . The method of claim 1 , further comprising:

receiving patient data related to the patient, the patient data including nonimaging-based data, and the first input data including the patient data.

4 . The method of claim 3 , wherein the patient data comprises at least one of:

a status of at least one lymph node of the patient,

a presence of pulmonary emphysema within the patient,

a metastasis pattern related to the patient,

an age of the patient,

a muscle and fat distribution of the patient, or

a smoking history of the patient.

5 . The method of claim 1 , wherein the first trained function includes at least one fully connected layer, or includes a convolutional neural network.

6 . The method of claim 1 , wherein at least one of the second trained function is a region-based convolutional neural network, the third trained function is a convolutional neural network, or the fourth trained function is a convolutional neural network.

7 . The method of claim 1 , wherein the molecular data relates to the biomarker within the at least one of the genome of the patient, the transcriptome of the patient, the proteome of the patient or the metabolome of the patient by comprising a numerical value, the numerical value indicating at least one of a probability of a presence of the biomarker or a level of the biomarker.

8 . The method of claim 1 , wherein the molecular data relates to at least one of an expression of programmed death-ligand 1 (PD-L1) within the proteome of the patient or to an mutation of epidermal growth factor receptor (EGFR) within the genome of the patient.

9 . A computer-implemented method for providing a first trained function, the method comprising:

receiving a computed tomography image of at least a part of a lung of a patient, the computed tomography image depicting at least one lung nodule;

determining feature data by processing the computed tomography image with a fourth trained function;

determining estimated molecular data by processing first input data with the first trained function, the first input data including the nodule-based data and the feature data;

receiving actual molecular data of the patient, the actual molecular data relating to the biomarker within at least one of the genome of the patient, the transcriptome of the patient, the proteome of the patient, or the metabolome of the patient;

adapting at least one parameter of the first trained function based on a comparison of the actual molecular data and the estimated molecular data; and

providing the first trained function.

10 . A providing system for providing molecular data, the system comprising:

a memory storing instructions; and

processing circuitry configured to execute the instructions to cause the system to,

receive a computed tomography image of at least a part of a lung of a patient, the computed tomography image depicting at least one lung nodule,

determine a nodule patch by processing the computed tomography image with a second trained function, the second trained function configured to localize lung nodules depicted in computed tomography images;

determine nodule-based data by processing the nodule patch with a third trained function, the nodule-based data relating to a biomarker within at least one of a genome of the patient, a transcriptome of the patient, a proteome of the patient, or a metabolome of the patient;

determine feature data by processing the computed tomography image with a fourth trained function;

determine the molecular data by processing first input data with a first trained function, the first input data including the nodule-based data and the feature data, and the molecular data relating to the biomarker within at least one of the genome of the patient, the transcriptome of the patient, the proteome of the patient, or the metabolome of the patient, and

provide the molecular data.

11 . A non-transitory computer program product comprising instructions which, when executed by a computer, cause the computer to carry out the method of claim 1 .

12 . A non-transitory computer-readable storage medium comprising instructions which, when executed by a computer, cause the computer to carry out the method of claim 1 .

Assignments (2)

ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Dec 20, 2023

From: SIEMENS HEALTHCARE GMBH

To: SIEMENS HEALTHINEERS AG

Reel/Frame 066267/0346 →

ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Sep 19, 2023

From: ADIYOSO, ARNAUD ARINDRA; AICHERT, ANDRE; TEICHMANN, MARVIN; HEIMANN, TOBIAS

To: SIEMENS HEALTHCARE GMBH

Reel/Frame 064947/0786 →

Priority Claims (1)

EP 22152264 · Jan 19, 2022 · regional

Continuity (1)

Related Publication 20230230704A1 · Jul 20, 2023

References Cited (21)

US 11741604B2 · Kapur · 2023 [cited by examiner]

US 20220185858A1 · Li et al. · 2022 [cited by applicant]

CN 110055224A · 2019 [cited by applicant]

CN 111553892A · 2020 [cited by applicant]

WO WO2021118918A1 · 2021 [cited by applicant]

Jiang Mengmend et al:; “Assessing PD-L1 Expression Level by Radiomic Features From PET/CT in Nonsmall Cell Lung Cancer Patients: An Initial Result”, Mengmeng Jiang MD et al., Academic Radiology, vol. 27, No. 2, pp. 171-… [cited by applicant]

Zhang, Baihua et al: “Deep CNN Model Using CT Radiomics Feature Mapping Recognizes EGFR Gene Mutation Status of Lung Adenocarcinoma”; Frontiers in Oncology; Feb. 12, 2021 (Feb. 12, 2021), XP55932796. [cited by applicant]

Chen Y. et al.: “Prediction of EGFR mutations by conventional CT-features in advanced pulmonary adenocarcinoma,” Eur. J. Radiol., vol. 112, No. Aug. 2018, pp. 44-51, 2019, doi: 10.1016/j.ejrad.2019.01.005. [cited by applicant]

Rossi G. et al.: “Radiomic Detection of EGFR Mutations in NSCLC,” Cancer Res., vol. 81, No. 3, pp. 724-731, 2021, doi: 10.1158/0008-5472.CAN-20-0999. [cited by applicant]

Girshick et al., “Rich feature hierarchies for accurate object detection and semantic segmentation” ar-Xiv:1311.2524. [cited by applicant]

Nair J.K.R. et al.: “Radiogenomic Models Using Machine Learning Techniques to Predict EGFR Mutations in Non-Small Cell Lung Cancer” in Can Assoc Radiol J. Feb. 2021;72(1):109-119. doi: 10.1177/0846537119899526. [cited by applicant]

Mu W. et al.: “Non-invasive decision support for NSCLC treatment using PET/CT radiomics,” Nat. Commun., vol. 11, No. 1, 2020, doi: 10.1038/s41467-020-19116-x. [cited by applicant]

Zhu .Y et al.: “A CT-derived deep neural network predicts for programmed death ligand-1 expression status in advanced lung adenocarcinomas,” Ann. Transl. Med., vol. 8, No. 15, pp. 930-930, 2020, doi: 10.21037/atm-19-469… [cited by applicant]

Zhang Y. et al.: “Improving prognostic performance in resectable pancreatic ductal adenocarcinoma using radiomics and deep learning features fusion in CT images,” Sci. Rep., vol. 11, No. 1, pp. 1-12, 2021, doi: 10.1038/… [cited by applicant]

Zhang Y. et al.: “Simultaneous identification of EGFR, KRAS, ERBB2, and TP53 mutations in patients with non-small cell lung cancer by machine learning-derived three-dimensional radiomics,” Cancers (Basel)., vol. 13, No.… [cited by applicant]

Liu Q. et al.: “Predicting EGFR mutation subtypes in lung adenocarcinoma using 18F-FDG PET/CT radiomic features,” Transl. Lung Cancer Res., vol. 9, No. 3, pp. 549-562, 2020, doi: 10.21037/tlcr.2020.04.17. [cited by applicant]

Saltz J. et al.: “Spatial Organization and Molecular Correlation of Tumor-Infiltrating Lymphocytes Using Deep Learning on Pathology Images,” Cell Rep., vol. 23, No. 1, pp. 181-193.e7, 2018, doi:10.1016/j.celrep.2018.03.… [cited by applicant]

Uijlings J.R. et al., “Selective search for object recognition” Int. J. Comp. Vision, 104(2), 154-171 (2013). [cited by applicant]

Wang, Chengdi et al: “Deep Learning to Predict EGFR Mutation and PD-LI Expression Status in Non-Small-Cell Lung Cancer on Computed Tomography Images”, Journal of Oncology; vol. 2021, Dec. 31, 2021 (Dec. 31, 2021), pp. 1… [cited by applicant]

Wang S. et al.: “Predicting EGFR mutation status in lung adenocarcinoma on computed tomography image using deep learning,” Eur. Respir. J., vol. 53, No. 3, 2019, doi: 10.1183/13993003.00986-2018. [cited by applicant]

Coudray N. et al.: “Classification and mutation prediction from non-small cell lung cancer histopathology images using deep learning,” Nat. Med., vol. 24, No. 10, pp. 1559-1567, 2018, doi:10.1038/s41591-018-0177-5. [cited by applicant]