Combination of features from biopsies and scans to predict prognosis in SCLC

Inventors

Madabhushi, AnantBarrera, CristianKhorrami, MohammadhadiJain, PranteshDowlati, Afshin

Assignees

Case Western Reserve UniversityUS Department of Veterans AffairsUniversity Hospitals Cleveland Medical Center

Publication Number

US-12159403-B2

Publication Date

2024-12-03

Expiration Date

2042-02-14

Interested in licensing this patent?

MTEC can help explore whether this patent might be available for licensing for your application.


Abstract

The present disclosure relates to a non-transitory computer-readable medium storing computer-executable instructions that, when executed, cause a processor to perform operations, including generating an imaging data set having both scan data and digitized biopsy data from a patient with small cell lung cancer (SCLC). Scan derived features are extracted from the scan data and biopsy derived features are extracted from the digitized biopsy data. A radiomic-pathomic risk score (RPRS) is calculated from one or more of the scan derived features and one or more of the biopsy derived features. The RPRS is indicative of a prognosis of the patient.

Core Innovation

Small cell lung cancer (SCLC) is a highly aggressive form of lung cancer characterized by rapid growth and early dissemination, often diagnosed after metastasis, which results in poor prognosis and limited recovery prospects. SCLC frequently develops resistance to standard chemotherapy treatments, such as platinum-based chemotherapy, leading to substantially diminished survival rates.

Existing challenges include a lack of accurate and consistent predictive biomarkers for chemotherapy response, making treatment planning difficult. This invention addresses this gap by combining radiomic features from radiological imaging and pathomic features from digital biopsy images to generate a radiomic-pathomic risk score (RPRS) indicative of patient prognosis.

The disclosed methods extract shape and texture features from both scan data (including multiple mask regions such as lesional and perilesional areas) and digitized biopsy data. A subset of predictive features with significant prognostic impact is identified using machine learning classifiers, such as random forest algorithms. A risk score calculation model, often a survival regression model like an elastic net regularized Cox regression, integrates these predictive features to compute the RPRS. This RPRS enables stratification of patients into risk groups correlating with different survival outcomes and treatment responses, thereby guiding clinical decision-making for improved SCLC patient management.

Claims Coverage

The patent discloses three independent claims focused on a computer-readable medium, a method, and a prognostic apparatus, each encompassing inventive features related to combining scan-derived and biopsy-derived features to predict SCLC prognosis.

Computer-readable medium storing instructions for combined feature extraction and prognosis prediction

Stores computer-executable instructions performing operations that generate an imaging dataset combining scan data and digitized biopsy data from a patient with SCLC, extract scan derived features and biopsy derived features, and calculate a radiomic-pathomic risk score (RPRS) indicative of prognosis.

Use of machine learning to identify predictive features and calculate RPRS

Utilizes a machine learning classifier to identify predictive scan derived features from the scan data and predictive biopsy derived features from digitized biopsy data, and calculates the RPRS from these predictive features.

Mask-based feature extraction from scan data

Identifies and employs first lesional and perilesional masks in scan data to extract scan derived features specifically from these defined regions.

Method of prognosis prediction integrating radiologic imaging and biopsy data

Obtains radiological images and tissue samples from a patient with SCLC, digitizes biopsy samples, extracts scan and biopsy features, identifies predictive feature subsets, and calculates an RPRS integrating these predictive features to indicate prognosis.

Prognostic apparatus integrating feature extraction and classification circuitry

Includes a memory storing scan and biopsy data, feature extraction circuitry extracting scan derived and biopsy derived features, a risk score calculation circuit calculating an RPRS from these features, and a classification circuit grouping patients into risk categories based on the RPRS.

These inventive features encompass the integration of radiomic and pathomic features using machine learning methods to calculate a prognostic risk score for SCLC patients, employing specific masking techniques for feature extraction, and utilizing apparatus and computer-readable media to implement the prediction system.

Stated Advantages

Improved accuracy in predicting prognosis for patients with small cell lung cancer by combining both radiomic and pathomic data.

Ability to identify patients' response to chemotherapy, enabling better personalized treatment decisions.

Facilitation of patient stratification into risk groups with distinct prognostic outcomes.

Use of machine learning methods enhances identification of the most predictive imaging and biopsy features.

Documented Applications

Predicting prognosis and chemotherapy response in patients with small cell lung cancer using combined radiological scans and digitized biopsy data.

Guiding clinical treatment decisions for SCLC patients by stratifying them into risk groups based on the radiomic-pathomic risk score.

Training and validating prognostic models using preparatory imaging and biopsy datasets from multiple patients.

Implementing diagnostic support in prognostic apparatus employing imaging devices and biopsy digitization.

JOIN OUR MAILING LIST

Stay Connected with MTEC

Keep up with active and upcoming solicitations, MTEC news and other valuable information.