Home Medizin Studie legt nahe, dass CT-Bildgebung mit automatisiertem KI-System den EGFR-Genotyp vorhersagt und den Mutationsstatus kostengünstig und nicht-invasiv identifiziert

Studie legt nahe, dass CT-Bildgebung mit automatisiertem KI-System den EGFR-Genotyp vorhersagt und den Mutationsstatus kostengünstig und nicht-invasiv identifiziert

von NFI Redaktion

A recent study published in The Lancet Regional Health-Southeast Asia revealed the development of an Artificial Intelligence (AI) based Prediction System (AIPS) for early detection of lung cancer by combining radiological, clinical, and genetic data.


Studie: KI-basierte Pipeline für die Früherkennung von Lungenkrebs: Integration radiologischer, klinischer und genomischer Daten.  Bildnachweis: Jobs erstellen 51/Shutterstock.com









Studie: KI-basierte Pipeline für die Früherkennung von Lungenkrebs: Integration radiologischer, klinischer und genomischer Daten. Bildnachweis: Jobs erstellen 51/Shutterstock.com

Hintergrund

Lung cancer prognostics have improved due to molecular targets and targeted treatments, including the therapy of mutations of the somatic epidermal growth factor receptor (EGFR). However, Next-Generation-Sequencing (NGS) is not available in resource-constrained countries such as India. A KI-based pipeline is necessary to recognize lung nodule features in computed tomography (CT) scans and predict the probability of an EGFR mutation, enabling near-optimal care and therapy.

Über die Studie

In the present study, the researchers demonstrated the APIS pipeline for identifying and predicting the presence of EGFR-mutated lung nodules based on CT scans in resource-constrained situations like India.

Ergebnisse

The deep learning model could replace machine learning models by adding AIPS-N results, improving the ML performance. The AIPS-Nodule model could automatically detect and characterize lung nodules with a mean AP50 score of 70%. The AIPS-M system combined the AIPS-Nodule results with clinical parameters to predict the EGFR genotype, resulting in Area Under the Receiver Operating Characteristic Curve (AUC) values between 0.6 and 0.9.

The study encompassed an Indian male smoker aged 71 who was diagnosed with squamous cell carcinoma and carried the EGFR gene mutation. The AIPS-Nodule model accurately identified the features and position of the detected nodules and assigned spiculation and sphericity to class 1. AIPS-Mutation models, trained on Cohort 1, were used to determine the EGFR gene status of the male patient, clinically set as mutated. All six machine learning algorithms predicted the EGFR status as mutated, providing a „true positive“ outcome.

Related Posts

Adblock Detected

Please support us by disabling your AdBlocker extension from your browsers for our website.