PET/CT (positron emission tomography/computed tomography)-based lung adenocarcinoma and squamous carcinoma diagnosis model training method and device

Abstract

The invention provides a PET / CT (positron emission tomography / computed tomography)-based lung adenocarcinoma and squamous carcinoma diagnosis model training method and device, and aims to assist in training a PET / CT image-based diagnosis classification neural network by using a multi-task learning method and extracting pathological features through a neural network obtained by diagnosis classification training based on pathological images. According to the method, the lung cancer diagnosis classification precision based on the PET / CT image is improved, and meanwhile, the pathological image is only used as priori knowledge in the training process and does not need to be used as network input in practical application. According to the method, through the concept of multi-scale fusion, the precision of the PET / CT image for lung cancer diagnosis classification is improved, the PET / CT image can be further popularized and applied as a means for early lung cancer diagnosis, and help is provided for diagnosis of a clinician on a patient and a subsequent treatment scheme; and meanwhile, the pathology image is used as a priori knowledge auxiliary scheme, the interpretability of the pathology section is further improved, and pathology doctors can further extract pathology features.

Description

technical field [0001] The present invention relates to the fields of medical imaging and deep learning, in particular to a fully automatic intelligent diagnosis model training method and device for lung adenocarcinoma and squamous cell carcinoma based on PET / CT and pathological slices. Background technique [0002] Positron emission tomography (PET) is a functional imaging device at the molecular level. The radioactive tracer needs to be injected into the patient before scanning, and the tracer decays and annihilates in the patient's body, producing a pair of emission directions about 180 o In contrast to the 511keV gamma photons, the detector collects information about where and when these gamma photons hit the crystal. By using image reconstruction algorithms to reconstruct and post-process the acquired information, the metabolism and uptake of the reaction tracer in the patient can be obtained. According to the imaging results of PET / CT, doctors comprehensively analyze...

AI Technical Summary

Problems solved by technology

[0005] Regarding the existing PET / CT-based lung cancer diagnosis and classification model, due to the limitation of data scale and accuracy, no matter whether the model is trained on single-center data or multi-center data, its diagnostic classification accuracy cannot meet the practical requirements.

However, pathological slides, which are regarded as the "gold standard" for cancer diagnosis, are rarely used in early diagnosis because sampling often requires invasive or even invasive examinations on patients.

Images