Method and device for predicting non-small cell lung cancer brain metastasis

Abstract

The invention discloses a non-small cell lung cancer brain metastasis prediction method and device, and relates to the field of non-small cell lung cancer brain metastasis prediction. The method comprises the following steps: acquiring two-dimensional non-small cell lung cancer CT image data; inputting two-dimensional non-small cell lung cancer CT image data into the trained focus boundary segmentation model, and outputting a focus region segmentation image; the focus boundary segmentation model adopts an alternating network architecture based on a transformer block and a double-current Mangbar-transformer block; according to the focus area segmentation image, utilizing the trained non-small cell lung cancer brain metastasis prediction model to obtain a prediction result of whether the non-small cell lung cancer CT image data is a brain metastasis label or not according to the non-small cell lung cancer CT image data; the non-small cell lung cancer brain metastasis prediction model is a deep learning classification network of a heterogeneous axial cross attention mechanism with discrete wavelet transform processing. According to the method, the accuracy of predicting the brain metastasis result by using the non-small cell lung cancer CT image data can be improved.

Images