Abstract: A semiconductor device includes a substrate, first and second stacks of semiconductor nanosheets, a gate structure, first and second strained layers and first and second dielectric walls. The substrate includes first and second fins. The first and second stacks of semiconductor nanosheets are disposed on the first and second fins respectively. The gate structure wraps the first and second stacks of semiconductor nanosheets. The first and second strained layers are respectively disposed on the first and second fins and abutting the first and second stacks of semiconductor nanosheets. The first dielectric wall is disposed on the substrate and located between the first and second strained layers. The second dielectric wall is disposed on the first dielectric wall and located between the first and second strained layers. A top surface of the second dielectric wall is lower than top surfaces of the first and second strained layers.

Abstract: A method for predicting epidermal growth factor receptor (EGFR) mutations in lung adenocarcinoma is provided. The method utilizes a lung adenocarcinoma EGFR mutation classification model based on a deep learning model, and performs back-propagation training on the deep learning model by using whole-slide pathological images and corresponding pathological data. The trained lung adenocarcinoma EGFR mutation classification model can determine whether a to-be-classified slide-level image with lung adenocarcinoma features have EGFR mutations.