Image classification of bone metastases from lung cancer based on attention mechanism

SPECT imaging is one of the main functional imaging methods in the field of medical imaging, playing an important role in the detection and diagnosis of bone metastasis caused by diverse primary tumors. Bone metastasis is a common pathological manifestation of lung cancer, having typical symptoms including bone pain, fractures, and hypercalcemia. To classify SPECT images of lung cancer bone metastasis into categories of concerns, in this work, we propsoe a ResNet-based image classification model. Specifically, an attention mechanism is incorporated into the ResNet-18 model, and hyperparameters are adjusted to build an improved model. Finally, a set of SPECT images are used to test the proposed model, obtaining scores of 0.7525, 0.7705, 0.7213, and 0.7442 for accuracy, precision, recall, and F-1, respectively.