Table 7.

List of studies that used Grad‐class activation map.

Authors	Year	Objective	Data set(s)	Data type	Machine learning (ML)/Deep learning (DL)	Explained model
Adoui et al. [103]	2020	Predicting the breast cancer response to Neoadjuvant chemotherapy (NAC) based on multiple MRI inputs	Institute of Radiology in Brussels (A cohort of 723 axial slices extracted from 42 breast cancer patients who underwent NAC therapy)	MRI	DL	Based on convolutional neural network (CNN)
Hussain et al. [104]	2022	Developing DL multiclass shape‐based classification framework for the tomosynthesis of breast lesion images	Based on the previous study [105]	Digital breast tomosynthesis (DBT)	DL	VGG, ResNet, ResNeXt, DenseNet, SqueezeNet, MobileNet‐v2
Agbley et al. [106]	2023	Breast tumor detection and classification using different magnification factors on the Internet of Medical Things (IoMT)	BreakHis [91]	Microscopic images	DL	ResNet‐18, Federated Learning (FL) to preserve the privacy of patient data
Gerbasi et al. [107]	2023	Proposing a fully automated and visually explained model to analyze raw mammograms with microcalcifications	INbreast data set [108] (train and test), CBIS‐DDSM [109] (used to implement the classification algorithm)	Scanned film Mammography	DL	U‐Net, ResNet18