Table 2.

Representative studies focus on multimodal image process

Topic	Study (year)	Cancer	Modality	Backbone	Learning method	Dataset	Fusion strategy	Code availability	Data availability
Registration	Song et al. (2023) [56]	Prostate	MRI, TRUS	CNN, attention	Supervised	662 patients	Intermediate	https://github.com/DIAL-RPI/ Attention-Reg	Partial available
Registration	Haque et al. (2023) [57]	Prostate	MSI, WSI	CNN	Supervised	5 patients	Early	https://github.com/inzamam1190/HEtoMALDI	Need request
Registration and segmentation	Gu et al. (2023) [52]	Esophagus, lung	PET-CT	FCN	Supervised	53 patients	Early	No	Need request
Segmentation	Lee et al. (2023) [60]	Various cancers	Multimodal microscopy images	Transformer	Semi-supervised	7242 images	Early	https://github.com/Lee-Gihun/MEDIAR	Yes
Segmentation	Zhao et al. (2018) [58]	Glioma	Multimodal MRI	FCN, RNN	Supervised	465 patients	Late	No	Yes

Note: CNN, convolutional neural network; TRUS, transrectal ultrasound; MSI, mass spectrometry image; WSI, whole slide image; FCN, fully convolutional network; RNN, recurrent neural network.