Table 9.

Performance-based comparative analysis of the proposed method with single/multiple neural networks for multimodality in detection of breast cancer detection.

Studies	Approach	Modalities	Performance
79	Two 3D ResNet-50 were combined for multimodal feature extraction and fusion	High‑dimensional MRI features and clinical information	AUC = 0.827
80	Integration of residual block with inception block to form a single CNN architecture	B-mode ultrasound, elastic ultrasound, pure elastic ultrasound, and H-channel images	Classification accuracy rates of breast lump detection is 94.76%
81	A single CNN architecture on B-mode and SE-mode ultrasound image	B-mode and elastography ultrasound images	sensitivity of 100 ± 0.00% and specificity of 94.28 ± 7.00%
33	A single neural architecture model for extracting stacked features using a sigmoid gated attention, and dense layer for bi-modality	Text-based, gene expression data and copy number alteration (CNA) data	Reported performance improvement for AUC, accuracy, precision, and sensitivity at 0.5%, 8.6%, 9.2% and 34.8% respectively
82	A single CNN architecture applied independently for extraction of multimodal features	Grey-scale images samples	Obtained 96.55%, 90.68%, and 91.28% on MIAS, DDSM, and INbreast datasets
This proposed study	A TwinCNN and binary optimization algorithm framework for multimodal classification using histology and mammography digital images	RGB-image and grey scale image samples	Classification accuracy for histology modality = 0.977, mammography modality = 0.913, and fused multimodalities = 0.684