A review and comparative study of cancer detection using machine learning: SBERT and SimCSE application

BMC Bioinformatics

Table 4 This table gives a summary of recent work that has been executed in colorectal cancer detection/survival using machine learning and deep learning algorithms as discussed in Sect. 2.4

References	Feature extraction	Data	ML/DL	Acc, AUC, ROC, or AUPRC (%)
[86] 2018	VGG16	TMA, Whole slide images	1-d LSTM, SVM, LR, Naive Bayes	61–69
[91] 2019	Normalization	EHR	CNN	92
[92] 2020	Macenko method [93]	H &E slide images [74, 94,95,96,97]	ShuffleNet	96
[90] 2020	–	Colonoscopic images	169-layer dense CNN	86.7–88.2
[89] 2021	Thresholding and normalization	Whole slide images [74, 98]	MobileNetV2	78–93
[99] 2021	Contrast-Limited Adaptive
	Histogram Equalization (CLAHE)	Warwick-QU dataset [100]	ResNet-18 & ResNet-50	73–88
[101] 2021	Normalization and data labeling	Numeric and clinical data	FNNs, SVMs, LR, LDA	77
[85] 2022	Faster-RCNN	Whole slide images	Gradient-boosted decision tree	91
[87] 2021	VGG16	Whole slide images [102]	MLP	99
[88] 2022	Aachen protocol [103] and,
	Macenko normalisation[93]	Whole slide images and
		clinical pathological data	ShuffleNet	56–73

ISSN: 1471-2105