|
Reference number | Year of publication | AI method | Number of units | Classification domain | Performance |
|
[81] | 1993 | A backpropagation ANN and a learning vector quantizer | 392 cases | Diagnosis of thyroid function | Overall accuracy on test data subsets was in the range of 96.4–99.7%, when extreme values used for training the overall accuracy were in the range of 92.7–98.8% |
|
[27] | 1996 | Backpropagation algorithm (three layers) | 51 patients | Cell nuclei and patients | Overall accuracy was 90.6% for nuclei classification and 98% on individual patients |
|
[61] | 1999 | LVQ classifier | 198 patients | Benign from malignant thyroid lesions. | Overall accuracy: 97.8% |
|
[31] | 2006 | Four methods: (1) a linear classifier, (2) a two-layer feedforward ANN, (3) a combined two-layer feedforward ANN generated by the AdaBoost method, and (4) the k nearest neighbor classifier (a method with many similarities with LVQ) | 157 patients | Benign from malignant thyroid lesions. | (1) 65.17%, (2) 73.20%, (3) 73.20%, and (4) 74.69% |
|
[82] | 2007 | Backpropagation algorithm (three layers) | 197 smears | Follicular carcinomas vs. follicular adenomas | Sensitivity: 97% |
|
[83] | 2004 | Multilayer perceptron 15 nodes in the input, 1 hidden layer of 15 units, and an output layer | 453 patients | High vs. low risk for cancer | Sensitivity: 90.6%, specificity: 62.2% |
|
[84] | 2006 | Two-layer ANN having inputs as cytological images | 30 images from 10 patients for training and 45 patients with follicular adenoma and 39 patients with follicular carcinoma for testing | Follicular carcinomas vs. follicular adenomas | Overall accuracy: 96% |
|
[86] | 2008 | Multiclassifier system | 115 cases | Benign vs. malignant nodules | Overall accuracy: 95.7% |
|
[58] | 2011 | LVQ ANN | 335 cases | In follicular neoplasms suspicious for malignancy and in Hürthle cell tumors | Overall accuracy: 94% |
|
[87] | 2014 | Optimal transport-based linear embedding for segmented nuclei | 94 patients | Distinguishing between follicular lesions | OA LOT-100% except FVPC vs. FC 87% |
|
[89] | 2013 | Supervised learning-based template matching for segmenting cell nuclei | Microscopy images to segment nuclei | Texture and shape variations of the nuclear structures | Not applicable, used for nuclei segmentation |
|
[90] | 2018 | ANN model to differentiate FA versus FC on the FNAC material | Microscopy images of FA–FC (26 and 32 cases respectively) | Follicular carcinomas vs. follicular adenomas | Overall accuracy of 93% on image analysis and an overall accuracy of 96% in automatic image classification to differentiate FA and FC |
|
[91] | 2018 | Convolutional neural network | 174 microscopy images | Papillary vs. nonpapillary | Sensitivity: 90.8%, specificity: 83.3% |
|
[92] | 2020 | Deep learning algorithm for whole slide images (WSIs) | 908 whole slide images | Malignancy prediction | Sensitivity: 92%, specificity: 90.5% |
|