undergo two views analysis The annular or
undergo two views analysis. The annular or search region is
Similarly, the results on DDSM dataset are depicted in Table 5.
dynamic and is based on the length of minor axis of the lesions
755 participating in the pairing process. Hence, the missing of any 5.4. Case-based analysis 791
756 lesion due to static parameter of region width is avoided which
757 results in appropriate and limited pairing. There are two Once the TP–TP and TP–FP pairs are finalized, the task is to 792 758 experiments performed in this INCB018424 clinical study. First, all classify the pair of benign mass from the pair of malignant 793 759 suspicious lesions on CC view are selected one by one and tumours. This is nothing but identifying correct TP–TP pairs. 794 760 are paired with every other lesion from MLO view and The 155 TP–TP pairs from CC to MLO are now divided into 4 sets 795 761 subsequently the correspondence score of the pair is calculat- of 31 pairs for training and set of 31 for testing using five-fold 796 762 ed. The lesions on the MLO view which are lying outside the cross validation using SVM classifier. Similarly, 163 TP–TP 797 763 annular regions defined by the radial distances of all the pairs from MLO to CC are divided into 4 sets of 32 pairs for 798 764 lesions on CC view are discarded. Thus, there are limited training and set of 35 for testing using five-fold cross validation 799 765 number of pairs from which one with highest correspondence using SVM classifier. The correspondence score of the lesion 800 766 score is selected. This highest score actually belongs to the on CC view extended by combined features set forms a new 801 767 most likely abnormal lesion from suspicious lesions on CC feature vector. The same experiment for lesions on MLO view 802 768 view. In the second experiment, all suspicious lesions on MLO is performed. The performances of a SVM classifier to 803 769 view are selected one by one and are paired with every other distinguish the suspicious lesions on CC and MLO view are 804 770 lesion from annular region on CC view. Subsequently, the depicted in Table 6. 805 771 correspondence score of the pair is calculated. All the possible The average case-based lesions diagnosis sensitivity is 806 772 pairs of lesions from CC to MLO view are divided into 4 sets of 75.91% (from 110 cases) at the cost of 0.69 FPs/I on an average. 807 773 104 pairs for training and remaining set of 104 pairs for five- The sensitivity here is calculated based on the number of TP– 808 774 fold cross validation using SVM classifier. SVM classified 155 of TP pairs classified correctly either as benign mass or malignant 809 775 all the pairs as TP–TP and 365 as TP–FP. All the possible pairs of tumour from actual 110 pairs. The specificity in monomer case is not 810 776 lesions from MLO to CC view are divided into 4 sets of 112 pairs meaningful as we have not selected true negatives (TN) for 811
Please cite this article in press as: Sapate S, et al. Breast cancer diagnosis using abnormalities on ipsilateral views of digital mammograms.
Table 7 – Performance of SVM classifier using fusion scheme on DDSM dataset.
True malignant False malignant True benign False benign Total pairs Sensitivity FPs/I Kappa
Fig. 8 – Performance of SVM classifier using (a) ROC plot and (b) FROC plot.
812 classification and hence not considered. Similarly, the results trials are statistically significant. In our experiment, the p 842 813 on DDSM dataset are depicted in Table 7. value is found less than 0.02. Hence the proposed fusion-based 843 814 The average case-based lesions diagnosis sensitivity is scheme is superior to the single view approach. 844 815 73.65% (from 110 cases) at the cost of 0.72 FPs/I on an average.