Skip to main content
. 2022 Jul 20;14(14):3515. doi: 10.3390/cancers14143515

Figure 7.

Figure 7

Examples feature of wavelet—HHL_ gldm_ Large Dependence High Gray Level Emphasis (LDHGLE) in non-pCR and pCR. (A) Images from a non-pCR breast cancer patient (aged 57 years old with invasive ductal carcinoma of the triple negative subtype. The first line is: (a) the image of pre-NAC, (b) ROI, and (c) a LDHGLE map of tumor ROI (mean LDHGLE value = 4268). The second line is: (d) the image of the first cycle of NAC(1st-NAC), (e) ROI, and (f) a LDHGLE map of tumor ROI (mean LDHGLE value = 4240). (B) Images from a pCR breast cancer patient (aged 51 years old with invasive ductal carcinoma of luminal B subtype). The first line is: (g) the image of pre-NAC, (h) ROI, and (i) a LDHGLE map of tumor ROI (mean LDHGLE value = 16,530). The second line is: (j) the image of 1stNAC, (k) ROI, and (l) a LDHGLE map of tumor ROI (mean LDHGLE value = 10,343). (C) Boxplot represents the feature of wavelet-HHL_ gldm_ LDHGLE distribution among patients of non-pCR and pCR in the training cohort. (D) The heatmap of selected feature in delta-radiomics model based on early phase. Demonstrates overall distribution of key delta-radiomics features among patients with pCR and non-pCR in the training cohort, which shows the obvious difference between the two groups. LDHGLE, large dependence high gray level emphasis.