Classification of laterality and mastectomy/lumpectomy for breast cancer patients for improved performance of deep learning auto segmentation

Figures & data

Figure 1. Pipeline design from the CT-scan to the contour of the final volumes. It includes normal tissue delineation, laterality classification, target delineation by the lumpectomy and mastectomy networks for the chosen tumour side and finally mastectomy/lumpectomy determination.

Table 1. Characteristics of the patients included in the study.

Parameter	Train/validation set	Test set
Number of patients	113	16
Age	Median: 56 yr. Range: 26–82 yr.	Median: 63 yr. Range: 24–76 yr.
Dose prescription:
50 Gy / 25 fx.	33 (29.20%)	1 (6.25%)
50 Gy / 25 fx. + SIB: 57 Gy / 25 fx. (EQD2: 50 Gy + 10 Gy)	3 (2.66%)	0 (0%)
51.52 Gy / 28 fx. + SIB: 63 Gy / 28 fx. (EQD2: 50 Gy + 16 Gy)	1 (0.89%)	0 (0%)
40 Gy / 15 fx. (EQD2: 45 Gy)	63 (55.75%)	13 (81.25%)
40 Gy / 15 fx. + SIB: 45.75 Gy / 15 fx. (EQD2: 45 Gy + 10 Gy)	5 (4.42%)	0 (0%)
42.3 Gy / 18 fx + SIB: 52.2 Gy / 18 fx. (EQD2: 45 Gy + 16 Gy)	8 (7.08%)	2 (12.5%)
Unilateral	113 (100%)	16 (100%)
Previous surgery or RT in the side being treated	0 (0%)	0 (0%)
Category
Right mastectomy	16 (14.16%)	2 (12.50%)
Left mastectomy	42 (37.17%)	5 (31.25%)
Right lumpectomy	16 (14.16%)	5 (31.25%)
Left lumpectomy	39 (34.51%)	4 (25.00%)

Figure 2. a: Head angle determination: performing a linear regression of the voxels (with bone HU window) of the last CT slices inside a bounding box that includes the head, b: Volumes segmented by the lumpectomy model and c: Volumes segmented by the mastectomy model.

Data availability statement

Due to the nature of the research, due to legal reasons the supporting data is not available.