The impact of Digital Breast Tomosynthesis on BIRADS categorization of mammographic non-mass findings

Figures & data

Table 1. Distribution of the studied cases according to final diagnosis of histopathology

			Histopathology & follow up
	Final pathological diagnosis	N	FNAC	Core needle biopsy	Excisional biopsy	Follow Up	Percentage of patients
Benign (n = 96)	Fibroadenoma	8	3	4	1	0	4.4%
	Atypical duct hyperplasia	4	0	4	0	0	2.2%
	Granulomatous mastitis	6	6	0	0	0	3.33%
	Papilloma	1	1	0	0	0	0.55%
	Peri-ductal mastitis	18	16	2	0	0	10%
	FCC	10	10	0	0	0	5.%
	Sclerosing adenosis	5	0	5	0	0	2.7%
	Fibro adenomatoid hyperplasia	2	0	2	0	0	0.55%
	Traumatic fat necrosis	16	14	2	0	0	8.8%
	Radial scar	1	0	0	1	0	0.55%
	hamartoma	5	5	0	0	0	2.7%
	Pseudo angiomatous stromal hyperplasia	3	0	3	0	0	1.66%
	Keloid	1	0	0	0	1	0.55%
	Glandular superimposition	16	0	0	0	16	8.8%
Malignant (n = 84)	IDC	60	8	52	0	0	33.8%
	ILC	8	2	6	0	0	4.4%
	Paget disease	3	1	0	2	0	1.66%
	DCIS	12	0	12	0	0	6.66%
	Intra cystic Papillary Carcinoma	1	0	1	0	0	1.11%
	Total	180	117	42	4	17	100

Table 2. Distribution of to the Non-Mass findings in correlation with histopathology

				Histopathology				Chi-square test (Fisher test)
		Total(n=180)		Benign (n=96)		Malignant (n=84)
	Full field digital mammography	No.	%	No.	%	No.	%	X^2	P
Asymmetry	Type Focal asymmetry Asymmetry Global Developing None	68 2 30 4 76	37.8% 1.1% 16.7% 2.2% 42.2%	44 2 13 1 35	24.4% 1.1% 7.8% 0.6% 19.4%	24 0 17 3 41	13.3% 0% 8.9% 1.7% 22.8%	-	0.06
Calcification	Distribution Diffuse Regional Grouped Linear /ductal Segmental None	15 4 22 8 20 111	8.9% 2.8% 12.2% 4.4% 10.6% 60.6%	14 2 9 0 0 69	7.8% 2.2% 5% 0% 0% 38.3%	2 2 13 8 20 40	1.1% 0.6% 7.2% 4.4% 10.6% 22.2%	-	<0.0001*
	Morphology Linear/linear branching Pleomorphic Coarse heterogeneous Amorphous Round skin Coarse calcification None	20 13 10 6 14 3 3 111	11.1% 7.2% 5.6% 3.3% 7.8% 1.7% 2.2% 65.6%	0 2 1 2 14 3 3 69	0% 1.1% 0.6% 1.1% 7.8% 1.7% 2.2% 38.3%	20 11 9 4 0 0 0 40	11.1% 6.1% 5% 2.2% 0% 0% 0% 22.2%	62.5	<0.0001*
Architecture distortion	Presence Present Absent	35 145	19.4% 80.6%	17 79	9.4% 43.9%	18 66	10% 36.7%	0.19	0.66
Assoc Features	Presence Present Absent	62 118	34.4% 65.6%	30 66	16.7% 36.7%	32 52	17.8% 28.9%	0.65	0.42

X²: Chi square test MC: Monte Carlo FE: Fisher Exact

p: p value for comparing between the studied categories.

*: Statistically significant at p ≤ 0.05.

Table 3. Distribution of the studied cases according to FFDM BIRADS category (n = 180)

FFDM BIRADS	No.	%
2	13	7.2
3	103	57.2
4a	25	13.9
4b	18	10
4c	21	11.6
Provisional diagnosis
Benign	116	64.4%
Malignant	64	35.6%

Table 4. Distribution of Tomosynthesis findings and BIRADS categories

Tomosynthesis	No.	%
Finding
Superimposition of glandular tissue	17	7.2
Suspicious mass	27	11.5
Multi focal masses	25	10.6
Benign mass	35	14.9
Interconnected tubular ducts	7	2.9
Post-operative distortion	19	4.7
Suspicious distortion	16	5.9
Suspicious calcifications	37	15.8
Benign calcifications	25	10.6
Skin and trabecular thickening	36	15.3
BIRADS 1	17	9.4%
2	30	16.6%
3	45	25.0
4	83	46.1
5	5	2.8

Table 5. Change in individual breast lesion grading on account of digital breast tomosynthesis (DBT) compared to digital mammography (DM)

DM	DBT
	BIRADS 1	BIRADS 2	BIRADS 3	BIRADS 4 a	BIRADS 4 b	BIRADS 4 c	BIRADS 5	Total
BI-RADS 2	0 (0%)	12 (6.7%)	0 (0%)	1 (0.6%)	0 (0%)	0 (0%)	0 (0%)	13 (7.2%)
BI-RADS 3	17 (9.4%)	16 (18.3%)	42 (23.3%)	19 (10.6%)	0 (0%)	9 (5%)	0 (0%)	103 (57.2%)
BI-RADS 4a	0 (0%)	1 (0.6%)	3 (1.7%)	17 (9.4%)	0 (0%)	4 (2.2%)	0 (0%)	25 (13.8%)
BI-RADS 4 b	0 (0%)	1 (0.6%)	0 (0%)	1 (0.6%)	9 (5%)	5 (2.8%)	2 (1.1%)	18 (10%)
BI-RADS 4 c	0 (0%)	0 (0%)	0 (0%)	1 (0.6%)	0 (0%)	17 (9.4%)	3 (1.7%)	21 (11.7%)
BI-RADS 5	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
Total	17(9.4%)	30 (16.6%)	45 (25%)	39 (21.7%)	9 (5%)	35 (19.4%)	5 (2.8%)	180 (100%)

Table 6. Significance of category changes after revising DBT according to lesion type

	Total No.(%)	No categorchange No.(%)	Category increased cases No. (%)	Category decreased cases. No. (%)	Chi square test/fisher test
					X^2	p-value
Asymmetry	104 (100%)	39 (37.5%)	31 (29.8%)	34 (32.7%)	39.92	<0.0001*
Calcification	69 (100%)	61 (88.6%)	5 (7.1%)	3 (4.3%)	-	0.99
Distortion	35 (100%)	28 (80%)	5 (14.3%)	2 (5.7%)	57.91	<0.0001*

Table 7. The diagnostic performance of full-field digital mammography, Digital Breast Tomosynthesis, and combined tomosynthesis and mammography for confident breast cancer diagnosis considered histopathology as the reference standard

Criterion	Mammography % (95% CI)	Tomosynthesis % (95% CI)	Combined % (95% CI)
Accuracy	77.2%	94%	95.6%
Sensitivity	64% (53.0–74.0%)	96% (90–99%)	96% (90–99%)
Specificity	88.5% (80.4–94.1%)	93% (86–97%)	95% (88–98%)
Positive likelihood ratio	5.61 (3.15–10.0)	13.22 (6.47–27.02)	18.51 (7.88–43.5)
Negative likelihood ratio	0.4 (0.30–0.54)	0.04 (0.01–0.12)	0.04 (0.01–0.11)
Disease prevalence	47% (39.0–54.0%)	47% (39–54%)	47% (39–54%)
PPV	83.1% (72.0–91.0%)	92% (84–97%)	94% (89–98%)
NPV	73.9% (65.0%-82.0%)	97% (91–99%)	97% (91–99%)

Figure 1. Comparison of the ROC areas of FFDM, DBT and combined FFDM and DBT for breast cancer diagnosis evidenced by histopathology as the reference standard

Figure 2. 49 years old case, with positive family history of breast cancer, presented for screening mammography (a) Craniocaudal and (b) Medio lateral oblique DM images of both breasts revealed suspected distortion noted at the inner portion of the left breast. (c) and (d) Craniocaudal and mediolateral oblique Tomosynthesis revealed an area of suspicious architectural distortion.(e) Ultrasound revealed an ill-defined hypo-echoic mass lesion with posterior shadowing. US guided core biopsy was performed and revealed infiltrating ductal carcinoma

Figure 3. 54 years old patient complaining of left breast lump. (a) Craniocaudal, (b) mediolateral oblique DM images of both breasts revealed an area of focal asymmetry at the left upper inner quadrant, (c) Thin cuts tomosynthesis revealed internal fat density areas with central lucency. (d) Ultrasound showed a small hyper echoic area confined to the pre mammary region, with central cystic changes which was proved to be focal traumatic fat necrosis

Figure 4. 32 years old lactating patient complaining of fever and right sided inflammation of the breast; (a) Medio-lateral oblique DM images of both breasts revealed right sided diffuse skin and trabecular thickening with suspicious pleomorphic segmental calcifications, (b &c) Cranio-caudal and medio-lateral oblique DBT images of right breast revealed ill-defined sizable heterogeneous mass lesion with focal retraction (tenting) of the posterior retro mammary fat and suspicious segmental calcifications, (d) CESM reveals a heterogeneously enhancing ill-defined mass lesion with related non-mass enhancement surrounding the index lesion, (e&f) Ultrasound revealed multiple ill-defined hypo-echoic masses with overlying calcifications with intervening suspicious dilated ducts. US core biopsy revealed multifocal infiltrating ductal carcinoma