Participant characteristics
A total of 172 patients were enrolled, with an average age of 41.3 years old (SD±13.3). Breast cups A, B, C, and D and above accounted for 26%, 40%, 27%, and 7%, respectively. LN values of 0, 1, 2, and 3 and more were distributed as present in the 8%, 23%, 19%, and 50%, respectively, of the patients (Table 1).
In this study, 15 cases had no lesions confirmed by ultrasound during the follow-up, while the remaining 157 women had a total of 555 lesions. Of them, 132 women underwent surgery or puncture and counted a total of 163 lesions that were: 1) malignant (N=51): 36 lesions of invasive breast carcinoma (non-special type, NST), 6 lesions of ductal carcinoma in situ(DCIS), 3 lesions of mucinous carcinoma, 5 lesions of papillary carcinoma, and 1 lesion of adenoid cystic carcinoma; and 2) benign (N=112): 81 lesions of fibroadenomas, 4 lesions of inflammatory lesions, 6 lesions of intraductal papillomas, 18 lesions of adenopathy, 1 lesion of lipoma and 2 lesions of fat calcified nodules. There were 392 lesions without pathological results. Of them, 70 lesions were confirmed as benign by MRI. The remaining 322 lesions were confirmed as benign by ultrasound in the follow-up (Table 1).
Table 1
Basic characteristics of enrolled patients
Group
|
Enrolled cases (n=172)
|
Benign (n=128)
|
Malignant (n=44)
|
Lesion number
|
|
|
|
0
|
15(8%)
|
15(12%)
|
0(0%)
|
1
|
41(23%)
|
18(14%)
|
23(52%)
|
2
|
32(19%)
|
23(18%)
|
9(20%)
|
3
|
28(16%)
|
22(17%)
|
6(14%)
|
4
|
19(11%)
|
15(12%)
|
4(9%)
|
≥5
|
37(23%)
|
35(27%)
|
2(5%)
|
Breast cup size
|
|
|
|
A
|
44(26%)
|
32(25%)
|
12(27%)
|
B
|
69(40%)
|
54(42%)
|
15(34%)
|
C
|
47(27%)
|
34(27%)
|
13(30%)
|
D and above
|
12(7%)
|
8(6%)
|
4(9%)
|
Examination time with AI-Breast and HHUS
Scan times of Groups A, B, and AI were 237.5±110.3 s(TA), 281.2±86.1 s(TB), and 262.15±40.4 s (TAI scan), respectively. Group AI interpretation time was 225.6±43.6 s (TAI interpretation). TA was significantly shorter than TB and TAI scan (Both P༜0.05), and TAI scan was relatively shorter than TB (P=0.057) (Fig. 1).
In the non-lesion groups, the scan times were 173.7±34.9 s (TA), 217.5±50.6 s (TB), and 239.5±30.5 s (TAI scan), while TAI interpretation was 193.7±34.1 s. For those with LN 1-5, times were 223.5±75.9 s (TA), 275.8±77.4 s (TB), 259.8±35.4 s (TAI scan), and 223.7±38.8s (TAI interpretation). In the LN ༞5 cases, times were 340.6±190.9 s (TA), 342.6±106.1 s (TB), 286.3±55.7 s (TAI scan), and 251.1±56.1 s (TAI interpretation). The scan times of patients with lesions were significantly longer than those of patients without lesions. In particular, Groups A and B had a positive linear correlation between the scan time and the LN (ra =0.512, rb =0.397). The scan time in Group AI did not correlate with LN (rAI scan=0.156). When LN was 1-5, there was no statistical significance (P>0.05) (Table 2 and 4).
Furthermore, scan times significantly increased with breast size in the three groups. In particular, TAI scan had a positive linear correlation with breast size (rAI scan=0.745) (Table 2 and 4). On the other hand, the scan times of benign cases from Group A and group B were significantly longer than those of malignant cases (P< 0.05). TAI scan did not correlate with benign or malignant lesions, and there were no significant differences (P = 0.688) (Table 2 and 4).
Lesion detection rate with AI-Breast and HHUS
In this study, the average detection rates of breast lesions were: Group A (95.0±13.6%), Group AI (92.8±17.0%) and Group B (85.0±22.9%). No statistical significance was noted between Group AI and Group A (P=0.367). However, differences were found when comparing Group A and Group AI, respectively, with Group B (both P<0.05).
Detection rates in Group A and Group AI did not correlate with LN, and there was no statistical significance. For its part, the detection rate in Group B had a negative linear correlation with the LN (rb= -0.436, P༜0.001). In Group B, patients with LN≥2 and C cup and above had a decreased detection rate that was significantly lower than those in Group A and Group AI (P<0.05) (Table 3 and 4). Moreover, detection rates throughout the three groups did not correlate with breast size (P>0.05) (Table 3 and 4).
For benign lesions, missed diagnosis rates in Group A, Group AI, and Group B were 7%, 8%, and 20%, respectively. Group B had higher rates than Group A and Group AI (both P<0.05) while there were no statistical differences between Group AI and Group A (P=0.629). For malignant lesions, missed diagnosis rates in Group A, Group AI, and Group B were 4%, 8%, and 14%, respectively. There were no statistical differences between them (P>0.05) (Table 5).
Table 2
Impact of LN, breast size, and pathological properties on the examination time in the three groups
Group
|
HHUS
|
AI-Breast
|
TA (s)
|
TB (s)
|
TAI scan (s)
|
TAI interpretation (s)
|
Lesion Number
|
|
|
|
|
0
|
173.7±34.9
|
217.5±50.6
|
239.5±30.5
|
193.7±34.1
|
1≤LN ≤5
|
223.5±75.9
|
275.8±77.4
|
259.8±35.4
|
223.9±38.8
|
>5
|
340.6±190.9
|
342.6±106.1
|
286.3±55.7
|
251.1±56.1
|
P-value
|
<0.001
|
<0.001
|
<0.001
|
<0.001
|
1
|
183.5±61.7
|
254.4±63.7
|
265.9±37.5
|
228.1±37.9
|
2
|
210.9±67.7
|
254.0±69.3
|
254.6±35.0
|
220.3±37.0
|
3
|
258.8±71.9
|
296.4±80.6
|
262.1±35.4
|
225.1±44.2
|
4
|
248.1±84.2
|
295.0±81.1
|
250.0±27.4
|
215.8±34.7
|
5
|
282.2±59.9
|
338.9±91.8
|
263.0±42.0
|
230.6±41.0
|
P-value
|
<0.001
|
0.003
|
0.474
|
0.758
|
Breast cup size
|
|
|
|
|
A
|
205.0±84.1
|
248.3±71.2
|
230.6±20.3
|
193.7±25.0
|
B
|
228.4±75.3
|
284.7±88.6
|
250.1±20.8
|
215.2±26.8
|
C
|
231.9±72.2
|
293.6±89.1
|
286.8±27.8
|
248.8±35.8
|
D and above
|
431.8±242.1
|
332.8±76.1
|
350.4±43.9
|
311.3±47.9
|
P-value
|
<0.001
|
0.007
|
<0.001
|
<0.001
|
Pathological properties
|
|
|
|
|
Number of malignant lesions (n=44)
|
220.0±122.4
|
252.4±64.5
|
266.0±42.6
|
228.8±43.3
|
Number of benign lesions (n=113)
|
243.6±105.7
|
291.1±90.5
|
260.9±39.8
|
224.5±43.8
|
P-value
|
0.027
|
0.038
|
0.688
|
0.770
|
Table 3
Impact of LN and breast size on the detection rates in the three groups
|
HHUS
|
Group AI
|
Comparison of P-value between AI and HHUS
|
|
Group A
|
Group B
|
P
|
Group A
|
Group B
|
Lesion Number
|
|
|
|
|
|
|
0
|
100.0±0.0
|
100.0±0.0
|
1.0000
|
100.0±0.0
|
1.0000
|
1.0000
|
1
|
100.0±0.0
|
95.1±21.8
|
0.7038
|
95.1±21.8
|
0.7038
|
1.0000
|
2≤LN ≤3
|
92.5±79.8
|
79.8±26.2
|
0.0152
|
90.6±17.7
|
0.5816
|
0.0450
|
≥4
|
92.6±14.4
|
79.1±18.3
|
0.0001
|
91.5±13.8
|
0.5924
|
0.0002
|
P-value
|
0.054
|
<0.001
|
NA
|
0.385
|
NA
|
NA
|
Breast size
|
|
|
|
|
|
|
A
|
93.5±14.0
|
86.1±20.7
|
0.1572
|
93.8±14.0
|
0.8806
|
0.1247
|
B
|
95.9±13.7
|
87.8±20.9
|
0.0216
|
94.3±15.4
|
0.4970
|
0.0812
|
C
|
93.8±14.5
|
82.8±24.5
|
0.0031
|
88.2±22.0
|
0.3260
|
0.0461
|
D and above
|
100.0±0.0
|
73.3±31.9
|
0.0153
|
98.5±5.2
|
0.7290
|
0.0262
|
P value
|
0.416
|
0.197
|
NA
|
0.193
|
NA
|
NA
|
Table 4
Correlation between LN, breast size, examination time, and detection rate in the three groups
|
Lesion Number
|
Breast size
|
Examination time
|
|
|
Scan time of Group AI
|
r =0.156 (P=0.042)
|
r = 0.745 (P< 0.001)
|
Interpretation time of Group AI
|
r =0.195 (P=0.011)
|
r = 0.665 (P< 0.001)
|
Group A
|
r = 0.512 (P< 0.001)
|
r =0.267 (P< 0.001)
|
Group B
|
r = 0.397 (P< 0.001)
|
r =0.268 (P< 0.001)
|
Detection rate
|
|
|
Group AI
|
r =-0.248 (P=0.001)
|
r =-0.0756 (P=0.218)
|
Group A
|
r = -0.275 (P< 0.001)
|
r =-0.0943 (P=0.536)
|
Group B
|
r = -0.436 (P< 0.001)
|
r =-0.0473 (P=0.218)
|
Table 5
Influence of benign and malignant lesions on the examination time and missed detection rate in the three groups
|
HHUS
|
Group AI
|
Comparison of P-value between AI and HHUS
|
|
Group A
|
Group B
|
P
|
Group A
|
Group B
|
Missed diagnosis rate (LN)
|
|
|
|
|
|
|
Number of benign lesions (n=504)
|
35(7%)
|
102(20%)
|
< 0.001
|
40(8%)
|
0.6290
|
<0.001
|
Number of malignant lesions (n=51)
|
2(4%)
|
7(14%)
|
0.156
|
4(8%)
|
0.6646
|
0.515
|
Chi-square test P
|
0.186
|
0.398
|
NA
|
0.787
|
NA
|
NA
|