Study population
From January 2018 to December 2021, 286 female patients were included in this study, of whom 64 were excluded. At last, a total of 222 patients aged from 23 to 78 (Median, 50; IQR, 42-56) years old with 223 target ALNs were eligible for this study, of whom one had bilateral breast cancer. According to the surgical histopathology results, 44.84% (n=100) of ALNs were metastatic, whereas 55.16% (n=123) were benign. Of metastatic axilla, 53 were with less than 3 metastatic ALNs and 47 were with 3 or more metastatic ALNs. Table 1 details clinical and pathological characteristics of the patients and lesions.
Table 1. Clinical and pathological characteristics of 223 lesions in 222 breast cancer patients.
Characteristics
|
Number of patients or lesions (%)
|
Age
|
|
< 40 years old
|
38(17.12)
|
≥ 40 years old
|
184(82.88)
|
Tumor location
|
|
Left
|
124(55.86)
|
Right
|
97(43.69)
|
Bilateral
|
1(0.45)
|
Clinical T stage
|
|
T1
|
94(42.15)
|
T2
|
129(57.85)
|
Histologic type
|
|
Invasive ductal carcinoma
|
174(78.03)
|
Invasive lobular carcinoma
|
4(1.79)
|
Others
|
45(20.18)
|
ALN status
|
|
Metastatic
|
100(44.84)
|
Non-metastatic
|
123(55.16)
|
ER status
|
|
Negative
|
59(26.46)
|
Positive
|
163(73.09)
|
Not available
|
1(0.45)
|
PR status
|
|
Negative
|
75(33.63)
|
Positive
|
147(65.92)
|
Not available
|
1(0.45)
|
HER2 status
|
|
Negative
|
144(64.57)
|
Positive
|
77(34.53)
|
Not available
|
2(0.90)
|
Ki-67
|
|
≥ 14%
|
199(89.24)
|
< 14%
|
24(10.76)
|
Axillary surgery
|
|
SLNB
|
100(44.84)
|
ALND
|
123(55.16)
|
Lymphovascular invasion
|
|
Present
|
73(32.74)
|
Absent
|
130(58.29)
|
Not available
|
20(8.97)
|
Histologic grade
|
|
I
|
2(0.90)
|
II
|
96(43.05)
|
III
|
95(42.60)
|
Not available
|
30(13.45)
|
ALN, axillary lymph node; ER, estrogen receptor; PR, progesterone receptor; HER2, human epidermal growth factor receptor; The age and tumor location were described based on number and percentage of patients. The remaining clinical and pathological characteristics were described based on number and percentage of lesions.
The performance of axillary Conventional US, SWE and Conventional US Combined with SWE
The Conventional US characteristics of benign and metastatic TALNs were shown in Table 2. The performance of Conventional US, SWE and Conventional US combined with SWE were shown in Table 3 and Figure 2.
Table 2. Conventional US characteristics of benign and metastatic Axillary lymph nodes.
Characteristics
|
Benign (123)
|
Metastatic (100)
|
p Value
|
Cortical thickness (mm)
|
2.6 (2.0-3.5) *
|
4.0 (3.0-6.9)
|
< 0.001
|
Long-diameter (mm)
|
14.4 (9.7-19.5)
|
13.8 (11.3-19.9)
|
= 0.305
|
Short-diameter (mm)
|
6.2 (4.7-7.2)
|
8.0 (6.2-10.7)
|
< 0.001
|
Short / Long diameter ratio
|
0.44 (0.34-0.56)
|
0.56 (0.46-0.71)
|
< 0.001
|
Hilum
|
|
|
< 0.001
|
Present
|
117 (95.1%)
|
57 (57.0%)
|
|
Absent
|
6 (4.9%)
|
43 (43.0)
|
|
Blood Flow
|
|
|
< 0.001
|
Non- or Normal
|
116 (94.3)
|
54 (54.0)
|
|
NHBF
|
7 (5.7%)
|
46 (46.0)
|
|
* median (interquartile range [IQR]); US, Ultrasound; NHBF, Non-hilum Blood Flow
Table 3. Diagnostic performance of conventional US, SWE, US and SWE, US or SWE and FNAB in predicting ALN status.
Method
|
Sensitivity %
|
Specificity %
|
PPV %
|
NPV %
|
Accuracy %
|
AUC
|
US
|
78.00
[68.61-85.67]
|
60.98
[51.77-69.64]
|
61.90
[56.00-67.47]
|
77.32
[69.66-83.50]
|
68.61
|
0.70 [0.63-0.76]
|
SWE
|
65.00
[54.82-74.27]
|
66.67
[57.60-74.91]
|
61.32
[54.30-67.90]
|
70.09
[63.56-75.88]
|
65.92
|
0.66
[0.59-0.72]
|
US and SWE
|
57.00
[46.71-66.86]
|
85.37
[77.86-91.09]
|
76.00
[66.67-83.37]
|
70.95
[65.83-75.58]
|
72.65
|
0.71
[0.65-0.77]
|
US or SWE
|
86.00
[77.63-92.13]
|
42.28
[33.42-51.51]
|
54.78
[50.52-58.96]
|
79.45
[68.66-86.30]
|
61.89
|
0.64
[0.57-0.70]
|
US, ultrasound; SWE, shear wave elastography; US or SWE, US positive or SWE positive; US and SWE, US positive and SWE positive; PPV, positive predictive value; NPV, negative predictive value; AUC, area under the receiver operating characteristic curve.
Data in brackets are the 95% confidence intervals.
The diagnostic performance of SWE parameters in evaluating the ALN status with reference to surgical staging was shown in Appendices. SWVmax with the superior sensitivity (65.00%) was chosen to be the performance of SWE, was used to combine the conventional US and guide the FNAB of axilla (Figure 3).
Two different combination methods of US and SWE were used to explore the diagnostic performance for metastatic ALNs, which were defined as “US or SWE” and “US and SWE”. In the “US or SWE” method, ALN was considered as suspicious if either the result of conventional US or SWE was positive. While in the “US and SWE” method, ALN was considered as suspicious if both the results of conventional US and SWE were positive. The method of “US or SWE” showed the highest sensitivity compared with conventional US alone (86.00% vs. 78.00%, p< 0.001), SWE alone (86.00% vs. 65.00%, p< 0.001) and the method of “US and SWE” (86.00% vs. 57.00%, p< 0.001), while the “US and SWE” showed the highest specificity among the conventional US alone (85.37% vs. 60.98%, p< 0.001), SWE alone (85.37% vs. 66.67%, p< 0.001) and method of “US or SWE” (85.37% vs. 42.28%, p< 0.001).
The performance of FNAB in different groups
The performance of FNAB in different groups was shown in Table 4. Out of the 75 individuals who had positive results from both conventional US and SWE, the sensitivity, specificity, PPV, NPV, and accuracy of FNAB were 96.49%, 88.89%, 96.49%, 88.89%, and 94.67% respectively. And of 55 true positive cases, 63.64% were high-burden patients. Out of the 31 patients who tested negative on conventional US but positive on SWE, 8 were confirmed to have ALN metastasis through surgical pathology. Among these 8 patients, 4 cases (50.00%) were identified to have ALN metastasis through FNAB, and 3 of 4 cases (75.00%) was high burden disease.
Table 4. The performance of FNAB in different groups.
Group
|
n
|
Sensitivity %
(No. of cases)
|
Specificity %
(No. of cases)
|
PPV %
(No. of cases)
|
NPV %
(No. of cases)
|
Accuracy
%
|
ALL
|
223
|
73.00(73/100)
|
98.37(121/123)
|
97.33(73/75)
|
81.76(121/148)
|
87.00
|
US +
|
126
|
87.18 (68/78)
|
95.83 (46/48)
|
97.14 (68/70)
|
82.14 (46/56)
|
90.48
|
US + SWE +
|
75
|
96.49 (55/57)
|
88.89 (16/18)
|
96.49 (55/57)
|
88.89 (16/18)
|
94.67
|
US + SWE –
|
51
|
61.90 (13/21)
|
100.00 (30/30)
|
100.00 (13/13)
|
78.94 (30/38)
|
84.31
|
US –
|
97
|
22.73 (5/22)
|
100.00 (75/75)
|
100.00 (5/5)
|
81.52 (75/92)
|
82.47
|
US – SWE +
|
31
|
50.00 (4/8)
|
100.00 (23/23)
|
100.00 (4/4)
|
85.18 (23/27)
|
87.10
|
US – SWE –
|
66
|
7.14 (1/14)
|
100.00 (52/52)
|
100.00 (1/1)
|
80.00 (52/65)
|
80.30
|
SWE +
|
106
|
90.77 (59/65)
|
95.12 (39/41)
|
96.72 (59/61)
|
86.67 (39/45)
|
92.45
|
FNAB, Fine-needle aspiration biopsy; US, ultrasound; SWE, shear wave elastography; PPV, positive predictive value; NPV, negative predictive value; +, positive; –, negative;Data in brackets are present as the number of FNAB/ Surgery cases.
The performance of different methods to guide FNAB of axilla and FNR
The performance of different methods to guide FNAB was shown in Table 5 and Figure 4. The AUC of the US or SWE-guided FNAB [0.85 (95% CI, 0.80-0.90)] was significantly higher than that of US-guided FNAB [0.83 (95% CI, 0.78-0.88), p = 0.042], SWE-guided FNAB [0.79 (95% CI, 0.73-0.84), p = 0.001], and US and SWE-guided FNAB [0.77 (95% CI, 0.71-0.82), p < 0.001]. The FNR of the US or SWE-guided FNAB was lowest with a value of 28.00%, and the FNR of US-guided FNAB, SWE-guided FNAB and US and SWE-guided FNAB were 32.00%, 41.00% and 45.00%, respectively. Four patients were detected by the US or SWE-guided FNAB but missed by the method of US-guided FNAB, 3 of whom had high burden disease (≥3 lymph node metastases).
Table 5. The performance of different modality-guided FNAB in predicting ALN status.
different modality-guided FNAB
|
|
Metastatic Axilla
(N=100)
|
Non-metastatic Axilla
(N=123)
|
AUC
|
Sensitivity
%
|
Specificity
%
|
PPV%
|
NPV%
|
Accuracy
%
|
US-guided FNAB
|
Positive
(N=70)
|
68
|
2
|
0.83a
[0.78-0.88]
|
68.00
[57.92-76.98]
|
98.37
[94.25-99.80]
|
97.14
[89.52-99.27]
|
79.08
[73.95-83.43]
|
84.75
|
Negative
(N=153)
|
32
|
121
|
SWE-guided FNAB
|
Positive
(N=61)
|
59
|
2
|
0.79b
[0.72-0.84]
|
59.00
[48.71-68.74]
|
98.37
[94.25-99.80]
|
96.72
[88.08-99.16]
|
74.69
[69.97-78.89]
|
80.72
|
Negative
(N=162)
|
41
|
121
|
US and SWE-guided FNAB
|
Positive
(N=57)
|
55
|
2
|
0.77c
[0.71-0.82]
|
55.00
[44.73-64.97]
|
98.37
[94.25-99.80]
|
96.49
[87.30-99.10]
|
72.89
[68.38-76.98]
|
78.92
|
Negative
(N=166)
|
45
|
121
|
US or SWE-guided FNAB
|
Positive
(N=74)
|
72
|
2
|
0.85
[0.80-0.90]
|
72.00
[62.13-80.52]
|
98.37
[94.25-99.80]
|
97.30
[90.06-99.31]
|
81.21
[75.92-85.55]
|
86.55
|
Negative
(N=149)
|
28
|
121
|
FNAB, Fine-needle aspiration biopsy; US, ultrasound; SWE, shear wave elastography; AUC, area under the receiver operating characteristic curve; PPV, positive predictive value; NPV, negative predictive value; +, positive; –, negative;Data in brackets are the 95% confidence intervals.
a indicates p = 0.042, Delong et al. in comparison with US or SWE-guided FNAB.
b indicates p = 0.001, Delong et al. in comparison with US or SWE-guided FNAB.
c indicates p < 0.001, Delong et al. in comparison with US or SWE-guided FNAB.
Preliminary exploration of independent predictors of axillary burden in early breast cancer patients
Table 6. Univariate and multivariate logistic analyses to determine the independent predictors associated with axillary burden.
Characteristics
|
Cut-off value
|
Univariate logistic
Regression
|
Multivariate logistic
Regression
|
Odds Ratio
(95% CI)
|
P value
|
Odds Ratio
(95%CI)
|
P value
|
Age
|
|
|
0.055
|
|
|
cT stage
|
|
|
0.112
|
|
|
Number of suspicious ALNs at US
|
≤2
|
1.0
|
|
1.0
|
|
>2
|
27.632
(9.580-79.649)
|
<0.001
|
23.204
(5.290-101.777)
|
<0.001
|
SWE
|
≤2.46m/s
|
1.0
|
|
|
0.588
|
>2.46m/s
|
7.931
(3.495-17.995)
|
<0.001
|
|
|
FNAB
|
Negative
|
1.0
|
|
1.0
|
|
Positive
|
48.375
(16.203-144.430)
|
<0.001
|
39.300
(8.353-184.898)
|
<0.001
|
Characteristics of TALN
|
|
|
|
|
|
Long axis
|
≤15.9mm
|
|
0.839
|
|
|
>15.9mm
|
|
|
|
|
Short axis
|
≤6.6mm
|
1.0
|
|
|
0.166
|
>6.6mm
|
5.820
(2.652-12.770)
|
<0.001
|
|
|
Cortical thickness
|
≤3.0mm
|
1.0
|
|
|
0.659
|
>3.0mm
|
11.312
(4.270-29.965)
|
<0.001
|
|
|
Blood flow
|
Non- or normal
|
1.0
|
|
|
0.235
|
Peripheral or mixed
|
6.006
(2.980-12.108)
|
<0.001
|
|
|
Hilum
|
Normal
|
1.0
|
|
|
0.367
|
Shift or disappear
|
6.303
(3.094-12.838)
|
<0.001
|
|
|
US, ultrasound; ALNs, axillary lymph nodes; TALN, target axillary lymph node; CI, confidence interval; SWE, shear wave elastography; FNAB, fine-needle aspiration biopsy.
The univariate and multivariate logistic regression analyses based on clinical and ultrasound characteristics were performed to explore the independent predictors of axillary burden in early breast cancer patients. As shown in Table 6, the number of suspicious ALNs (>2, P<0.001), SWE (>2.46 m/s, P<0.001), positive FNAB (P<0.001) and characteristics of TALN (short axis [>6.6 mm, P<0.001], Hilum [lymphatic hilum shifted or disappeared, P<0.001], Blood signal [mixed-blood flow or peripheral blood flow, P<0.001], Cortical thickness [>3.0 mm, P<0.001] were detected to be significantly associated with axillary burden. Furthermore, multivariable analysis showed that the enrolled patients with positive FNAB result (odds ratio [OR],39.300, 95% CI, 8.353-184.898; p < 0.001) and more than 2 suspicious ALNs (odds ratio [OR],23.204, 95% CI, 5.290-101.777; P < 0.001) were observed to be the independent factors of axillary burden.