Does axillary lymph node size predict better metastatic involvement than apparent diffusion coefficient (ADC) value in women with newly diagnosed breast cancer?

Abstract

Background

It has been demonstrated that the number of metastatic axillary lymph nodes (mALNs) influence disease-free and overall survival in patients with breast cancer.

Purpose

To determine if the ALN size is more accurate than the ALN apparent diffusion coefficient (ADC) value to predict metastatic involvement in newly diagnosed breast cancer.

Material and Methods

A total of 44 patients with breast cancer were included. Magnetic resonance imaging (MRI) examinations were performed on a 1.5-T system with sagittal T1-weighted fast spin-echo non-fat saturated, sagittal T2-weighted fast spin-echo non-fat saturated, axial diffusion-weighted imaging echo-planar (b values of 0 and 700 s/mm²), and non-contrast axial VIBRANT sequences. The size and the ADC value were obtained for ALN ipsilateral and contralateral to breast cancer. The reference standard was the histopathologic lymph node status.

Results

mALN had a greater cortical thickness compared to contralateral non-mALN (10.3 ± 5.32 mm vs. 4 ± 1.17 mm, P ≤ 0.001). The threshold of ≥6.7 mm for predicting axillary metastatic involvement had a sensitivity and a specificity of 80.0% and 97.7%, respectively. The ADC value of mALN was significantly higher than the contralateral non-mALN (0.90 ± 0.12 × 10⁻³mm²/s vs. 0.78 ± 0.12 × 10⁻³mm²/s; P = 0.001). The threshold of ≥0.86 × 10⁻³mm²/s had a sensitivity and a specificity of 66.7% and 76.7%, respectively.

Conclusion

Our results indicate that the cortical thickness has a better diagnostic performance in the differentiation of metastatic and non-metastatic lymph nodes than the lymph node ADC.

Keywords

Breast cancer axillary lymph node axillary staging magnetic resonance imaging magnetic resonance diffusion apparent diffusion coefficient

Introduction

Since the 1980s, it has been demonstrated that the number of metastatic axillary lymph nodes (mALN) influence disease-free and overall survival in patients with early breast cancer (1). For many years, the axillary lymph node dissection (ALND) was the gold standard procedure for both LN staging and local disease control (2). However, the evolution of axillary surgery has been modified to a staging role, leaving radical axillary surgical dissection in cases with locally advanced breast cancer or bulky disease (3). The sentinel lymph node biopsy (SLNB) is the accepted procedure for accurate LN staging, leaving radical ALND for patients with confirmed LN metastasis (4 –6).

For a long time, the preoperative evaluation of the ALN was based on clinical examination; however, the accuracy of this method was low, since up to 40% of the patients with negative clinical examination have metastases (7, 8) and 30% of the patients with palpable nodes have a result of negative LN (7). Metastatic cancer cells reach the LN by afferent lymphatic channels that drain into subcapsular space and cortical sinusoids (9). Metastatic cell growth produces cortex thickening and eventually progresses to complete obliteration of the hilum. In addition, metastatic involvement tends to increase disproportionately to the minor axis of the LN, producing a spherical morphology (10). These changes can be evaluated by imaging. Ultrasound is the most common method used for axillary evaluation in breast cancer. However, it is an operator-dependent evaluation with a diagnostic accuracy that depends on the morphological criteria used (11).

In the field of magnetic resonance imaging (MRI), Luciani et al. (12) and Arslan et al. (13) reported that the minor axis of the ALN is a good parameter to differentiate between mALN and non-mALN. Diffusion-weighted imaging (DWI) is a non-contrasted MRI sequence that provides a quantitative parameter of the isotropic diffusion of water in tissues, generally used for the characterization of breast masses (14). The role of DWI in the evaluation of LN, has been explored by different authors like Scaranelo et al. (15) and Razek et al. (16).

Scaranelo et al. (15) found that the mALN have lower values of the apparent diffusion coefficient (ADC) than non-mALN (15). Other studies have presented similar results (17 –19). However, other investigators have found that mALN have higher ADC values than benign LN (20), and others found no significant difference between the ADC value of benign and metastatic LN (21, 22).

To our knowledge, the uniqueness of this study is that it demonstrates differences in size and ADC value between the ALN ipsilateral to the tumor against the contralateral LN in patients with breast cancer. The main aim of the present study was to determine whether the ALN size is more accurate than the ALN ADC value to predict metastatic involvement in newly diagnosed breast cancer, determined by the gold standard of histopathology.

Material and Methods

Patients

This prospective descriptive study was conducted in accordance with approval from the Ethics and Research Committee of the Hospital Universitario “Dr. José Eleuterio González”, in Monterrey, Mexico (registration code: RA15-002). All patients provided written informed consent. Patients not willing to participate or who were physically unable to tolerate the prone position needed for breast MRI were excluded.

A simple breast MRI was performed in 65 patients with breast cancer before a tumor or lymph node biopsy. The enrollment period was from February 2015 to November 2015.

Two patients with a diagnosis of bilateral breast cancer were eliminated as well as 10 patients with a history of previous contralateral breast cancer treated with mastectomy and axillary dissection. Nine patients whose imaging studies could not be recovered from the picture archiving and communication system (PACS) were also eliminated. A total of 44 patients were included.

MRI acquisition

Examinations were performed in the prone position on a 1.5-T MR system (EXCITE HDX-Y, General Electric, Wilmington, MA, USA), with a bilateral eight-channel breast coil. Our MRI protocol for this work was as follows: sagittal T1-weighted (T1W) fast spin-echo (FSE) non-fat saturated sequence; sagittal T2-weighted (T2W) FSE non-fat saturated sequence; axial DWI echo-planar sequence; and non-contrast axial VIBRANT sequence. DWI was performed using an axial echo-planar imaging sequence, with b values of 0 and 700 s/mm²; TR = 4825 ms (range = 3000–6000); TE = 87.9 ms; bandwidth = 62.5–250 MHz; matrix size = 128 × 128; field of view = 32 cm; slice thickness =3 mm; intersection gap = 0 mm. Total acquisition time was 12–15 min.

Post-processing: size and ADC measurements

All images were transferred to a workstation (IntelliSpace Portal version 7, Philips, Best, The Netherlands). We selected one LN in both axillae of each patient: ipsilateral and contralateral to breast cancer. In the ipsilateral axilla, the most suspicious LN of having metastasis (e.g. cortical thickening, obliteration of the fatty hilum, or mass appearance) was chosen; if there was no suspicious LN, the largest LN was selected. In the contralateral axilla, a normal morphology LN (with thin cortical and preservation of the fatty hilum) was selected.

The cortical thickness, the major axis, and the minor axis of the ALN were measured on the non-contrast axial VIBRANT sequence. In addition, we calculated a LN size ratio as follows: major axis/minor axis.

The DWI sequence was post-processed with a commercial software (MR Diffusion, Philips, Best, The Netherlands) to obtain ADC maps and ADC measurements.

To achieve standardized conditions for the analysis, our method for measuring the ADC consists of marking the region of interest (ROI) using three small ellipses (mean area = 6.1 mm² with interval 3.3–11.7 mm² for ipsilateral axilla; mean area = 5.1 mm² with interval 2.3–8.7 mm² for contralateral axilla) placed randomly at different sites in the cortex of the LN, trying to avoid the fatty hilum. Finally, the mean of the ADC of the three ROI measurement was obtained. Representative images are shown in Fig. 1. In addition, we calculated an ADC ratio as follows: ipsilateral ALN ADC/contralateral ALN ADC.

Fig. 1.

Images of a 44-year-old woman. (a) Axial DWI (b = 700) zoom showing a LN with cortical thickening and compression of the fatty hilum in the right axilla. (b) ADC map at the same level that shows one of the three ROIs into the LN with an ADC value of 1.10 × 10⁻³mm²/s. (c) Non-contrast axial VIBRANT zoom shows the same right ALN, with a cortical thickness of 9.5 mm. (d) Axial DWI (b = 700) zoom showing a normal morphology LN in the left axilla. (e) ADC map at the same level that shows one of the three ROIs into the LN with an ADC value of 0.59 × 10⁻³mm²/s. (f) Non-contrast axial VIBRANT zoom shows the same left ALN, with a cortical thickness of 2.9 mm. According to the average of the three ROIs, the ADC on the right axilla was 0.94 × 10⁻³mm²/s and on the left axilla was 0.60 × 10⁻³mm²/s. This patient has an infiltrating ductal carcinoma poorly differentiated in the right breast. The right ALN was positive to metastases in ultrasound-guided core biopsy previous to surgical treatment. ADC, apparent diffusion coefficient; ALN, axillary lymph node; DWI, diffusion-weighted imaging; ROI, region of interest

Histopathological assessment

Patients with suspicious ALN in the conventional imaging evaluation before MRI (n = 35) had an ultrasound-guided biopsy of the LN: fine needle aspiration (FNA) in four patients and core needle biopsy (CNB) in 31 patients. It was decided not to perform an ultrasound-guided biopsy in nine patients, due to conventional imaging evaluation.

Of the 44 patients, SLNB was performed in 12 patients (in seven patients, SLNB was the first histological assessment) and ALND was performed in 29 patients (in two patients ALND was the first histological assessment). We do not have information about surgical procedures in seven patients who received their treatment in another hospital after the ultrasound-guided biopsy.

Twenty patients received neoadjuvant chemotherapy (NAC); they all had a histological assessment of the LN before the NAC (FNA in three patients, CNB in 14 patients, and SLNB in three patients).

For the purposes of this study, we considered mALN if invasive tumor cells were found in LN by any of the following methods: FNA; CNB; SLNB; and ALND. If tumor cells were not found, the LN was considered non-mALN.

All the participants had a conventional negative imaging (ultrasound and mammography) and clinical evaluation of the breast and axilla contralateral to cancer; in this manner, for the purposes of this study, we assumed that all ALN on the contralateral side of breast cancer also were non-mALN.

Statistical analysis

Relevant clinical and imaging information was collected. The statistical analysis was done using SPSS version 20.0 (SPSS, Inc., Armonk, NY, USA) and MedCalc version 17.9.7 (Applied Math, Mariakerke, Belgium). In the case of qualitative variables, frequencies and percentages were obtained. To compare categorical variables, we performed Pearson’s χ² test or Fisher’s exact test for 2 × 2 tables. We conducted a descriptive statistical analysis for quantitative variables, measures of central tendency, and dispersion.

To compare the means of LN size and LN ADC according to LN histopathology, we used Student’s t-test for previous corroboration of normal distribution of the data. First, we compared patients with mALN versus patients with non-mALN (Comparison 1). In second place, only in patients with mALN, we compared the ipsilateral axilla versus the contralateral axilla (Comparison 2). To counteract the problem of multiple comparisons we used the Bonferroni correction.

A multivariate logistic regression analysis was performed to identify significant variables to predict positive lymph nodes. Odds ratios (OR) and 95% confidence intervals (CI) were calculated.

Sensitivity and specificity were calculated to measure the performance of LN size and LN ADC according to LN histopathology. The optimal threshold was calculated with Youden index and conventional receiver operating characteristic (ROC) curve analysis. A P value ≤0.05 was considered statistically significant, except in cases where Bonferroni correction was used.

Results

The mean age of patients was 48 years (age range =25–76 years) at diagnosis. The breast tumor histopathology was: 40 patients with infiltrating ductal carcinoma (IDC); one patient with infiltrating lobular carcinoma; one patient with IDC mucinous type; one patient with infiltrating papillary carcinoma; and one patient with intracystic papillary carcinoma. Mean tumor size by MRI at diagnosis was 32.5 mm (range = 11–70 mm) for the 44 tumors. The ipsilateral ALN had a mean size by MRI of 17 mm (range =8.2–38.5 mm) and contralateral ALN of 14 mm (range = 8.1–31.5 mm). According to the histopathology, there were 30 patients with mALN and 14 patients with non-mALN. The flow of the patients through the study is presented in Fig. 2. Table 1 shows the clinicopathological characteristics and the conventional imaging features of the patients.

Fig. 2.

Study flow. The diagram shows the flow of patients through the study. ADC, apparent diffusion coefficient; mALN, metastatic axillary lymph node; MRI, magnetic resonance imaging; non-mALN, non-metastatic axillary lymph node; PACS, picture archiving and communication system; SLNB, sentinel lymph node biopsy.

Table 1.

Clinicopathological characteristics and imaging features of the patients with breast cancer.

Characteristics	Breast cancerpatients(n = 44)	Patients non-mALN (n = 14)	Patients mALN (n = 30)	P values
Age (years)	50.5 ± 11.6	51.8 ± 11.3	50 ± 12	0.63
Menopausal status				0.99
Premenopausal	35 (79.5)	11 (78.6)	24 (80)
Postmenopausal	9 (20.5)	3 (21.4)	6 (20)
Clinical history				0.62
Negative	34 (77.3)	11 (78.6)	23 (76.7)
Non-first-degree family with breast cancer	7 (15.9)	3 (21.4)	4 (13.3)
First-degree family with breast cancer	2 (4.5)	0 (0)	2 (6.7)
Ovarian cancer	1 (2.3)	0 (0)	1 (3.3)
Indication for examination				0.58
Asymptomatic	4 (9.1)	2 (14.3)	2(6.7)
Palpable breast mass	40 (90.9)	12 (85.7)	28 (93.3)
Axillae by mammography				0.31
Negative	24 (54.5)	10 (71.4)	14 (46.7)
Indeterminate	15 (34.1)	3 (21.4)	12 (40)
Positive	5 (11.4)	1 (7.1)	4 (13.3)
LN ultrasound morphology				0.07
Normal	6 (13.6)	4 (28.6)	2 (6.7)
Abnormal	38 (86.7)	10 (71.4)	28 (93.3)
LN ultrasound-specific morphology				0.02
Normal	6 (13.6)	4 (28.6)	2 (6.7)
Focal or diffuse cortical thickening	18 (40.9)	8 (57.2)	10 (33.3)
Compression of the fatty hilum	7 (15.9)	1 (7.1)	6 (20)
Mass or perinodal infiltration	13 (29.6)	1 (7.1)	12 (40)

Values are given as n (%) or mean ± SD.

P values calculated by Pearson’s χ² or Fisher’s exact tests for categorical variables and independent Student’s t-test for numerical variables.

LN, lymph node; mALN, metastatic axillary lymph node.

Comparison 1

In the analysis of LN size by MRI with the Bonferroni correction, we found that the LN size ratio of patients with mALN was shorter than in patients with non-mALN, with a significant statistical difference (1.39 ± 0.23 mm vs. 1.64 ± 0.29 mm, P = 0.004). We also found that the cortical thickness of the mALN was significantly larger than in patients with non-mALN (10.4 ± 5.32 mm vs. 5.9 ± 2.56 mm, P = 0.001). A descriptive analysis of the quantitative variables according to LN histopathology for comparison between patients is provided in Table 2.

Table 2.

Chart of Comparison 1, between LN size and LN ADC of patients with mALN and patients with non-mALN.

Parameters	Patients non-mALN (n = 14) (mm or × 10^–3mm²/s)	Patients mALN (n = 30) (mm or x 10^-3mm²/s)	P value*	P value^†	Exp β^‡
Ipsilateral LN major axis	14.2 ± 3.5	18.3 ± 8.5	0.028
Ipsilateral LN minor axis	8.9 ± 2.5	13.3 ± 5.5	0.001**	0.855	0.967
Ipsilateral LN size ratio	1.64 ± 0.29	1.39 ± 0.23	0.004**	0.060	0.040
Cortical thickness	5.9 ± 2.56	10.4 ± 5.32	0.001**	0.207	1.267
Ipsilateral LN ADC	0.85 ± 0.14	0.90 ± 0.12	0.24	0.350	31.595
ADC LN ratio	1.05 ± 0.32	1.18 ± 0.26	0.17

Values are given as mean ± SD.

*Calculated by Student’s t-test. With the Bonferroni correction the significance level to each variable must be α ≤ 0.007. The P values that meet this criterion are marked with **.

^†Calculated by logistic regression.

^‡Odds ratio.

ADC, apparent diffusion coefficient; LN, lymph node; mALN, metastatic axillary lymph node.

In juxtaposition of ipsilateral ALN ADC (as an independent value) of the 30 patients with mALN versus the 14 patients with non-mALN, we found that the ADC of patients with mALN was higher than patients with non-mALN, but we did not find any significant statistical difference (0.90 ± 0.12 × 10⁻³mm²/s vs. 0.85 ± 0.14 × 10⁻³mm²/s, P = 0.24).

When analyzing the normalized LN ADC, the 30 patients with mALN had a higher value than the 14 patients with non-mALN, but we did not find any significant statistical difference (1.18 ± 0.26 ×10⁻³mm²/s vs. 1.05 ± 0.32 × 10⁻³mm²/s, P = 0.17).

Comparison 2

When comparing the ipsilateral axilla versus the contralateral axilla, only in patients with mALN and with the Bonferroni correction, we found that the LN size ratio of the ipsilateral axilla was shorter than in the contralateral axilla, with a significant statistical difference (1.39 ± 0.23 mm vs. 1.93 ± 0.44 mm, P ≤ 0.001). We also found that the cortical thickness of the mALN was significantly larger than contralateral non-mALN (10.4 ± 5.32 mm vs. 4 ± 1.17 mm, P < 0.001). A descriptive analysis of the quantitative variables according to LN histopathology for comparison between axillae in patients with metastasis is provided in Table 3.

Table 3.

Chart of Comparison 2, between LN size and LN ADC of the axilla ipsilateral to breast cancer and the axilla contralateral to breast cancer only in patients with mALN.

Parameters	Contralateral negative LN (n = 30) (×10^–3mm²/s)	Ipsilateral positive LN (n = 30) (×10^–3mm²/s)	P value*
LN major axis	14.2 ± 4.9	18.3 ± 8.5	0.024
LN minor axis	7.4 ± 1.9	13.3 ± 5.5	≤0.001*
LN size ratio	1.93 ± 0.44	1.39 ± 0.23	≤0.001*
Cortical thickness	4 ± 1.17	10.3 ± 5.32	≤0.001*
LN ADC	0.78 ± 0.12	0.90 ± 0.12	0.001*

Values are given as mean ± SD.

*Calculated by Student’s t test. With the Bonferroni correction the significance level to each variable must be α ≤0.01. The P values that meet this criterion are marked with **.

ADC, apparent diffusion coefficient; LN, lymph node; mALN, metastatic axillary lymph node.

We observe that, in patients with mALN, the ADC value of the ipsilateral ALN was significantly higher than in the contralateral axilla (0.90 ± 0.12 ×10⁻³mm²/s vs. 0.78 ± 0.12 × 10⁻³mm²/s, P = 0.001). The distribution of the LN ADC for both axillae (ipsilateral and contralateral) in the 30 patients with mALN is presented in Fig. 3.

Fig. 3.

Box and whisker plot. Box and whisker plot shows the distribution of the LN ADC, for the ipsilateral axilla and the contralateral axilla in the 30 patients with mALN with a 95% confidence interval. The median for the ipsilateral LN was 0.90 × 10⁻³mm²/s and for the contralateral LN was 0.79 × 10⁻³mm²/s, and the minimum and maximum values were 0.67–1.18 x 10⁻³mm²/s for the ipsilateral LN and 0.58–1.11 × 10⁻³mm²/s for the contralateral LN. ADC, apparent diffusion coefficient; LN, lymph node.

Logistic regression analysis

With the logistic regression model, we do not identify significant variables for metastasis prediction. The variable closest to the significance was the ipsilateral LN size ratio (P = 0.06), which could mean the higher the size ratio the more likely the LN was negative for metastases. The ipsilateral LN ADC value was far from significant with controlled variables (P = 0.35), but in this case we could observe that the higher ADC value is 30 times more likely to have a positive LN. The P values of this analysis are shown in Table 2.

Accuracy analysis

Regarding the LN size ratio, the sensitivity and specificity were 76.7% and 86.7%, respectively, with an area under the ROC curve (AUC) of 0.906 (95% CI = 0.802–0.966, P ≤ 0.001). An optimal threshold of ≤1.56 (Youden index J of 0.633) was established for differentiating between mALN and non-mALN. With this cut-off, seven positive LN were misdiagnosed.

For the cortical thickness, the sensitivity and specificity were 80.0% and 97.7%, respectively, with an AUC of 0.929 (95% CI = 0.856–1.000, P ≤ 0.001). The ROC curve is presented in Fig. 4. An optimal threshold of ≥6.7mm (Youden index J of 0.777) was established for differentiating between mALN and non-mALN. With this cut-off, six positive lymph nodes were misdiagnosed.

Fig. 4.

ROC curve. The ROC curve shows the diagnostic performance of the ALN cortical thickness using Comparison 2. The area under the curve was 0.929 (95% CI = 856–1.000) with a significance level of P ≤ 0.001. A cut-off point of LN size ratio of ≥ 6.7mm (Youden index J 0.777) corresponded to a sensitivity of 80.0% and specificity of 97.7%. ALN, axillary lymph node; CI, confidence interval; ROC, receiver operating characteristic.

Only in the patients with mALN, the LN ADC has a sensitivity of 66.7%, a specificity of 76.7%, and an AUC of 0.751 (95% CI = 0.627–0.874, P = 0.001). The ROC curve is presented in Fig. 5. An optimal threshold of ≥0.86 × 10⁻³mm²/s (Youden index J of 0.434) was established for differentiating between ipsilateral mALN and contralateral non-mALN. With this threshold, 10 positive LN were misdiagnosed. The measures of diagnostic accuracy for the LN size and the LN ADC in the differentiation of mALN and non-mALN are provided in Table 4.

Fig. 5.

ROC curve. The ROC curve shows the diagnostic performance of the ALN ADC using Comparison 2. The area under the curve was 0.751 (95% CI = 0.627–0.874) with a significance level of P = 0.001. A cut-off point of ADC of ≥0.86 × 10⁻³mm²/s (Youden index J 0.434) corresponded to a sensitivity of 66.7% and specificity of 76.7%. ALN, axillary lymph node; CI, confidence interval; ROC, receiver operating characteristic.

Table 4.

Threshold, sensitivity, specificity, and AUC.

Parameter	Threshold (mm or ×10^–3mm²/s)	Sensitivity (%)	Specificity (%)	AUC	P value
LN major axis C1	≥14	66.7	50	0.613	0.23
LN minor axis C1	≥10.2	70	64.3	0.733	0.014
LN size ratio C1	≤1.42	63.3	85.7	0.755	0.002
Cortical thickness C1	≥7.45	76.7	71.4	0.775	0.003
LN major axis C2	≥13.1	76.7	66.7	0.672	0.022
LN minor axis C2	> 9.1	76.7	90	0.868	≤0.001
LN size ratio C2	≤1.56	76.7	86.7	0.906	≤0.001
Cortical thickness C2	≥6.7	80	97.7	0.929	≤0.001
LN ADC C2	≥0.86	66.7	76.7	0.751	0.001

ADC, apparent diffusion coefficient; AUC, area under the curve; C1, Comparison 1; C2, Comparison 2; LN, lymph node.

Discussion

In the present study, we found that the cortical thickness is more accurate than the ADC value of the ALN to identify metastatic involvement, with a sensitivity of 80.0% and a specificity of 97.7% when we use the cut-off point ≥6.7 mm.

Currently, bulges or thickening of the ALN cortex on ultrasound is well recognized as a predictor of metastasis (23), but with some limitations such as high false-negative rates and the difficulty of visualizing all the LN in the axilla (24).

The possibility of performing a simple and short breast MRI protocol with the measurement of the ALN cortical thickness would allow us to select patients who would benefit from an ultrasound-guided biopsy before the surgery. We found that the cortical thickness of the ALN by MRI was greater in metastatic LN, similar to the findings observed by Luciani et al. (12) and Arslan et al. (13). It is important to mention that the cortical thickness averages reported by Luciani et al. (12) and Arslan et al. (13) (4.5 mm and 4.11 mm, respectively) were lower than ours (10.3 mm). These findings could be related to the fact that the mean of the minor axis (13.3 mm) of our population was greater than that reported by Luciani et al. and Arslan et al. (7.3 mm and 8 mm, respectively), and the fact that the proportion of positive LN in the population of Luciani et al. was lower than ours (14/165).

The correlation between the ADC and the breast cancer prognostic factors has been previously studied by Razek et al. (25). They found that tumors with lower ADC were more frequently associated with mALN. However, unlike our study, they focused on tumor ADC and not on LN ADC.

Regarding the ALN ADC, we found a relatively good diagnostic performance in differentiating mALN from non-mALN, with a sensitivity of 66.7% and a specificity of 76.7%.

Some authors have reported that the LN ADC value of patients with metastatic LN was significantly lower than in patients with non-metastatic LN (15 –17). On the other hand, Kamitani et al. (20) found the opposite results, observing a significantly greater ADC in metastatic LN versus benign LN. We found a difference similar to that reported by Kamitani et al.

In other publications, the ALN ADC of patients with LN metastases is compared against patients without LN metastases. In our analysis, we did not find any significant difference when comparing patients with mALN versus patients with non-mALN. We must point out the importance of comparing the ipsilateral ALN against the contralateral LN because the LN ADC can vary due to each individual’s physiological circumstances (26).

Razek et al. (16) found that mALN had a lower ADC value than reactive LN; they attributed it to high cellularity among other histopathologic characteristics. However, it is important to note that they included only suspected LN, and the ROI was drawn around the inner margin of the LN, without clarifying whether they included the fatty hilum. We studied LN with suspected morphology against apparently normal morphology LN; the ROI was drawn in the cortex, avoiding the fatty hilum.

Zhang et al. (27) studied the difference in ADC values in metastatic and reactive LN in mice models and they proved that the ADC of metastatic LN were greater than the ADC of reactive LN. They explained that these findings were due to the lower cellularity of the metastatic tumors compared to the cellularity of the surrounding lymphatic tissue (27).

The value of the ADC as a radiomic tool has shown good results in different areas of breast cancer research, such as the prediction of malignancy in breast tumors or the prediction of the tumor response to NAC (28). However, according to our findings, the ADC would not have the same performance in predicting the metastatic involvement of ALN; we found that the size of the LN has a better diagnostic performance.

Morphological MRI criteria such as short axis, irregular contours, or cortical thickness of the ALN could suggest metastatic involvement through the difference in means (12,13). In our study, the difference in means with the Bonferroni correction was statistically significant; however, the logistic regression analysis did not find any significant criterion. This last finding is probably due to the size of our population and the few cases of non-mALN.

The present study has some limitations. First, the size of the patient cohort is small, which could be related to the fact that our institution is a referral hospital that receives patients with previous interventions (percutaneous or surgical biopsies) performed in other centers, in which a pre-interventional breast MRI was not performed. Another possible limitation is the fact that two-thirds of the patients resulted with axillary metastatic involvement, which differs from other works; this shows the unfortunate reality of our country, where up to 70.2% of diagnoses are made in late stages (29).

Currently, the benefits of knowing the LN status before the surgery is controversial, mainly in the context of patients with T1–2 breast cancer tumors (30). However, in the context of patients with breast cancer > 5 cm (especially those with triple-negative or HER2-positive) (31) or in patients with smaller tumors with ≥3 affected LN (30), the identification of LN with metastases before surgery could be decisive in choosing the most appropriate treatment.

In conclusion, this study demonstrates that the difference in means of the LN size, more specifically the cortical thickness, has a better diagnostic performance than the LN ADC value, for discriminating between mALN and non-mALN. It is necessary to conduct more studies in the future with larger populations in order to validate our findings. This could provide a more accurate and non-invasive way of staging ALN disease in selected patients with breast cancer and therefore in treatment planning.

Footnotes

Authors' note

Carla Melissa Ferrara-Chapa's current affiliation is: Department of Radiology and Imaging, Hospital General de Zona #33, Instituto Mexicano del Seguro Social, Monterrey, Nuevo León, México.

Acknowledgements

The authors thank Dr. Néstor Alejandro Rangel-Ovalle and Dr. Emma Bertha García-Quintanilla for their help in the language editing of the manuscript.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) declare not having received any payment or support in kind in any aspect of the submitted work.

ORCID iD

Yazmín Aseret Ramírez-Galván

References

Fisher

Bauer

Wickerham

, et al. Relation of number of positive axillary nodes to the prognosis of patients with primary breast cancer. An NSABP update. Cancer 1983; 52:1551–1557.

Orr

The impact of prophylactic axillary node dissection on breast cancer survival–a Bayesian meta-analysis. Ann Surg Oncol 1999; 6:109–116.

Brackstone

Fletcher

Dayes

, et al. Locoregional therapy of locally advanced breast cancer: a clinical practice guideline. Curr Oncol 2014; 22:54.

Lyman

Giuliano

Somerfield

, et al. American Society of Clinical Oncology guideline recommendations for sentinel lymph node biopsy in early-stage breast cancer. J Clin Oncol 2005; 23:7703–7720.

Giuliano

Hunt

Ballman

, et al. Axillary dissection vs no axillary dissection in women with invasive breast cancer and sentinel node metastasis: a randomized clinical trial. JAMA 2011; 305:569–575.

Giuliano

Ballman

Mccall

, et al. Effect of axillary dissection vs no axillary dissection on 10-year overall survival among women with invasive breast cancer and sentinel node metastasis: the ACOSOG Z0011 (Alliance) randomized clinical trial. JAMA 2017; 318:918–926.

Pamilo

Soiva

Lavast

Real-time ultrasound, axillary mammography, and clinical examination in the detection of axillary lymph node metastases in breast cancer patients.

J Ultrasound Med 1989; 8:115–120.

Chevinsky

Ferrara

James

, et al. Prospective evaluation of clinical and pathologic detection of axillary metastases in patients with carcinoma of the breast. Surgery 1990; 108:612–617.

Rosen

PP.

Pathologic examination of breast and lymph node specimens, including sentinel lymph nodes. In: PP.

Rosen

Rosen’s Breast Pathology. 3rd ed. Hong Kong: Lippincott Williams & Wilkins, 2009:1034–1103.

10.

Stavros

AT.

Evaluation of regional lymph nodes in breast cancer patients. In: AT

Stavros

. Breast ultrasound. 1st ed. Philadelphia, PA: Lippincott Williams & Wilkins, 2004:834–876.

11.

Bedi

Krishnamurthy

, et al. Cortical morphologic features of axillary lymph nodes as a predictor of metastasis in breast cancer: In vitro sonographic study. AJR Am J Roentgenol 2008; 191:646–652.

12.

Luciani

Pigneur

Ghozali

, et al. Ex vivo MRI of axillary lymph nodes in breast cancer. Eur J Radiol 2009; 69:59–66.

13.

Arslan

Altintoprak

Yirgin

, et al. Diagnostic accuracy of metastatic axillary lymph nodes in breast MRI. Springerplus 2016; 5:735.

14.

Abdel Razek

AAK

Zaky

Bayoumi

, et al. Diffusion tensor imaging parameters in differentiation recurrent breast cancer from post-operative changes in patients with breast-conserving surgery. Eur J Radiol 2019; 111:76–80.

15.

Scaranelo

Eiada

Jacks

, et al. Accuracy of unenhanced MR imaging in the detection of axillary lymph node metastasis: study of reproducibility and reliability. Radiology 2012; 262:425–434.

16.

Razek

AAKA

Lattif

Denewer

, et al. Assessment of axillary lymph nodes in patients with breast cancer with diffusion-weighted MR imaging in combination with routine and dynamic contrast MR imaging. Breast Cancer 2016; 23:525–532.

17.

Chung

Youk

Kim

J-A

, et al. Role of diffusion-weighted MRI: predicting axillary lymph node metastases in breast cancer. Acta Radiol 2014; 55:909–916.

18.

Rautiainen

Könönen

Sironen

, et al. Preoperative axillary staging with 3.0-T breast MRI: Clinical value of diffusion imaging and apparent diffusion coefficient. PLoS One 2015; 10:e0133111.

19.

Sui

Chen

Peng

, et al. The diagnosis of metastatic axillary lymph nodes of breast cancer by diffusion weighted imaging: a meta-analysis and systematic review. World J Surg Oncol 2016; 14:155.

20.

Kamitani

Hatakenaka

Yabuuchi

, et al. Detection of axillary node metastasis using diffusion-weighted MRI in breast cancer. Clin Imaging 2013; 37:56–61.

21.

Schipper

Paiman

Beets-Tan

, et al. Diagnostic performance of dedicated axillary T2- and diffusion-weighted MR imaging for nodal staging in breast cancer. Radiology 2015; 275:345–355.

22.

Rahbar

Conlin

Parsian

, et al. Suspicious axillary lymph nodes identified on clinical breast MRI in patients newly diagnosed with breast cancer: can quantitative features improve discrimination of malignant from benign? Acad Radiol 2015; 22:430–438.

23.

Ecanow

Abe

Newstead

, et al. Axillary staging of breast cancer: what the radiologist should know. RadioGraphics 2013; 33:1589–1612.

24.

Luparia

Campanino

Cotti

, et al. Role of axillary ultrasound in the preoperative diagnosis of lymph node metastases in patients affected by breast carcinoma. Radiol Medica 2010; 115:225–237.

25.

Razek

AAKA

Gaballa

Denewer

, et al. Invasive ductal carcinoma: Correlation of apparent diffusion coefficient value with pathological prognostic factors. NMR Biomed 2010; 23:619–623.

26.

Yamaguchi

Schacht

Nakazono

, et al. Diffusion weighted images of metastatic as compared with nonmetastatic axillary lymph nodes in patients with newly diagnosed breast cancer. J Magn Reson Imaging 2015; 42:771–778.

27.

Zhang

Zhu

Huang

, et al. Differentiation of reactive and tumor metastatic lymph nodes with diffusion-weighted and SPIO-enhanced MRI. Mol Imaging Biol 2013; 15:40–47.

28.

Abdel Razek

AAK

Gaballa

Denewer

, et al. Diffusion weighted MR imaging of the breast. Acad Radiol 2010; 17:382–386.

29.

Martínez-Montañez

Uribe-Zúñiga

Hernández-Ávila

Políticas públicas para la detección del cáncer de mama en México. Salud Publica Mex 2009; 51:s350–s360.

30.

Mamtani

Barrio

A V

King

, et al. How often does neoadjuvant chemotherapy avoid axillary dissection in patients with histologically confirmed nodal metastases: results of a prospective study. Ann Surg Oncol 2017; 23:3467–3474.

31.

Pilewskie

Jochelson

Gooch

, et al. Is preoperative axillary imaging beneficial in identifying clinically node-negative patients requiring axillary lymph node dissection? J Am Coll Surg 2016; 222:138–145.