Background
Axillary lymph node metastasis (ALNM) is an important prognostic factor for invasive breast cancer, and axillary lymph node dissection (ALND) may improve patient survival and decrease local recurrence [
1,
2]. Thus, ALND should be performed in patients with macrometastasis (Mac), as they have a poorer prognosis than patients with micrometastasis (Mic). Over half of patients with Mac develop non-sentinel lymph node (SLN) metastasis [
3-
6]; thus, the number of metastatic lymph nodes is critical in therapeutic planning, including chemo-endocrine therapy, surgery, and radiation therapy. ALND is important for disease control and breast cancer staging, but it may cause numerous complications such as lymph edema and sensory or motor disturbances in the upper limb. To precisely diagnose ALNM and perform minimally invasive ALND for breast cancer, sentinel lymph node biopsy (SLNB) is recommended in clinically node-negative (Neg) cases because it offers the same prognosis as ALND [
7]. If the SLNB is positive, then ALND is required.
Even with additional irradiation or chemo-endocrine therapy, ALND does not improve the prognosis of patients with Mac [
8,
9]. In contrast, a minimally invasive procedure such as SLNB has fewer complications compared to ALND and is therefore commonly performed in clinically node-negative breast cancer patients. As indicated by the superior clinical outcomes in several clinical trials including the ACOSOG Z0011, IBCSG 23–01, and AATRM 048/13/2000 trials, small metastatic lesions that include Mic may not require additional ALND [
9-
11]. Thus, as therapy improves, ALND may be deemed unnecessary for select patients. To accomplish this goal, axillary evaluation is critical. ALND and SLNB provide postoperative clinical staging; therefore, preoperative diagnostic findings should be fully assessed in breast cancer patients. Ultrasonography (US), computed tomography (CT), and magnetic resonance imaging (MRI) are common diagnostic modalities, with sensitivities of 72% to 78%, 78%, and 67%, and specificities of 77% to 78%, 75%, and 78%, respectively [
12-
14]. Given their modest sensitivities and specificities, additional diagnostic tests are necessary as the diagnostic utility of these modalities is insufficient.
Recent therapeutic trends suggest that the tumor burden within axillary lymph nodes may impact surgical planning and prognosis; therefore, quantitative axillary assessment may be required to discriminate macrometastasis (Mac) from micrometastasis (Mic) and node-negative (Neg) cases. Common diagnostic modalities such as MRI, CT, and US can morphologically identify the metastatic lymph node, but quantification of the tumor burden and metastasis is difficult. Several methods of metastatic prediction have been described. Among them, nomogram is one representative predictive model used to identify sentinel or additional lymph node metastasis at several facilities. Nomogram is performed preoperatively and is based on clinicopathological data, including age, tumor size, location, lymphovascular invasion, and hormonal receptor activity in biopsy samples. However, the accuracy of nomogram is limited according to previous reports, with an area under the curve (AUC) in the range of 0.688 to 0.721 [
15,
16]. Recently, two methods of intraoperative detection were reported. The first, one-step nucleic acid amplification (OSNA), is a unique quantitative method that amplifies cytokeratin 19 messenger ribonucleic acid (mRNA). This method has quickly become widespread because it produces results similar to those of histopathologic staining with a >77.5% sensitivity and >95.8% specificity, and provides easy quantitative prediction [
17,
18]. The second technique, rapid double staining method with hematoxylin & eosin (HE) stain and immunohistochemistry, has decreased the false negative rate from 33.3% to 16.7% even in patients with Mic [
19]. These methods can be useful in facilities possessing the required specialized equipment and pathology expertise, but also require greater concentration to perform due to time constraints.
In breast cancer medicine, fluorodeoxyglucose positron emission tomography/computed tomography (FDG-PET/CT) may be an acceptable alternative for detecting distant metastases [
20,
21]. In ALNM evaluation of breast cancer patients, PET/CT is less sensitive (20% to 37%) but more specific (>95%) than other modalities [
22-
25]. It also functionally detects abnormal glucose metabolism; a high maximum standard uptake value (SUVmax) indicates tumor activity within the axillary lymph node. For staging and estimating prognosis, the SUV Lymph node/tumor ratio (NT ratio), defined as the SUVmax ratio between the axillary lymph node (SUV-LN) and the primary tumor (SUV-T), is useful in detecting nodal malignancy in patients with non-small cell lung cancer [
26].
Recent clinical trends indicate that additional lymphadenectomy is not required in patients with micrometastasis in the axillary lymph nodes; therefore, preoperative quantification of axillary disease is required to discriminate Mac from Mic and Neg. The present study investigates the utility of the NT ratio and SUV-LN as assessed by PET/CT for quantifying axillary lymph node metastasis in patients with invasive breast cancer.
Discussion
For minimally invasive surgery of breast cancer, we strongly believe in the importance of preoperatively evaluating the axillary lymph node. We previously investigated the utility of navigation surgery based on composite PET/CT and US images, specifically in cases of axillary neoplasia, and found that PET/CT is a valuable tool for breast surgery [
31]. The PET/CT-based NT ratio was selected for preoperative quantitative evaluation because SUV measurement is simple and easy to perform. Furthermore, this method can be used to evaluate both the SLN and the non-SLN in the entire axillary space. The technique was initially applied to non-small-cell lung cancer (NSCLC) as a universal predictor of mediastinal node malignancy and showed a 0.56 cutoff value, 94% sensitivity, and 72% specificity [
26]. These previous data motivated us to use the NT ratio for axillary evaluation in breast cancer patients. Recently, one report described the utility of the NT ratio for predicting ALNM in breast cancer patients [
32]. The mean NT ratio was 0.3, and the cutoff value was 0.2; when the diagnostic performance was assessed, the technique showed a 71.4% sensitivity and a 77.3% specificity. The AUC for the NT ratio was 0.776 and was superior to the AUC for the SUV-LN (0.705).
In the present study, the mean NT ratio for Mac (0.571) was significantly higher than observed for Mic (0.227) and node-negative lesions (0.316). Potentially, the NT ratio may reflect the tumor burden, though the number of Mic cases was small. Notably, 96% of patients with a high SUV-LN (≥1.5) had Mac, but a low SUV-LN (0.5-1.5) may interfere with evaluation. In 30 patients with Mac and a low SUV-LN, 17 (57%) had a high NT ratio (>0.288), suggesting that the NT-ratio may predict patients with Mac. However, node-negative cases may show false positive results, as shown in Figure
1. A total 42 of 46 (91%) node-negative cases with a low SUV-T (<2.5) showed a seemingly high NT ratio (mean: 0.54, range: 0.319-1.061). This discrepancy prompted a reexamination of the NT ratio accuracy in patients with a high SUV-T (≥2.5; n = 118), and no statistical difference was found between the AUCs of the NT-ratio (0.757) and SUV-LN (0.797; Figure
2b). Under these conditions, the sensitivities for Mac were 68.9% using the NT ratio and 75.6% using the SUV-LN. However, combining these methods increased the sensitivity to 84.4%. The NT ratio appears to be particularly reliable in patients with an SUV-T greater than 2.5. Our results suggest that the NT ratio is one option for the preoperative quantification of axillary lymph node metastasis. Combining the NT ratio and the SUV-LN may be important for minimizing false positive cases. However, tumors with a low SUV, including low-grade malignancies or benign inflammatory lesions, also warrant attention. Both morphological evaluation by conventional modalities and assessment by functional modalities such as the NT ratio and SUV-LN using PET/CT are required to quantitatively diagnose the axillary lymph node. Notably, all the data analyzed in this study were obtained in patients prior to treatment.
Whether the NT ratio is efficacious in patients receiving chemo, endocrine, and molecular target therapies is unclear. Reportedly, in tumors that were vulnerable to chemotherapy, the SUV-T significantly decreased after the second cycle of neoadjuvant chemotherapy, resulted in low sensitivity (66.7% to 68%) and high specificity (75% to 96.4%) for predicting the pathological complete response (pCR) [
33,
34]. At present, there are no data on lymph node assessment during neoadjuvant chemotherapy. The therapeutic response at the metastatic site, including lymph nodes, may differ from that for the primary tumor, which may make the NT ratio complex and difficult to understand. When evaluating the effect of neoadjuvant chemotherapy, changes in both the SUV-T and SUV-LN should be considered.
Axillary evaluation is crucial for staging and therapeutic planning in patients with invasive breast cancer. Although the current data was retrospectively analyzed at a single institution, our findings suggest that both the SUV-LN and the NT ratio obtained by PET/CT may help predict preoperative Mac in axillary lymph nodes. A prospective large cohort study is recommended to validate the NT ratio as a reliable predictor of Mac.
Competing interests
The authors declare that they have no competing interests, financial or otherwise.
Authors’ contributions
MF designed the study, collected and interpreted the data, and drafted the manuscript. AT measured the SUV-T and SUV-LN and analyzed the data. KK performed the pathological evaluation in each axillary node. KM, MN, KM, AM, and RM performed SLNB and ALND. KY analyzed the data and finalized the manuscript. All authors read and approved the final manuscript.