Lymphoscintigraphy, a key component of sentinel lymph node biopsy (SLNB), plays a crucial role in visualizing lymphatic drainage and locating sentinel nodes (SLNs) preoperatively. Although various factors can influence the success of lymphoscintigraphy visualization, one well-established determinant is the tumor burden within the regional lymphatic system. Vargas et al. reported that patients with extensive axillary metastases had a significantly higher rate of SLN detection failure than did those with fewer metastatic nodes (1). Similarly, Sener et al. reported that patients with more than 10 involved lymph nodes experienced a significantly greater incidence of mapping failure than did node-negative patients (2). These findings suggest that extensive nodal involvement can disrupt normal lymphatic drainage, thereby impeding tracer transport and contributing to sentinel node mapping failure in patients with breast cancer.

Although early breast cancer typically has a favorable prognosis, with 10-year survival rates ranging from 92% to 95% in Japan (3), complete survival is not guaranteed. A subset of patients experience recurrence despite treatment, suggesting the presence of more aggressive disease. Although a few studies from Western countries have investigated whether nonvisualized (nonvLSG) correlates with poorer prognosis, their conclusions have been contradictory even within the same region (4-7). Besides, recent studies have shown that lymphovascular invasion (LVI) is a strong independent risk factor for non-sentinel lymph node metastasis in SLN-positive breast cancer patients (8,9). Given that LVI involves tumor emboli infiltrating lymphatic channels, it can potentially obstruct normal lymphatic drainage. This may manifest clinically as failed lymphoscintigraphy, where radiotracer migration to the sentinel node is impeded. We hypothesize that such failure may serve as a surrogate indicator of occult LVI and, by extension, portend a worse prognosis.

In this study, we therefore investigated whether the absence of sentinel node visualization on preoperative lymphoscintigraphy - a phenomenon that may signify lymphatic obstruction caused by tumor cells or stromal changes - correlates with poorer overall survival (OS) and relapse-free survival (RFS). This research aims to shed light on the prognostic implications of lymphoscintigraphy visibility in patients with early breast cancer undergoing SLNB.

Study population. This single-center retrospective study reviewed data from 357 patients who underwent SLNB between June 1999 and May 2003. The study included patients with early breast cancer (cTis - 2 and below, cN0) and excluded those with synchronous or metachronous bilateral breast cancer to avoid confounding survival outcomes. Patients involving subtumoral radiocolloid injections (10), a nonstandardized procedure, and those lacking lymphoscintigraphy status information were also excluded. All received adjuvant treatment therapy based on protocols at the time, aligned with the 8th St. Gallen Meeting 2003 (11) guidelines.

Lymphoscintigraphy procedure. A 30-80 MBq dose of 99mTc-tin colloid was injected intradermally (peritumorally or periareolarly) under local anesthesia the day before surgery. Lymphoscintigraphy was conducted one to two hours after the injection, and anterior-oblique projections were captured via a large-field stationary scintillation camera. Focal accumulations of radioactivity, known as "hot spots", indicate potential lymphatic drainage sites in the axilla, parasternal, or extraregional node regions, with or without a visible drainage tract, and are classified into the visualized lymphoscintigraphy (vLSG) group. Absence of a "hot spot" is categorized as the nonvLSG group.

Sentinel node procedure. The biopsy procedure involved the use of a hand-held gamma detection collimated probe (Navigator, US Surgical Co., Norwalk, CT, USA) to detect radioactivity. A dual method for SLNB was employed, combining radiocolloid detection with isosulfan blue dye (1%, Lymphazuria, US Surgical Co.). During the operation, the blue dye was injected peritoumorally or periareolarly (at three, six, nine, and 12 o'clock positions) before making the incision, followed by a 5-min massage to facilitate lymphatic drainage.

The excised nodes were assessed ex vivo for radioactivity (counts per second), with nodes considered "hot" if they measure 400% of the axillary background. SLNs were also identified as blue if they were partially or completely stained with the dye. If SLNs cannot be identified or if intraoperative frozen section analysis revealed cancer metastases, axillary lymph node dissection (ALND) was performed; otherwise, only SLNB is conducted.

Sentinel lymph node histology examination. SLNs were sampled for frozen section analysis using a 2 mm thick cut from the largest surface. Remaining tissue was fixed in 10% buffered formalin, processed overnight, sliced into 2mm sections, embedded in paraffin, and examined via H&E staining and immunohistochemistry (IHC; avidin-biotin-peroxidase) using anticytokeratin AE1/3 (Histofine, Nichirei Co., Tokyo, Japan). Metastasis was defined as clusters of immunopositive cells; isolated single cells were not considered metastatic. SLNs were classified as positive if metastasis was detected by frozen section, H&E, or IHC. Cases positive on frozen section but negative on H&E or IHC were considered false positives. For non-SLNs, one representative H&E paraffin section per node was examined.

Statistical analysis. Statistical analysis was conducted to compare the vLSG and nonvLSG groups. RFS and OS were analyzed via the Kaplan‒Meier method, with group comparisons performed via the log-rank test. Collinearity diagnostics were conducted to evaluate multicollinearity among variables, and Cox regression was used to identify independent factors influencing OS and RFS. All statistical analyses with a significance level of p<0.05 were considered statistically significant. Statistical analyses were conducted via JMP Pro (version 17.2.0, SAS Institute, Inc., Cary, NC, USA) and RStudio (version 2024.12.0+467, Posit Software, PBC, Boston, MA, USA).

A total of 357 patients underwent SLN mapping via lymphoscintigraphy with Technetium-99. Among these patients, 110 were excluded for various reasons: 24 had bilateral breast cancer (synchronous or metachronous), 10 presented with cN1 disease, four had cT3 tumors, 69 underwent nonstandard subtumoral injections, and three lacked lymphoscintigraphy data. After exclusions, 247 patients were included in the final analysis. Among them, 223 patients (90.3%) had vLSG, whereas 24 patients had nonvLSG imaging. During intraoperative SLN detection, one patient from the vLSG group and two patients from the nonvLSG group had undetectable SLNs (Figure 1).

Clinical and tumor characteristics between the visible and nonvisible lymphoscintigraphy groups. A comparative analysis of clinical and tumor characteristics between the vLSG and nonvLSG groups revealed no statistically significant differences in most variables except hormone receptor status (Table I). The mean age was slightly lower in the vLSG group (53.5±10 years) than in the nonvLSG group (57±12 years, p=0.14). Menopausal status, body mass index (BMI), tumor type, tumor size, tumor focality, tumor location, pathological nodal (pN) status and clinical T stage (cT) were not significantly different between the groups. A significant difference was observed in hormone receptor status, with 62.5% of the nonvLSG group being hormone receptor positive compared with 80.3% in the vLSG group, with a p value of 0.03.

Impact of the association between nodal burden and lymphoscintigraphy visibility (vlsg and nonvlsg). To further investigate whether the lymphoscintigraphy status was associated with the number of nodal metastases, a subgroup analysis was performed. A linear negative binominal LASSO regression model was used. The effect of nonvLSG on number of positive node was not statistically significant (estimate=-0.2784, 95%CI=-0.59 to 0.04, p=1.0).

Survival of patients with visualized versus nonvisualized sentinel nodes. Among the vLSG patients (n=223), 47 (21.1%) experienced recurrence; 12 were local recurrences, five regional, 22 distant recurrences, and eight had both regional and distant. In the nonvLSG group (n=24), five (20.8%) recurred: one local, one regional, two distant, and one with both regional and distant metastases. There were 28 deaths (12.4%) in the vLSG group - 20 from disease progression, one from treatment complications, and seven from unrelated causes. In the nonvLSG group, three patients (12.5%) died, all due to breast cancer disease progression, with no deaths from treatment complications or unrelated causes. The Kaplan‒Meier survival analyses revealed no significant differences in OS or RFS between the groups (Figure 2 and Figure 3). At 10 years, the OS was 85% for the vLSG group and 88% for the nonvLSG, whereas the RFS probability was 80% in both groups. At 20 years, the OS was 75% for the vLSG group and 70% for the nonvLSG group, and the RFS was approximately 70% for both groups. Wide confidence intervals in the nonvLSG group reflect the smaller sample size in this cohort. The log-rank test yielded p=0.927 for OS and p=0.762 for RFS, indicating no statistically significant differences.

Collinearity testing showed no multicollinearity (GVIF <2). Factors for Cox regression were selected using the PREDICT model, a validated recurrence prediction tool for early breast cancer developed on the basis of multiple studies (12). As shown in Table II, the only statistically significant predictor of survival in the model was the number of metastatic nodes. A lower number of involved lymph nodes (≤3 nodes) was associated with a better prognosis [hazard ratio (HR)=0.28, p=0.01].

Patient outcomes of axillary node positive for malignancy (pn+) by lymphoscintigraphy status. Among pN+ patients (n=57), mortality was 16.7% in the nonvLSG group (n=6) and 13.7% in the vLSG group (n=51), with an overall rate of 12.6%. The odds ratio (OR) for death in nonvLSG vs. vLSG was 1.74 [95% confidence interval (CI)=0.13-17.0, p=0.60], showing no significant difference. Breast cancer-related events (recurrence or death) occurred in 33.3% of nonvLSG and 27.5% of vLSG patients (OR=1.31, 95%CI=0.11-10.42, p=1.00). Kaplan-Meier analysis showed no significant differences in OS (p=0.49) or RFS (p=0.96), indicating lymphoscintigraphy status does not significantly impact survival in pN+ patients.

With respect to survival outcomes, only Hellingman et al. (4) and Verheuvel et al. (7) conducted retrospective cohort studies exploring the relationship between lymphoscinti-graphy visibility and survival outcomes in patients with breast cancer undergoing SLNB. Hellingman et al. (4) included 2,042 patients with breast cancer who underwent SLNB, including patients with both unilateral and bilateral breast cancer, while ipsilateral recurrent disease was excluded. The median follow-up was six years (range less than one to nine and a half years), with survival data available for 10 years. They reported no significant difference in OS or the distant-metastasis free interval (DMFI) between the vLSG and nonvLSG groups (SLN positivity: 16.0% vs. 18.0%, p=0.593). In contrast, Verheuvel et al. (7) analyzed a significantly larger cohort of 76,472 patients with clinically node-negative breast cancer across all cT stages, with a median follow-up of 3.3 years and a maximum follow-up of nine years. They reported that patients with nonvLSG had a greater risk of nodal involvement and worse OS (HR=1.18, p=0.015). Compared with these studies, our study included a smaller cohort of 247 patients with early-stage breast cancer, excluding those with bilateral disease, cT3, and cT4 cases, but had a considerably longer follow-up duration, with a median of 12.2 years (range=<1-24.5 years) and over 20 years of survival data available. Like in previous studies, we compared the vLSG and nonvLSG groups and measured OS and RFS. Our findings indicated no significant difference in OS or RFS based on lymphoscintigraphy status. Hellingman et al. (4) reported no impact of lymphoscintigraphy visualization on OS, and Verheuvel et al. (7) found worse OS for patients with nonvLSG, suggesting that nonvLSG may indicate a higher-risk group. Our study, with a much longer follow-up, reinforces that pN status remains a stronger predictor of survival than does lymphoscintigraphy visualization alone. Additionally, we excluded bilateral disease because it may have distinct tumor biology compared with that of unilateral breast cancer, potentially affecting prognosis. In cases of bilateral disease, it is also challenging to determine which side predominantly contributes to a worse outcome, making it difficult to analyze the prognostic impact of the lymphoscintigraphy status accurately.

We observed a significant difference in hormone receptor positivity between the vLSG and nonvLSG groups (80.3% vs. 62.5%, p=0.03). These findings suggest that hormone receptor status may be associated with SLN visualization, potentially influencing lymphatic drainage patterns. Nowikiewicz et al. (6) similarly showed that luminal B HER2-negative breast cancer was significantly more common in nonvSLN patients (p=0.0012). In contrast, Hellingman et al. (4) and Verheuvel et al. (7) reported no significant differences in estrogen receptor (ER) or HER2 status between SLN visualization groups. Our findings suggest a potential association between hormonal receptor status and SLN visualization, which may merit further investigation into how tumor biology affects lymphatic mapping outcomes.

While this study offers a reliable analysis of long-term survival trends and provides a valuable perspective specific to a Japanese cohort, certain contextual factors should be considered when interpreting the findings. The smaller sample size of the nonvLSG group (n=24) may affect the precision of some estimates, particularly in subgroup analyses for survival for the pN+ group by lymphoscintigraphy status. The interpretation of these findings should be approached cautiously, as the small sample size, particularly in the nonvLSG group, may limit the statistical power. Finally, the single-center design, while ensuring procedural consistency, may limit the direct applicability of the findings to populations in other settings. These considerations, however, do not detract from the study's value in addressing an important gap in the literature and providing understanding of lymphoscintigraphy visualization in early breast cancer.

Lymphoscintigraphy visualization status was not found to significantly impact OS or RFS in patients with early breast cancer undergoing SLNB. These findings indicate that nonvLSG does not compromise survival outcomes, suggesting that SLNB remains a dependable staging and treatment approach irrespective of visualization status. This finding clinically supports the reliability of SLNB even in cases of nonvisualized lymphoscintigraphy, which reassures patients and clinicians that survival outcomes are not adversely affected.

The Authors have no relevant financial or non-financial interests to disclose.

All Authors contributed to the study's conception and design. Khoo Hui Ying prepared materials, collected data, performed the analysis, and drafted the initial manuscript. Masami Tsukabe, MD, PhD, managed the ethical board application and extensively reviewed the study. Yoshiaki Sota, MD, PhD, provided significant input on statistical analysis. Kenzo Shimazu, MD, PhD, served as the main reviewer and was instrumental in shaping the study’s conception. All Authors reviewed and approved the final manuscript.

The Authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

The Authors used Curie AI solely for proofreading and improving the English language of the manuscript prior to submission. No AI tools were used for data analysis, interpretation, or generation of scientific content. The final manuscript was reviewed and approved by all authors.