This study has statistically compared the regional location of breast SLNs identified by PT versus SA injections. The analysis has included the largest number of patients from a single-centre and was further strengthened by incorporating precise anatomical localisation of both breast tumours and SLNs from SPECT/CT. No other study has incorporated tumour locations discretised into clockface breast regions or had SLNs identified in as many axillary subregions as well as including interval nodes.

Findings showed there were no significant differences between the regional location of axillary SLNs identified by the injection types, even with multiple subregions including axillary level I (covering anterior, central, posterior, lateral, interpectoral groups), axillary level II, and axillary level III. This agrees with previous studies by Noushi et al. [10], although they grouped all axillary SLNs into one region, and Uren et al. [5], who did incorporate axillary subregions but did not discretise the breast into regions. Findings in this study also showed that identification of internal mammary SLNs was lower when using SA injections compared to PT injections, which was concordant with previous studies [2, 5, 10, 18], and was statistically significant in all clockface breast regions except those corresponding with the upper outer quadrant. The only other breast region with significant differences between injection sites was the 5 o’clock breast region, for identifying SLNs classified as interval nodes, which has not been reported previously. Interval nodes include intramammary and intercostal SLNs, which are often overlooked, even though they may have prognostic significance and show advanced pathological features when malignant [19].

Recent anatomical studies of the lymphatic system in the breast confirm these findings and support PT injections being more suitable than SA injections for accurately mapping SLNs [4]. Initially, SA injections were supported by historical anatomical studies that described lymphatics in the breast parenchyma converging to the subareolar plexus and then switching to the lymphatic vessels running to the axillary lymph nodes [20]. However, publications from the same group also depicted the lymphatic vessels in the upper torso connecting to the axillary lymph nodes but the relationship with these lymphatic vessels from two different origins was not clarified [4].

Anatomical studies using a microinjection technique to visualise lymphatic vessels recently reported that lymphatics in the breast drain to the lymph nodes via three different lymphatic pathways [4]. This includes lymphatic vessels originating from the upper torso, lymphatic vessels originating from the areola, and perforating lymphatic vessels running with the perforating arteries to the internal mammary artery and connecting to the internal mammary lymph nodes (Fig. 3). This explains why PT injections are more likely to identify the internal mammary lymph nodes and may identify different SLNs from SA injections.

Lymphatic anatomy of the breast: A lymphatic vessels originating from the upper torso, B a lymphatic vessel originating from the areola and C perforating lymphatic vessels connecting to the internal mammary lymph nodes

Metastatic involvement of internal mammary nodes is known to be higher in patients with medial tumours and positive axillary SLNs [21]. Furthermore, the metastatic status of internal mammary nodes is included in the American Joint Committee on Cancer staging criteria where it has prognostic importance similar to axillary node involvement [22] and undiagnosed internal mammary metastasis has been associated with significantly worse survival outcomes [23]. Therefore, ensuring accurate detection of SLNs in the internal mammary node field would logically be important for informing and optimising patient treatment, including surgery and radiation therapy options. If it is too risky to biopsy an internal mammary sentinel node, then the staging process will be incomplete and regional node irradiation to the internal mammary nodes should be considered. A recent review article by Lenihan et al. [24] reported four clinical trials which considered adjuvant internal mammary node irradiation when axillary nodes were positive. Overall, trial findings supported regional nodal irradiation including internal mammary nodes, with one Danish trial confirming it provided a survival advantage [25].

PT injections carried out in the retroareolar (0 o’clock) region and SA injection sites are both located behind the areola. However, PT injections are located at the depth of the tumour, which is deeper than SA injections located less than 0.5 cm from the skin surface. Hence, the statistical comparisons between retroareolar PT injections and SA injections suggest that deeper injections increase the drainage proportion to internal mammary node fields. These results contradict Noushi et al. [10], who found that the depth of PT injection did not have a significant effect on discordance, albeit with a smaller number of patients. The PT injection depth was not recorded in this study, and comparisons between patients may be difficult due to differences in breast shape and deformation. Further research to investigate the effect of injection depth on lymphatic drainage patterns is therefore warranted.

There are some limitations to this study, including the low number of patients with drainage in some breast regions, which meant the two-proportion z-test could not be carried out for all node fields. Additionally, SA injections were administered when patients were scheduled for a prophylactic mastectomy, had DCIS, or had tumours or DCIS not visualised on ultrasound. Consequently, SA patients were typically at a comparatively earlier cancer stage and were younger than PT patients (Table 1) which may impact lymphatic drainage, as studies by Kawase et al. [26], Kong et al. [27], and Lukesova et al. [28] have shown that younger patients have a higher likelihood of having internal mammary SLNs. Controlling for the effects of age may increase the difference between injection types for internal mammary drainage, but it was not carried out in this study. Furthermore, the number of injections and hence the volume of tracer varied between patients in both PT and SA groups. An increase in injection volume likely increases the drainage observed, so future analyses which take this into account would allow for greater confidence in the results.