- Top of page
- Materials and methods
- Author contributions
Current understanding of the lymphatics draining the breast is controversial, despite its known importance in the spread of breast cancer. Similarly, knowledge regarding the spatial distribution of primary tumours in the breast is limited. This study sought to test commonly held assumptions in this field, including: (i) that breast lymphatic drainage and tumour prevalence are symmetric between the left and right sides of the body, (ii) that males and females have the same drainage patterns and tumour prevalences, and (iii) that lymphatic drainage in the breast occurs independently to different node fields. This study has used lymphoscintigraphy data from 2304 breast cancer patients treated at the RPAH Medical Centre, Sydney, Australia. Symmetry of lymphatic drainage and tumour distribution as well as gender differences were tested using Fisher’s exact test. Drainage independence was assessed using Fisher’s exact test, and a multivariate probit model was used to test for drainage correlations. Results showed that the breasts are likely to have symmetric lymphatic drainage and tumour prevalence, and that there is no significant difference between males and females. Furthermore, results showed that direct lymphatic drainage of the breasts is likely to be independent between node fields. Collectively, these results serve to further our understanding of lymphatic anatomy and the distribution of tumours in the breast.
- Top of page
- Materials and methods
- Author contributions
Over the years, there have been relatively few detailed studies to characterise the lymphatics of the breast (Suami et al., 2009). Current knowledge of the breast lymphatics is based mainly on the work of Sappey (1874), who performed the first gross anatomical mappings of the lymphatic system by injecting mercury into the lymphatic vessels of human cadavers. His technique was derived from the early work of Cruikshank (1786), who investigated the lymphatics by injecting mercury through the nipple. Other important early work includes that of Delamere et al. (1903) who, based on these prior results, created composite diagrams describing the lymphatics of the breast. More recently, cadaveric studies have been performed which have effectively delineated the lymphatics of the breast using more reliable methods (Suami et al., 2008).
Techniques used to aid the treatment of breast cancer have also shed light on our knowledge of breast lymphatics. Sentinel lymph node biopsy (SLNB), which utilises lymphoscintigraphy (LS) imaging, was initially developed to map the lymphatics of the skin for patients with melanoma (Morton et al., 1992). It is now also used to map the lymphatics of the breast in patients with breast cancer (Uren et al., 1999). During LS, the breast is injected with a radiocolloid, which is then tracked through the lymphatic vessels to draining lymph nodes in surrounding node fields. Nodes receiving direct lymphatic drainage from the primary tumour site are termed ‘sentinel nodes’ (SNs). Each SN is then surgically excised to check for metastatic cancer. Examples of LS images are shown in Fig. 1 which clearly delineate the lymphatics of the breast.
Figure 1. Anterior and lateral lymphoscintigraphy images of a patient with left breast cancer. Initial and delayed imaging shows two collectors draining to two SNs in the left axilla and a single faint collector draining to a single faint SN in the left internal mammary chain.
Download figure to PowerPoint
Collectively, these studies have shown a number of node fields directly drain the breasts, with the most prominent being the axillary node field. The axillary lymphatic pathway was first located by Cruikshank (1786), and Sappey (1874) later concluded that the axillary node field received lymphatic drainage from the entire breast via the subareolar plexus. It has now become common knowledge that the axillary node field almost always drains the breast, with clinicians commonly performing axillary dissection during breast cancer treatment (Benson et al., 2007). In addition to the axilla, there are other node fields that directly drain the breasts, including the interpectoral, infraclavicular, supraclavicular, intercostal and internal mammary node fields. Occasionally, patients can show drainage to lymph nodes in the breast tissue between the injection site and a predefined node field, which are termed interval nodes (Uren et al., 1999).
Accumulated LS studies for breast cancer have provided important information regarding patterns of breast lymphatic drainage. A number of clinical studies have presented tabulated results and graphical displays to summarise their findings (Borgstein et al., 1998; Byrd et al., 2001; Uren et al., 2001; Estourgie et al., 2004). To the authors’ knowledge, there has never been a thorough statistical analysis of breast LS data. For example, consideration has never been given to whether breast lymphatic drainage patterns are independent between node fields. That is, whether drainage to one node field influences the likelihood of drainage to another. It has also never been quantified whether lymphatic drainage or tumour prevalence is symmetric between the right and left sides of the body, or whether there is a difference between males and females. We sought to address these issues, building upon previous work that statistically analysed lymphatic drainage of the skin (Reynolds et al., 2010).
This study used LS data collected between October 1992 and November 2009 from breast cancer patients treated at the RPAH Medical Centre in Sydney, Australia. During LS, four peritumoural injections of Technetium-99m-antimony sulphide colloid (99mTcSb2S3) were administered around the primary tumour. Injections were given under ultrasound guidance, at the depth of the centre of the tumour at the 12, 3, 6, and 9 o’clock positions, and the draining SNs were located (Uren et al., 1995). The region of each primary tumour and the node field locations of each SN were then recorded in a database.
Each patient’s primary tumour location was recorded according to its quadrant location in the breast (see Fig. 2). These regions were defined as the upper inner (UI), lower inner (LI), lower outer (LO) and upper outer (UO) quadrants, with an additional region defined behind the nipple (BN). To provide a more accurate primary tumour location, when the tumour overlapped two quadrants, it was recorded as located in the upper inner outer (UIO), upper lower inner (ULI), lower outer inner (LOI) or upper lower outer (ULO) regions. For example, a primary tumour was classified in the UIO region if it was located in both the UO and UI quadrants. Thus, each patient’s primary tumour was recorded in one of nine possible regions.
Figure 3 illustrates the approximate location of each node field that directly drained the breast, which included the axillary, interpectoral, internal mammary, infraclavicular and supraclavicular node fields. Intercostal nodes also directly drained the breasts; however, they have been recorded as interval nodes as it was difficult to determine their exact anatomical position during LS. In total there were 2363 patients in the LS database, 57 of whom (2.4%) did not display drainage to a SN, and two patients only displayed drainage to interval nodes. Note that drainage to interval nodes, and thus also intercostal nodes, was not included during testing as these nodes do not have a consistent anatomical location. This study has analysed the remaining 2304 patients, 2284 of whom were female and 20 of whom were male.
Table 1 details the number of cases with drainage to each combination of node fields on each side of the body. It can be seen that drainage to the axillary [n = 2263 (98.2%)] and internal mammary [n = 813 (35.3%)] node fields were markedly more frequent compared with drainage to the interpectoral [n = 15 (0.7%)], infraclavicular [n = 25 (1.1%)] and supraclavicular [n = 70 (3.0%)] node fields. Patients usually had drainage to one node field (63.6%); however, it was possible for drainage to occur to multiple node fields (36.4%) from a primary tumour site. There were no cases of contralateral drainage to any node field (i.e. drainage to the other side of the body); however, drainage was observed to occur across the nipple centre line of the breast. On average, patients displayed drainage to 1.4 node fields.
Table 1. Number of patients with lymphatic drainage to different node fields, on either side of the body. Note that drainage to interval nodes has been excluded from testing, as they do not have a defined anatomical position.
|Draining node fields||No. patients|
|Axilla & internal mammary||371||351|
|Axilla & supraclavicular||12||9|
|Axilla & infraclavicular||11||9|
|Axilla & interpectoral||4||7|
|Axilla, internal mammary & supraclavicular||25||17|
|Axilla, infraclavicular & internal mammary||8||5|
|Axilla, infraclavicular, internal mammary & supraclavicular||0||1|
|Internal mammary only||11||18|
|Internal mammary & supraclavicular||2||0|
|Internal mammary & infraclavicular||1||2|
|Internal mammary & interpectoral||1||0|
|Internal mammary, supraclavicular & interpectoral||0||1|
|Supraclavicular & interpectoral||1||0|
|Total||1157 (50.2%)||1147 (49.8%)|
- Top of page
- Materials and methods
- Author contributions
This study has quantitatively assessed previous assumptions regarding breast lymphatic drainage and tumour prevalence. First, a sequence of contingency tables was used to test whether lymphatic drainage and tumour distribution were symmetric and/or gender invariant. Secondly, the assumption that lymphatic drainage is independent between node fields was tested using both Fisher’s exact test and a Bayesian MVP model.
It is not surprising that the lymphatic drainage of the breasts was found to be symmetric, as the underlying anatomy is also largely symmetric. The ribs (12 on each side) and musculature (pectoralis minor and pectoralis major, serratus anterior, external oblique muscles and the intercostal muscles) are the same on each side, as is the arterial vascular supply and the venous drainage (Macea & Fregnani, 2006). Furthermore, the overall structure of the lymphatic vessels tends to follow the structure of the major veins, which are themselves largely symmetric (Uren et al., 1999). This anatomical symmetry is consistent with the symmetric patterns of lymphatic drainage quantified in this study.
It is significant to consider these results in light of a recent cadaveric study by Pan et al. (2009) which analysed the breasts and anterior upper torso lymphatics of a female human cadaver. Interestingly, their study showed the cadaver had asymmetric lymphatic vessels draining the breast. Whilst our results indicate there is symmetry in lymphatic drainage of the breasts across a population, there clearly are asymmetries in lymphatic drainage of the breast in individual patients. Hence, symmetry cannot be assumed in single patients and an individualised approach must still be employed during breast cancer treatment.
Whilst previous studies have shed light on lymphatic drainage symmetry, to the authors’ knowledge, symmetry in the spatial distribution of breast tumours has never been assessed. It has been observed, however, that breast mass and size asymmetry are predictors of breast cancer development (Manning et al., 1992; Scutt et al., 1997, 2006) and that the left breast is more likely to develop tumours than the right breast (Dane et al., 2008). It has also been found that tumour location is an important prognostic factor in determining survival outcomes, with medial tumour patients experiencing diminished survival rates compared to lateral tumour patients (Gaffney et al., 2003). It is significant then, in light of these studies, that we found symmetric tumour prevalence in the breast.
From an anatomical perspective, the structure of the breast is the same between males and females, although women usually have more glandular tissue and fat. Hence, it is also not surprising that our study concluded that lymphatic drainage, and for the most part tumour prevalence, of the breasts is the same between sexes. A recent study by Suami et al. (2008) also investigated the lymphatic drainage patterns of the breast between the sexes by assessing five male and nine female cadavers. They injected the breast lymphatic vessels of each cadaver with a tracer and followed the drainage patterns to draining lymph nodes. It was shown that the breast lymphatic vessels had no significant structural differences between males and females (Suami et al., 2008); this further supports our finding that drainage patterns are invariant of gender.
The result which showed males were more likely to experience tumours behind the nipple was interesting. This could be due to the fact that male breasts typically have very little mammary tissue, and what tissue is present is mostly behind the nipple. Note that with only 20 male cases in our sample (0.87%), any conclusions with regard to gender are tentative due to low counts. To be more certain in future, a larger sample size would be required.
Previous studies have not assessed the possibility of breast lymphatic drainage independence between node fields, that is, that the appearance of drainage to one node field does not affect the likelihood of drainage to any other node field. We have addressed this assumption by testing against the null hypothesis of independence between node fields, both using an exact test and by quantifying the drainage correlations between node fields. By using two different paradigms of analysis we have provided additional insights and validation of findings. Most node field combinations were found to be insignificant for interdependence, whilst those that were significant were conditioned upon upper regions and involved the internal mammary node field. However, it is difficult to draw any definite conclusions regarding possible deviations from drainage independence due to limitations in the LS dataset.
As the analysis was systematically conducted on subsets of the data, there were several breast regions which displayed very few counts. This was typically observed for the infraclavicular, supraclavicular and interpectoral node fields. In addition, the LS injection technique utilised four separate injections around the primary tumour, which can cause a large region of breast tissue to contain radioactive tracer, thus making observations less precise. In future, to conclude whether there is indeed interdependence between node fields, more data would be required which had ideally been collected in a more controlled manner. Overall, however, it appears likely that breast lymphatic drainage is independent between node fields.
The analyses carried out here have allowed testing to be conducted by considering nine separate breast regions. This has the advantage of providing a highly spatially refined analysis; however, it has also meant that some regions contained a sparse amount of data. As more data is collected over time, this limitation will be minimised. In addition, the quantified symmetry between both sides of the body and, to a somewhat limited extent, gender invariance provides justification for reflecting the data to one side of the body and grouping the genders for future studies. This would effectively double the sample size, thus allowing more precise inferences to be made. The finding that node field drainage likelihoods are predominantly independent of each other has implications whereby the probability of drainage to a given node field can be considered in isolation from all other node fields.
In summary, this study has provided the first quantitative assessment of symmetry and gender difference of lymphatic drainage and tumour prevalence in the breast. Using a large LS dataset, we have quantified that although individual patients can exhibit asymmetric drainage, drainage patterns are likely the same in both left and right breasts across the population. In addition, we have quantified that tumour prevalence is symmetric and that males and females show the same patterns of drainage and for the most part similar tumour distributions. Furthermore, we have found that direct lymphatic drainage is largely independent between node fields. As more data is collected in the future, we may be able to determine whether drainage between node fields exhibits a marked deviation from independence. These results serve to further our understanding of breast lymphatics by quantifying characteristics of this vitally important system.