Factors Affecting Visualization Rates of Internal Mammary Sentinel Nodes During Lymphoscintigraphy

MATERIALS AND METHODS

A retrospective review of 82 group 1 patients with successful visualization of axillary SNs was performed and compared with 61 consecutive group 2 patients using the new injection protocol. As long as the location of the primary lesions could be reasonably determined, lymphoscintigraphy was performed for both groups. The methods used to guide injections included palpation, lumpectomy or biopsy sites, wire localization, or ultrasound.

RESULTS

Examples of patients with IM SNs are presented in Figure 1. It shows the wide variation that can be present in IM SN intensity, position, and number, as well as the dramatic effect of areolar-cutaneous-junction injections on SN counts originally derived from perilesional injections.

In group 1, the IM SN was visualized in 4 of 82 patients (4.9%). In 1 patient, 2 IM SNs were noted. In group 2, the IM SN was visualized in 14 of 61 patients (23.0%). Of those 14, 5 had 1 IM SN detected, 3 had 2 IM SNs detected, 5 had 3 IM SNs detected, and 1 had 4 IM SNs detected. When the groups as a whole were compared, the difference in IM SN visualization rate between groups was significant (Table 1). The mean relative count per IM SN corrected for perilesional dose for both groups is depicted in Table 1. It is higher for group 2 patients as a whole: 16.6 for group 1 versus 26.4 for group 2. More IM SNs were seen per patient in the second group: 1.2 for group 1 versus 2.1 for group 2.

Breast size versus the number of IM SN–positive patients for a specific size is depicted in Table 2. In group 2, a clear trend is noted, with IM SNs much more likely to be visualized in women with smaller breasts: 46.2% (6/13) for small breasts, 21.1% (8/38) for medium breasts, and 0% for large breasts (0/10) (P < 0.017). The location of primary lesions in the breasts for both groups and the number of IM SNs associated with the location of the primary lesion are depicted in Figure 2. In group 1, there was a 10% rate of IM SN visualization when the primary lesion was clearly medial (2/20) and 4.2% when clearly lateral (2/48) (P = NS). In group 2, the rate of IM SN visualization was 38.9% (7/18) for clearly medial lesions and 16.2% (6/37) for clearly lateral lesions (P = 0.066, Fisher exact test). The few lesions at the 12- and 6-o’clock positions were not included in the analysis (n = 14 for group 1 and n = 6 for group 2), as the position to place them into was easily influenced by minor breast displacement or position differences. Group 2 also showed differences in the frequency of IM SN visualization depending on whether the primary lesion was clearly in the lower or upper portion of the breast—6 of 15 (40.0%) versus 7 of 36 (19.4%) for inferior and superior lesions, respectively (excluding 3- and 9-o’clock lesions, n = 10)—but these differences did not reach statistical significance. In group 2 patients with clearly medial or lateral lesions, the frequency of IM SN visualization was 50% and 44% in the small breast and 42% and 10% in the medium breast for medial and lateral lesions, respectively (P = NS).

Table 3 shows the rate of IM SN visualization for perilesional doses above and below the mean and median for both groups. Though there were 3 times as many IM SNs seen with doses above the mean and median in group 1, the differences in the rate of IM SN visualization did not reach statistical significance in either group.

DISCUSSION

During the middle of the last century and beyond, IM SN dissection was routinely performed as part of extended radical mastectomy. During the 1980s, IM SN dissection had fallen from favor and was viewed as having low diagnostic value or limited therapeutic benefit (7). Both IM SN irradiation and dissection were increasingly viewed as providing no effective increase in survival based on analysis of patient outcome studies, though the current role of radiation is more favorably regarded (7). A poorer prognosis in patients with both IM SN and axillary disease continues to be suggested.

The debate on IM SNs has been intensified again with the advent of lymphoscintigraphy and SN harvesting. Reasonable arguments can be presented that the only way to be truly “accurate” in staging the disease is to include harvesting of IM SNs when they are detected during lymphoscintigraphy or with the probe. Information from this increased accuracy in staging could in theory be used to further divide patients into those with positive IM SNs and a negative axilla, who might benefit from chemotherapy or parasternal irradiation, and those with a negative axilla and parasternal regions, for whom a decision to refrain from chemotherapy or irradiation therapy might be made (8). A review several years ago of national practice patterns suggested that most surgeons in the United States did not view the effort and morbidity associated with IM SN harvesting as worthwhile; 72% of surgeons in that survey did not routinely remove IM SNs showing drainage (9). Nevertheless, the issue is extremely controversial on many fronts, with proponents on both sides of the fence (8,10–14). Indeed, there have been several reports of IM SNs occasionally containing tumor in the absence of axillary disease (15,16). In a 1,273-patient study using “traditional” perilesional-injection techniques, in only 2.4% of patients was the IM SN detected with the probe, but of these, 10% had IM SN disease exclusively, with no axillary involvement (13). This would suggest that exclusive IM SN disease (negative axilla) is rare when considering all SN biopsy patients overall (0.24%). Yet, when considering the subgroup of patients in whom IM SNs are found, the frequency of exclusive IM SN disease is much higher (10%) and could change the patient’s stage if or when these IM SNs are harvested.

In recent publications from Europe, an IM SN visualization rate of 65.6% for inner-quadrant lesions and 10% for outer-quadrant lesions was noted using deeper perilesional (subtumoral) injections of nanocolloid below the tumor. These rates were in contrast to a 2.1% and 1% IM SN visualization rate for superficial injections in patients with inner- and outer-quadrant lesions, respectively (17,18). In this same study, 2% of patients with inner-quadrant lesions had a significant change in stage based on the results of IM SN harvesting; that is, in 3.4% of patients whose IM SN nodes were harvested (65.6% of all patients), the IM SNs were the only nodes with disease. An additional 3% of patients with inner-quadrant lesions were restaged upward; that is, in 5.1% of patients whose IM SNs were harvested, disease was detected in the IM SN basin as well as in the axilla (17,18). It also was suggested that when the primary lesion is in the inner quadrants, resection of the IM SNs can be attempted through the same single incision used to remove the primary tumor, with little additional morbidity.

Our results with ^99mTc-SC appear to demonstrate a similar effect. The depth of perilesional injections and, to a lesser extent, the location of the primary lesion play a role in the rate of IM SN visualization, with a statistically significant difference in IM SN visualization rate demonstrated between groups as a whole: 4.9% for group 1 versus 23% for group 2 (P < 0.003). Injection below the lesion is as valid as beside and above the lesion, as drainage can occur anywhere along the sphere surface of the tumor.

Differences within groups based on doses above and below the mean or median values were not statistically significant (Table 3). The effect of breast size on IM SN visualization rate was significant, with almost half the patients with small breasts demonstrating IM SNs whereas no IM SNs were detected in large-breasted women, a striking difference. The smaller the breast size, the higher the chances that IM SNs would be detected (Table 2). This could be explained by distance effects in large breasts (Fig. 3). A lesion in the center of a large breast (and the injected perilesional dose) would be much farther from the chest wall and the deeper lymphatic channels that course to the parasternal IM SNs, whereas a lesion in the center of a small breast would be much closer to the chest wall (Fig. 3), that is, an injection site proximity effect.

The location of the primary lesion has been suggested by some to influence IM SN detection, and such an influence could also be explained by a proximity effect. As noted in Figure 2, medial group 2 primary lesions had a higher rate of IM SN visualization (P < 0.066). Other investigators have also noted higher rates of IM SN detection when the primary lesion was medially located (13), whereas still others have noted no effect (19).

The effect of breast size on IM SN visualization rate has not been widely described and could represent an important factor for tumor spread in patients with small breasts with primary tumors located anywhere in the breast, that is, regardless of medial or lateral location of the primary tumor. No differences in the high IM SN visualization rate were noted for medial or lateral primary tumors in small breasts in our data; however, this could be secondary to the small number of patients involved.

Though differences within groups based on dose did not reach statistical significance, one could theorize that a higher dose would tend to make more tracer available to the IM SNs by a mass activity effect. This type of effect has been suggested for perilesional injections in respect to axillary SNs (20–22). For similar reasons, a possible threshold effect could explain the low IM SN visualization rate in the group 1 patients when perilesional doses less than the mean were used; at these doses, the noise level of the system could have been a factor (2.2% below the mean, 8.1% above the mean, P = NS) (Table 3).

Factors that have influenced delineation of axillary SNs other than injection location (perilesional vs. surface injections) have included increasing age (20,21,23,24) and large breast size (20,25). Both factors have been associated with decreased visualization of the SN. In addition, excisional biopsy has been associated with reduced rates of SN visualization (20,26), but this effect has not been consistently reported. Lesions in the upper lateral regions of the breast have also been associated with decreased detection of the axillary SN, possibly from diffusion zone masking of the injection site (27). Massage (28) and a larger dose (20–22) both have been associated with improved detection of the axillary SN. Filtration, injection volume, and fatty breasts (postmenopausal), among others factors, have influenced axillary SN detection, but the exact effects and benefits are controversial (29–33). Most authors agree that smaller particles tend to delineate a greater number of nodes “downstream” from the SN along the lymphatic channels, that is, echelon nodes, as activity passes through the SN to the next node (30).

When one combines the results from this article and others, some of the factors that influence IM SN visualization include medially located primary lesions (13,17,18), injection depth (17,18), breast size, and possibly dose and massage. Leppanen et al. noted that the IM SN was more likely to be visualized in patients who were younger, had a lower body mass index, had primary lesions in lower breast segments, and had a nonpalpable tumor (27). Dermal injections do not delineate the IM SNs to any extent, as has been widely reported by multiple authors (30). In a recent study from Japan by Shimazu et al., there was a higher chance that axillary and IM SNs were going to be involved with tumor when the primary lesion was deep in the breast (34). In addition, it was reported that even areolar injections are capable of delineating IM SNs in those rare instances in which the axillary SNs are extensively replaced (blocked) by tumor (34).

The terms peritumoral and perilesional are themselves ambiguous. Do injections labeled as such include only laterally placed deposits, or does the tissue space anywhere near the surface sphere of the tumor (including above and beneath) also qualify? The exact technical details of peritumoral injections are sufficiently vague in most reports that variations in methodology could partly account for the variations in IM SN detection rates in the literature. It is easily conceivable that some of the injections could have a significant “subtumoral” component. When lesions are very small, volumes of injection very large, and the exact position of the tumor and needle tip unclear, depositing activity beneath the tumor is certainly possible even if not originally planned.

Could exclusively subtumoral injections replace perilesional injections? Using only subtumoral injections of radiocolloid, only 50% of axillary SNs would have been detected by probe alone, as referenced to the use of intradermal blue dye concurrently (34). This same patient group during lymphoscintigraphy had a 38% IM SN and 70% axillary SN visualization rate. Among patients with visualized IM SNs, only 20% of those with inner-quadrant lesions had axillary SNs visualized along with IM SNs. All patients with SNs visualized both in the IM SN and axillary SN basins had outer-quadrant tumors (34). This is in contrast to the 85% of axillary SNs detected by probe from perilesional injections of radiocolloid in another group in the same study, though perilesional injections of dye were used as a reference. Nevertheless, the suggestion is that exclusively subtumoral injections (especially in patients with inner-quadrant lesions) could underestimate SNs in the axilla if performed alone. Because metastases can conceivably spread from anywhere along the surface of the tumor, a multidepth perilesional injection (including an injection below the level of the tumor) could provide better overall coverage for SN delineation in the axilla SN and IM SN basins, as would hybrid combination-injection techniques of concurrently administered radiocolloid at subtumoral and intradermal sites.

Radiocolloid administrations at dual injection sites were performed by Roumen et al. but not sequentially on the same day (1). Most other investigators have compared dual-injection techniques using a combination of radiocolloid and dye. Initially, we used a combination of simultaneous perilesional and intradermal radiocolloid injections over the lesion (35). Eventually, we replaced the intradermal injection with an injection at the areolar cutaneous junction while retaining concurrent perilesional injections (4). This hybrid combination-injection technique proved to be an even better method of delivering the radiotracer to the SN, with more of the injected dose reaching the SN, and was especially useful for patients undergoing next-day surgery (Fig. 1). We continue to use perilesional injections to retain the ability to visualize the IM SNs and to increase the probability that the correct nodes are assigned SN status.

CONCLUSION

The debate over IM SNs as regards both delineation and the eventual disposition of the nodes during surgery is controversial. Until a greater consensus is achieved, measures to delineate them should not be abandoned. Our data suggest that injection depth and breast size are important factors, among others, in determining IM SN visualization and can be explained by a simple proximity effect. Centers wishing to delineate IM SNs should examine their perilesional-injection protocols for adequate injection depth. Injections that extend deeper, to below the lesion, are physically closer to the deeper lymphatic channels that drain into the parasternal nodes and more likely to be taken up by these lymphatics. Realization of the major effect of these factors on IM SN visualization allows more control over the process and could explain some of the difference noted in the literature regarding this issue.