Evaluation of New Tracers in the Identification of Sentinel Lymph Node in Patients with Early Breast Cancer

Abstract

Purpose: 

Sentinel node mapping is the standard of care for evaluation of axilla for women with early node negative breast cancer. Validation of a new tracer for sentinel node biopsy requires full axillary lymph node dissection to establish its performance indicators. This exposes about 70% of women to unnecessary axillary dissection with its attendant morbidity.

Aims and Objective: 

To investigate the predictive value of identification of sentinel lymph nodes by a tracer for knowing its sensitivity and false-negative rates (FNR).

Methods: 

A linear regression on data from a network meta-analysis was carried out, and the correlation between identification and sensitivity and its predictive value were ascertained.

Results: 

A strong linear relationship was observed between identification and sensitivity of sentinel node biopsy (correlation coefficient r = 0.97). The sensitivity and false negativity can be predicted by the identification rate. An identification rate of 93% corresponds to sensitivity = 90.51% and a FNR = 9.49%. The current literature on newer tracers has been succinctly reviewed.

Conclusion: 

The linear regression demonstrated a very high predictive value of identification rate for ascertaining the sensitivity and FNRs of sentinel node biopsy. A new tracer for sentinel node biopsy can be introduced in clinical practice, if it achieves an identification rate of 93% or more.

Introduction

The standard guidelines for the management of axilla for patients with early breast cancer, presenting with clinically and sonographically negative axilla, are to perform sentinel lymph node biopsy (SLNB). The current recommendation is to use dual tracers,[¹] a combination of a blue dye and radioisotope tagged with a large particulate matter such as sulfur colloid, antimony, or nanocolloid albumin. In this review, we present the present state of the knowledge about new tracer techniques for sentinel node biopsy in early breast cancer.

The sentinel node mapping is based on the principle of laws of fluid hydrodynamics regulating the flow of interstitial fluid in the tissue spaces and lymphatics. The particles smaller than a 1–2 nm in diameter mostly penetrate the blood capillary membrane, whereas larger particles can enter the lymphatic capillaries and be transported to lymph nodes. The particles larger than 25 nm in diameter are drained by lymphatics rather than blood capillaries. The optimal colloidal size for lymphatic mapping is approximately 50–70 nm. Initial studies used Tc-99m sulfur colloid tagged to a large-particle sulfur colloid, a radiopharmaceutical initially approved for liver and spleen scintigraphy. This radiocolloid is still used, with the US Food and Drug Administration approving its indication for SLNB. In Europe, albumin colloids and a preformed Tc-99m Sulfur colloid are the approved agents. The technetium may be tagged to antimony, which is cheaper than sulfur colloid. In Australia and in Sanjay Gandhi Postgraduate Institute of Medical Sciences, Lucknow, antimony tagged technetium is preferred radiotracer, whereas Tc-99m calcium phytate is the tracer used in Japan.

Donald Morton first described lymphatic mapping in 1992 for truncal cutaneous melanoma with the use of radioactive gold isotope.[²] David Krag used a radiotracer Tc-99m tagged with sulfur colloid and gamma probe for the identification of sentinel lymph node (SLN) mapping in breast cancer in 1993.[³] Dr. Giuliano demonstrated sentinel nodes in breast cancer with Isosulfan blue dye in 1994.[⁴]

Problems associated with radioisotope as a tracer

Although the identification rate is high with isotope, this procedure has several limitations. The half-life of Tc-99m sulfur colloid is 6 h, which limits its availability and use in the small centers, where radioactive isotope handling is nonexistent. The technetium-99m is the decay product of Mo-99. The Mo-99 decays rapidly and therefore needs to be supplied to hospital nuclear medicine departments every 2 weekly and is produced in only in a few reactors across the world. The 6 h half-life of the isotope restricts scheduling of surgery because the isotope is injected by the nuclear medicine department. Furthermore, patients may not like the idea of being exposed to radiation, especially if they are pregnant or lactating. The radioisotope is very expensive and not available in many hospitals. It also needs a gamma ray detection probe which costs about 2 million Indian rupees. The constraints of combined dual tracer-based SLN mapping have led to the development of alternative tracers.

Novel tracers for sentinel node biopsy

To overcome the problems with radioisotope, numerous novel techniques have been developed, namely, fluorescent dye like indocyanine green (ICG), carbon nanoparticles, superparamagnetic iron oxide (SPIO) nanoparticles, and contrast-enhanced ultrasound (CEAU) scan using microbubbles.[⁵]

These newer tracing techniques with their performance indicators are summarized as follows:

Fluorescent dyes

ICG has been evaluated by a number of authors. In a randomized study, Tong et al. compared the identification of SLN with ICG + patent blue with patent blue alone. ICG + patent blue could demonstrate SLN in 96.9% (93/96), with an accuracy of 98.9% (92/93) and false-negative rate (FNR) = 3.4% (1/29). The patent blue alone detected the SLN only in 84.9% (62/73), with an accuracy = 96.8% (60/62) and FNR = 11.1% (2/18). The ICG + patent blue group showed significantly superior results compared to the patent blue alone group for SLN detection (P = 0.005) with a greatly reduced FNR.[⁶] Valente et al., in 2019, reported an identification rate of SLNB with ICG = 98.9%, compared with radioisotope = 97.8%.[⁷] Qin et al. demonstrated identification with ICG + methylene blue = 100% compared to methylene blue dye = 96.7% and with carbon nanoparticle = 98.3%.[⁸] A potential benefit of the ICG fluorescence is the ability to visualize the lymphatics and the nodes on the surface of skin before the incision. Thus, a very small skin incision may be placed exactly over the florescent node.

A study in Cambridge, UK, on the role of ICG in 100 women reported a sensitivity for ICG = 100%.[⁹] A major drawback of ICG is the high cost of the infrared device needed to excite the dye and the near-infrared camera to detect the fluorescence. The Spy camera costs Rs 2 crores, while the cheapest device available, photodynamic-eye costs Rs 35 lacs. The ICG dye (Aurogreen) costs Rs 800 per vial as compared to Rs 50 for an ampoule of fluorescein.

Another fluorescent dye, fluorescein, has been employed for sentinel node biopsy in breast cancer by a number of authors. We were the first to report the sentinel node mapping in breast cancer with fluorescein.[¹⁰] Later, several authors have demonstrated a very high identification rate with fluorescein ranging from 90% to 95% and FNR of 6%–8%.[^11,12]

Ahmed et al. published a systematic review to assess three techniques for SLNB that are not radioisotope dependent, namely, ICG fluorescence, contrast-enhanced ultrasound using microbubbles, and SPIO nanoparticles. Their systematic review suggested that these new methods for SLNB have clinical potential, but give high false-negative results. They also found that identification with ICG is similar to radioisotope.[¹³]

Cox et al. performed a study on 347 patients using phospholipid stabilized microbubbles containing sulfur hexafluoride gas with a mean diameter of 2·5 mm. Between 0.2 ml and 0.5 ml of ultrasound contrast agent was injected (up to three consecutive injections) intradermally in the periareolar region in the upper outer quadrant. The ultrasound examinations were performed with a Sequoia scanner providing: conventional gray scale, pulse-inversion harmonic gray scale, contrast-specific sonographic imaging with live dual images of tissue only and contrast agent image (Cadence contrast pulse sequencing). A high frequency 14-MHz linear array probe was used. After 10–30 s of massage, lymphatic channels were visualized on contrast pulse sequencing and followed into the axilla. Areas of contrast accumulation were then imaged with gray scale to identify the lymph nodes. Once identified, they were confirmed on fine-needle aspiration cytology or core biopsy. In this study, FNR was 13·3% with an identification rate of 87·7%.[¹⁴]

A multicenter equivalence trial was conducted to evaluate the magnetic particles, SentiMAG, in 160 patients with invasive, and in situ breast malignancy. This trial compared the radioisotope with or without blue dye against the magnetic technique (magnetic tracer and handheld magnetometer). All patients received a periareolar or peritumoral injection of radioisotope preoperatively and 2 mL magnetic tracer (27 mg iron/mL) diluted in 3 mL of normal saline perioperatively. SLNs were identified in 95% of patients with the standard technique and in 94.4% with the magnetic technique. The magnetic technique was concluded to be noninferior to the standard technique.[¹⁵]

Professor Umberto Veronesi of Institute of Oncology Milan studied 1446 cases of breast cancer patients and found that the spread of breast cancer to the axilla follows a regular pattern, Level I being involved first, followed by Level II and then Level III. He reported only 1.3% skip metastasis to Level II and Level III without involving Level I.[¹⁶]

Sentinel node is the first node to receive the lymphatic drainage from the tumor, and the tumor does not spread to other area before reaching the sentinel node. Hence, absence of tumor in the sentinel nodes denotes absence of metastasis in other nodal basin in the majority of cases.

The National Surgical Adjuvant Breast and Bowel Project Trial protocol B-32 on 5611 patients with early breast cancer used the combination of blue dye and radioisotope and compared SLN and axillary dissection to SLN alone, having axillary dissection only for patients with positive sentinel nodes. The B-32 trial demonstrated identification in 97% and confirmed nodal status in 96% of patients with a 9.8% FNR. After 8 years of follow-up, overall survival, disease -free survival, and regional control were all statistically equivalent across both the groups.[¹⁷]

Veronesi et al. compared total axillary dissection to SLNB with axillary dissection only if metastasis was found in the SLN. They randomized 516 patients with cancers <2 cm in diameter to each of these groups and used Tc-99m sulfur colloid-labeled particles of colloidal human albumin either peritumoral or subdermal depending on the location of tumor. In the axillary group, the overall accuracy was 96.9%, the sensitivity was 91.2%, and the specificity 100% with a FNR of 8.8%. In the sentinel group, positive SLNB was present in 35.5%, and in the ALND group, it was 32.%. There were less pain and better arm mobility in the patients who underwent sentinel node biopsy alone, compared to those who also underwent axillary dissection.[¹⁸] Professor Mansel et al. of Cardiff conducted a multicenter randomized controlled ALMANAC trial in the United Kingdom.[¹⁹] This trial demonstrated less arm and shoulder morbidity, including swelling, sensory loss, and mobility loss after SLNB along with better quality of life scores. The risk of lymphedema for the SLNB group compared to the standard axillary treatment group at 12 months was 5% and 13%, respectively. Similarly, sensory loss was 11% versus 31% in SLNB and standard axillary treatment, respectively.

Wu et al. evaluated carbon nanoparticles in SLNB. Carbon nanoparticles are synthetic tracers with an average diameter of 150 nm. Carbon nanoparticles selectively enter the lymphatic vessels rather than blood capillaries due to their large molecular size and permeability. Upon injection into the tissues around the tumor, carbon nanoparticles are rapidly engulfed by macrophages and then pass through the lymphatic vessels to the SLNs, thus staining them black. In this study, SLNs were identified in all patients (100%) using carbon nanoparticles method and all SLNs were stained black by carbon nanoparticles. None of the 83 patients experienced adverse effects in response to carbon nanoparticles. Of the 83 SLNB procedures using blue dye, 73 had SLNs identified successfully at an identification rate of 88% (73/83). There was a decrease in the FNR, from 15.8% with blue dye to 11.1% using carbon nanoparticles.[²⁰]

The performance indicators for combined radioisotope + blue dye are summarized in Table 1. Table 2 presents the identification and false negativity with newer tracers.

Table 1: Performance indicators of radio-isotope+blue dye for sentinel node biopsy in breast cancer

Table 2: Performance indicators for newer tracers in sentinel node biopsy in breast cancer

With the growing technology, scientists are developing new tracers for identifying the SLNs. The identification rate and FNR of the new tracer needs to be evaluated against acceptable standards before adopting it in routine clinical practice. The identification rate can be determined by performing sentinel node biopsy alone in eligible patients. However, FNR of the new tracer can only be determined if patients undergo sentinel node biopsy, followed by axillary node dissection at the same operation. The histological examination of both the sentinel nodes and the “rest of the axilla” is carried out to calculate the FNR in identifying the metastasis in the sentinel nodes. It is recommended to perform the validation of a new tracer on at least 30 patients with axillary node-negative breast cancer.

It, thus, requires surgeons to perform full axillary dissection in at least 30 patients, in whom about 70% (70% out of 30 = 21 patients) will prove to be negative for axillary node metastasis. For these 21 women, performance of full axillary dissection is not only useless, but harmful. It subjects these patients to unnecessary risk of lymphedema, shoulder stiffness, anesthesia, or paresthesia in the upper arm and armpit. Many of these patients with early breast cancer are well informed about the literature and adverse effects of axillary node dissection and, hence, are unlikely to agree and consent for the validation study.

A test with high sensitivity has low FNR (FN rate = 1 – sensitivity). We explored whether it is possible to use identification rate alone to evaluate new tracers and avoid unnecessary axillary node dissection and its associated arm morbidity.

We carried out regression analysis extracting data from published studies of sentinel node biopsy in early breast cancer and found that there is a strong relationship between the identification rate of sentinel nodes and the sensitivity of the sentinel node biopsy. Mok et al. pooled the data from 35 studies which reported the identification and sensitivity of six tracers: blue dye alone, blue dye along with radioisotope technetium, ICG, radioisotope technetium alone, SPIO, and CEAU imaging.[²⁸] They performed a network meta-analysis on these 35 studies and computed the identification and sensitivity for the aforementioned tracers.

Methods

We extracted the data from a network meta-analysis published by Mok et al.[²⁸]

The data of this meta-analysis are presented in Table 3. The identification rate and sensitivity of the sentinel node biopsy were plotted on a scatter diagram [Figure 1]. A “straight line relationship” between identification and sensitivity was observed. The correlation coefficient between sensitivity and identification was computed. A simple linear regression was carried out to obtain the regression equation.

Figure 1: Showing the linear relationship between the Identification on X axis and Sensitivity on Y axis based on data from Mok et al.^[28] Network meta-analysis

Table 3: Identification sensitivity and false negativity of sentinel node mapping with different tracers from the data of network meta-analysis by Mok et al ^[28]

Results

A linear relation with a strong correlation was found between sensitivity and identification with a correlation coefficient, r = 0.9778. The simple linear regression yielded the following equation:

Y = A + BX

where Y = sensitivity; A = intercept =; B = slope=; and X = identification rate.

Details of linear regression: Best fit values: slope = 1.500 ± 0.1607 (95% confidence interval [CI] 1.054–1.946); Y intercept = −48.99 ± 15.23 (95% CI −91.27 to −6.723); X-intercept = 32.65 (95% CI 6.374–46.91); goodness of fit R² = 0.9561, S_y,x = 1.531; F = 87.18; degree of freedom DFn, DFd = 1,4; P = 0.0007.

Substituting the values of A and B, we obtained the following regression equation:

Y = 1.5 X– 48.99; A (intercept) = -48.99 and B (slope) = 1.5

We computed the sensitivity and false negativity for varying identification rates as follows:

Note: Sensitivity % + FNR % = 100%

So FNR % = 100 – sensitivity%

Applying this equation, we got the following results:

• For identification = 93% the sensitivity = 90.51%; False negative = 9.49%
• For identification = 95% the sensitivity = 93.51%; False negative = 6.49%
• For identification = 97% the sensitivity = 96.51% False negative = 3.49%
• For identification = 99% the sensitivity = 99.51%; False negative = 0.49%
• For identification = 100% the sensitivity = 101%; False negative = 0%.

Our analysis suggests that if a surgeon achieves identification of 93% or more for a new tracer, acceptable sensitivity and FNR can be obtained obviating the need of performing “full axillary dissection” as part of a “validation program” for a new tracer for sentinel node mapping.

Conclusion

The regression analysis suggests a strong predictive value of identification of sentinel node in predicting the sensitivity of the sentinel node biopsy in patients with early breast cancer. A new tracer for sentinel node biopsy can be introduced in clinical practice if it achieves an identification rate of 93% or more.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.