168: Sentinel Node Mapping in Lung Cancer

Early in the nineteenth century, Virchow implicated lymph nodes in the process of the local spread of solid tumors to a more widespread systemic disease. The node identified by Virchow is specifically located in the left supraclavicular region. In the modern vernacular, the term sentinel lymph node (SLN) describes the first lymph node to receive drainage from any solid tumor in any anatomic region. It was not until 1989 that the concept of SLN biopsy, as it is currently known, was first made popular based on studies by Morton et al.¹ A feasibility study using blue dye was translated into a successful clinical trial in patients with melanoma. Their results indicated that biopsy and analysis of SLNs accurately reflected the tumor status of the lymph node basin. SLN biopsy was introduced in breast cancer patients shortly thereafter.² Further studies supported SLN biopsy as a way to identify patients at highest risk for locoregional recurrence and metastatic spread, and, therefore, most likely to benefit from adjuvant therapy. SLN mapping aids in the identification of lymph nodes at highest risk of metastasis and allows for more detailed analysis to detect early metastatic disease thus identifying patients who may benefit from adjuvant therapy. While SLN biopsy is standard in both breast cancer and melanoma patients, a number of factors have led to slower adoption in patients with non-small cell lung cancer (NSCLC).

Lung cancer is the leading cause of cancer deaths among both men and women in the US, with an estimated 170,000 new cases and 160,000 deaths reported each year.³ Even in patients with stage I disease thought to have undergone curative resection, recurrence rates remain high at nearly 40%.⁴ This high incidence of recurrent disease in stage I lung cancer patients suggests that these patients are currently understaged and undertreated. In fact, retrospective analysis has demonstrated that nearly 15% to 20% of N0 patients harbor “occult” metastatic disease when nodes are histologically scrutinized in the manner done for breast cancer and melanoma.⁵ Not surprisingly, patients with occult nodal disease exhibit poorer survival and increased rates of recurrence. Despite all the recent clinical advances in lung cancer therapy, the best predictor of patient outcome following surgical resection remains the presence or absence of metastatic disease to lymph nodes. To improve survival of patients with surgically resectable lung cancer, detailed intraoperative lymph node evaluation is needed to more accurately stage patients and identify candidates for early adjuvant therapy.

SLN mapping seeks to identify the first lymph node to harbor metastatic disease from a nearby tumor and has become an integral part of patient selection for adjuvant treatment in solid malignancies such as breast cancer and melanoma. The two primary benefits of SLN mapping in breast cancer and melanoma have been (1) limiting the extent of lymph node dissection in the axilla or groin and (2) focused pathologic assessment of specific “at risk” nodes. In lung cancer surgery, the morbidity of mediastinal node dissection or sampling is low; however, lymph node drainage patterns are complex and variable, and thus the primary utility of SLN mapping is to identify those nodes at highest risk for metastatic disease and to facilitate the focused assessment of such nodes. The ability to conduct more detailed histologic and molecular analysis on the identified SLN would allow for better patient selection for subsequent adjuvant therapy in hopes of improving survival and decreasing locoregional recurrence. Although the goal of detecting micrometastatic disease prospectively via SLN mapping in lung cancer is sound, conventional SLN mapping techniques using radioisotopes or blue dye have not been successfully translated to the thorax, and currently this important therapeutic intervention is not clinically available for the care of lung cancer patients.

Several factors unique to the lung, including the presence of intrapulmonary and anthracotic (black) lymph nodes, have made identification of SLN with blue dye more difficult in lung cancer patients. Furthermore, “shine through” of radioactivity from the tumor injection site to nearby nodes in the mediastinum and hilum has hampered reliable SLN identification using radioisotopes in the thorax. In this chapter, we review prior attempts at SLN identification in non-small cell lung cancer and discuss recent research findings using indocyanine green (ICG) dye and near-infrared (NIR) fluorescence imaging to improve SLN mapping in patients with early-stage lung cancer.

SLN biopsy is currently the standard of care in both melanoma and breast cancer patients, but these methods have not been successfully translated to NSCLC. Unfortunately, only approximately 45% of surgeons completely remove lymph nodes from the hilum and mediastinum at the time of surgical resection.⁶ In addition, lymphatic drainage patterns are highly variable in the lung, making it difficult to predict the draining lymph node basin that should be sampled, and it is unknown whether the specific SLN for a patient is even included in the conventional surgical lymphadenectomy specimens. Furthermore, skip metastases to the N2 nodal station have been shown to bypass the nearby N1 stations and are present in an estimated 20% of cases.⁷ By removing only the nodes included with the resection specimen and failing to sample hilar and mediastinal nodes, surgeons are missing these skip metastases.

In addition to potentially missing nodal disease during surgical resection, current conventional histologic analysis may also miss micrometastatic disease in removed nodes. This prospect is particularly worrisome as nearly 16% of patients with histologically node-negative lung cancer and up to 27.5% of patients with subcentimeter adenocarcinomas are found to harbor evidence of occult micrometastatic disease or disseminated tumor cells within sampled nodes when these nodes are analyzed retrospectively with more time-intensive “SLN” histologic analysis techniques. Furthermore, Liptay et al. recently demonstrated in an analysis of 104 patients with lung cancer, and Takizawa in a series of 157 lobectomies, that the SLN was the only positive node present in 36% to 37% of node-positive patients.^7,8 This is important information, since failure to remove this node or to identify metastatic disease histologically leaves the patient with untreated occult micrometastatic disease which has been shown to correlate with a threefold increase in recurrence and a significant decrease in patient survival.⁸ To impose more focused histologic analysis on all lymph nodes removed during a lung resection is impractical. However, if SLN mapping could identify a small subset of nodes for the pathologists to scrutinize, focused histological and molecular examination could be performed on these nodes to better identify micrometastasis.

Conventional means of SLN mapping using blue dye and radioactive have been attempted in NSCLC but have remained unreliable secondary to physical properties of both the tracer and the lung. The initial attempt at SLN biopsy in lung cancer was performed in 1999 by Little et al.⁹ Thirty-six patients underwent intraoperative injection of blue dye with a SLN identified in only 47% of patients. Low success rates were attributed to the learning curve associated with the injection technique in presence of anthracotic nodes. The authors acknowledged an unacceptably low rate of SLN identification, but advocated for the addition of radiolabeled tracers to improve sensitivity and specificity. Unfortunately, however, this combination proved unsuccessful.

Intraoperative radioactive tracers were first used for SLN mapping in lung cancer by Liptay et al. in 2000.¹⁰ Intraoperative injection of technetium 99 resulted in successful radioisotope migration in 81% of patients and SLN identification in almost 87%. Roughly a quarter of the SLNs identified were located in the mediastinum representing skip metastasis. Furthermore, SLN analysis on previously deemed negative nodes detected occult micrometastatic disease in 24% of these patients. The SLN was the only metastatic node in just over a third of patients. As a result of this initial Phase I study, a multicenter Phase II trial under the direction of Cancer and Leukemia Group B (CALGB) research cooperative focused on intraoperative injection of technetium 99 in patients with suspected stage I NSCLC. Unfortunately, accrual was less than 50% and of those enrolled, only 51% of patients had an SLN identified. Poor accrual and the low success rate was attributed to logistical issues of organizing nuclear medicine, surgery, and pathology for intraoperative injection and analysis, and cumbersome regulations for radioactivity handling in addition to a difficult learning curve of the injection technique.¹¹ Less than optimal results obtained with either blue dye or radioisotopes led to different variations of the technique in an effort to increase the sensitivity and specificity of SLN identification in lung cancer. Owing to the fact that lymphatic vessels may be disrupted during intraoperative incision and dissection, Nomori et al. hypothesized that preoperative CT-guided injection and intraoperative injection via a transbronchial approach may improve results, but findings were similar to the initial studies by Liptay et al.¹² with successful SLN identification reported in only 81% of patients. Shine-through effect, residual radioactivity, and possible decreased lymphatic density or impaired lymphatic flow in COPD patients were also thought to contribute to these less than optimal results. Even more worrisome, however, was the increased morbidity noted with this technique. The preoperative injection was associated with complications of bleeding, pneumothorax, and potential tumor seeding along the injection track. These studies demonstrate that SLN mapping with single-agent blue dye or radioisotope is not optimal, but that a combined approach may be warranted.

Two separate groups attempted combining both blue dye and radioisotopes to aid in the detection of SLN in patients with lung cancer. Schmidt et al. used intraoperative injection of both blue dye and Tm-99, achieving a SLN identification rate of 81%.¹³ Tiffet et al. was the second group to evaluate intraoperative injection of both markers, but again demonstrated suboptimal results, identifying the SLN in only 13 of 24 patients (54%).¹⁴ The technical difficulties associated with each of these approaches included visualization of blue dye within anthracotic nodes, intrathoracic use of the Geiger counter to identify radioisotopes, and anatomical differences in patients, all of which were thought to contribute to the low accuracy in SLN identification. In addition to these technical issues, potential risks of radioactivity to the surgeon, operating room personnel, patient, and pathologist raise legitimate concerns that hampered adoption of this technique.

The search for a better technique has led surgical groups to innovate by exploring nonradioactive agents that not only promise more reliable and accurate SLN identification but also the potential for therapeutic intraoperative targeting of nodal metastases (Table 168-1).

Recent large animal and Phase I/II clinical trials have investigated the use of Near-Infrared (NIR) imaging for the identification of SLNs in breast, skin (melanoma), stomach, colon, and lung.^15,16 These studies have used optimized NIR lymphatic tracers, the most common of which is ICG dye, since it is FDA-approved for intravenous use in the calculation of cardiac outputs and ophthalmologic angiography. The success rates for SLN identification in these studies have approached 100% with doses of peritumoral ICG ranging from 0.1% to 5% of the conventional dose used intravenously. Peritumoral injection of ICG can quickly and reliably migrate in real time to draining lymph node basins permitting rapid and accurate SLN identification in an operative setting. Invisible NIR light penetrates deep into tissue allowing detection of fluorescent signal at a depth of 1 cm in solid tissue. Unique NIR cameras are now commercially available with the spectral sensitivity to provide safe NIR excitation with simultaneous intraoperative imaging of visible reflected light and the NIR fluorescence emitted from the ICG fluorophore. Separation of the visible and NIR fluorescent light allows simultaneous acquisition of color and NIR fluorescence images with high signal to background ratios with subsequent overlay of the two images (Fig. 168-1). This approach results in a single, intraoperative procedure whereby SLNs can be visualized but the surgical field remains unaltered by radioactivity or discoloration. This approach is particularly attractive because it allows direct videoscopic images of the necessary anatomic landmarks within the surgical field and extremely low background signal from biologic tissues thus permitting accurate surgical dissection. In large animal studies, NIR–SLN mapping has demonstrated a high success rate in identification of a single SLN station (>90%) with only occasional identification of a split lymphatic channel to two separate SLNs (<10%). Furthermore, this approach is nontoxic, without exposure risk to the patient or hospital personnel or the need for special biologic or disposal precautions.

One of the first studies to look at the use of NIR-imaging for SLN mapping in the lung was performed by Soltesz et al. in a swine model using NIR-fluorescent quantum dots.¹⁷ Quantum dots are semiconductor nanocrystals that contain an inorganic core of cadmium telluride, an inorganic shell of cadmium selenide, and an outer organic coating of solubilizing oligomeric phosphines. The intraparenchymal injection of quantum dots did result in SLN identification in all animals tested regardless of the lobe of the lung evaluated, and confirmed that SLN identification was also possible for the lung. Unfortunately, although quantum dots offer excellent lymphatic mapping and SLN identification, they have not been aggressively pursued for clinical use because of concerns over the possibility of heavy metal toxicity in patients. Therefore, additional preclinical studies were performed to establish the feasibility of SLN mapping in lung tissue using the organic FDA-approved fluorophore ICG. These large animal pulmonary studies demonstrated evidence of an NIR fluorescent signal within the SLN within 10 minutes of intraparenchymal ICG injection, establishing the potential of NIR imaging as a rapid, safe, and accurate means of SLN mapping for the lung.

Based on these successful preclinical studies in swine, we and others have assessed the safety and feasibility of NIR imaging using ICG for SLN identification in patients with early-stage non-small cell lung cancer. In a study by Ito et al., whereby ICG was used at high concentrations to visibly see migration with the naked eye, an SLN was identified in only 7 of 38 patients with lung cancer (<20%) as detection was limited to evidence of stained lymph nodes.¹⁸ However, Yamashita et al. demonstrated that when ICG detection utilized intraoperative fluorescent imaging, SLNs were identified in 80.3% of the lung cancer patients studied using 10 mg of ICG.¹⁹ Our current Phase I clinical trial has utilized both an open platform (FLARE) and thoracoscopic (Novadaq SpyScope) NIR system to conduct a dose escalation study to optimize intraoperative mapping of SLN in lung cancer. The current technique consists of intraparenchymal, peritumoral injection in each of four quadrants around the tumor, followed by a short period of ventilation to improve lymphatic flow of ICG (Fig. 168-1). To date, successful SLN identification in 80% of patients undergoing lung resection has been obtained with the injection of only 1 mg ICG in an ICG–albumin mixture. Future goals of this study are to optimize the ICG dose and determine whether the SLN identified using NIR imaging would have been missed during a conventional lymphadenectomy, and if focused analysis of the identified SLN will identify occult metastatic disease in patients currently classified as “node negative” by non-SLN techniques.

The best predictor of survival for patients with surgically resectable disease is the presence or absence of metastatic disease in regional lymph nodes. However, even in the setting of complete surgical resection, 30% to 50% of patients with early-stage lung cancer will develop recurrent disease, proving that current approaches are inadequate to accurately stage nodal disease in patients with lung cancer. SLN identification with concomitant focused histologic analysis for micrometastatic disease could identify patients that may benefit from adjuvant therapy and thus improve patient survival. Despite limited success with standard approaches for SLN identification, recent investigations suggest that SLN mapping using NIR imaging may prove to be a safe, rapid, and feasible approach to improve staging in lung cancer patients. Future studies are required to further optimize the ICG dose and technique, and to assess the efficacy and clinical impact of SLN analysis on the use of adjuvant therapy and subsequent patient outcomes.

The future of SLN mapping in lung cancer is on the horizon and will become feasible for all with future advances in technology. The hope is that through better staging and treatment planning, SLN mapping will improve outcomes for patients with lung cancer.