Radioisotope Lymph Node Mapping in Nonsmall Cell Lung Cancer: Can It Be Applicable for Sentinel Node Biopsy?

Transcription

1 Radioisotope Lymph Node Mapping in Nonsmall Cell Lung Cancer: Can It Be Applicable for Sentinel Node Biopsy? Kazuhiro Ueda, MD, Kazuyoshi Suga, MD, Yoshikazu Kaneda, MD, Hisashi Sakano, MD, Toshiki Tanaka, MD, Masatarou Hayashi, MD, Tao-Sheng Li, MD, and Kimikazu Hamano, MD First Department of Surgery and Department of Radiology, Yamaguchi University School of Medicine, Ube Yamaguchi, Japan Background. Previous studies on intrathoracic lymph node mapping have focused on the validity of a sentinel node concept, but not on the usefulness for sentinel node biopsy. Methods. The subjects were 15 patients clinically diagnosed with N0 nonsmall cell lung cancer. Technetium- 99m tin colloid was injected into the peritumoral area 1 day preoperatively and a time course of tracer migration was monitored by scintigraphy. A hand-held gamma probe counter was used to count the intrathoracic lymph node stations. Resected nodes were also counted to assess the accuracy of the intrathoracic counting. Results. Serial scintigraphies showed that the tracer migrated through airways and the appearance resembled hot nodes. On intrathoracic counting, 50% of the nodal stations appeared positive; however, only 23% of these apparently positive nodal stations were ultimately shown to be truly radioactive. The true positive and true negative rates of detecting intrathoracic hot nodes were 100% and 56%, respectively. Because the counts of the nodal stations could include the counts from the hot primary tumor ( shine-through ) or airway radioactivity, legitimate hot nodes were identified after dissecting all the apparently positive nodal stations. Two of the 9 patients in whom hot nodes were identified had nodal metastatic disease and actually had tumor cells within the hot nodes. The only complication related to the preoperative injection of technetium-99m was a minor pneumothorax. Conclusions. Although radioisotope intrathoracic lymph node mapping is safe, it appears to be unsuitable for sentinel node biopsy because shine-through and the airway-migrated radioactive tracer complicated the intrathoracic counting. Only serial scintigraphy could distinguish hot nodes from airway migration. (Ann Thorac Surg 2004;77:426 30) 2004 by The Society of Thoracic Surgeons Sentinel lymph node biopsy is standard practice in the surgical management of breast cancer and melanoma [1, 2]. Moreover clinical trials validating this practice have been conducted on other major carcinomas [3 5]. Although much effort is being made to practice minimally invasive surgery for patients with nonsmall cell lung cancer, the efficacy of sentinel node biopsy has not yet been confirmed for this disease. Lymph node mapping using the blue dye method [6] has some disadvantages. First, the blue dye can rapidly infiltrate the sentinel node and pass through into the distal drainage nodes, compromising the identification of a true sentinel node [7]. Second, every stained lymph node should be visually assessed by dividing the pulmonary vessels or even by incising the mediastinal pleura. Third, anthracotic black lymph nodes can be difficult to distinguish among the blue and negative nodes. The advantages are that radiolabeled nodes can be Accepted for publication July 3, Address reprint requests to Dr Ueda, First Department of Surgery, Yamaguchi University School of Medicine, Minami-Kogushi, Ube Yamaguchi , Japan; kaueda@yamaguchi-u.ac.jp. quantitatively detected by a hand-held gamma probe even when they are not visible, allowing for a mediastinal node search without manipulation. Several investigators have attempted sentinel lymph node mapping after the intraoperative local injection of various kinds of nanocolloid particles (less than 80 nm in diameter and labeled by technetium-99m) [8 10]. In these trials, successful lymph node labeling was done, achieving a sentinel lymph node identification ratio in more than 80% of patients; however, there is still concern that such small particles could pass through into the distal drainage lymph nodes [11] and that the intraoperative lymphatic drainage could be artificial. Technetium-99m tin colloid is about 400 to 1000 nm in diameter and can slowly migrate to the sentinel lymph node and, on rare occasions, pass through into the next drainage lymph node even 1 day after the injection [11, 12]. In terms of accurate lymph node mapping under physiologic lymphatic conditions, a preoperative injection of technetium-99m tin colloid appears to be ideal. Thus, we conducted a clinical trial to determine the safety of lymph node mapping in patients with nonsmall cell lung cancer after a 1-day preoperative local injection of technetium-99m tin colloid. We also evaluated the accu by The Society of Thoracic Surgeons /04/$30.00 Published by Elsevier Inc doi: /s (03)

2 Ann Thorac Surg UEDA ET AL 2004;77: RADIOISOTOPE LYMPH NODE MAPPING racy of conducting an intrathoracic search for radiolabeled lymph nodes to determine the usefulness of this practice for selective lymph node dissection. Patients and Methods This human study was approved by the Yamaguchi University Hospital Institutional Review Board in July From August 2001 to July 2002, 15 consecutive patients with a pathologic diagnosis of nonsmall cell lung cancer classified clinically as N0 were enrolled in this study after giving informed consent. None of the patients had received prior adjuvant therapy or undergone mediastinoscopy. All patients had a final pathologic diagnosis of nonsmall cell lung cancer, differentiated as either adenocarcinoma in 12 patients and as squamous cell carcinoma in 3. There were 10 men and 5 women with a mean age of 65 years (range, 54 to 78). Lesions were located in the right upper lobe in 5 patients, the right middle lobe in 2, the right lower lobe in 3, the left upper lobe in 3, and the left lower lobe in 2. The mean tumor size was 2.2 cm (range, 1.2 to 3.5 cm). Postoperative pathologic examination confirmed that all patients were free of visceral pleural invasion. We performed lobectomy in 13 patients and segmentectomy in 2. The pathologic disease stages were stage IA in 9 patients, stage IB in 2, stage IIA in 1, stage IIB in 1, stage IIIA in 1, and stage IIIB in 1 (which was due to an intrapulmonary metastasis [pm1] diagnosed postoperatively). Tracer Injection and Time Course Study Technetium-99m tin colloid was used for the radioactive tracer, which was injected 1 day preoperatively. Before injection, the location and depth of the tumor was identified by computed tomography (CT) and the puncture site was determined in relation to the position of the tumor. After the local anesthesia with a total of 5 ml of 1% lidocaine hydrochloride, one (n 2) or two (n 13) injections of technetium-99m tin colloid, a dose of 2 mci (74 MBq) suspended in 1.5 ml per injection, was given into each peritumoral area using a 23-gauge needle attached to a 2.5-mL syringe. Immediately after the injection, scintigraphy was done to confirm that there was no pleural seeding of the tracer. Chest CT scanning was also done to check for pneumothorax or hemothorax and to confirm whether the injection had been successful. In the first 10 patients, the time-course of tracer migration from the injected site was monitored by serial lymphoscintigraphies 5 minutes after and then 1, 2, 4, and 19 hours after the injection. Intraoperative Counting for Tracer Migration Counting was done routinely using the hand-held gamma probe counter (Navigator; Auto Suture Japan, Tokyo, Japan) for each ipsilateral nodal station: stations 1 to 4 and 7 to 9 (N2 stations) and stations 10 to 13 (N1 stations) for right-sided tumors; and stations 4 to 9 (N2 stations) and 10 to 13 (N1 stations) for left-sided tumors. Anatomic landmarks for each lymph node station were 427 determined according to the definition adopted in 1997 by the American Joint Committee on Cancer and the Union Internationale Contre le Cancer [13]. The counts per 5 seconds for tumor tissue, nodal stations, trachea, and intrathoracic background were recorded before and after anatomical pulmonary resection. When counting each nodal station adjacent to the airway, a gamma probe was carefully directed tangential to the airway to avoid counting the airway radioactivity. Positive radioactivity for a nodal station was defined if the counts for that nodal station were more than three times the intrathoracic background values. We then performed systematic lymph node dissection for each ipsilateral nodal station, namely stations 1 to 4 and 7 to 9 for right-sided tumors and stations 4 to 9 for left-sided tumors. After the specimens were resected from the operative field, counting for the resected specimens was also done separately to assess the accuracy of the intrathoracic search for radioactive lymph nodes. Because intrathoracic counts for nodal stations could include counts from hot primary tumors or airway radioactivity, significant lymph node radiolabeling (hot nodes) was confirmed if the counts for the resected node were more than three times the intrathoracic background value or greater than the counts for resected mediastinal fatty tissue surrounding these lymph nodes. Pathologic Nodal Assessment The dissected lymph nodes were examined histologically using formalin-fixed and paraffin-embedded sections with hematoxylin and eosin staining. Results The only complication related to the preoperative tracer injection was a pneumothorax requiring transient tube drainage in 1 patient, although minimal free air space was detected by routine chest CT scanning in 2 other patients who were asymptomatic (Fig 1). There were no other respiratory or systemic adverse reactions throughout the examination. Radioisotope injections were accurately delivered in all patients (Fig 1). Scintigraphy was used to confirm that no radioisotope seeding into the pleural cavity had occurred. The average duration of tracer injection and thoracotomy was 20 hours (range, 19 to 20.5). Preoperative lymphoscintigraphy did not show significant lymph node migration during 19 hours of observation (n 10). In 7 patients, radioisotope migration through the airway was seen just after the tracer injection (Fig 2). In 4 patients, hot spots were seen in the hilar or mediastinal region 1 hour to 4 hours after injection and had an appearance similar to intrathoracic hot nodes (Fig 2). However, on serial scintigraphies, these hot spots moved along the airway and seemed to be from airway isotopes rather than from lymph nodes. Intraoperatively, 47 of 60 N1 stations in the 15 patients had positive radioactivity, whereas after the resection of tumor tissue and N1 stations, only 15 of 47 radioactive N1 stations had hot nodes because the counts for N1 stations included the counts from the hot primary tumor ( shine- GENERAL THORACIC

3 428 UEDA ET AL Ann Thorac Surg RADIOISOTOPE LYMPH NODE MAPPING 2004;77: Table 1. Accuracy of Intrathoracic Detection of Hot Node Stations Intrathoracic Counting Accuracy N1 stations (60/15 patients) With hot node 15 Positive: 15 True positive Negative: 0 rate 100% Without hot node 45 Positive: 32 Negative: 13 True negative rate 29% N2 stations (100/15 patients) With hot node 3 Positive: 3 True positive Negative: 0 rate 100% Without hot node 97 Positive: 30 Negative: 67 True negative rate 69% Fig 1. Chest computed tomography scan 30 minutes after the injection of technetium-99m tin colloid. The tracer was accurately injected into two sites of the peritumor (arrows). Note the minimal air space although this patient did not complain of any symptoms. through ) and the intrabronchial radioactivity. Routine countings for each ipsilateral N2 station revealed positive radioactivity in 33 of 100 N2 stations in 15 patients, whereas after the dissection, only 3 of 33 radioactive N2 stations had hot nodes because the counts for the N2 stations included the counts from residual intratracheal radioactivity. We did not find any hot nodes within negative radioactive nodal stations. Thus, the true positive and true negative rates of the intrathoracic detection of N1-hot nodes were 100% and 29%, respectively, whereas the true positive and the true negative rates of the intrathoracic detection of N2-hot nodes were 100% and 69%, respectively (Table 1). True hot nodes were finally identified after dissection of all the apparently radioactive N1 and N2 stations. A final counting for the resected specimens showed 18 hot nodes in 9 of the 15 (60%) patients: 2 had only one hot node, 5 had two hot nodes, and 2 patients had three hot nodes. A hot node was located in station 13 in 3 patients, station 12 in 7 patients, station 11 in 3 patients, station 10 in 2 patients, station 7 in 1 patient, station 4 in 1 patient, and station 5 in 1 patient (Table 2). No perilymphatic tissues were positive for accumulation in this patient series. All the patients with an identified hot node had at least one hot node within the N1 stations and 3 patients had hot nodes in both the N1 and N2 stations. Lymph node metastasis was proven in 4 of the 15 patients (27%) after serial pathologic examination. Hot nodes were identified in 2 of these 4 patients, 1 of whom was found to have metastasis only in the hot node (station 13) whereas the other was found to have metastasis in both the hot node (station 11) and the nonhot node (station 9). There was no false negative nodal histology in this series. Comment In 2000, Liptay and coworkers reported successful intraoperative radioisotope lymph node mapping [8], since Table 2. Location of Hot Node Station Hot Node (18/9 Patients) Fig 2. Serial scintigraphies after the injection of technetium-99m tin colloid. Note that the tracer migrated through the airway immediately after the injection (5 minutes) and that this airway radioactivity was weaned thereafter. One hour after the injection a hot spot found in the hilar region that resembled a hot lymph node was transported along the airway to the upper mediastinum 4 hours after the injection. (min minutes; h hours.) 4 Lower paratracheal 1 5 Aortopulmonary window 1 7 Subcarinal 1 10 Hilar 2 11 Interlobar 3 12 Lobar 7 13 Segmental 3

4 Ann Thorac Surg UEDA ET AL 2004;77: RADIOISOTOPE LYMPH NODE MAPPING when several validation studies of this practice have been conducted [9, 10, 14, 15]. However, the best methods of radioactive tracing remain controversial. We believe that injecting Technetium-99m tin colloid 1 day preoperatively could be ideal for intraoperative radioisotope lymph node mapping for the following reasons: preoperative injection methods may be more reflective of the physiologic lymphatic state than intraoperative methods; as shown by the present serial scintigraphies, the tracer migrates initially through the airways rather than through the lymphatics when injected into the lung tissue, thereby complicating an accurate lymph node count, whereas airway tracers would be cleared timedependently; and tin colloid may slowly migrate into the first drainage lymph node but it will rarely pass through into the following nodes because the colloid consists of relatively large particles (about 400 to 1000 nm in diameter). According to a time-course study of breast lymphoscintigraphy, hot sentinel nodes persisted for up to 1 day after the injection of tin colloid without passing through to the following nodes [12]. Therefore, previous investigators defined a sentinel node as any hot node identified using tin colloid and found an average 1.7 to 2.0 sentinel nodes per patient [14, 15], which is comparable with our results. On the contrary, some nanocolloid particles will rapidly migrate to a sentinel lymph node, then pass through and be distributed widely to the following lymph nodes [11]. These colloid particles have mainly been used in the intraoperative migration method to find the sentinel node just as the tracer reaches the node [8 10]. However, concern remains that sentinel nodes might be incorrectly identified even with the intraoperative use of these nanocolloid particles. The true sentinel node can be missed during anatomical resection because it may be hidden by a hot primary tumor or airway radioactivity until the nodes are resected. Our study confirmed the safety of preoperative radioisotope injections. Scintigraphy showed successful radioisotope injection into lung parenchyma without radioisotope seeding into the pleural cavity and CT showed that the injection site was indeed a peritumor lesion located within the same subsegment as the tumor (Fig 1). Furthermore, in our series, the incidence of complications resulting from preoperative tracer injections was well within acceptable limits. We found hot nodes in 60% of the patients, which seemed a relatively low percentage compared with those of previous studies using tin colloid for the identification of breast [16] or gastric [17] sentinel lymph nodes. In our series, 5 of the 6 patients in whom hot nodes were not identified had moderate to severe anthracotic lymph node silicosis and pulmonary emphysema due to heavy smoking. Nomori and colleagues found that patients without lymph node radiolabeling had complications associated with obstructive respiratory dysfunction [13]. Either attenuated density of lymphatics at the injection site or lymphatic dysfunction may have existed in these patients and ours due to degenerative pulmonary diseases such as emphysema. We clarified the accuracy of the intrathoracic detection 429 of hot nodes to determine whether this practice is appropriate for so-called sentinel node biopsy. As a result, we identified all the hot nodes without overlooking any, although overestimation frequently occurred. In fact, 44% of nonhot nodes were overestimated to be hot because the hot primary tumor and airway radioactivity complicated the search for true hot nodes. Although the hot primary tumor and airway radioactivity seem to be a common and specific pitfall in sentinel node assessment of the respiratory tract, few reports address this issue. Nomori and associates studied the correlation between the intrathoracic and extrathoracic detection of hot nodes [14]. They overlooked only one hot node by intrathoracic detection in their series, but found greater discrepancy between the intrathoracic and extrathoracic detection of hot nodes when counting N1 stations rather than N2 stations, probably because the hot primary tumor and airway radioactivity resulted in false intrathoracic counting. Although these results would appear to be similar to ours, they are not comparable because they did not show the accuracy of intrathoracic detection in terms of the overall frequency of overestimated lymph node stations. This seems to be important for assessing whether this practice is useful for sentinel node biopsy because in the present series, all the overestimated nodal stations had to be dissected before the true hot node was able to be identified. The high incidence of airway radioactivity in our study may not be directly related to the size of the tracer. Regardless of molecular size, some parts of the intraalveolarly administered tracer are assumed to inherently migrate into the tracheobronchus due to coughing or respiratory lung motion even when it is properly injected into the peritumoral area under CT guidance. However, one might speculate that even the airway-migrating tracer could contribute to incorrect lymph node radiolabeling or that airway migration may leave insufficient tracer in the injected site, resulting in inaccurate lymph node radiolabeling. To dispel these concerns, we analyzed the linear dependence of the tracheal radioactivity and the highest lymph node radioactivity by linear regression analysis, and found no significant correlation between these two values (n 15, Y to 0.243X, r 0.004, p 0.826; Fig 3). Furthermore we believe that the tracer migrated correctly to the hot node through the lymphatic route because none of the background tissues surrounding the hot nodes showed significant radioactivity. Concerning skip metastasis, previous lymph node mapping studies found N2-sentinel nodes in 20% to 35% of patients [8, 14]. In our series, an N2-hot node was identified in 3 of 9 patients (30%) with any identified hot nodes. These 3 patients also had identified N1-hot nodes, similar to those reported by Nomori and associates who identified sentinel lymph nodes using tin-colloid [14]. A total of two or three hot nodes were identified in each of these 3 patients, and we believe that all of these nodes drained directly from a primary tumor. The results of the Nomori s study supported the notion that a tumor can drain directly into a N2 lymph node and be related to two GENERAL THORACIC

5 430 UEDA ET AL Ann Thorac Surg RADIOISOTOPE LYMPH NODE MAPPING 2004;77: Fig 3. Relationship between the tracheal radioactivity and the highest lymph node radioactivity. There was no significant correlation between the radioactivity levels (n 15, r 0.004, p 0.826). (Dotted line hot node; straight line Y X.) or more lymphatic vessels. Because plural sentinel lymph nodes can be located in various stations, lymph node status may not necessarily be determined by location (N1, N2, or N3), or even by the number of metastatic lymph nodes. Thus, only lymph node mapping would be able to determine accurate lymph node status. In conclusion, although radioisotope intrathoracic lymph node mapping was found to be safe, it appears to be unsuitable for sentinel node biopsy because shinethrough and the airway-migrating radioactive tracer complicated the intrathoracic counting. Only serial scintigraphy can truly distinguish hot nodes from airway migration. References 1. Mariani G, Moresco L, Viale G, et al. Radioguided sentinel lymph node biopsy in breast cancer surgery. J Nucl Med 2001;42: Lenisa L, Santinami M, Belli F, et al. Sentinel node biopsy and selective lymph node dissection in cutaneous melanoma patients. J Exp Clin Cancer Res 1999;18: Joosten JJ, Strobbe LJ, Wauters CA, Pruszczynski M, Wobbes T, Ruers TJ. Intraoperative lymphatic mapping and the sentinel node concept in colorectal carcinoma. Br J Surg 1999;86: Sherif A, De La Torre M, Malmstrom PU, Thorn M. Lymphatic mapping and detection of sentinel nodes in patients with bladder cancer. J Urol 2001;166: Sliutz G, Reinthaller A, Lantzsch T, et al. Lymphatic mapping of sentinel nodes in early vulvar cancer. Gynecol Oncol 2002;84: Little AG, DeHoyos A, Kirgan DM, Arcomano TR, Murray KD. Intraoperative lymphatic mapping for non-small cell lung cancer: the sentinel node technique. J Thorac Cardiovasc Surg 1999;117: Hansen NM. Blue versus hot: learning the techniques with dye and isotopes. Ann Surg Oncol 2001;8:64S 66S. 8. Liptay MJ, Masters GA, Winchester DJ, et al. Intraoperative radioisotope sentinel lymph node mapping in non-small cell lung cancer. Ann Thorac Surg 2000;70: Melfi FM, Chella A, Menconi GF, et al. Intraoperative radioguided sentinel lymph node biopsy in non-small cell lung cancer. Eur J Cardiothorac Surg 2003;23: Schmidt FE, Woltering EA, Webb WR, Garcia OM, Cohen JE, Rozans MH. Sentinel nodal assessment in patients with carcinoma of the lung. Ann Thorac Surg 2002;74: Boxen I, McCready D, Ballinger JR. Sentinel node detection and definition may depend on the imaging agent and timing. Clin Nucl Med 1999;24: Yeung HW, Cody HS III, Turlakow A, et al. Lymphoscintigraphy and sentinel node localization in breast cancer patients: a comparison between 1-day and 2-day protocols. J Nucl Med 2001;42: Mountain CF, Dresler CM. Regional lymph node classification for lung cancer staging. Chest 1997;111: Nomori H, Horio H, Naruke T, Orikasa H, Yamazaki K, Suemasu K. Use of technetium-99m tin colloid for sentinel lymph node identification in non-small cell lung cancer. J Thorac Cardiovasc Surg 2002;124: Sugi K, Fukuda M, Nakamura H, Kaneda Y. Comparison of three tracers for detecting sentinel lymph nodes in patients with clinical N0 lung cancer. Lung Cancer 2003;39: Sato K, Uematsu M, Saito T, et al. Sentinel lymph node identification for patients with breast cancer using large-size radiotracer particles: technetium-99m-labeled tin colloids produced excellent results. Breast J 2001;7: Kitagawa Y, Fujii H, Mukai M, Kubota T, Otani Y, Kitajima M. Radio-guided sentinel node detection for gastric cancer. Br J Surg 2002;89:604 8.