Contrast-Enhanced Ultrasound-Guided Sentinel Lymph Node Biopsy of the Ocular Conjunctiva

Transcription

1 The Laryngoscope VC 2014 The American Laryngological, Rhinological and Otological Society, Inc. Contrast-Enhanced Ultrasound-Guided Sentinel Lymph Node Biopsy of the Ocular Conjunctiva Mindy R. Rabinowitz, MD; Daniel A. Merton, BS, RDMS; Ji-Bin Liu, MD; Shivam Saxena, BS; John Pluta, BS; John R. Eisenbrey, PhD; Adam L. Baker, MD; Michael P. Rabinowitz, MD; Sara Lally, MD; David Cognetti, MD; Barry B. Goldberg, MD; Edmund A. Pribitkin, MD; Joseph M. Curry, MD Objectives/Hypothesis: Sentinel lymph node biopsy (SLNB) has been utilized for cutaneous melanoma and other malignancies arising from the eye and ocular adnexa. Currently, SLNB requires blue dyes and/or radiopharmaceuticals; both of which have significant shortcomings. This study sought to evaluate the feasibility of SLNB with the use of real-time, contrastenhanced ultrasound (CEUS) as an alternative technique for tumors arising in the conjunctiva. Study Design: Prospective feasibility study in a porcine model. Methods: Twelve experiments were performed on six non tumor-bearing Yorkshire swine. An ultrasound contrast agent, Sonazoid (GE Healthcare, Oslo, Norway), 99m technetium ( 99m Tc), and methylene blue (MB) (Covidien, Mansfield, MA) were injected in the ocular conjunctiva. Sentinel lymph nodes (SLNs) were localized with CEUS and findings were compared to that of MB and 99m Tc. Fisher exact test was used. Results: Contrast-enhanced SLNs were identified within an average of 6.2 minutes from time of injection of Sonazoid. A total of 17 SLNs were identified by at least one of the three techniques. Correlation between Sonazoid and 99m Tc was 94.1% (16/17 SLNs). Correlation between 99m Tc and MB was 88.2% (15/17). One SLN that was positive for 99m Tc but negative for Sonazoid and was considered to be a false positive (1/17); findings were similar for MB (1/17). Differences between the three techniques were not significant (P 5.886). Conclusions: CEUS-guided injection of conjunctiva for SLNB is technically feasible and correlates well with standard detection techniques. This technique shows promise for rapid, real-time, intraoperative imaging for SLNB, using a widely available imaging modality and avoiding the need for radiopharmaceuticals. Key Words: Sentinel lymph node biopsy, Sonazoid, conjunctival melanoma, ultrasound contrast. Level of Evidence: NA Laryngoscope, 124: , 2014 INTRODUCTION Conjunctival melanoma is a rare but life- and sightthreatening condition that accounts for 2% to 5% of ocular malignancies. 1 Classically, lymph node metastasis has been reported as the single worst prognostic factor for these patients, decreasing survival by nearly 50%. 2,3 Sentinel lymph node biopsy (SLNB) has been a widely used technique in cutaneous melanoma for detection of From the Department of Otolaryngology-Head and Neck Surgery (M.R.R., A.L.B., D.C., E.A.P., J.M.C.), Department of Radiology (D.A.M., J.-B.L., J.R.E., B.B.G.), and Thomas Jefferson University School of Medicine (S.S.), Thomas Jefferson University, Philadelphia, Pennsylvania; the Department of Radiology (J.P.), University of Pennsylvania, Philadelphia, Pennsylvania; and the Department of Ophthalmology (M.P.R., S.L.), Wills Eye Institute, Thomas Jefferson University Hospital, Philadelphia, Pennsylvania, U.S.A. Editor s Note: This Manuscript was accepted for publication May 27, Edmund Pribitkin, MD, is a consultant for the Stryker Corporation. The authors have no other funding, financial relationships, or conflicts of interest to disclose. Send correspondence to Mindy Rabinowitz, MD, 925 Chestnut St., 6th Floor, Philadelphia, PA mfigures@gmail.com DOI: /lary subclinical micrometastasis in patients with no known metastatic disease. Conjunctival melanomas may also benefit from nodal assessment, although few studies have explored SLN in this subsite. 1 Traditionally, prophylactic treatment includes lymph node dissection or radiation to the head and neck for malignancies that pose a >15% to 20% risk of regional metastasis. 2 The frequency of regional lymph node metastases in ocular melanoma is reported to be between 26% and 40%, and although not all metastatic disease is preceded by regional disease, postponement of lymph node dissection until patients develop evidence of regional involvement may allow spread to multiple lymph nodes and distant sites resulting in a substantial drop in long-term survival. 1,3 For this reason, SLNB may be a valuable technique in ocular melanoma patients without clinically evident nodal disease. Current indications for SLNB in conjunctival melanomas include disease 2 mm in histologic thickness and/or ulceration. 4 Traditional SLNB has been performed with the use of vital blue dyes (such as methylene blue or isosulfan), which provide visual identification of the sentinel lymph nodes (SLNs) intraoperatively, and the use of radiopharmaceuticals (such as technetium), which can be used to 2531

2 and in the oral cavity, this UCA s ability to detect SLNs when injected into the ocular conjunctiva is unknown. 2,9,10 Previous studies have shown that Sonazoid is trapped in first-echelon lymph nodes (i.e., SLNs) and is safe in humans for intravenous injection. 2,6,9 Once injected, it is phagocytized by normal lymph node macrophages. Reticuloendothelial specific uptake of Sonazoid results in prolonged enhancement of SLNs (up to 4 hours), even after the lymph nodes are surgically excised. 11 We previously described the use of CEUS to demonstrate feasibility of SLNB in other sites in the head and neck, with the purpose of application to head and neck squamous cell carcinoma. 2 Additionally, Goldberg et al. have demonstrated utility and accuracy of this technique in a porcine melanoma model. 5 The goal of this study was to evaluate the technical feasibility of CEUS SLNB in an animal model for conjunctival melanoma as compared to traditional blue dye and lymphoscintigraphy. Fig. 1. A 0.1-mL injection of Sonazoid is introduced subconjunctivally in four sites (for a total of 0.4 ml per eye), around the eye in a clockwise direction. [Color figure can be viewed in the online issue, which is available at MATERIALS AND METHODS Experimental Design The objective of this study was to evaluate the feasibility of using CEUS with conjunctival injections of Sonazoid to localize SLNs and compare its reliability to that of conventional SLNB techniques. End points included correlation between Technetium-99, blue dye and Sonazoid, as well as time to identification of SLNs with CEUS. detect the SLNs pre- and/or intraoperatively with a handheld gamma probe. 5 However, although these techniques have proven to be effective in the identification of SLNs, they both have limitations. Most notably, blue dyes will often stain not only the conjunctiva but also second echelon nodes, as the dyes small size allows them to pass through the first echelon lymph nodes into multiple nodal levels, which may lead to more extensive dissection. 5 7 Radiocolloids small particles can similarly pass through the first echelon lymph nodes. 5 Furthermore, lymphoscintigraphy (use of radiopharmaceuticals) may exhibit the shine through effect that occurs when a signal from the primary injection site interferes with a signal from a closely approximated SLN. 2,6 Therefore, timing can be critical, requiring injection several hours before surgery with lymphoscintigraphy and surgery within the appropriate time window for the specific radiocolloid; this time requirement can place constraints on the patient, surgeon, and operating room schedule. Finally, staining of the ocular tissues secondary to blue dye injection and cataract formation or radiation retinopathy from conjunctival injection of radiolabeled sulfur colloid may theoretically occur. 3,6 Sonazoid (GE Healthcare, Oslo, Norway), is a lipidstabilized perfluorobutane microbubble-based ultrasound contrast agent (UCA). 8 This UCA has been injected around naturally occurring melanoma tumors in swine and can be detected using contrast-enhanced ultrasound (CEUS) in both the lymphatic channels (LCs) as well as the SLNs. 2 Although useful in breast and liver tumors 2532 Animals, Anesthesia, and Monitoring The animal study protocol was approved by the Animal Care and Use Committee of our institution and was performed under the supervision of laboratory animal personnel. Six Yorkshire swine, weighing ( kg, mean 25 kg), were used for the experiment in a humane and ethical fashion. The animals were fasted overnight and underwent general anesthesia prior to any examination or procedures. Sedation was carried out with 2 to 4 mg/kg tiletamine zolazepam (Telazol) via intramuscular injection, followed by endotracheal intubation. Anesthesia was maintained with isoflurane in a concentration of 3% to 4%. A single dose of buprenorphine (0.01 mg/kg, intramuscular) was provided for analgesia. Pulse oximetry was employed in addition to CO 2 monitoring/o 2 monitoring with a DATEX machine (Datex-Ohmeda, Inc., Madison, WI) intraoperatively. Intravenous saline was administered at approximately 10 to 15 ml/kg per hour throughout anesthesia. The swine were placed on a warming blanket for the duration of the experiments. Ultrasound and Experimental Procedure Sonazoid was reconstituted with 2 ml of sterile water (estimated 10 9 microbubbles per ml). In the first pig, 0.25 ml of Sonazoid was injected at four sites into the conjunctiva and subconjunctival space of one eye, for a total of 1 ml injection per eye. However, after injection, it was realized that this volume was likely much larger than needed for the conjunctiva. Therefore, in the remaining five pigs, the dose of Sonazoid was decreased to 0.1 ml for each site, with 0.4 ml in total per eye (Fig. 1). This volume appeared to be more appropriate for the conjunctiva with less associated chemosis. After approximately 1 minute of gentle massage at the injection site, CEUS was performed using a broad bandwidth (8 15 MHz) linear array

3 Fig. 2. An approximately 3- to 4-cm incision was made about 1 to 2 fingerbreaths below the mandibular angle. After the neck was opened, the gamma probe and visualization of the blue dye were used to identify the sentinel lymph node. [Color figure can be viewed in the online issue, which is available at transducer and a Sequoia 512 US system (Siemens Medical Solutions, Mountain View, CA) in an attempt to identify contrastenhanced LCs and SLNs. Cadence pulse sequencing contrast imaging mode was used at low acoustic power to minimize contrast agent microbubble destruction. Following the CEUS imaging evaluation 10 to 15 minutes before the animal was to be sacrificed, 0.1 ml (1 mg/ml) blue dye (1% Lymphazurin; Ben Venue Labs, Inc., Bedford, OH) (except for first pig using 1.0 ml) and 0.1 ml (1.85 MBq) of filtered technetium 99 sulfur colloid ( 99m Tc) was injected into the same four sites as was Sonazoid (0.4 ml in total), and the areas were massaged as previously described by Goldberg et al. 5 Given time constraints of veterinary anesthesia, the swine were then sacrificed with an overdose of a euthanasia agent containing the barbiturate sodium pentobarbital (Buthanasia) given intravenously at a minimum of 10 to 15 ml per 100 pounds. Surgical dissection was then performed to remove the SLNs. An approximately 3- to 4-cm incision was made about 1 to 2 fingerbreaths below the mandibular angle. After the neck was opened, the gamma probe and visualization of the blue dye were used to identify the SLN. Only after the lymph node was removed was CEUS used to correlate with the gold standard of blue dye and technetium (Fig. 2). During dissection, the blue dye was visually identified in the LCs and followed to the SLNs. A handheld gamma probe (Neoprobe Corp., Dublin, OH) was used to survey the nodal basins in vivo for 99m Tc radioactivity. A baseline radioactivity count was performed at the injection site for comparison. Comparison of the blue dye and 99m Tc with CEUS findings were made during dissection. Once the SLNs were exposed, CEUS was performed again to determine the presence or absence of Sonazoid in the lymph nodes. The SLNs were then removed and ex vivo tested for radioactivity and contrast enhancement on each node. The neck was then examined visually using the gamma probe for residual radioactivity, as well as with CEUS, in an attempt to identify any Fig. 3. Ultrasound identification of Sonazoid. (a d) The intensity of the ultrasound contrast agent increases over time, correlating to increased Sonazoid sequestration within the cluster. (d) The Sonazoid contrast agent is seen in a few sentinel lymph nodes (cross-like symbols) within the sentinel lymph node cluster. [Color figure can be viewed in the online issue, which is available at

4 Fig. 4. Ex vivo sentinel lymph node (SLN) cluster. Methylene blue dye is easily visible within a few SLNs (*) within the cluster of nodes. [Color figure can be viewed in the online issue, which is available at remaining contrast-enhanced SLNs. Results were considered negative when counts were equivalent to background and when no contrast-enhanced blue dye nodes were identified. SLNs were identified within an average of 6.2 minutes from time of last injection (range, 2 13 minutes). The locations of the LCs and SLNs detected sonographically, including their distance beneath the skin and relationships to other anatomical landmarks, were noted to permit correlation with the results of surgical dissection. Average lymph node (LN) size detected sonographically was 8.4 mm (range, 5 11 mm) in maximal diameter. When a cluster of LNs were detected, the contrast agent could easily be distinguished as residing within one or two single LNs within the cluster (Fig. 3), which correlated to the blue-stained LN within a cluster seen at time of dissection (Fig. 4). After those nodes were dissected from the cluster, the presence of Sonazoid and radioactivity was confirmed in all cases (Table I). As visualized with CEUS, after contrast uptake, the SLNs became progressively more echogenic (Fig. 3), and the LCs could be seen as hyperechoic linear structures leading from the injection site to the SLN, providing a road map that could be followed noninvasively from the injection site to the SLNs (Fig. 5). Statistical Analysis A contingency table was created with each treatment type as the columns and the number correct/incorrect identifications as the rows. Fisher exact test was used for statistical comparison of techniques using 99m Tc as gold standard and presumed true positive. RESULTS Ultrasound Identification of Sonazoid at the injection site was confirmed using CEUS in all cases. Contrast-enhanced Sentinel Lymph Node Biopsy Under ultrasound, 99m Tc and blue-dye guided dissection, SLNs were found in the parotid, periparotid, and upper neck region in all pigs (Table I). An average of 1 to 2 LNs were found on each side. The presence of 99m Tc was confirmed with Neoprobe (Mammotome, Cincinnati, OH) counts greater than the surrounding background. The presence or absence of blue dye and Sonazoid was evaluated for each dissected LN by direct visualization or CEUS, respectively. Completion neck dissection was performed in all pigs. No LNs with methylene blue, 99m Tc, and/or Sonazoid were identified TABLE I. Sentinel Lymph Node Biopsy. Pig No. Side LN No. LN Location Geiger Sonazoid Blue Dye 1 Right 1 Parotid Left 1 Superficial Deep Right 1 Parotid Upper neck Left 1 Parotid Right 1 Parotid Left 1 Parotid Right 1 Superficial parotid Left 1 Upper periparotid cluster Slightly anterior to LN Right 1 Deep periparotid Left 1 Superficial periparotid Deep periparotid Right 1 Superficial Periparotid, superior neck, deep Left 1 Periparotid, superior neck LN 5 lymph node. 2534

5 Fig. 5. Contrast-enhanced lymphatic channel. Contrast agent is seen at the right of the image collecting in the subconjunctiva after injection. Toward the left of the image, the contrast agent can be seen traveling from the injection site through a lymphatic channel. [Color figure can be viewed in the online issue, which is available at after completion neck dissection in any case. The average greatest dimension of the SLN size was 16.4 mm. Statistical Analysis The Fisher exact test was not significant (P 5.886, one-tailed), implying that there is no statistically significant difference between the detection methods. Correlation between Sonazoid and 99m Tc was 94.1% (16/17 SLNs). Correlation between 99m Tc and methylene blue was 88.2% (15/17) (Table II). DISCUSSION This study suggests the technical feasibility of conjunctival injections for CEUS for SLNB and high reliability and correlation between standard techniques and CEUS. SLNB has been most widely studied in cutaneous melanoma, where for medium thickness disease (Breslow thickness of mm) in all skin locations it exhibits a positivity of about 15%. 3 The rare nature of conjunctival melanoma has restricted robust study in the utility of SLNB as compared to its cutaneous counterpart. However, given a relative rate of regional lymph node metastasis of 26% to 40% for conjunctival melanomas, many affected patients may benefit from early detection of involved SLNs. The current gold standard in SLN localization includes the use of radiopharmaceuticals that are detected pre- or intraoperatively with handheld gamma probes and/or vital blue dyes that provide visual identification of the draining LCs and SLNs. 3 In general, these techniques are accurate, but each has limitations. For example, due to their small size, blue dyes and radiopharmaceuticals will often lead to uptake in not only SLNs but also adjacent and second echelon LNs, and this can lead to more extensive dissection than necessary. Additionally, the use of radiopharmaceuticals with a nonsectional imaging tool results in a shine-through effect from the primary injection site, which may obscure the ability to detect SLNs residing nearby. These traditional techniques may also harm the patient. There exists a small, theoretical risk of permanent blue staining of the conjunctiva after injection of vital blue dyes. 7 Additionally, radioactive tracer exposure can cause radioactive damage to the lens and retina leading to cataracts and radiation retinopathy. 3,7 Furthermore, additional time is required to administer injections and perform imaging (lymphoscintigraphy) prior to surgery. Many surgeons are becoming more facile with the use of ultrasound technology due to its widespread availability and growing applications in medicine and surgery. CEUS-guided SNLB can potentially decrease scheduling constraints on the patient and surgeon, as it can be both injected and immediately evaluated intraoperatively in real time by the operating physician or other appropriately skilled professional. Additionally, Sonazoid injected intraoperatively, while the patient is anesthetized, may reduce patient anxiety as well as the discomfort associated with conscious ocular injections. Furthermore, intraoperative injection reduces time that may be lost between preoperative injection and travel to the operating room. In our study, only one LN was found with blue dye that had no radioactivity or contrast enhancement. Presumably this LN represented a second echelon LN. In another swine, one LN was found with Sonazoid and radioactivity, but it contained no blue dye, suggesting that it may have been a SLN that was missed by the blue dye. Finally, one LN was found with radioactivity and blue dye but no Sonazoid. This may represent a second echelon LN, because in this experiment an additional SLN containing all three was identified. 99m Tc and blue dye, due to their small particle size, may pass onto the second echelon, and Sonazoid, with its larger size and propensity to remain in the first echelon node, may not have travelled into this node. 5 Alternatively, this LN may have been missed secondary to technical factors surrounding injection of contrast or 99m Tc; for the purposes of this experiment, given the absence of true pathology (nodes containing cancer) to serve as a control, it was treated as a false-negative result. During the completion neck dissection, no other LNs were detected carrying any of the three agents. This resulted in a presumed single false negative (1/17), which was encountered in the 99m Tc/blue dye (1/17) comparison as well. CEUS has many potential benefits. Given the large particle size of the agent (a median diameter between 2.4 and 3.5 mm) and the agent s ability to be trapped in the SLN, CEUS-guided SLNB can eliminate secondechelon nodal uptake and limit the possibility of unnecessary LN dissection. Unlike single photon TABLE II. Correlation Data. Correlation % Sonazoid vs 99m Tc 16/ Sonazoid vs blue 14/ Sonazoid vs 99m Tc or blue 16/ m Tc vs blue 15/

6 emission computed tomography (SPECT), which also provides excellent anatomic information, this technique does not require a radiotracer, and thereby avoids the attendant exposure risks to patients and personnel as well as the cost associated with handling, storage, and disposal of radioactive materials. Additionally, unlike SPECT, there are no renal contraindications to Sonazoid, and there is no wait time for scanner availability as for computed tomography. The mean time to identification of the SLN after injection was 6.2 minutes, allowing for the technique s use in real time. If applied to human use, CEUS-guided SLNB should eliminate the need for lymphoscintigraphy s preoperative injection and imaging. Moreover, the agent can be traced from the site of injection to the SLN, improving the surgeon s anatomic localization of the SLN. Unlike blue dye, CEUS provides threedimensional localization and allows the surgeon to identify surrounding anatomical structures. As previously shown in a tumor-bearing swine model, the CEUS technique is capable of localizing microfoci of tumor within the SLN. 5 As the contrast enters the node, a hypoechoic region is seen where the tumor is located. This ability of CEUS to identify a macrofocus of cancer may allow for CEUS-guided fine-needle aspiration to preoperatively stage the neck and eliminate unnecessary surgery or allow the patient to be prepared for full nodal dissection. Our results, although limited in the number of swine studied, suggest that Sonazoid-enhanced CEUS is at least as effective as blue dye and 99m Tc for SLNB from the ocular conjunctiva. Thus, Sonazoid may offer an alternative to blue dye and 99m Tc for SLNB in the future. Real-time Sonazoid-enhanced CEUS can allow for the detection and mapping of SLNs from their LCs preoperatively, which can be confirmed intraoperatively. Although this study shows promise for the use of Sonazoid for SLNB for ocular conjunctival melanoma, additional investigations are required. The appropriate volume of contrast agent to be injected into the primary lesion must be determined. This would likely be dependent upon the primary site, as some locations such as the conjunctive would be less tolerant of higher volumes of injection. We were able to empirically decrease injection volumes from 1 ml used in the first swine to to 0.25 ml in the remaining animals, secondary to what appeared to be volume overload in the first animal, and this change did not appear to affect the ability to detect SLNs. Additionally, given that euthanasia was performed prior to the end of the procedure, this may have prevented further spread of smaller Sonazoid particles into second- and third-echelon LNs. However, previous studies have shown that Sonazoid is trapped in the first echelon nodes. 2,6,9 A tumor-bearing model would be helpful to validate the technique compared to 99m Tc directly, with histopathology on the SLN and lymph node packet used to represent the absolute standard of accuracy. Finally, introduction of human trials would represent a key step in applying this technique to the conjunctive and other sites to determine its technical feasibility and accuracy. CONCLUSION CEUS SLNB after Sonazoid injection into the conjunctiva is technically feasible and correlates well with blue-dye and 9m Tc-guided SLNB. This technique shows promise for simplifying the technique of SLNB, improving anatomic localization, and avoiding the need for radiopharmaceutics. However, further investigation is required. BIBLIOGRAPHY 1. Lim LA, Madigan MC, Conway RM. Conjunctival melanoma: a review of conceptual and treatment advances. Clin Ophthalmol 2013;6: Curry JM, Bloedon E, Malloy KM, et al. Ultrasound-guided contrastenhanced sentinel node biopsy of the head and neck in a porcine model. Otolaryngol Head Neck Surg 2007;137: Esmaeli B. Sentinel node biopsy as a tool for accurate staging of eyelid and conjunctival malignancies. Curr Opin Ophthalmol 2002;13: Pfeiffer ML, Savar A, Esmaeli B. Sentinel lymph node biopsy for eyelid and conjunctival tumors: what have we learned in the past decade? Ophthal Plast Reconstr Surg 2013;29: Goldberg BB, Merton DA, Liu JB, et al. Sentinel lymph nodes in a swine model with melanoma: contrast-enhanced lymphatic US. Radiology 2004;230: Kawai, Y, Ajima K, Nagai T, et al. Real-time imaging of the lymphatic channels and sentinel lymph nodes of the stomach using contrastenhanced ultrasonography with Sonazoid in a porcine model. Cancer Sci 2011;102: Esmaeli B, Reifler D, Prieto VG, et al. Conjunctival melanoma with a positive sentinel lymph node. Arch Ophthalmol 2003;121: Sontum PC. Physicochemical characteristics of Sonazoid, a new contrast agent for ultrasound imaging. Ultrasound Med Biol 2008;34: Nanashima A, Tobinaga S, Abo T, et al. Usefulness of Sonazoidultrasonography during hepatectomy in patients with liver tumors: a preliminary study. J Surg Oncol 2011;103: Omoto K, Matsunaga H, Take N, et al. Sentinel node detection method using contrast-enhanced ultrasonography with Sonazoid in breast cancer: preliminary clinical study. J Ultrasound Med Biol 2009;35:8: Goldberg BB, Merton DA, Liu JB, et al. Contrast-enhanced ultrasound imaging of sentinel lymph nodes after peritumoral administration of Sonazoid in a melanoma tumor animal model. J Ultrasound Med 2011; 30: