Original article. Lymphatic mapping to tailor selective lymphadenectomy in cn0 tongue carcinoma: beyond the sentinel node concept

Transcription

1 Original article Lymphatic mapping to tailor selective lymphadenectomy in cn0 tongue carcinoma: beyond the sentinel node concept C. De Cicco 1, G. Trifirò 1, L. Calabrese 2, R. Bruschini 2, M. E. Ferrari 3, L. L. Travaini 1, M. Fiorenza 1, G. Viale 4, F. Chiesa 2, G. Paganelli 1 1 Nuclear Medicine, European Institute of Oncology, Via Ripamonti 435, Milan, Italy 2 Head and Neck Surgery Division, European Institute of Oncology, Via Ripamonti 435, Milan, Italy 3 Medical Physics Division, European Institute of Oncology, Milan, Italy 4 Pathology Division, European Institute of Oncology and University of Milan, Milan, Italy Received: 19 September 2005 / Accepted: 16 January 2006 / Published online: 8 April 2006 Springer-Verlag 2006 Abstract. Purpose: Cervical lymph node status is the most important pathological determinant of prognosis and decision making in head and neck squamous cell carcinoma (SCC). The aim of this study was to demonstrate that lymphoscintigraphy (LS) can supply a complete map of the lymphatic drainage before surgery, allowing planning of the type of intervention and serving to guide lymphadenectomy. Methods: The study population comprised 14 patients with T2 4 SCCs of the tongue and clinically negative lymph nodes in the neck (cn0) who were scheduled to undergo tumour resection and selective level I IV neck dissection extended to level V. LS was performed in all patients following the injection of 99m Tc-colloidal sulphide in three aliquots around the primary lesion. Dynamic, static and tomographic images of the head and neck were acquired. The operative specimens were subjected to lymphoscintigraphic evaluation. Preoperative and postoperative imaging results were compared with the pathological findings. All nodes were examined using haematoxylin-eosin staining. Results: Preoperative LS was successful in all patients. Preferential pathways of lymphatic drainage were identified: level II of the neck was the most common lymphatic drainage pattern, followed by levels IV and III. Contralateral drainage occurred in 11 patients and in two of them metastatic nodes were found on the contralateral side. Metastases were observed only in radioactive lymph nodes. Conclusion: LS is able to supply a complete map of the lymphatic drainage before surgery, making it possible to tailor selective neck dissection to each individual patient based on the results of preoperative mapping, thereby C. De Cicco ()) Nuclear Medicine, European Institute of Oncology, Via Ripamonti 435, Milan, Italy direzione.mnu@ieo.it Tel.: , Fax: sparing healthy lymphatic tissue and reducing surgeryrelated morbidity. Keywords: Tongue carcinoma Radiocolloid Lymphoscintigraphy Lymphatic mapping Neck dissection Eur J Nucl Med Mol Imaging (2006) 33: DOI /s Introduction The tongue is the commonest site of origin of primary squamous cell carcinoma (SCC) in the oral cavity in Western countries. Regional nodal metastasis is the most important prognostic predictor for local recurrence and distant metastases and consequently for survival. The incidence of occult cervical metastases in patients with SCC of the tongue and clinically negative lymph nodes in the neck (cn0) has been found to be at least 21% [1]. The presence of metastases undetected by non-invasive methods [ultrasound, computed tomography (CT), magnetic resonance imaging (MRI)] significantly downgrades the patient s prognosis. 18 F-fluorodeoxyglucose positron emission tomography (FDG PET) is a promising imaging tool in this field, but its sensitivity is still insufficiently high for it to replace surgical lymph node staging [2]. Surgical management of the neck in these patients is controversial; some authors suggest a watchful waiting policy [3] while others prefer to dissect or irradiate all N0 necks [4, 5]. Sentinel node biopsy (SNB) can reliably predict the status of regional lymph nodes in several solid cancers such as melanoma and breast cancer [6 8]. The procedure seems useful in the head and neck region, too [9 14]. However, there are still controversial aspects of the method and a question mark over its applicability in SCC when a combined approach is required, owing to the great

2 complexity of the lymphatic pathways of the oral cavity and the tongue in particular. Moreover, as in melanoma patients, neck dissection is often an essential part of potentially curative treatment of oral cancers. The aim of this study was to demonstrate that LS can supply a complete map of the lymphatic drainage before surgery and consequently allow planning of the type of intervention and guidance of lymphadenectomy through selection of the latero-cervical levels to be dissected. If LS has this ability, it will be possible to tailor selective neck dissection to each individual patient on the basis of the results of preoperative mapping. Materials and methods Patients From February 2002 to November 2004, 14 patients (12 males, 2 females; mean age 51.8 years, range 28 74) were enrolled in the study. Patients were selected from among those who had ct2 T4 SCC of the mobile tongue margin, with suspected invasion of the extrinsic muscle of the tongue, and were scheduled to undergo tumour resection and bilateral neck dissection. Eligibility criteria were: age years; histological diagnosis of SCC of the tongue; absence of clinical and/or imaging evidence of cervical metastases after CT and/or MRI; and no evidence of distant metastases as assessed by clinical examination, chest X-rays, abdominal ultrasound and bone scan. Pregnant or lactating women and those patients with a history of surgery on or radiotherapy to the oral cavity or neck were excluded from the study. Informed consent was obtained from each patient. The mean follow-up was 31.6 months (range months). position of the lymph node draining the injection area, SPECT-CT was performed after delayed static images employing a hybrid device (Millennium-Hawkeye, GE Medical System, Milwaukee, WI, USA) in which CT and SPECT systems are combined. SPECT images were acquired in a pixel matrix, obtaining 120 views (30 s per view). To improve SPECT quality, we applied scatter and attenuation correction. Scatter contribution was evaluated by acquiring statistics in the scatter energy window (120 kev ±10%); attenuation correction was carried out using the attenuation map from the CT images. The system allows simultaneous acquisition of anatomical and functional information. The Functional Anatomical Fusion software is used to generate and review slices created from the co-registered functional and anatomical imaging data. Planar images of the specimens were acquired immediately after surgery, about 20 h post injection, in order to mark each lymph node which had retained radiocolloids, ordering them according to level (Fig. 2). These images were compared with preoperative LS and with the pathological findings. Surgery 901 Within 20 h of radiocolloid injection, patients underwent surgery. All patients underwent a composite resection of the primary tumour, followed by bilateral selective level I IV neck dissection extended to the lymph nodes of level V, with preservation of the cervical branches of the cervical plexus. Lymphoscintigraphy On the day before surgery, colloidal sulphide particles with a size of <50 nm (Lymphoscint) from Amersham-Nycomed-Sorin (Saluggia, Italy) were labelled with freshly eluted 99m Tc. Radiocolloids were checked for free technetium according to the manufacturer s instructions in order to assure a radiochemical purity of >95%. Each patient received a maximum total activity of 40 MBq in three injections around the primary lesion, with an injected volume of 0.1 ml for each aliquot. After injection, patients were instructed to rinse their mouths thoroughly with water to remove any residue of radiocolloid. Scintigraphic imaging of the head and neck was performed using a single-head gamma camera (Starcam 4000 GE Medical System, Milwaukee, WI, USA) for dynamic and static images. A double-head gamma camera (Millennium GE Medical System, Milwaukee, WI, USA) was used to obtain tomographic images. Gamma cameras were equipped with a low-energy high-resolution collimator. Dynamic acquisition was started immediately after administration of radiopharmaceuticals and continued for 15 min in the anterior view (30 s/frame). Static images of the head and neck in the anterior and lateral views were acquired 30 min and 2 h after injection, collecting 100 kcounts; these images were acquired with the patient lying in the supine position, with the head extended. A 57 Co point source was used to place an ink marker on the cutaneous projection of the hot spots. Exact localisation of all hot spots was checked by a gamma probe (C-Trak System, Care-Wise, CA, USA). Anatomical localisation of the lymph nodes was reported according to the Robbins classification (Fig. 1). To correctly determine the anatomical Fig. 1. Anatomical localisation of the latero-cervical lymph node levels according to the Robbins classification. m mandible, c clavicle, i hyoid bone, 1 angle of the mandible, 2 venter posterior musculi digastrici, 3 hyoglossus muscle, 4 mylohyoid muscle, 5 venter anterior musculi digastrici, 6 sternocleidomastoid muscle, 7 venter superior musculi omohyoidei, 8 sternohyoid muscle, 9 trapezium, 10 venter inferior musculi omohyoidei

3 902 Pathological examination All of the nodes were formalin fixed and paraffin embedded. Isolated hot nodes greater than 0.5 cm in the major axis were bisected and both halves were examined; nodes smaller than 0.5 cm were processed uncut. Sections were stained with haematoxylin and eosin. Results Fig. 2. Method for localisation in the specimen of the radioactive lymph nodes visualised on the delayed static scintigraphic images. The specimen, including the tongue (arrow head) and bilateral latero-cervical levels I V, is positioned on the gamma camera detector surface, facing upwards. Needles are placed on two hot spots, one on the left side and one on the right (white arrows). Use of a 57 Co pen (black arrow) allows identification of the labelled lymph nodes revealed by the gamma camera Patients and tumour characteristics are shown in Table 1. Preoperative LS was successful in all patients. Dynamic and early static images allowed the identification of 41 nodes (mean 2.9, range 1 6). Delayed static images localised an additional 30 nodes, so that in total 71 nodes were localised. The mean number of nodes detected per patient was 5.0 (range 1 8). SPECT/CT images gave the same results as the static images alone regarding both the number and the localisation of the lymph nodes. Bilateral drainage occurred in 11 out of 14 patients. Contralateral drainage was documented by early dynamic Table 1. Patient characteristics and scintigraphic and pathological findings Pt. no. Gender Age (yrs) Side pt LN Dyn Level LN Stat Level (n) LN Spec Level LN Excis Tot pn+r/l Level 1 M 58 R 4 1R IV 3R I-IV (2) 2R I-IV 29 1R IV 2L II-IV 2L II-IV 1L IV 2 M 68 R 2 0R 4R II (2)-III-IV 2R II-IV L II 2L II-III 1L II 3 M 52 R 2 1R II 1R II 1R II L III-IV 2L III-IV 3L III-IV(2) 4 M 60 L 3 1R II 1R II 1R II 42 1R II 1L II 1L II 2L II (2) 5 M 30 L 2 1R IV 3R II-IV (2) 2R II-IV L II-III 3L II-III-IV 1L IV 6 M 30 L 2 1R IV 1R IV 0R L II 5L II (2)-IV (2)-sl 5L II (2)- IV (2)-sl 7 F 28 L 4 1R II 1R II 2R II(2) L II(2) 3L II(2)-IV 4L II(2)-IV(2) 8 M 69 L 4 0R 2R II (2) 2R II (2) L II-III(2)-IV(2) 5L II-III (2)-IV (2) 4L II-III-IV (2) 9 M 74 L 3 0R 1R II 1R II 26 3L II (2)-III 1L II 4L II (2)-III-IV 3L II (2)-III 10 M 70 R 3 6R II (3)- IV (3) 6R II (3) - IV (3) 2R II-IV L II-IV 2L II-IV 0L 11 M 52 R 2 2R II (2) 6R II (2)-III (2)-IV (2) 2R II-IV L 0L 0L 12 M 47 L 2 0R 0R 0R L II 4L II-III-IV -sl 3L II-IV-sl 13 M 39 L 2 1R III 1R III 1R III 59 1R III 3L II-IV (2) 4L II-III-IV (2) 2L II-IV 14 F 49 R 2 3R I-II-IV 4R I-II-III-IV 2R II (2) 54 3R II-IV (2) 0L 0L 0L pt pathological T, LN Dyn number of lymph nodes detected by dynamic scintigraphy, LN Stat number of lymph nodes detected by static scintigraphy, Level laterocervical levels, (n) number of lymph nodes, LN Spec number of lymph nodes detected by lymphoscintigraphy of the specimen, LN Excis Tot number of lymph nodes removed, pn+ number of metastatic nodes, R right side, L left side, sl sublingual

4 images in 9 of 14 patients. LS of the specimen revealed 49 nodes (mean 3.5, range 1 6). The total number of nodes removed was 774 (mean 55.2, range ). The main lymphatic pathways were represented by levels II, IV and III, according to the frequency of visualisation. Pathological examination staged eight cases as T2, three as T3 and three as T4. Well-differentiated SCCs were found in two cases, G2 SCCs in five cases and poorly differentiated SCCs in seven cases. The total number of metastatic lymph nodes was nine in five patients; all of them were revealed on static and SPECT lymphoscintigraphy, while dynamic study failed to detect metastatic nodes in two patients. Contralateral lymph nodes were involved in two cases (patients no. 4 and 13). LS of the specimen performed after surgery usually detected fewer lymph nodes (a total of 49) than in vivo scan (Fig. 3). During the follow-up one patient died because of lung metastases; the remaining 13 patients were in good clinical condition, without evidence of disease. Dosimetry During SPECT-CT, the patient was exposed to an additional radiation dose owing to the X-ray device. The Hawkeye is a low-output system compared with conventional CT, delivering a very low dose to the patient. In fact, the CTDIw (computed tomography dose index weighted), representing the average dose in the scan plan, was 5 mgy for Hawkeye (140 kv, 2.5 ma, 13 s) and 50 mgy for a conventional CT (140 kv, 300 ma, 0.8 s). Fig. 3. a c Static lymphoscintigraphy acquired 2 h post injection in right lateral (a), left lateral (b) and anterior (c) views. Images show injection sites (arrows) and two lymph nodes at levels II and IV on each side of the neck. d Scan of the specimen acquired 26 h post injection revealed only three lymph nodes Discussion 903 The presence of cervical lymph node metastasis is the most important prognostic factor for regional recurrence and death in patients with tongue carcinoma. As for other neoplastic diseases like melanoma and breast cancer, the sentinel node (SN) procedure has been investigated in N0 head and neck carcinoma, with encouraging results [9, 13 15]. Particularly in tongue carcinoma the use of LS and a gamma probe seems promising for the identification of occult metastases within the cn0 neck [10 12, 14, 16]. The procedure helps in the decision making process for patients with tumours that require a transoral surgical approach, according to the recently endorsed SN guidelines [17]. Surgical management of these patients after SNB remains controversial, especially regarding the type of elective lymphadenectomy to be performed in cases of metastatic involvement of the SN. Selective neck dissection has become widely used, replacing modified radical neck dissection. However, assessment of the risk of metastatic involvement for groups of lymph nodes at each level is often the subject of debate [18]. Moreover, an obstacle to the more widespread application of the SN procedure in these tumours is its apparent inability to offer an adequate solution for in-transit metastases due to tumour emboli [19]. This prospective study was not designed with the intent of localising SNs and excising them for diagnostic purposes. Our aim was to obtain a complete lymphatic map of the tumour area by means of preoperative LS in a selected series of patients with T2 4 SCC of the tongue. The aim of this work was to verify the hypothesis that the removal of latero-cervical nodes based on a lymphatic scan could represent an exhaustive approach in the neck management of cn0 patients. In spite of the absence of clinically suspected neck lymph nodes, a selective (I IV) neck dissection extended to level V was planned in these patients owing to the primary tumour characteristics (lesion not more than 4 cm in its greatest dimension, but with suspected deep invasion of the extrinsic muscle). This selection explains the small number of subjects enrolled in the present study. Since complete lymphatic mapping was needed, unlike in our previous studies on SNB, sulphide colloids with a particle size of less than 50 nm were used, allowing a more complete visualisation of the lymphatic pathway. We decided to inject radiocolloid in three aliquots in order to standardise the procedure. The results demonstrated that LS was always successful in detecting lymphatic drainage, and in 11 out of 14 cases drainage was bilateral. These results are probably not correlated with tumour size, since 8 of the 14 primary tumours were pt2. Both the success of LS in detecting lymphatic drainage and the high rate of bilateral drainage are due to the small size of colloid used, which allows a wider diffusion through the interstitial tissues compared with Nanocoll, the generally used colloid for the localisation of SNs. Some preferential pathways have been identified in lymphatic spread: level II of the neck was the most common lymphatic drainage pattern, followed by levels IV and III. Our data seem to be

5 904 at odds with the traditional concept that cervical nodal metastases occur in an orderly progression from level I to subsequent levels [20, 21]. This concept led conventional surgical treatment to consist in a modified radical neck dissection (removal of levels I to V) or a selective procedure removing levels I to III or IV [18, 22, 23]. However, it has been reported that about 4 15% of patients who undergo a selective dissection have skip metastases at level IV [14, 20] (skip metastases being defined as nodes involved with cancer which are outside the pattern of orderly progression of the disease). Owing to this potential problem, many surgeons prefer to remove the level IV cervical nodes in all patients with SCC of the tongue. Several studies in head and neck carcinoma have demonstrated that LS frequently delineates unexpected nodal levels and that SNs are sometimes found in unpredictable sites [24, 25]. Moreover, there are a few lymph nodes close to tongue (or floor of the mouth) tumours, namely the sublingual lymph nodes, that are most likely to be involved by metastases [26, 27]. Because of their small size they escape clinical and imaging evaluation even when they harbour cancer; these nodes are difficult to localise by LS, too. In our small series a sublingual lymph node (found to be disease free at pathological examination) was detected by static LS (Fig. 4) in two patients. Due to the lateral projections, static LS showed better sensitivity than the dynamic study. In contrast with the opinion of other authors [28], we found that additional acquisition of co-registered SPECT/CT images has little impact on image interpretation. In our series the diagnostic gain in anatomical localisation of the lymph nodes and in particular of sublingual nodes was negligible as compared with the procedure of marker placement in real time. This was Fig. 4. Lymphoscintigraphy in patient no. 12. Static images in the left lateral (a) and anterior (b) projections show the injection sites (arrows), lymph nodes at left latero-cervical levels II, III and IV, and a sublingual node (open arrows). c Sagittal view of CT, SPECT and SPECT/CT fused images of the same patient confirmed the results of the planar study probably due to the poor resolution power of CT in the hybrid system (Hawkeye GE is a first-generation CT). Furthermore, during SPECT-CT the patient is exposed to additional radiation dose due to the X-ray device. Although the Hawkeye system is a low-output machine, this additional exposure does not offer diagnostic benefit for the patient and consequently is not justified from the radioprotection point of view. It is also to be noted that the procedure of attenuation and scatter correction of SPECT images does not help to separate the activity of the injection site from the hot spots related to sublingual lymph nodes. For this purpose we obtained better results by placing a lead plate on the injection site. In general, delayed static images allowed the localisation of a large number of lymph nodes compared with the dynamic study (71 versus 41), while all metastatic nodes were detected by static acquisitions. These data suggest the necessity of acquiring delayed static images, while a dynamic study and SPECT/CT acquisition seem not to be very useful. The decreased number of nodes found at specimen scan can be explained by the isotope decay: on average 20 h elapsed between the time of injection and the end of the operation, this interval being due to our internal organisation. Shortening of this time would be an easy solution to the problem; injection of a higher tracer dose is not suggested since such high activity might mask lymph nodes close to the injection site during LS. In the present study we omitted serial slicing and immunohistochemistry from lymph node analysis. In our opinion these procedures are suitable when analysing one or a few SNs; this was not the aim of the study, and consequently we examined all the removed lymph nodes using the same technique. However, in cases with equivocal findings for tumour presence, cytokeratin stain was performed. Although statistical evaluation is not possible with 14 cases, the results suggest good prediction of latero-cervical node levels involved by cancer. In fact, LS was able to indicate nodes with the highest risk of metastases because it detected any intra-individual variation in metastatic node distribution for the same tumour site. In this patient population all the involved nodes (11 in five patients) took up the radiocolloid; these data seem to confirm that our method of injection does not underestimate the number of SNs visualised. Moreover, contralateral metastatic nodes were found in two patients with primary tumours classified as pt3 G3 and pt2 G2, respectively, showing that there is a risk of nodal metastases to the contralateral side in tumours that are not locally advanced. The incidence of N+ (5/14 patients) does not seem low if we consider the prevalence of pt2 patients in this series. In this context, LS may be of great benefit in elucidating the functional anatomy of lymphatic drainage, giving the surgeon the opportunity to plan the intervention on the basis of the preoperative lymphatic mapping results and to carry out a selective neck dissection guided by the gamma probe. In conclusion, these preliminary results obtained in a small, selected group of patients suggest that LS for the purpose of lymphatic mapping may minimise the extent of

6 the operation and thus reduce surgery-related morbidity while ensuring adherence to the principle of oncological radicality through the selection of neck levels potentially at risk of harbouring cancer. Acknowledgements. This work was supported by grants from the Italian Association for Cancer Research (AIRC). The authors thank Deborah Console for editing the manuscript. References 1. Woolgar JA. Pathology of the N0 neck. Br J Oral Maxillofac Surg 1999;37: Ng SH, Yen TC, Liao CT, Chang JT, Chan SC, Ko SF, et al. 18 F-FDG PET and CT/MRI in oral cavity squamous cell carcinoma: a prospective study of 124 patients with histologic correlation. J Nucl Med 2005;46: Weiss MH, Harrison LB, Isaacs RS. Use of decision analysis in planning a management strategy for the stage N0 neck. Arch Otolaryngol Head Neck Surg 1994;120: Zwetyenga N, Majoufre-Lefebvre C, Siberchicot F, Demeaux H, Pinsolle J. Squamous-cell carcinoma of the tongue: treatment results and prognosis. Rev Stomatol Chir Maxillofac 2003;104: Haddadin KJ, Soutar DS, Oliver RJ, Webster MH, Robertson AG, MacDonald DG. Improved survival for patients with clinically T1/T2, N0 tongue tumours undergoing a prophylactic neck dissection. Head Neck 1999;21: Morton DL, Wen DR, Foshag LJ, Essner R, Cochran A. Intraoperative lymphatic mapping and selective cervical lymphadenectomy for early-stage melanomas of the head and neck. J Clin Oncol 1993;11: Schwartz GF, Giuliano AE, Veronesi U. Consensus Conference Committee. Proceedings of the consensus conference on the role of sentinel lymph node biopsy in carcinoma of the breast, April 19 22, 2001, Philadelphia, Pennsylvania. Cancer 2002;94: Veronesi U, Paganelli G, Viale G, Luini A, Zurrida S, Galimberti V et al. A randomized comparison of sentinelnode biopsy with routine axillary dissection in breast cancer. N Engl J Med 2003;349: Shoaib T, Soutar DS, Prosser JE, Dunaway DJ, Gray HW, McCurrach GM, et al. A suggested method for sentinel node biopsy in squamous cell carcinoma of the head and neck. Head Neck 1999;21: Chiesa F, Mauri S, Grana C, Tradati N, Calabrese L, Ansarin M et al. Is there a role for sentinel node biopsy in early N0 tongue tumors? Surgery 2000;128: Mozzillo N, Chiesa F, Botti G, Caraco C, Lastoria S, Giugliano G, et al. Sentinel node biopsy in head and neck cancer. Ann Surg Oncol 2001;8(Suppl 9):103S 105S 12. Ionna F, Chiesa F, Longo F, Manola M, Villano S, Calabrese L, et al. Prognostic value of sentinel node in oral cancer. Tumori 2002;88:S18 S Nieuwenhuis EJ, Pijpers R, Castelijns JA, Snow GB. Lymphoscintigraphic details of sentinel lymph node detection in 82 patients with squamous cell carcinoma of the oral cavity and oropharynx. Nucl Med Commun 2003;24: Ross GL, Soutar DS, MacDonald DG, Shoaib T, Camilleri I, Roberton AG, et al. Sentinel node biopsy in head and neck cancer: preliminary results of a multicenter trial. Ann Surg Oncol 2004;11: Civantos FJ, Gomez C, Duque C, Pedroso F, Goodwin WJ, Weed DT, et al. Sentinel node biopsy in oral cavity cancer: correlation with PET scan and immunohistochemistry. Head Neck 2003;25: Barzan L, Sulfaro S, Alberti F, Politi D, Pin M, Savignano MG, et al. An extended use of the sentinel node in head and neck squamous cell carcinoma: results of a prospective study of 100 patients. Acta Otorhinolaryngol Ital 2004;24: Stoeckli SJ, Pfaltz M, Ross GL, Steinert HC, MacDonald DG, Wittekind C, et al. The second international conference on sentinel node biopsy in mucosal head and neck cancer. Ann Surg Oncol 2005;12: Robbins KT, Clayman G, Levine PA, Medina J, Sessions R, Shaha A, et al. American Academy of Otolaryngology-Head and Neck Surgery Neck dissection classification update: revisions proposed by the American Head and Neck Society and the American Academy of Otolaryngology-Head and Neck Surgery. Arch Otolaryngol Head Neck Surg 2002;128: Meijer S, van den Brekel MW, Nieuwenhuis EJ. A suggested method for sentinel node biopsy in SCC of the head and neck. Head Neck 2000;22: Byers RM, Weber RS, Andrews T, McGill D, Kare R, Wolf P. Frequency and therapeutic implications of skip metastases in the neck from squamous carcinoma of the oral tongue. Head Neck 1997;19: Werner JA, Dunne AA, Myers JN. Functional anatomy of the lymphatic drainage system of the upper aerodigestive tract and its role in metastasis of squamous cell carcinoma. Head Neck 2003;25: Spiro JD, Spiro RH, Shah JP, Sessions RB, Strong EW. Critical assessment of supraomohyoid neck dissection. Am J Surg 1988;156: Medina JE, Byers RM. Supraomohyoid neck dissection: rationale, indications, and surgical technique. Head Neck 1989;11: Shoaib T, Soutar DS, MacDonald DG, Camilleri IG, Dunaway DJ, Gray HW, et al. The accuracy of head and neck carcinoma sentinel lymph node biopsy in the clinically N0 neck. Cancer 2001;91: Pitman KT, Johnson JT, Brown ML, Myers EN. Sentinel lymph node biopsy in head and neck squamous cell carcinoma. Laryngoscope 2002;112: Dutton JM, Graham SM, Hoffman HT. Metastatic cancer to the floor of mouth: the lingual lymph nodes. Head Neck 2002;24: DiNardo LJ. Lymphatics of the submandibular space: an anatomic, clinical and pathologic study with applications to floor-of-mouth carcinoma. Laryngoscope 1998;108: Even-Sapir E, Lerman H, Lievshitz G, Khafif A, Fliss DM, Schwartz A, et al. Lymphoscintigraphy for sentinel node mapping using a hybrid SPECT/CT system. J Nucl Med 2003;44: