If the primary tumor meets criteria described above, the next issue is whether the neck is grossly involved. While the sentinel node technique is an excellent technique for detecting micrometastases, it is less useful for detecting nonpalpable but grossly involved LNs. This appears to be particularly true with squamous cell carcinoma. It is postulated that when a large percentage of the LN is replaced by cancer, physiologic obstruction occurs and alternative patterns of lymphatic drainage develop. It is important to detect the presence of such gross disease on preoperative imaging and avoid applying this technique to that group of patients, in order to avoid false positives. Generally contrast-enhanced computerized tomography (CT) and magnetic resonance imaging (MRI) (if iodine allergic) are the imaging modalities used. These should be strictly interpreted, and patients should be excluded if there are nodes >1.5 cm in size for levels I and II; >1.2 cm in size for levels III, IV, V, and VI; with central necrosis, irregular enhancement, or a poorly defined or irregular capsular border; or with groups of three or more asymmetrically located LNs, with a minimal axial diameter of 8 mm or more, in the suspected tumor drainage basin (Fig. 1.1). The role of positron emission tomography (PET) remains to be delineated and is plagued by false positives, but this may ultimately also prove useful in ruling out such gross disease. It should be kept in mind that the sentinel node technology represents an excellent technique for detection of micrometastases in patients felt to have a reasonably low risk of metastases, but is not as accurate at detecting grossly involved nodes.

Imaging studies, including CT with iodine contrast, MRI with gadolinium contrast, and ultrasonography, should be used to better identify grossly involved, nonpalpable nodes. CT with contrast is standard in North America, though ultrasound, if done in a detailed fashion, is a reasonable alternative. MRI with gadolinium contrast is a good option in iodine allergic patients. Central necrosis of the lymph nodes, although highly predictive, is a late finding. The uptake of 2-deoxy-2-[18F]fluoro-d-glucose as measured by PET has been reported as significantly more sensitive and only slightly less specific than MRI. However, foci of cancer smaller than 1 cm are below the resolution of PET as with CT and MRI.

The injection is performed prior to the surgical procedure, generally on the morning of surgery, in the radiology suite. Injection is also sometimes performed late the day before, although the effect of this on the success rate of sentinel node identification remains to be delineated. While awake injection and imaging in radiology are the most commonly used techniques, as we extend this procedure to endoscopically accessible oropharyngeal, supraglottic, and hypopharyngeal lesions, it is likely that cooperative efforts with the nuclear radiologist and the use of portable cameras will allow for intraoperative endoscopic injection, with or without radiologic imaging, and gamma probe-guided sentinel node biopsy without the need for uncomfortable injections in an awake patient. Theoretical advantages of injecting under general anesthesia include better exposure of the primary and avoidance of motion of the patient related to discomfort. This may eventually further increase the reliability of this method. Taking into account that the radiolocalization of the detected hot spots does not represent the drainage of the primary, but the drainage of the tracer deposits, which are supposed to mimic the lymphatic drainage of the primary, the impact of a thorough and representative tracer injection becomes evident. Due to density and direction of the head and neck lymphatics, the primary may drain into several alternative lymphatic pathways, all representing first draining “sentinel” LNs.

Nevertheless, due to regulatory issues related to the injection of radioactive substances and the lack of widely available portable nuclear imaging, awake injection remains the most commonly used technique at present. It is important to ensure that the patient is comfortable so that an adequate preoperative injection is obtained. I use topical anesthetic, mild oral sedation, and lingual, inferior alveolar, and/or sphenopalatine nerve blocks to ensure patient comfort during manipulation and injection of the primary tumor. Direct injection of the tumor with local anesthetic should not be performed as it may affect uptake of the radionuclide and reportedly may even cause it to precipitate in the tissues. The injection technique involves narrow injection with a fine 25-gauge needle circumferentially encompassing the leading edge of the lesion and an additional injection in the center of the lesion (Figs. 1.2 and 1.3). Five tuberculin syringes with 1-mL aliquots of technetium-99 sulfur colloid, with a total radioactivity of 400 millicuries, would represent a standard dose for the morning of surgery. A slightly higher dose would be used the night before. These dosages are extrapolated from the practice for melanoma and have worked well for cancer of the oral cavity, but formal comparative evaluation of dosages and volumes for use in the oral cavity have yet to be performed.

I have used this technique for visible oral lesions. For cutaneous lesions, it is well documented that a scar from a previous excisional biopsy can be injected to allow accurate sentinel node biopsy. Whether a previously excised oral lesion could undergo sentinel node excision by injection of an intraoral scar has yet to be determined. It is important to inject narrowly and not to inject the deep tissues. The radionuclide will extravasate more widely in the oral cavity around the site of injection than occurs in the skin and will usually go to the neck more quickly. There is no benefit to trying to inject a margin around the tumor, as this will lead to an unmanageable excess of radioactive nodes. I prefer to use unfiltered technetium-99 sulfur colloid. The presence of larger particles allows for retention of radioactivity in the proximal lymphatics. Retained particles at the site of primary injection are not a major issue as we recommend removal of the primary tumor first. However, it is also possible to obtain good results using filtered technetium sulfur colloid, and if there is a strong preference for addressing the lymphatics first, then this may be preferable as there is a better clearance of background radioactivity from the primary site. The more rapid migration of the filtered agent may also make it advantageous if injections of radionuclide are performed on the operating room table at the start of the procedure rather than prior to the surgical procedure.