Multimodal image registration for localization of sentinel nodes in head and neck squamous cell carcinoma

The aim of this clinical study was to propose a new method of preoperative 3-dimensional localization of the neck sentinel node in clinically and radiographically N0 squamous cell carcinoma of the oral cavity.

This prospective study was carried out between August 2002 and August 2003. Ten patients with staged oral squamous cell cancer underwent this method of preoperative localization of the sentinel lymph node. At the present time, lymphoscintigraphy is recognized as the investigation of choice for the preoperative detection of sentinel nodes. Although we found it to be highly sensitive (100% in our study), it does not allow precise localization of the sentinel node in the anatomic context of the neck. To overcome this drawback, we used multimodal image registration applied to the cervical region in association with the imaging modalities of scintigraphy (single-photon emission computed tomography [SPECT] and gamma transmission) and computed tomography (CT). SPECT and gamma transmission (barium-133 sources) were carried out with a triple-head Philips-Marconi camera (Philips Medical System, Cleveland, OH), and CT was carried out with a General Electric console (GE Medical Systems, Milwaukee, WI). Gamma transmission was used to place the SPECT images in a reference cervical anatomic context (CT images). Registration involved the use of a manual method based on definition of homologous volume structures.

Preoperative image registration effectively localized the cervical sentinel node in 9 of the 10 patients, enabling a guided surgical approach in 90% of cases. The sentinel node was localized and resected using a hand-held gamma probe.

This original study adapts a new tool (multimodal image registration) to obtain precise preoperative localization of the cervical sentinel node in N0 oral squamous cell carcinoma. We confirmed the feasibility of this method in this indication. Although this method is a novel one, we believe that it will become extremely useful once a consensus has been reached on exclusive excision of the cervical sentinel node in oral cancers, as is the case for melanoma or breast cancer.