Sentinel node biopsy for oral and laryngopharyngeal squamous cell carcinoma: a retrospective study of 177 patients in Japan.

OBJECTIVE: Sentinel node (SN) biopsy in the head and neck region has not been widely used in Japan, except at a few facilities. However, almost all these facilities perform preoperative localization and intraoperative diagnosis by frozen section analysis of SN to select patients who must undergo neck dissection in a one-stage procedure. The objective of this study was to determine the actual status of SN biopsy at those facilities in Japan that have actively conducted this procedure, and to elucidate the usefulness and drawbacks of this technique in head and neck cancer.

METHODS: We retrospectively reviewed 177 patients who had undergone SN biopsy at 7 facilities. The underlying pathology was laryngeal or hypopharyngeal cancer in 20 patients from one hospital, while the remaining 157 patients had oral cancer. Preoperative localization of SN was determined using conventional lymphoscintigraphy with or without single photon emission computed tomography with CT (SPECT-CT). Intraoperative localization and diagnosis of SN were performed by gamma probe and frozen section analysis.

RESULTS: Conventional lymphoscintigraphy detected a mean of 2.6 SNs per patient in 137 patients with oral cancer, compared to 2.7 in 71 patients using SPECT-CT and 2.9 in 154 patients using the gamma probe. No significant differences were apparent between techniques. Forty of the 520 SNs (7.7%; 33 in oral cancer and 7 in laryngeal or hypopharyngeal cancer) were pathologically positive in the final diagnosis. Of these, 3 were not processed for frozen sectioning and were diagnosed only with hematoxylin and eosin staining. Among the others, 32 (86.5%) were diagnosed intraoperatively as showing metastasis. In terms of the false-negative rate, 144 patients were determined by SN biopsy to have no positive SNs. Of these, 2 patients had non-SN metastases found in their dissected neck and 8 patients without neck dissection showed late nodal recurrence. The false-negative rate was thus 6.9%.

CONCLUSION: Frozen section analysis, particularly multislice sectioning, offers a relatively reliable intraoperative diagnostic method. We were able to perform immediate neck dissection based on the results of multislice sectioning as a single-stage procedure.