How sensitive is single-photon emission computed tomography/computed tomography (SPECT/CT) in sentinel lymph node biopsy (SLNB) for cutaneous malignancies of the head and neck?
Background: Early identification of metastatic disease followed by node dissection has been shown to provide a survival benefit for many cutaneous malignancies. Recently, SPECT/CT has been introduced as another adjunct to improve SLNB. No study has quantified the test’s reliability in terms of sensitivity, specificity, and predictive value, and it is unknown if SPECT/CT reliability varies by primary tumor subsite.
Explore this issue:April 2015
Study design: Blinded retrospective imaging review of 47 patients (92 imaging reads) with a head and neck cutaneous malignancy who underwent pre-operative SPECT/CT followed by sentinel lymph node biopsy with a gamma probe between May 2010 and April 2013.
Setting: Department of Otolaryngology, Massachusetts Eye and Ear Infirmary, Boston.
Synopsis: SPECT/CT sensitivity for all patients was 73%. Primary ear lesions had a sensitivity at 85%, while eyelid lesions had a sensitivity at 63%. SPECT/CT specificity was 92%. The overall negative predictive value was 96%, with similar rates among all subsites. Overall positive SPECT/CT predictive value was 54%, with a high number of false-positive basins. Reviewers predicted the hottest single basin on pre-operative SPECT/CT 52% of the time. SPECT/CT was significantly better at predicting the hottest basin for eyelid lesions than for lesions of the neck or forehead. SPECT/CT predictive ability to identify the basin or an adjacent basin containing the single hottest lymph node was 92%. SPECT/CT consistently overestimated the number of nodal basins affected by an average of one basin. Wilcoxon signed-rank test demonstrated a statistically and clinically significant overestimation of involved nodal basins on SPECT/CT when compared to surgically sampled basins for primary lesions of the cheek, scalp, and ear. Limitations included a small cohort, tumor histologic heterogeneity, possible SPECT/CT scans that were misclassified, and classification functions that may have been subject to bias based on sample size.
Bottom line: SPECT/CT has high sensitivity, specificity, and negative predictive value but may overestimate relevant nodal basins in sentinel lymph node biopsy.
Citation: Remenschneider AK, Dilger AE, Wang Y, Palmer EL, Scott JA, Emerick KS. The predictive value of single-photon emission computed tomography/computed tomography for sentinel lymph node localization in head and neck cutaneous malignancy. Laryngoscope. 2015;125:877-882.
—Reviewed by Amy Hamaker