Combined SPECT/CT improves detection of initial bone invasion and determination of resection margins in squamous cell carcinoma of the head and neck compared to conventional imaging modalities

Abstract

Purpose: Knowledge of the presence and extent of bone infiltration is crucial for planning the resection of potential bone-infiltrating squamous cell carcinomas of the head and neck (HNSCC). Routinely, plain-film radiography, multislice computed tomography (MSCT) and magnetic resonance imaging (MRI) are used for preoperative staging, but they show relatively high rates of false-positive and false-negative findings. Scintigraphy with (99m)Tc-bisphosphonate has the ability to show increased metabolic bone activity. If combined with anatomical imaging (e.g. (SPECT)/CT), it facilitates the precise localization of malignant bone lesions. The aim of this study was to analyse the indications and advantages of SPECT/CT compared with standard imaging modalities and histology with regard to specificity and sensitivity

Methods: A longitudinally evaluated group of 30 patients with biopsy-proven HNSCC adjacent to the mandible underwent (99m)Tc-bisphosphonate SPECT/CT, MRI, MSCT and conventional radiography before partial or rim resection of the mandible was performed. Bone infiltration was first evaluated with plain films, MSCT and MRI. In a second reading, SPECT/CT data were taken into account. The results (region and certainty of bone invasion) were evaluated among the different imaging modalities and finally compared with histological specimens from surgical resection as the standard of reference. For a better evaluation of the hybrid property of SPECT/CT, a retrospectively evaluated group of 20 additional patients with tumour locations similar to those of the longitudinally examined SPECT/CT group underwent SPECT, MSCT and MRI. To assess the influence of dental foci on the specificity of the imaging modalities, all patients were separated into two subgroups depending on the presence or absence of teeth in the area of potential tumour-bone contact.

Results: Histologically proven bone infiltration was found in 17 patients (57 %) when analysed by conventional imaging modalities. SPECT/CT data revealed bone infiltration in two additional patients (7 %), who both showed discrete cortical bone erosion not visible by MSCT or MRI. There were no false-positive or false-negative findings on SPECT/CT. The quality criteria for detecting bone involvement in HNSCC by SPECT/CT were as follows: sensitivity 100 % (lower 95 % confidence interval limit 80 %), specificity 100 % (75 %), positive predictive value 100 % (80 %) and negative predictive value 100 % (75 %). Corresponding data for MRI were 95 % (76 %), 94 % (73 %), 95 % (76 %) and 94 % (73 %), and for MSCT were 89 % (71 %), 100 % (85 %), 100 % (86 %) and 88 % (69 %). In the retrospective evaluation SPECT showed results similar to SPECT/CT.

Conclusion: Hybrid SPECT/CT has a high specificity as it can provide additional information about the existence and local extent of malignant bone infiltration of the mandible. Although the sensitivity of conventional SPECT is similar to that of SPECT/CT, the latter provides a much better delineation of the local tumour-bone contact area. Based on this information, surgical intervention of the rim versus partial resection can be planned and performed more precisely. Patient outcome can be improved by avoiding undertreatment and unnecessary or overextended bone resections.