Yukiko Kami1, Toru Chikui2, Osamu Togao3, Shintaro Kawano4, Shinsuke Fujii5,6, Masahiro Ooga7, Tamotsu Kiyoshima5, Kazunori Yoshiura2. 1. Department of Oral and Maxillofacial Radiology, Faculty of Dental Science, Kyushu University, 3-1-1, Maidashi, Higashi-ku, Fukuoka, 812-8582, Japan. kami@rad.dent.kyushu-u.ac.jp. 2. Department of Oral and Maxillofacial Radiology, Faculty of Dental Science, Kyushu University, 3-1-1, Maidashi, Higashi-ku, Fukuoka, 812-8582, Japan. 3. Department of Molecular Imaging & Diagnosis, Graduate School of Medical Sciences, Kyushu University, Fukuoka, Japan. 4. Section of Oral and Maxillofacial Oncology, Division of Maxillofacial Diagnostic and Surgical Sciences, Faculty of Dental Science, Kyushu University, Fukuoka, Japan. 5. Laboratory of Oral Pathology, Division of Maxillofacial Diagnostic and Surgical Sciences, Faculty of Dental Science, Kyushu University, Fukuoka, Japan. 6. Dent-craniofacial Development and Regeneration Research Center, Faculty of Dental Science, Kyushu University, Fukuoka, Japan. 7. Department of Medical Technology, Kyushu University Hospital, Fukuoka, Japan.
Abstract
OBJECTIVES: To compare the delineation of mandibular cancer by 3D T1 turbo field echo with compressed SENSE (CS-3D-T1TFE) images and MDCT images, and to compare both sets of images with histopathological findings, as the gold standard, to validate the accuracy and clinical usefulness of CS-3D-T1TFE reconstruction. METHODS: Twenty-four patients with mandibular squamous cell carcinoma (SCC) who underwent MRI including CS-3D-T1TFE and MDCT examinations before surgery were retrospectively included. For both examinations, 0.5-mm-thick coronal plane images and 0.5-mm-thick plane images perpendicular and parallel to the dentition were constructed. Two radiologists rated bone invasion in three categories indexed by cortical bone, cancellous bone, and mandibular canal (MC), and inter-rater agreement was assessed by weighted kappa statistics. In 20 of the 24 patients who underwent surgery, the correlation of bone invasion with the histopathological evaluation by pathologists was assessed using Pearson's correlation coefficient. Soft-tissue invasion was assessed by diagnosing the presence of invasion into the mylohyoid muscle, gingivobuccal fold, and masticator space, and inter-rater agreement was assessed by kappa statistics. RESULTS: The interobserver agreement for bone invasion assessment was almost perfect with CS-3D-T1TFE and substantial with MDCT. The image evaluations by both observers agreed with the pathological evaluations in 15 of the 20 cases, showing high correlation (r > 0.8). CS-3D-T1TFE also showed higher inter-rater agreement than MDCT for all measures of soft-tissue invasion. CONCLUSIONS: CS-3D-T1TFE reconstructed images were clinically useful in accurately depicting the extent of mandibular cancer invasion and potentially solving the problem of lesion overestimation associated with conventional MRI. KEY POINTS: • Reconstructed CS-3D-T1TFE images were useful for the diagnosis of mandibular cancer. • CS-3D-T1TFE images showed higher inter-rater agreement than MDCT and high correlation with pathological findings. • CS-3D-T1TFE images may solve the problem of overestimation of the tumor extent, which has been associated with MRI in the past.
OBJECTIVES: To compare the delineation of mandibular cancer by 3D T1 turbo field echo with compressed SENSE (CS-3D-T1TFE) images and MDCT images, and to compare both sets of images with histopathological findings, as the gold standard, to validate the accuracy and clinical usefulness of CS-3D-T1TFE reconstruction. METHODS: Twenty-four patients with mandibular squamous cell carcinoma (SCC) who underwent MRI including CS-3D-T1TFE and MDCT examinations before surgery were retrospectively included. For both examinations, 0.5-mm-thick coronal plane images and 0.5-mm-thick plane images perpendicular and parallel to the dentition were constructed. Two radiologists rated bone invasion in three categories indexed by cortical bone, cancellous bone, and mandibular canal (MC), and inter-rater agreement was assessed by weighted kappa statistics. In 20 of the 24 patients who underwent surgery, the correlation of bone invasion with the histopathological evaluation by pathologists was assessed using Pearson's correlation coefficient. Soft-tissue invasion was assessed by diagnosing the presence of invasion into the mylohyoid muscle, gingivobuccal fold, and masticator space, and inter-rater agreement was assessed by kappa statistics. RESULTS: The interobserver agreement for bone invasion assessment was almost perfect with CS-3D-T1TFE and substantial with MDCT. The image evaluations by both observers agreed with the pathological evaluations in 15 of the 20 cases, showing high correlation (r > 0.8). CS-3D-T1TFE also showed higher inter-rater agreement than MDCT for all measures of soft-tissue invasion. CONCLUSIONS: CS-3D-T1TFE reconstructed images were clinically useful in accurately depicting the extent of mandibular cancer invasion and potentially solving the problem of lesion overestimation associated with conventional MRI. KEY POINTS: • Reconstructed CS-3D-T1TFE images were useful for the diagnosis of mandibular cancer. • CS-3D-T1TFE images showed higher inter-rater agreement than MDCT and high correlation with pathological findings. • CS-3D-T1TFE images may solve the problem of overestimation of the tumor extent, which has been associated with MRI in the past.
Authors: A Imaizumi; N Yoshino; I Yamada; K Nagumo; T Amagasa; K Omura; N Okada; T Kurabayashi Journal: AJNR Am J Neuroradiol Date: 2006-01 Impact factor: 3.825