Single-centre evaluation and staging of rectal carcinoma on a 3-Tesla magnetic resonance imaging and correlation with histological profile.

ABSTRACT

PURPOSE: To assess magnetic resonance imaging (MRI) features of rectal carcinoma in correlation with pathology, and to evaluate the diagnostic accuracy of 3-Tesla MRI. MATERIAL AND METHODS: A total of 86 patients were included. 3T MR imaging was conducted pre-operatively, with imaging data correlated to pathology results. High-resolution, 2-dimensional, T2-weighted fast-spin echo sequences in the sagittal, axial, and coronal planes used to tumour staging. Diffusion-weighted images were used to increase the accuracy of tumour evaluation. RESULTS: Rectal carcinoma was staged as T3 in 45.3% of all patients (n = 39), without involvement of the mesorectal fascia in 31.4% (n = 27), and with a possible or obvious invasion in 14% of patients (n = 12). The diagnostic accuracy of 3T MRI was 97.6% for the T1 stage, 92.1% for T2, 89% for T3, and 90% for T4 tumours. MR-derived extramural vascular invasion (EMVI) was found in 16.2% (n = 14), with an estimated diagnostic accuracy of 95%. Diffusion-weighted images and apparent diffusion coefficient were estimated for the different histology types of rectal carcinoma. The average apparent diffusion coefficient for adenocarcinoma was 0.846 ± 0.17, for mucinous adenocarcinoma it was 1.17 ± 0.08, and for signet cell and squamous carcinomas it was 0.91 ± 0.11 and 0.796 ± 0.21 mm2/s, respectively. CONCLUSIONS: 3T MRI enables high levels of diagnostic accuracy in local rectal carcinoma staging, including assessment of mesorectal fascia infiltration and EMVI-status with high accuracy.