3D MRI segmentation and 3D circumferential resection margin evaluation for a standard rectal cancer assessment

Aim. Recent studies focused on rectal cancer suggested that a 3D imaging segmentation obtained from MRI data could contribute in the definition of the circumferential resection margin (CRM) and in the assessment of the tumor regression following neo-adjuvant treatments. Here, we propose a method for defining and visualizing the circumferential margins using 3D MRI segmentation; this methodology was tested in a clinical study comparing 3D CRM assessment vs standard MRI imaging. Patients and methods. MRI scans performed before neo-adjuvant treatments were selected and reviewed 3D mesorectal/tumor segmentations were obtained using Digital Imaging and COmmunications in Medicine (DICOM) data; CRMs were calculated using 3D volumes plus a color scale for the closest distances. Results. 3D reconstructions were possible in all selected cases and 3D images implemented by the color scale were positive for immediate CRM visualization. Statistical analyses comparing standard radiology disclosed that the degree of consistency, the reliability of ratings, the correlation and precision were optimal considering the overall cases, but lower in the CRM>0 mm sub-group. Conclusions. This new method is not inferior comparing standard radiology; moreover, the imaging segmentation we obtained was highly promising and could be helpful in defining a standard CRM measurement, thus it could improve clinical practice.