Diagnostic Certainty in Characterizing Liver Lesions in Rectal Cancer: Abbreviated Liver MRI versus CT.

Abstract

Early diagnosis of metastases is crucial but routine staging with contrast-enhanced multidetector computed tomography (ceMDCT) is suboptimal. A total of 20% will have indeterminate or too small to characterize (TSTC) liver lesions on CT, requiring formal characterization by magnetic resonance imaging (MRI). This UK cross-sectional study reports our experience undertaking routine abbreviated liver MRI (MRI). A total of 99 patients with rectal cancer had ceMDCT, abbreviated liver MRI, and rectal MRI at diagnosis. Liver imaging was scored for liver metastases, benign or indeterminate/TSTC lesions on a per patient basis. Primary rectal cancer was risk scored on MRI. A total of 42/99 (42%) had liver lesion(s) on ceMDCT versus 55/99 (56%) by MRI, and 46/99 (46%) had high-risk rectal cancer. ceMDCT showed 5 patients with liver metastases, 14 with benign lesions, and 23 with indeterminate/TSTC lesions. MRI showed 6 with liver metastases, 45 with benign lesions, and 4 with indeterminate/TSTC lesions. All liver metastases were in high-risk rectal cancer, OR 17.18 (p = 0.06), with 12.5% conversion rate of TSTC lesions to metastases in high-risk rectal cancer and 0% in low-risk rectal cancer. Diagnostic certainty of the liver findings was achieved in 93% of patients by MRI compared with 45% by ceMDCT (p < 0.0001). Abbreviated liver MRI diagnosed fewer indeterminate/TSTC lesions and provided greater diagnostic certainty than ceMDCT, p < 0.0001. High-risk rectal cancer is associated with a higher conversation rate of TSTC lesions to metastases than low-risk rectal cancers. Risk stratified; routine abbreviated liver MRI sequences should be investigated as part of the patient pathway for high-risk rectal cancer.