Utility of an automatic adaptive iterative metal artifact reduction AiMAR algorithm in improving CT imaging of patients with hip prostheses evaluated for suspected bladder malignancy

Purpose To compare the utility of a novel metal artifact reduction algorithm to standard imaging in improving visualization of key structures, diagnostic confidence, and patient-level confidence in malignancy in patients with suspected bladder cancer. Methods Patients with hip implants undergoing CT urography for suspected bladder malignancy were enrolled. Images were reconstructed using 3 methods: (1) Filtered Back Projection (FBP), (2) Iterative Metal Artifact Reduction (iMAR), and (3) Adaptive Iterative Metal Artifact Reduction (AiMAR) strength 4. In multiple reading sessions, three radiologists graded visualization of critical anatomic structures and artifact severity (6-point scales, lower scores desirable), and diagnostic confidence in blinded fashion. They also graded patient-level confidence in malignancy based on imaging findings in each patient. Results Thirty-two patients (8 females) with a mean age of 74.5 ± 8.5 years were included. The median (range) visualization scores for FBP, iMAR, and AiMAR were 3.6 (1.1–4.9), 1.6 (0.3–2.8), and 1.6 (0.3–2.6), respectively. Both iMAR and AiMAR had anatomic visualization and artifact scores better than FBP ( P 0.05). Structures with the most improvement in visualization score with the use of metal artifact reduction algorithms included the obturator internus muscle, internal and external iliac nodal chains, and vagina. iMAR and AiMAR improved diagnostic confidence ( P