Diagnostic value of procedural images in CT-guided interventions

Purpose: We hypothesize procedural images of CT-guided interventions may contain diagnostic findings not present in the reference images. Method: A retrospective review of CT-guided interventions performed at our hospital, from 01 April 2017 to 08 May 2020. Two radiologists independently reviewed the procedural CT images for presence of diagnostic findings in comparison to the reference images (CT, MRI, or PET/CT). ACR RADPEER score was assigned to all findings. The Findings were categorized into new finding, characterization of prior finding or change of prior finding. The results of biopsy and drainage samples were also reviewed. Results: The prevalence of diagnostic findings in procedural CT images was found to be 6.1% (81/1336); 32 new, 8 characterization, and 41 change findings. Having CT as reference image, procedure in the chest and having drainage were associated with presence of findings (p < 0.05). Increase time interval between the reference image and the procedure increases the odds of having diagnostic findings (p < 0.001). Age, sex, or whether inpatient or out-patient, malignant pathology result or infectious collection were not related to presence of findings (p > 0.05). The majority of findings were likely clinically significant (73%) and the majority were not documented in the procedure report (63%). Conclusion: Clinically relevant diagnostic findings in procedural images of CT-guided interventions are not uncommon and are underreported. Time delay between the reference image and the procedure is the most significant factor associated with presence of diagnostic findings.