Image review for COVID-19 on CBCT vs MIP's

Abstract

Purpose or Objective Thoracic CT features of COVID-19 include ground glass opacity or lung consolidation and can also be identified on Cone Beam Computed Tomography (CBCT). The authors developed a reporting tool using maximum intensity projections (MIP); enabling rapid review by therapeutic radiographers of longitudinal changes in successive CBCT images suggestive of COVID-19 infection in asymptomatic patients during radiotherapy. The aim of this work was to assess if the reporting tool required fewer departmental resources than standard CBCT review process for COVID-19 related changes. Materials and Methods The tool - run overnight for all thoracic CBCTs - provides a single report containing sequentially positioned lung MIP images, multiple slices of the last CBCT to show the most up-to-date anatomy in more detail. Additionally, a longitudinal plot of the right and left lung mean density over time. Lung density values (in mean Hounsfield Units (HU)) are also shown in the MIP images. The MIP images clearly show lung infections and their progression (Fig. 1). To demonstrate the efficiency of this tool and quantify the additional burden related to COVID-monitoring, two experienced observers independently reviewed thoracic CBCTs and MIP reports from 22 patients. CBCT images were reviewed from a single day (13.02.21) and all MIPs were reviewed for all fractions available for each patient up to and including the day of the CBCT. The time taken to review all CBCTs and MIP reports was recorded in seconds (sec), each observer reviewed 11 patients. The observers checked accuracy through the review of patient’s clinical notes to confirm no symptoms of COVID-19. All reports were cross checked by both observers for lung density changes. Results Fig.2 shows time taken for observers to review one CBCT for typical COVID-19 lung density changes; 55 sec (range 35-86 sec, standard deviation (SD) 12.8 sec). The mean time taken for observers to review all MIPs together was 12 sec (range 9-16 sec, SD 1.8 sec). The MIP review included a mean of 9 fractions worth of images per patient (range 2-25). Neither observer reported COVID-19 changes using either method for the patients reviewed on this date. Although the observers reviewed different patient image sets, their mean review times for CBCT were similar at 56 and 61 seconds per CBCT respectively. The mean time for MIP reviews were also comparable at 13 and 12 seconds respectively. An additional benefit of the MIP method was that multiple CBCT images are available in a single “snap-shot” allowing progression to be more easily visualised Conclusion The use of MIPs facilitates more rapid assessment of COVID-19 related lung density changes than CBCT review resulting in less impact on clinical resource. This will provide minimal burden for screening of infections in radiotherapy departments.