ARRS 2022 Abstracts


1544. Don't Take it Lying Down: Distinguishing Features and Diagnostic Capability of Prone Radiographs
Authors * Denotes Presenting Author
  1. Geraldine Villasana Gomez; Montefiore Medical Center
  2. Rydhwana Hossain; University of Maryland
  3. Allison Herring; University of Maryland
  4. Jeffrey Levsky; Montefiore Medical Center
  5. Charles White; University of Maryland
  6. Linda Haramati *; Montefiore Medical Center
  7. Matthew Lazarus; Montefiore Medical Center
Prone positioning improves oxygenation and lung recruitment in critically ill patients, and has been widely adapted during the COVID-19 pandemic. Radiography of these patients is used to locate support devices, which is particularly important after patient repositioning, and to assess for changes in lung disease extent. However, literature on prone radiographs is sparse, and their reliability to identify lines and tubes is unknown. We assessed image quality, distinguishing features, and ability to identify support devices on prone radiographs of critically ill patients.

Materials and Methods:
Patients admitted from 3/2020 to 2/2021, with both supine and prone radiographs within 1 day, were included. IRB approval was obtained, HIPAA compliance was maintained. Radiograph quality was assessed by evaluation of obscuration/exclusion of a lung base or apex or presence of significant rotation. Specific imaging characteristics assessed were inferior scapula tip position with respect to the rib shadow, stomach gas bubble in the fundus, and medial/superior displacement of a breast shadow (in women). If an endotracheal tube, enteric tube, or central venous catheter was present, we noted if the position could be adequately assessed, and if it was properly positioned. Each radiograph pair was reviewed in consensus by at least 2 cardiothoracic radiologists, blinded to position. Fisher exact test was used.

Eighty-one patients were included in the study (63 years ±13 years old, 30% women). Adequate imaging of the lung bases (prone 68/81 [81%] vs. supine 73/81 [90%], p=0.35) and apices (prone 75/81 [93%] vs. supine 76/81 [94%], p=1) were comparable, and significant rotation was more common in prone radiographs (prone 29/81 [36%] vs. supine 15/81 [19%], p=0.021). Prone exams more frequently showed one or both scapula tips beyond the rib border (prone 72/81 [89%] vs. supine 12/81 [15%], p<0.001) and more frequently had a fundal stomach bubble (prone 36/81 [44%] vs. supine 8/81 [10%], p<0.001). In women, a displaced breast shadow was more common when prone (prone 11/24 [46%] vs. supine 2/24 [8%], p=0.008). To identify prone technique, scapula position was 89% sensitive and 85% specific, stomach gas bubble in the fundus was 44% sensitive and 90% specific, and displaced breast shadow was 46% sensitive and 92% specific. Both techniques reliably assessed line and tube position (prone 177/178 [99%] vs. supine 202/203 [99.5%]). Prone exams trended toward higher rate of malpositioned line or tube (21/178 [12%] prone vs. 13/202 [6%] supine, p=0.07).

Position-dependent differences can distinguish prone and supine radiographs. Scapular tip position was most closely associated with prone position. Stomach bubble in the fundus or displaced breast shadow were also associated with prone position, but were not always applicable. Most importantly, prone radiographs reliably visualize support devices. There was a trend toward more frequent line or tube malposition in prone patients, which underscores the need to evaluate these devices on prone radiographs.