Aims
Third-space endoscopy generates recognizable postoperative CT changes, yet distinguishing expected findings from true complications remains challenging; this study aimed to define the radiologic signature and identify features linked to clinically relevant events.
Methods
We performed a single-center retrospective analysis of patients who underwent third-space endoscopy and received standardized postoperative CT within 24 hours. CTs were reviewed using a structured template and quantified with accepted radiologic scales, and categorized as expected or suspected complications. Clinical variables included procedure type, duration, age, and complications requiring management change. Analyses were conducted in R 4.4.2.
Results
Of the 91 patients who underwent third-space endoscopy, 79 received postoperative CT imaging and were included. Baseline characteristics and CT findings are summarized in the Table. Thoracic abnormalities were frequent, with pneumomediastinum being the dominant finding. Pleural effusions were mostly small and bilateral; all moderate effusions occurred after esophageal POEM or gastric STER, resolved spontaneously, and were associated with longer procedure duration (90 vs 30 min, p = 0.0015). Pericardial effusion showed no association with procedure duration. Radiologic evidence of chronic aspiration was more common in older patients, and atelectasis was highly prevalent. No hydropneumomediastinum, mediastinal collections, or pleural/mediastinal abscess was identified, and all thoracic findings resolved on follow-up imaging. Abdominal findings were mild. Pneumoperitoneum was consistently limited in extent, with occasional retroperitoneal air. Subcutaneous emphysema was common, exclusively after esophageal procedures, and did not correlate with procedure duration. In achalasia cases, esophageal wall thickening at the procedure site showed no relationship to procedural duration, while intramural air and visualized submucosal tunnels were common and fluid-free. No contrast leak, hematoma, or localized collection was observed.A radiologist flagged 13 CTs as suspected complications, but only two reflected clinically relevant events: a persistent mucosal defect after POEM requiring stenting and a gastrografin leak after gastric STER successfully managed endoscopically, with no leak on next-day imaging.
|
Baseline Characteristics and Distribution of CT Findings Across Third-Space Endoscopic Procedures |
|
|
Demographics |
Gender: Male 43 (54.4%); Female 36 (45.6%); Age, median (range) 64 (17-87) |
|
Thoracic findings |
Pneumomediastinum: Z-POEM 23/29 (86.2%); POEM 40/42 (95.2%); Pneumothorax (volume <30%): Z-POEM 5/29 (17.2%); POEM 2/42 (4.7%); Pleural effusion: Z-POEM 6/29 (20.6%); POEM 25/42 (59.5%) G-STER: 2/3 (66.6%); Hydropneumothorax: Z-POEM 1/29 (3.4%); Pneumopericardium: Z-POEM 1/29 (3.4%); Pericardial effusion: Z-POEM 1/29 (3.4%); POEM 10/42 (23.8%); |
|
Pulmonary parenchymal changes |
Aspiration pneumonitis/ infiltrates: Z-POEM 9/29 (31.0%); POEM 17/42 (40.4%); G-POEM 3/3 (100%); Atelectasis: Z-POEM 13/29 (44.8%); POEM 36/42 (85.7%); E-STER 1/1 (100%); G-POEM 3/3 (100%) |
|
Peritoneal and retroperitoneal findings |
Pneumoperitoneum: Z-POEM 1/29 (3.4%); POEM 41/42 (97.6%); G-POEM 2/3 (66.6%); G-STER 3/3 (100%); Retroperitoneal air: POEM 8/42 (19.0%); G-POEM 1/3 (33.3%); Ascites (minimal - perihepatic fluid collection): POEM 1/42 (2.3%); Subcutaneous emphysema: Z-POEM 24/29 (82.7%); POEM 27/42 (64.2%); E-STER 1/1 (100%) |
Conclusions
Post–third-space endoscopy CT shows a consistent pattern of expected changes, while true complications are rare. Multicenter studies are needed to validate these imaging signatures and enhance their diagnostic utility.