Radiological Predictors of Prolonged Mechanical Ventilation Following Major Abdominal and Thoracic Surgeries: A Retrospective Study

This is the authors' abstract. We don't add key points for this paper.

Background: Prolonged mechanical ventilation (PMV) following major abdominal and thoracic surgery is a significant driver of postoperative morbidity, intensive care unit (ICU) resource utilisation, and mortality.Reliable preoperative or early postoperative predictors of PMV are lacking in the Indian tertiary care setting.Radiological parameters -obtained from routinely available preoperative chest X-rays, CT thorax/abdomen, and point-of-care ultrasound -offer objective, quantifiable biomarkers that may complement clinical scoring systems for predicting PMV.Objectives: To evaluate the association between radiological parameters -including preoperative CT-derived lung density indices and chest AP diameter, perioperative diaphragm excursion and thickening fraction on ultrasound, postoperative lung ultrasound score (LUS), CT-derived psoas muscle index (PMI), and chest X-ray findings (pleural effusion, atelectasis grade, cardiomegaly) -and the occurrence of PMV (>48 hours) in adults undergoing major abdominal or thoracic surgery.Methodology: A retrospective observational study was conducted from July 2025 to December 2025 at RDJMMCH, Muzaffarpur, Bihar.Medical records of 130 adult patients who underwent elective major abdominal or thoracic surgery under general anaesthesia were reviewed.Preoperative radiological data (CT thorax/abdomen, chest X-ray), perioperative diaphragm ultrasound, and postoperative chest X-ray findings were extracted and correlated with duration of mechanical ventilation.Multivariate logistic regression was used to identify independent radiological predictors of PMV.Data were analysed using SPSS v26.Results: PMV (>48 hours) occurred in 31 of 130 patients (23.8%).On multivariate logistic regression, diaphragm thickening fraction (DTF) <25% (OR 5.84; 95% CI 2.21-15.42;p<0.001), preoperative CT mean lung density <-850 HU (OR 4.17; 95% CI 1.58-11.03;p=0.004), postoperative LUS >8 (OR 3.92; 95% CI 1.47-10.43;p=0.006), and CT psoas muscle index (PMI) in the lowest sex-specific quartile (OR 3.28; 95% CI 1.19-9.02;p=0.022) were independent predictors of PMV.The combined radiological model yielded AUC 0.904 (95% CI 0.847-0.961).Conclusion: Radiological parameters -particularly diaphragm thickening fraction, preoperative lung density, postoperative lung ultrasound score, and CT-derived psoas muscle index -are significant independent predictors of PMV after major abdominal and thoracic surgery.Integration of these parameters into a radiological risk-stratification tool could improve perioperative planning and ICU resource allocation at tertiary care centres in north Bihar and comparable settings.