Objective: Postoperative acute interstitial pneumonia is a subset of post-surgical acute respiratory distress syndrome (ARDS) and is responsible for one third of in-hospital deaths following lung resection in patients with primary lung cancer. We evaluated the usefulness of computed tomography (CT) for detection of interstitial pneumonia (IP) as a risk factor of postoperative ARDS. Methods: Preoperative chest CT of patients who underwent thoracotomy for primary lung cancer was reviewed retrospectively and IP findings in the chest CT were detected. Results: A total of 1148 patients with primary lung cancer underwent thoracotomy. Fifteen patients (1.3%) developed postoperative ARDS. Eleven of these 15 patients died of ARDS. Three of 41 patients who received induction therapy developed postoperative ARDS. Induction therapy was a risk factor of postoperative ARDS (p ≪ 0.01). Eleven out of the 15 patients who developed postoperative ARDS had IP findings (10: localized, 1: diffuse) in their chest CT. Two of three patients who had postoperative ARDS after induction therapy also had IP findings. Chest CTs of 834 patients were retrospectively analyzed; 91 patients (10.9%) had IP-findings (diffuse 1.8%, localized 9.1%). Postoperative ARDS occurred in 8.8% of IP-positive patients, and in 0.4% of IP-negative patients (p ≪ 0.001). Conclusion: Detection of IP by chest CT is useful for the selection of high-risk patients who may have postoperative ARDS following thoracotomy.