Multimodality imaging features of primary pulmonary artery sarcoma and exploratory survival analysis: a multicenter retrospective study

Background Primary pulmonary artery sarcoma (PAS) is a rare malignancy often misdiagnosed as pulmonary thromboembolism. This multicenter study aimed to characterize the multimodality imaging features of PAS and perform an exploratory survival analysis of factors associated with overall survival (OS). Methods This multicenter study retrospectively included PAS patients from January 2015 to January 2025. Baseline CTPA, available MRI and PET/CT, surgical outcomes, and complications were assessed. An exploratory survival analysis was performed among patients with available follow-up data. OS was analyzed using the Kaplan–Meier method and log-rank test; univariable and multivariable Cox regression models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs). Results A total of 101 patients (55 female; median age 53 years) were included and comprised the CTPA imaging cohort. On CTPA, PAS predominantly showed complete intraluminal filling defects with expansive growth (93.1%) and proximal bulging/lobulated sign (91.1%). Distal aneurysmal dilatation occurred in 47.5%, while saddle-shaped filling defects were observed in 35.6%. Among 31 patients with available MRI, PAS showed median native T1 and T2 values of 1483 ms and 82 ms, heterogeneous enhancement (90.3%), and the Wall Eclipsing Sign (35.5%). Among 29 patients with available PET/CT, PAS showed avid 18F-FDG uptake (median SUVmax 8.10). Intimal sarcoma was the predominant pathological subtype (n = 91, 90.1%). Follow-up data were available for 33 of the 101 patients, of whom 22 underwent surgery; 21 achieved complete macroscopic resection (R0/R1, 95.5% of surgery cohort) and 1 underwent partial resection. Perioperative complications occurred in 36.4%, with 30-day mortality of 9.1%. In the exploratory survival analysis of this follow-up cohort, surgical treatment was associated with longer OS than non-surgical treatment (median OS, 56.1 vs. 14.1 months, log-rank P = 0.015), with consistent findings in univariable Cox analysis (HR 0.33, 95% CI 0.13–0.84, P = 0.021). However, baseline CTPA features were not significantly associated with OS in log-rank tests or univariable Cox analysis. Conclusions Integrating CTPA with MRI and PET/CT facilitates PAS recognition. Surgical treatment was associated with longer OS in this exploratory survival analysis of the follow-up cohort. Baseline CTPA features were not significantly associated with OS in this limited follow-up cohort, suggesting radiological complexity alone should not preclude surgical evaluation, though prospective validation is needed. Clinical trial number Not applicable.