Given the propensity for lung metastases (LM), NCCN guidelines recommend lung surveillance (LS) with either CXR or CT in high-grade soft tissue sarcoma (HG-STS). Considering survival, diagnostic sensitivity and cost, the optimal modality is unknown.The US Sarcoma Collaborative database (2000-2016) was reviewed for patients who underwent resection of a primary HG-STS. Primary outcome was overall survival (OS). Cost analysis was performed.Among 909pts, 83% had truncal/extremity, 17% had retroperitoneal (RP) tumors. Recurrence occurred in 48% of which 54% were LM. LS was performed with CT in 80% and CXR in 20%. Both groups were clinically similar although CT patients had more RP tumors and recurrences. Regardless of modality, 85-90% of LM were detected within the first 2yrs with a similar re-intervention rate. When considering age, tumor size, location, margin status, and receipt of radiation, LM was independently associated with worse OS(HR:4.26; p<0.01) while imaging modality was not(HR:1.01; p=0.97). CXR patients did not have an inferior 5-year OS compared to CT(71vs60%, p<0.01). When analyzing patients in whom no LM was detected, both cohorts had a similar 5-year OS(73vs74%, p=0.42), suggesting CXR was not missing clinically relevant lung nodules. When adhering to a guideline-specified protocol for 2018 projected 4,406 cases, surveillance with CXR for 5yrs results in savings of $5-8M/year to the US healthcare system.In this large multicenter study, LS with CXR did not result in worse overall survival when compared to CT. With considerable savings, a CXR-based protocol may optimize resource utilization for LS in HG-STS; prospective trials are needed.