Purpose: Dieulafoy first described a lesion (DL) more than 130 years ago as a large focal artery without ulceration of the stomach causing massive fatal UGI hemorrhage (UGIH). Endoscopically, DL is reported in patients with severe UGIH when major stigmata of recent hemorrhage (SRH) without ulceration are seen. For endoscopically diagnosed and treated UGI DLs, our purposes were: 1) to report prevalences of SRH, 2) to compare 30-day outcomes after thermal coagulation vs. hemoclipping (HC), and 3) to describe recent changes in hemostasis of DL with Doppler probe ultrasound (DUP) monitoring. Methods: 56 patients with severe UGIH were enrolled in prospective CURE hemostasis studies. Techniques evolved from monotherapy, to combination hemostasis, to use of DUP to complete treatments. Thirty two patients before 6/2006 were treated with thermal coagulation. Twenty four other patients after 8/2006 had hemostasis with hemoclipping. Endoscopic treatments usually included epinephrine pre-injection and tattooing after hemostasis. When it became available, DUP was used for arterial blood flow detection underneath SRH in nine patients, to map the arteries, and to check for blood flow ablation to complete hemostasis. All patients were followed up to 60 days. Data were prospectively collected, entered onto computer databases, and managed with SAS. Results: For all 56 patients, prevalences of SRH were 60.7% active bleeding, 17.9% adherent clot, 2.5% non-bleeding visible vessel (NBVV), and 8.9% spot. For thermal vs. HC treated patients, there were no significant differences in 30 day outcomes: re-bleeding 25% vs. 29%; surgery 16% vs. 17%; or death 6% vs. 0%. However, the non-DUP assisted DL hemostasis group had a significantly higher 30 day re-bleed rate than the DUP assisted patients: 33% (15/45) vs. 0% (0/9). Conclusion: For 56 patients with severe UGIH from Dieulafoy's lesions: 1) SRH were active bleeding 60.7%, clot 17.9%, NBVV 12.5%, and other SRH 8.9%. 2) 30-day outcomes of re-bleeding, surgery, death were similar for thermal & hemoclip treated patients. 3) However, the 30-day re-bleeding rate of DUP assisted hemostasis patients was significantly lower than those without DUP monitoring. Acknowledgements: Partially supported by a Clinical VA Merit Review grant and NIH CURE DDRC CURE grant-Human Studies Core (AM41301).