Purpose: Reported local failure rates for FIGO stage II-IVA vaginal cancer usually range between 20% and 60%, even in recent series. In this study we investigate the clinical feasibility and report about the first treatment outcomes of MRI image guided brachytherapy (IGBT) in these pts using the same strategy as recently developed for cervix cancer. Materials: Forty pts with vaginal cancer were treated at our institution from 1999-2006. Thirteen pts met the inclusion criteria for this study and received IGBT based on a prospective protocol. FIGO stage distribution was: stage II=4, stage III=5, stage IV=4. Five pts had positive pelvic lymph nodes and in 7 pts tumour size was >5cm. All pts received EBRT (40-50Gy) and 11 pts concurrent chemotherapy (9 cisplatin and 2 MMC+5-FU). IGBT dose prescription was either 4x7 Gy (HDR) or 35-40 Gy (0.5-0.8Gy; PDR) to the HR CTV. In 10 cases a combined intravaginal/interstitial application technique was used. Based on the recent developments in the treatment of cervical cancer pts with IGBT (Gyn GEC ESTRO Recommendations), systematic concepts for HR CTV, OAR, biological modelling, DVH-analysis, dosevolumeadaptation (D90, D2cc), and dose escalation were integrated into the treatment (planning) of all pts. Reported doses were converted to the equivalent dose in 2Gy (EQD2) applying the LQ-model (α/β of 10 for tumour and 3 for OAR). Clinical results were calculated using the KaplanMeier method. Results: Spatial relations between vaginal cylinder +/needle applicators, HR CTV, and OAR were clearly visible on MRI in all cases. Mean total prescribed dose was 80Gy EQD2 (1SD±8Gy).Mean D90 for HR CTV was 86Gy EQD2 (1SD±13Gy). For OAR the following mean D2cc were documented: 80Gy (1SD±20Gy) for bladder, 76Gy (1SD ±16Gy) for urethra, 70Gy (1SD ±9Gy) for rectum and 60Gy (1SD±9Gy) for sigmoid colon. All pts achieved complete remission at 3 months. Median follow up was 43 months (range: 1987 months). Two pts developed fistula (1 vesico-vaginal, 1 recto-vaginal) due to tumour-necrosis and 1 with distal vaginal involvement complete vaginal obliteration. Three recurrences were documented: 1 local in a patient with FIGO stage III and 2 distant. All these pts had positive pelvic lymph nodes. Actuarial local control, OS and CSS rates at three years were 92%, 85% and 85%. For FIGO stages II, III, and IV they were 100%, 80% and 75%, respectively. Conclusions: IGBT for vaginal cancer seems to be feasible similar to the concepts as introduced for cervix cancer. Clinical and dosimetric results of this first experience with IGBT for locally advanced vaginal cancer are promising. Local control and survival in this small series seems to be significantly higher than in previous reports about conventional X-ray based brachytherapy, associated at the same time with acceptable morbidity. However, further prospective studies with more pts are needed to confirm these results.