Purpose: To present the SALT group results using Linac radiosurgery (RS) for AVM in 169 evaluable patients treated from January 1990 thru December 1993. Methods and Materials: Median age was 33 years (range 6–68 years). Irradiation was the only treatment in 55% patients. Other treatment modalities had been used prior to RS in 45%: one or more embolizations in 36%, surgery in 6%, and embolization and surgery in 3% patients. Nidus were supratentorial in 94% patients, infratentorial in 6% patients. Circular 15 MV x-ray minibeams (6–20 mm) were delivered in coronal arcs by a GE-CGR Saturne 43 Linac. Patient set-up included a Betti arm-chair, a Talairach frame. Prescribed peripheral dose was 25 Gy on the 60%–70% isodose (max dose 100%). Arteriographic results were reassessed in December 1997 at 48 to 96 months follow-up. Results: The overall obliteration rate (OR) was 64% (108/169). AVM volumes ranged from 280 to 19,920 mm 3 , median 2460 mm 3 . OR was 70% for AVM ≤ 4200 mm 3 vs. 51% for AVM > 4200 mm 3 (p = 0.02). The largest nidus dimension ranged from 8 to 51 mm, median 22 mm. OR was 70% for nidus ≤ 25 mm vs. 54% for nidus > 25 mm (p = 0.04). OR was 71%, in the absence of embolization, vs. 54% for previously embolized nidus (p = 0.03). OR was 71% for monocentric RS vs. 54% for multi-isocenters (p = 0.03). Peripheral doses (Pd) ranged from 15 to 28Gy, median 25 Gy, OR was 52% in patients receiving Pd ≤ 24.1 Gy or > 25.9 Gy and 70% for 25 Gy (NS). Peripheral isodoses ranged from 50%–90%, median 70%: OR was 67% for peripheral isodoses of 65% and 70% vs. 61% for peripheral isodoses of 50%–60% (NS). The mean lesion doses (MLd) ranged from 14 to 36 Gy, median 29 Gy: OR was 72% for MLd > 28 Gy vs. 55% for values ≤ 28 Gy (p = 0.02). The mean lesion isodoses (MLi) ranged from 53 to 90 Gy, median 79%: OR was 75% for MLi > 79% vs. 57% for lower values (p = 0.03). The minimum lesion doses (mLd) ranged from 3.6 to 23, median 16 Gy: OR was 69% for mLd > 17 Gy, vs. 59% for mLd ≤ 16 Gy (p = 0.05). The minimum lesion iodoses (mLi) ranged from 9%–65%, median 45%: OR was 71% for mLi > 40%, vs. 54% for mLi ≤ 40% (p = 0.05). The coverage ratio (CR) ranged from 33%–100%, median 85%: OR was 68% for CR > 85% vs. 60% for CR ≤ 84% (NS). For patients treated according to our protocol, i.e., 24–26 Gy on the 60%–70% isodoses, OR was higher (68%) than for other patients (47%) (p = 0.02). After multivariate analysis, absence of previous embolization and mono isocentric-irradiation were independent factors predicting obliteration. Complications were: recurrent hemorrhage, 4 patients (1 patient died); brain necrosis on MRI, 2 patients; subsequent epilepsy, 4 patients; other subsequent neurologic deficits, 3 patients. Conclusion: Overall OR was 64% (48–96 months follow-up). After monovariate analysis higher ORs were associated with smaller volumes ≤ 4200 mm 3 , smaller nidus size ≤ 25 mm, absence of prior embolization, monoisocentric RS, higher values for mean and minimum lesion doses and compliance to our protocol. Higher values for the peripheral dose and isodose tended to give better results. Multivariate analysis showed that the absence of prior embolization and monoisocentric irradiation were independent factors predicting successfull irradiation.