One hundred ten consecutive cases (110 patients) of arthroscopically assisted rotator cuff repair through a limited, portal-extension approach were retrospectively reviewed. The average age of our patients was 58 years (range, 30 to 79). There were 35 women and 75 men. The dominant shoulder was affected in 67 patients (61%). All patients underwent a standard arthroscopic decompression. Acromioclavicular resections were performed in 15% of patients. The anterolateral portal was extended in the direction of Langer's lines to a total length of no more than 3 cm. The torn tendon was accessed through a small deltoid muscle split and repaired with nonabsorbable sutures. At an average follow-up of 35 months (range, 24 to 86), 106 patients (96%) had achieved excellent or satisfactory results. The average American Shoulder and Elbow Surgeons pain score improved from 7 preoperatively to 2 postoperatively. All but four patients were satisfied with the clinical result and reported significant improvement in active elevation and strength and a significant lessening of pain. Late acromioclavicular joint pain contributed to failure in three of the four patients with unsatisfactory results in this series. The results of this study suggest that, in selected patients with small to medium rotator cuff tears, arthroscopically assisted repair through an anterolateral portal-extension approach can produce excellent results.