To review the technical improvements and changes in management that have occurred over 21 years, which have made the minimally invasive repair of pectus excavatum safer and more successful. In 1997, we reported our 10-year experience with a new minimally invasive technique for surgical correction of pectus excavatum in 42 children. Since then, we have treated an additional 1173 patients, and in this report, we summarize the technical modifications which have made the repair safer and more successful. From January 1987 to December 2008, we evaluated 2378 pectus excavatum patients. We established criteria for surgical intervention, and patients with a clinically and objectively severe deformity were offered surgical correction. The objective criteria used for surgical correction included computed tomography (CT) scans of the chest, resting pulmonary function studies (spirometry and/or plethysmography), and a cardiology evaluation which included echocardiogram and electrocardiogram. Surgery was indicated if the patients were symptomatic, had a severe pectus excavatum on a clinical basis and fulfilled two or more of the following: CT index greater than 3.25, evidence of cardiac or pulmonary compression on CT or echocardiogram, mitral valve prolapse, arrhythmia, or restrictive lung disease. Data regarding evaluation, treatment, and follow up have been prospectively recorded since 1994. Surgical repair was performed in 1215 (51%) of 2378 patients evaluated. Of these, 1123 were primary repairs, and 92 were redo operations. Bars have been removed from 854 patients; 790 after primary repair operations, and 64 after redo operations. The mean Haller CT index was 5.15 ± 2.32 (mean ± SD). Pulmonary function studies performed in 739 patients showed that FVC, FEV1, and FEF25-75 values were decreased by a mean of 15% below predicted value. Mitral valve prolapse was present in 18% (216) of 1215 patients and arrhythmias in 16% (194). Of patients who underwent surgery, 2.8% (35 patients) had genetically confirmed Marfan syndrome and an additional 17.8% (232 patients) had physical features suggestive of Marfan syndrome. Scoliosis was noted in 28% (340). At primary operation, 1 bar was placed in 69% (775 patients), 2 bars in 30% (338), and 3 bars in 0.4% (4). Complications decreased markedly over 21 years. In primary operation patients, the bar displacement rate requiring surgical repositioning decreased from 12% in the first decade to 1% in the second decade. Allergy to nickel was identified in 2.8% (35 patients) of whom 22 identified preoperatively received a titanium bar, 10 patients were treated successfully with prednisone and 3 required bar removal: 2 were switched to a titanium bar, and 1 required no further treatment. Wound infection occurred in 1.4% (17 patients), of whom 4 required surgical drainage (0.4% of the total). Hemothorax occurred in 0.6% (8 patients); 4 during the postoperative period and four occurred late. Postoperative pulmonary function testing has shown significant improvement. A good or excellent anatomic surgical outcome was achieved in 95.8% of patients at the time of bar removal. A fair result occurred in 1.4%, poor in 0.8%, and recurrence of sufficient severity to require reoperation occurred in 11 primary surgical patients (1.4%). Five patients (0.6%) had their bars removed elsewhere. In the 752 patients, more than 1 year post bar removal, the mean time from initial operation to last follow up was 1341 ± 28 days (SEM), and time from bar removal to last follow-up is 854 ± 51 days. Age at operation has shifted from a median age of 6 years (range 1-15) in the original report to 14 years (range 1-31). The minimally invasive procedure has been successfully performed in 253 adult patients aged 18 to 31 years of age. The minimally invasive repair of pectus excavatum has been performed safely and effectively in 1215 patients with a 95.8% good to excellent anatomic result in the primary repairs at our institution.