We retrospectively generated IMRT plans for 14 NSCLC patients who had experienced grade 2 or 3 esophagitis (CTCAE version 3.0). We generated 11-beam and reduced esophagus dose plan types to compare changes in the volume and length of esophagus receiving doses of 50, 55, 60, 65, and 70 Gy. Changes in planning target volume (PTV) dose coverage were also compared. If necessary, plans were renormalized to restore 95% PTV coverage. The critical organ doses examined were mean lung dose, mean heart dose, and volume of spinal cord receiving 50 Gy. The effect of interfractional motion was determined by applying a three-dimensional rigid shift to the dose grid. For the esophagus plan, the mean reduction in esophagus V50, V55, V60, V65, and V70 Gy was 2.8, 4.1, 5.9, 7.3, and 9.5 cm(3), respectively, compared with the clinical plan. The mean reductions in LE50, LE55, LE60, LE65, and LE70 Gy were 2.0, 3.0, 3.8, 4.0, and 4.6 cm, respectively. The mean heart and lung dose decreased 3.0 Gy and 2.4 Gy, respectively. The mean decreases in 90% and 95% PTV coverage were 1.7 Gy and 2.8 Gy, respectively. The normalized plans' mean reduction of esophagus V50, V55, V60, V65, and V70 Gy were 1.6, 2.0, 2.9, 3.9, and 5.5 cm(3), respectively, compared with the clinical plans. The normalized plans' mean reductions in LE50, LE55, LE60, LE65, and LE70 Gy were 4.9, 5.2, 5.4, 4.9, and 4.8 cm, respectively. The mean reduction in maximum esophagus dose with simulated interfractional motion was 3.0 Gy and 1.4 Gy for the clinical plan type and the esophagus plan type, respectively. In many cases, the esophagus dose can be greatly reduced while maintaining critical structure dose constraints. PTV coverage can be restored by increasing beam output, while still obtaining a dose reduction to the esophagus and maintaining dose constraints.