Local tumor recurrence remains a major problem in patients with inoperable non–small-cell lung cancer undergoing radiotherapy. We investigated the theoretical gain in the estimated tumor control probability (TCP) using an individualized maximal tolerable dose (MTD) prescription, for both conventional and accelerated fractionation schemes.For 64 non–small-cell lung cancer patients, five treatment plans were compared, dependent on the normal tissue dose constraints for the lung and spinal cord. The first two used a classic fractionation (2 Gy/d, 5 d/wk) to a total dose of 60 Gy (QD classic ) or determined by the individualized MTD (QD MTD ). The third scheme assumed a hypofractionated schedule of 2.75-Gy fractions (QD hypofr ). The fourth and fifth assumed hyperfractionation and acceleration (1.8 Gy twice daily, either BID classic or BID MTD ). The TCPs for the groups of patients were estimated.The mean biologic equivalent dose in 2-Gy fractions for tumor, corrected for accelerated repopulation was significantly greater for the BID MTD scheme (62.1 Gy) than for any other scheme (QD classic , 47.5 Gy; QD MTD , 52.0 Gy; QD hypofr , 56.9 Gy; and BID classic , 56.9 Gy; p < 0.001). Although both dose-escalation (QD MTD ) and hypofractionation (QD hypofr ) resulted in an increase in the mean estimated TCP of 5.6% (p < 0.001) and 14.6% (p < 0.001), respectively, compared with QD classic , the combination of escalation and acceleration (BID MTD ) improved the mean estimated TCP by 26.4% (p < 0.001).The results of this planning study showed a large gain in the estimated TCP using an MTD scheme with 1.8-Gy fractions BID compared with other fractionation schedules. Clinical studies implementing this concept are ongoing.