The success of cancer treatment with external beam radiation therapy depends critically on the adequate coverage of the tumor. Treatment variation due to daily setup variation and organ motion must be taken into account. Typically, a treatment margin is added to make allowance for these variations. It is generally desirable to have a small treatment margin. A higher dose can then be prescribed without inflicting serious failure. However, improper reduction of the margin will greatly increase the risk of local failure. The problem of margin reduction is particularly important for treatment of tumors in the lungs and upper abdomen. The dismal prognosis of these diseases compels the use of higher dose for treatment. Yet, attempt to do so is severely hampered by the large margin needed to cover the tumor motion while the patient breathes. The applicants hypothesize that treatment of tumors in the thorax and abdomen is much more effective when breathing motion is minimized. To do so, the approach of active breathing control (ABC) is developed to immobilize breathing motion temporarily and reproducibly. At any pre-selected phase in the breathing cycle, an ABC apparatus temporarily restricts airflow to and from the patient, thereby immobilizing breathing motion. The period of active breath hold is that which can by comfortably maintained by the patient. With ABC, radiation is turned on and off manually, and only, during the period of active breath hold. The need of complex control of the treatment machine is thus avoided. The ABC procedure can be applied for three dimensional (3D) CT scanning and treatment such that the patient information used in planning reproduces that during treatment. In this proposal, an ABC apparatus will be used to acquire 3D CT scans of 40 patients with lung and liver cancer at 8 different phases in the breathing cycle. The merit of treatment with ABC is evaluated by comparing the dosimetry of a 3D treatment plan where breathing motion is immobilized at one respiratory phase with that were 3D organ motion exists, as modeled by the 8 scans. Clinical feasibility of treatment with ABC will also be tested on 25 patients undergoing conventional treatment by studying the reproducibility and acceptance of the procedure by the patients and the treatment personnel.