© International Federation for Medical and Biological Engineering 2007. Purpose: This study investigates feasibility of CBCT-based treatment planning. Methods and materials: Hounsfield unit (HU) values and profiles of phantoms and patients in CBCT images were compared to those in CT images. CBCT-based treatment plans for the phantoms and patients were compared to CT-based treatment plans dosimetrically. Results: Mean HU values of different materials in CBCT images were very close to those in CT images for Catphan. CBCT images included scatters and artifacts. As a consequent, the HU profiles of the homogeneous phantoms in CBCT images showed inhomogeneous HU distribution and the peripheral areas near the edge of the field-of-view showed reduced HU values. The HU profiles of the inhomogeneous phantom showed reduced HU values throughout the phantom and more significantly around the peripheral areas and lungs than other areas. The HU values in patients were also reduced in most tissue regions in CBCT images. Most plans based on CBCT with a bowtie filter showed good agreement with those based on CT with less than 1% of MU/cGy difference or 1 – 2 mm of isodose discrepancy. Large dosimetrical discrepancy occurred when CBCT was scanned without a bowtie filter, when a treatment beam passed through a significant amount of lung, and when a treatment beam was largely off-centered so as to pass through a body part imaged near the edge of the field-of-view. All IMRT plans based on CBCT and CT showed isodose distribution with very good agreement for patient cases except for one lung cancer patient. Conclusion: This study proves feasibility of CBCT-based treatment planning by comparing HU values, MU/cGy and isodose distributions to the gold standard CT-based treatment planning. CBCT could be used for treatment planning purposes for most cases except for the cases with large discrepancy shown in this study.