At the time of treatment planning it would be useful to know whether part of the treatment beam passes through the patient/couch support assembly before it passes through the patient. In the previous work of Yorke, the range of gantry angles leading to beam-couch intersection was found as a function of couch translation for symmetric field sizes and for zero couch rotation. Yorke's method has been extended to include couch rotation, dual independent jaws, and multi-leaf collimator (MLC) field shapes. In addition, the new method is also applicable in the situation of the couch top located above the isocenter. For a clinically treatable, 20 x 20 cm field configuration in a linac, the range of gantry angles leading to beam-couch intersection are different by 6.7 degrees for a couch rotation angle of 25 degrees when compared to no couch rotation. The new method agrees with data within the setup and measurement uncertainties for a variety of field sizes including an oval shaped MLC field, and various couch locations, couch, and collimator rotation angles.