“…Furthermore, given a set of solutions, the problem of choosing the "best" one from a clinical and patientspecific perspective still remains (Amols and Ling, 2002). A decision-making process that incorporates clinical reasoning was introduced by Meyer et al (Meyer et al, 2004), but it only works when the trade-offs between different plans are very large. Another natural approach is to cast IMRT as a constrained minimization problem (see, e.g., (Romeijn et al, 2003, Shepard et al, 1999), but this is still not the clinical standard, since efficiently solving such problems with DVH-based constraints is very difficult.…”