Purpose: To develop a new optimization algorithm to carry out true dose painting by numbers (DPBN) planning based on full Monte Carlo (MC) calculation.

Methods: Four configurations with different clustering of the voxel values from PET data were proposed. An optimization method at the voxel level under Lineal Programming (LP) formulation was used for an inverse planning and implemented in CARMEN, an in-house Monte Carlo treatment planning system.

Results: Beamlet solutions fulfilled the objectives and did not show significant differences between the different configurations. More differences were observed between the segment solutions. The plan for the dose prescription map without clustering was the better solution.

Conclusions: LP optimization at voxel level without dose-volume restrictions can carry out true DPBN planning with the MC accuracy.