A multi-FPGA accelerator for radiation dose calculation in cancer treatment

Radiation therapy (RT) is a major modern modality for cancer treatment. However, the use of a large number of radiation beams in treatment plans and image-guided on-line replanning in newly emerging RT planning and delivery systems present a daunting challenge to radiation dose calculation even on the state-of-the-art quad-core computers as the dose calculation time becomes unacceptably long for clinical applications. Based on a popular dose calculation algorithm, the Collapsed-Cone Convolution/Superposition (CCCS) algorithm, this paper presents a multi-FPGA accelerator design for the dose calculation problem. Our performance-driven design strategy yields a fully pipelined architecture, which includes a resource-economic raytracing engine and high-performance energy deposition pipeline. The design is capable of processing 3 transporting lines simultaneously and accomplishing one dose deposition in every 2 cycles while raytracing along the test phantom. The evaluation based on a set of clinical treatment planning cases confirms that our FPGA design almost fully utilizes the available external memory bandwidth and achieves close to the best possible performance for the CCCS algorithm while using less resource. With the support of a memory-rich multi-FPGA platform, our floating-point based design working at 90Mhz obtains a speedup of 20X over a commercial multi-threaded software on a quad-core system and 15X performance improvement over the closely related results.