Clinical feasibility study of a 2D ripple filter to improve the efficiency of carbon ion therapy

Background and purpose: The ripple filter (RiFi) is a passive energy modulator used in particle beam therapy to broaden the Bragg peak. The 1D-RiFi features a wavy structure that can broaden a monoenergetic carbon ion beam to 3 mm, while the 2D-RiFi employs a two-dimensional groove structure to achieve a 6 mm beam broadening. This study aimed to evaluate the potential advantages of the 2D-RiFi over the 1D-RiFi in terms of dose distribution optimization, treatment efficiency, and organ at risk (OAR) dose control by comparing water phantom and clinical patient plans. Methods: Carbon ion treatment plans were designed for water phantoms and 20 patients using both 1D-RiFi and 2D-RiFi. The water phantom plans targeted a cubic region of interest (80 mm×80 mm×80 mm) at ranges of 95, 105, 190 and 290 mm. From patients who underwent carbon ion therapy at Shanghai Proton and Heavy Ion Center, 20 cases were selected via simple random sampling with computer-generated random numbers, stratified by the proportion of different tumor sites (6 head and neck tumors, 4 prostate tumors, 4 lung tumors, 2 pancreatic tumors, 2 liver tumors and 2 shoulder tumors). Key dosimetric metrics, including homogeneity index (HI), conformity index (CI) and clinical target volume (CTV) coverage by 95% prescription dose (V95), were analyzed along with OAR doses. Energy layers, beam time, and irradiation time were compared between the two RiFi types. Statistical analysis was performed using the Wilcoxon rank-sum test, with a significance level of P<0.05. This study was approved by the ethics committee of Shanghai Proton and Heavy Ion Center (approval number: 240311EXP-01). Results: For water phantom plans, the 1D-RiFi plans achieved HI of 0.04±0.01, CI of 1.10±0.03, V95 of 99.92%±0.06% and flatness of 6.52%±0.61%, while the 2D-RiFi plans achieved HI of 0.04±0.01, CI of 1.11±0.04, V95 of 99.92%±0.06%, and flatness of 7.52%±0.81%. The mean doses to the distal and lateral block in 1D-RiFi plans were (1.34 Gy±0.43) Gy [relative biological effectiveness (RBE)]and (0.98±0.05) Gy (RBE), respectively, compared to (1.47±0.33) Gy (RBE) and (0.94±0.03) Gy (RBE) for 2D-RiFi plans. The use of 2D-RiFi reduced the average beam-on time by 43% and the number of energy layers by 48%. For clinical plans, the 1D-RiFi plans had HI of 0.07±0.04, CI of 1.94±0.67, and V95 of 98.81%±1.61%, compared to HI of 0.07±0.05, CI of 1.95±0.70, and V95 of 98.79%±1.69% for the 2D-RiFi plans, with no statistically significant differences (P=0.77, 0.65 and 0.66, respectively). OAR mean doses increased slightly with the 2D-RiFi plans (average increase of 0.8%, P=0.62) but remained within clinically acceptable limits. The 2D-RiFi plans reduced energy layers by 45%-50% (average 48%), beam time by 32%-49% (average 44%), and irradiation time by 28%-41% (average 36%). Conclusion: Treatment plans using the 2D-RiFi achieved comparable target coverage to those using the 1D-RiFi, with a slight but clinically acceptable increase in OAR doses. The application of the 2D-RiFi significantly reduced the number of energy layers, beam time and irradiation time in carbon ion therapy, enhancing treatment efficiency.