Plan evaluation tool for spatially fractionated radiation therapy for unresectable large tumors via spatial biological effective dose modeling in combination therapy.

We present the utility of a plan evaluation tool for a multi-course radiation treatment consisting of a highly heterogeneous SFRT plan followed by a course of curative radiation therapy for large and bulky unresectable tumors. For a more accurate plan assessment, this novel method calculates the voxelized biological effective dose (BED) spatially from each course and combines them into a single spatial BED distribution (s-BED). Ten previously treated head and neck (H&N) cancer patients with MLC-based 3D-conformal SFRT (15 Gy in 1 fraction) followed by a curative course of VMAT for 66-70 Gy in 33-35 fractions were retrospectively analyzed using this new s-BED method. The s-BED calculations were based on the standard linear-quadratic (LQ) model. Evaluations of mean BED using this s-BED method were compared to other methodologies that use each course's DVH, mean dose, and prescription dose. From this, tumor control probability (TCP) was calculated using these different methodologies. Lastly, doses to nearby organs at risk (OARs) were evaluated using the s-BED method and compared to each course's physical dose distribution. The OARs evaluated were the spinal cord, brainstem, optic pathway, cochlea, parotid glands, larynx, esophagus, and mandible. From the physical dose distributions, a s-BED distribution and a spatial EQD2 (s-EQD2) distribution were able to be calculated and visualized. The various methods of calculating mean BED using each course's dose prescription, mean dose, DVH, and from the s-BED resulted in varying mean BED: 121.8 Gy, 99.7 Gy, 88.3 Gy, and 100.6 Gy, respectively. In turn, this also gave varying predictions in tumor response: 100.0%, 98.2%, 92.3%, and 91.1%, respectively. Of the 8 H&N patients who received follow-up imaging, 7 (87.5%) had local tumor control. Reported toxicities of this cohort saw 2 cases of grade 3 toxicities (skin desquamation), 3 grade 1 toxicities (oral mucositis and odynophagia), and 1 grade 4 toxicity (necrotizing fasciitis). The composite s-BED distributions provided a means of better understanding the effective biological dose being delivered to both the target and nearby OARs spatially. Future utilization of this method during the treatment planning process may allow for more personalized treatment prescriptions for the SFRT course and the follow-up combination therapy, potentially enhancing therapeutic benefits in managing large and bulky unresectable tumors.