PurposeWe aimed to construct predictive models for the overall survival (OS), progression-free survival (PFS), and distant metastasis-free survival (DMFS) for nasopharyngeal carcinoma (NPC) patients by using CT-based radiomics.Materials and MethodsWe collected data from 197 NPC patients. For each patient, radiomic features were extracted from the CT image acquired at pretreatment via PyRadiomics. Feature selection was performed in two steps. First, features with high inter-observer variability based on multiple tumor delineations were excluded. Then, stratified bootstrappings were performed to identify feature combinations that most frequently achieved the highest (i) area under the receiver operating characteristic curve (AUC) for predicting 3-year OS, PFS, and DMFS or (ii) Harrell’s C-index for predicting time to event. Finally, regularized logistic regression and Cox proportional hazard models with the most frequently selected feature combinations as input were tuned using cross-validation. Additionally, we examined the robustness of the constructed model to variation in tumor delineation by simulating 100 realizations of radiomic feature values to mimic features extracted from different tumor boundaries.ResultsThe combined model that used both radiomics and clinical features yielded significantly higher AUC and Harrell’s C-index than models using either feature set alone for all outcomes (p < 0.05). The AUCs and Harrell’s C-indices of the clinical-only and radiomics-only models ranged from 0.758 ± 0.091 to 0.789 ± 0.082 and from 0.747 ± 0.062 to 0.767 ± 0.074, respectively. In comparison, the combined models achieved AUC of 0.801 ± 0.075 to 0.813 ± 0.078 and Harrell’s C-indices of 0.779 ± 0.066 to 0.796 ± 0.069. The results showed that our models were robust to variation in tumor delineation with the coefficient of variation ranging from 4.8% to 6.4% and from 6.7% to 9.3% for AUC and Harrell’s C-index, respectively.ConclusionOur results demonstrated that using CT-based radiomic features together with clinical features provided superior NPC prognostic prediction than using either clinical or radiomic features alone.

Background: The study of radiophotoluminescent glass dosimeters (RGDs) in the clinical usage of proton beams is limited. The aim of this study was to investigate the dosimetric characteristics of RGDs for pencil beam scanning proton therapy. The feasibility of using an RGD in end-to-end testing of intensity-modulated proton therapy (IMPT) plans at various treatment sites was also evaluated. Materials and methods: The dosimetric characteristics of the GD-302M type glass dosimeter were studied in terms of uniformity, short-term and long-term reproducibility, stability of the magazine position readout, dose linearity in the range from 0.2-20 Gy, energy response in 70-220 MeV, MU/spot, dose rate response, and fading effect. The reference conditions of the spot scanning beam from the Varian ProBeam Compact system were operation at 160 MeV, a 2 cm water equivalent depth in a solid water phantom, a 10×10 cm field size at the isocenter, and 2 Gy dose delivery. End-to-end testing of IMPT plans for the head, abdomen, and pelvis was verified by using the Alderson Rando phantom. The overall uncertainty analysis was confirmed in this study. Results: The relative response of RGDs for the uniformity test was within 0.95-1.05. The %CVs of the short-term and long-term reproducibility were 1.16% and 1.50%, respectively. The FGD-1000 automatic reader showed stable magazine position readout. The dose linearity was found to have an obviously good linear relationship, with R2 = 0.9988. The energy response relative to 160 MeV was approximately within 4.0%. The MU/spot and dose rate had less effect on the RGD readout. The fading effect was relatively stable for 10 weeks of storage, within 2.4%. For the end-to-end test, the maximum difference between the treatment plan and RGD measurement showed a very good result that was within 1.0%. The overall uncertainty of the RGD measurement for the proton beam was 4.6%. Conclusion: RGDs have confirmed the potential for proton dosimetry, including in end-to-end testing. The appropriate correction factor for the energy response can be applied for dose verification of scanning proton beams.

Background: The Monte Carlo (MC) simulation is an effective tool for determining the absorbed dose in small field sizes. To calculate accurate results, the MC simulation requires precise geometric and material descriptions of the linear accelerator head. Due to proprietary information issues, the description of the Varian TrueBeam™ linear accelerator (Varian Medical Systems, Palo Alto, CA) head geometry and material information are not available. Instead, the manufacturer provided a phase-space file just above the jaw for each photon energy level. Although several studies have validated the accuracy of this phase-space file, to the best of our knowledge, there are no reported data for a small field size (<2x2 cm2) of 6 MV photon beams. Objectives: The purpose of this study was to evaluate the Varian TrueBeam™ phase-space file of the 6 MV photon beam provided by the manufacturer for the Monte Carlo (MC) simulation in small field dosimetry. Materials and methods: The TrueBeam™ linear accelerator was simulated using an EGSnrc MC code with a Varian phase-space file as the input. The simulation was compared with the measurement using percent depth dose (PDD) and beam profile, and the field output factor (FOF) for the 0.6x0.6, 1x1, 2x2, 3x3, 4x4, 6x6, and 10x10 cm2 field sizes. Results: The agreement between the measurements and simulated PDD data was under 2.2% beyond the buildup region. The distance to agreement (DTA) in the buildup region was within 1.0 mm. The simulation data presented identical profiles with the measurement within 1.0% of the dose difference or 1.2 mm of the DTA. The mean dose difference in the radiation field was ≤1.5% for the ≥1x1 cm2 field size. The largest deviation was observed in the 0.6x0.6 cm2 inline beam profile. The deviation of the penumbra and full width at half maximum (FWHM) between simulation and measurement was <2 mm. The agreement of the simulated and measured FOF was within 1.0%, except for the 0.6x0.6 cm2 field size. Conclusion: Overall, the MC simulation demonstrates data that is consistent with the measurement for the ≥1x1 cm2 field sizes. These data assure that the 6 MV Varian phase-space file can be used as a radiation source for accurate MC dose calculation in a small field. However, a large discrepancy in beam profiles was observed at the 0.6x0.6 cm2 field size due to the different primary source sizes among TruebeamTM machines.

Background Ultrahypofractionation can shorten the irradiation period. This study is the first dosimetric investigation comparing ultrahypofractionation using volumetric arc radiation therapy (VMAT) and intensity-modulated proton radiation therapy (IMPT) techniques in postmastectomy treatment planning. Materials and methods Twenty postmastectomy patients (10-left and 10-right sided) were replanned with both VMAT and IMPT techniques. There were four scenarios: left chest wall, left chest wall including regional nodes, right chest wall, and right chest wall including regional nodes. The prescribed dose was 26 Gy(RBE) in 5 fractions. For VMAT, a 1-cm bolus was added for 2 in 5 fractions. For IMPT, robust optimization was performed on the CTV structure with a 3-mm setup uncertainty and a 3.5% range uncertainty. This study aimed to compare the dosimetric parameters of the PTV, ipsilateral lung, contralateral lung, heart, skin, esophageal, and thyroid doses. Results The PTV-D95 was kept above 24.7 Gy(RBE) in both VMAT and IMPT plans. The ipsilateral lung mean dose of the IMPT plans was comparable to that of the VMAT plans. In three of four scenarios, the V5 of the ipsilateral lung in IMPT plans was lower than in VMAT plans. The Dmean and V5 of heart dose were reduced by a factor of 4 in the IMPT plans of the left side. For the right side, the Dmean of the heart was less than 1 Gy(RBE) for IMPT, while the VMAT delivered approximately 3 Gy(RBE). The IMPT plans showed a significantly higher skin dose owing to the lack of a skin-sparing effect in the proton beam. The IMPT plans provided lower esophageal and thyroid mean dose. Conclusion Despite the higher skin dose with the proton plan, IMPT significantly reduced the dose to adjacent organs at risk, which might translate into the reduction of late toxicities when compared with the photon plan.

Background: Ultrahypofractionation can shorten the irradiation period. This study is the first dosimetric investigation comparing ultrahypofractionation using volumetric arc radiation therapy (VMAT) and intensity-modulated proton radiation therapy (IMPT) techniques in postmastectomy treatment planning. Materials and methods: Twenty postmastectomy patients (10-left and 10-right sided) were replanned with both VMAT and IMPT techniques. There were 4 scenarios: left chest wall, left chest wall including regional nodes, right chest wall, and right chest wall including regional nodes. The prescribed dose was 26 Gy (RBE) in 5 fractions. For VMAT, a 1-cm bolus was added for 2 in 5 fractions. For IMPT, robust optimization was performed on the CTV structure with a 3-mm setup uncertainty and a 3.5% range uncertainty. This study aimed to compare the dosimetric parameters of the PTV, ipsilateral lung, contralateral lung, heart, skin, esophageal, and thyroid doses. Results: The PTV-D95 was kept above 24.7 Gy in both VMAT and IMPT plans. The ipsilateral lung mean dose of the IMPT plans was comparable to that of the VMAT plans. In three of four scenarios, the V5 of the ipsilateral lung in IMPT plans was lower than in VMAT plans. The Dmean and V5 of heart dose were reduced by a factor of 4 in the IMPT plans of the left side. For the right side, the Dmean of the heart was less than 1 Gy for IMPT, while the VMAT delivered approximately 3 Gy. The IMPT plans showed a significantly higher skin dose owing to the lack of a skin-sparing effect in the proton beam. The IMPT plans provided lower esophageal and thyroid mean dose. Conclusion: Despite the higher skin dose with the proton plan, IMPT significantly reduced the dose to adjacent organs at risk, which might translate into the reduction of late toxicities when compared with the photon plan. Key words: proton therapy, ultrahypofractionation, postmastectomy, breast irradiation