Preliminary outcomes of boron neutron capture therapy for head and neck cancers as a treatment covered by public health insurance system in Japan: Real-world experiences over a 2-year period.

PURPOSE: Since June 2020, boron neutron capture therapy (BNCT) has been a health care service covered by health insurance in Japan to treat locally advanced or recurrent unresectable head and neck cancers. Therefore, we aimed to assess the clinical outcomes of BNCT as a health insurance treatment and explore its role among the standard treatment modalities for head and neck cancers. MATERIALS AND METHODS: We retrospectively analyzed data from patients who were treated using BNCT at Kansai BNCT Medical Center, Osaka Medical and Pharmaceutical University, between June 2020 and May 2022. We assessed objective response rates based on the Response Evaluation Criteria in Solid Tumors version 1.1, and adverse events based on the Common Terminology Criteria for Adverse Events, version 5.0. Additionally, we conducted a survival analysis and explored the factors that contributed to the treatment results. RESULTS: Sixty-nine patients (72 treatments) were included in the study, with a median observation period of 15 months. The objective response rate was 80.5%, and the 1-year locoregional control, progression-free survival, and overall survival rates were 57.1% (95% confidence interval [CI]: 43.9%-68.3%), 42.2% (95% CI: 30.1%-53.8%), and 75.4% (95% CI: 62.5%-84.5%), respectively. Locoregional control was significantly longer in patients with earlier TNM staging and no history of chemotherapy. CONCLUSIONS: BNCT may be an effective treatment option for locally advanced or recurrent unresectable head and neck cancers with no other definitive therapies. If definitive surgery or radiation therapy are not feasible, BNCT should be considered at early disease stages.

Feasibility study of optical imaging of the boron-dose distribution by a liquid scintillator in a clinical boron neutron capture therapy field.

BACKGROUND: Evaluation of the boron dose is essential for boron neutron capture therapy (BNCT). Nevertheless, a direct evaluation method for the boron-dose distribution has not yet been established in the clinical BNCT field. To date, even in quality assurance (QA) measurements, the boron dose has been indirectly evaluated from the thermal neutron flux measured using the activation method with gold foil or wire and an assumed boron concentration in the QA procedure. Recently, we successfully conducted optical imaging of the boron-dose distribution using a cooled charge-coupled device (CCD) camera and a boron-added liquid scintillator at the E-3 port facility of the Kyoto University Research Reactor (KUR), which supplies an almost pure thermal neutron beam with very low gamma-ray contamination. However, in a clinical accelerator-based BNCT facility, there is a concern that the boron-dose distribution may not be accurately extracted because the unwanted luminescence intensity, which is irrelevant to the boron dose is expected to increase owing to the contamination of fast neutrons and gamma rays. PURPOSE: The purpose of this research was to study the validity of a newly proposed method using a boron-added liquid scintillator and a cooled CCD camera to directly observe the boron-dose distribution in a clinical accelerator-based BNCT field. METHOD: A liquid scintillator phantom with 10 B was prepared by filling a small quartz glass container with a commercial liquid scintillator and boron-containing material (trimethyl borate); its natural boron concentration was 1 wt%. Luminescence images of the boron-neutron capture reaction were obtained in a water tank at several different depths using a CCD camera. The contribution of background luminescence, mainly due to gamma rays, was removed by subtracting the luminescence images obtained using another sole liquid scintillator phantom (natural boron concentration of 0 wt%) at each corresponding depth, and a depth profile of the boron dose with several discrete points was obtained. The obtained depth profile was compared with that of calculated boron dose, and those of thermal neutron flux which were experimentally measured or calculated using a Monte Carlo code. RESULTS: The depth profile evaluated from the subtracted images indicated reasonable agreement with the calculated boron-dose profile and thermal neutron flux profiles, except for the shallow region. This discrepancy is thought to be due to the contribution of light reflected from the tank wall. The simulation results also demonstrated that the thermal neutron flux would be severely perturbed by the 10 B-containing phantom if a relatively larger container was used to evaluate a wide range of boron-dose distributions in a single shot. This indicates a trade-off between the luminescence intensity of the 10 B-added phantom and its perturbation effect on the thermal neutron flux. CONCLUSIONS: Although a partial discrepancy was observed, the validity of the newly proposed boron-dose evaluation method using liquid-scintillator phantoms with and without 10 B was experimentally confirmed in the neutron field of an accelerator-based clinical BNCT facility. However, this study has some limitations, including the trade-off problem stated above. Therefore, further studies are required to address these limitations.

Out-of-field dosimetry using a validated PHITS model and computational phantom in clinical BNCT.

BACKGROUND: The out-of-field radiation dose for boron neutron capture therapy (BNCT), which results from both neutrons and γ-rays, has not been extensively evaluated. To safely perform BNCT, the neutron and γ-ray distributions inside the treatment room and the whole-body dose should be evaluated during commissioning. Although, certain previous studies have evaluated the whole-body dose in the clinical research phase, no institution providing BNCT covered by health insurance has yet validated the neutron distribution inside the room and the whole-body dose. PURPOSE: To validate the Monte Carlo model of the BNCT irradiation room extended for the whole-body region and evaluate organ-at-risk (OAR) doses using the validated model with a human-body phantom. METHODS: First, thermal neutron distribution inside the entire treatment room was measured by placing Au samples on the walls of the treatment room. Second, neutron and gamma-ray dose-rate distributions inside a human-body water phantom were measured. Both lying and sitting positions were considered. Bare Au, Au covered by Cd (Au+Cd), In, Al, and thermoluminescent dosimeters were arranged at 11 points corresponding to locations of the OARs inside the phantom. After the irradiation, γ-ray peaks emitted from the samples were measured by a high-purity germanium detector. The measured counts were converted to the reaction rate per unit charge of the sample. These measurements were compared with results of simulations performed with the Particle and Heavy Ion Transport code System (PHITS). A male adult mesh-type reference computational phantom was used to evaluate OAR doses in the whole-body region. The relative biological effectiveness (RBE)-weighted doses and dose-volume histograms (DVHs) for each OAR were evaluated. The median dose (D50% ) and near-maximum dose (D2% ) were evaluated for 14 OARs in a 1-h-irradiation process. The evaluated RBE-weighted doses were converted to equivalent doses in 2 Gy fractions. RESULTS: Experimental results within 60 cm from the irradiation center agreed with simulation results within the error bars except at ±20, 30 cm, and those over 70 cm corresponded within one digit. The experimental results of reaction rates or γ-ray dose rate for lying and sitting positions agreed well with the simulation results within the error bars at 8, 4, 11, 7 and 7, 4, 7, 6, 5, 6 out of 11 points, respectively, for Au, Au+Cd, In, Al, and TLD. Among the detectors, the discrepancies in reaction rates between experiment and simulation were most common for Au+Cd, but were observed randomly for measurement points (brain, lung, etc.). The experimental results of γ-ray dose rates were systematically lower than simulation results at abdomen and waist regions for both positions. Extending the PHITS model to the whole-body region resulted in higher doses for all OARs, especially 0.13 Gy-eq increase for D50% of the left salivary gland. CONCLUSION: The PHITS model for clinical BNCT for the whole-body region was validated, and the OAR doses were then evaluated. Clinicians and medical physicists should know that the out-of-field radiation increases the OAR dose in the whole-body region.

Automatic measurement of beam-positioning accuracy at off-isocenter positions.

PURPOSE: This study performed an automatic measurement of the off-axis beam-positioning accuracy at a single isocenter via the TrueBeam Developer mode and evaluated the beam-positioning accuracy considering the effect of couch rotational errors. METHODS: TrueBeam STx and the Winston-Lutz test-dedicated phantom, with a 3 mm diameter steel ball, were used in this study. The phantom was placed on the treatment couch, and the Winston-Lutz test was performed at the isocenter for four gantry angles (0°, 90°, 180°, and 270°) using an electronic portal imaging device. The phantom offset positions were at distances of 0, 25, 50, 75, and 100 mm from the isocenter along the superior-inferior, anterior-posterior, and left-right directions. Seventeen patterns of multileaf collimator-shaped square fields of 10 × 10 mm2 were created at the isocenter and off-axis positions for each gantry angle. The beam-positioning accuracy was evaluated with couch rotation along the yaw-axis (0°, ± 0.5°, and ± 1.0°). RESULTS: The mean beam-positioning errors at the isocenter and off-isocenter distances (from the isocenter to ±100 mm) were 0.46-0.60, 0.44-0.91, and 0.42-1.11 mm for the couch angles of 0°, ±0.5°, and ±1°, respectively. The beam-positioning errors increased as the distance from the isocenter and couch rotation increased. CONCLUSION: These findings suggest that the beam-positioning accuracy at the isocenter and off-isocenter positions can be evaluated quickly and automatically using the TrueBeam Developer mode. The proposed procedure is expected to contribute to an efficient evaluation of the beam-positioning accuracy at off-isocenter positions.

Exploration of the threshold SUV for diagnosis of malignancy using 18F-FBPA PET/CT.

BACKGROUND: The goal of the study was to evaluate the diagnostic ability of 18F-FBPA PET/CT for malignant tumors. Findings from 18F-FBPA and 18F-FDG PET/CT were compared with pathological diagnoses in patients with malignant tumors or benign lesions. METHODS: A total of 82 patients (45 males, 37 females; median age, 63 years; age range, 20-89 years) with various types of malignant tumors or benign lesions, such as inflammation and granulomas, were examined by 18F-FDG and 18F-FBPA PET/CT. Tumor uptake of FDG or FBPA was quantified using the maximum standardized uptake value (SUVmax). The final diagnosis was confirmed by cytopathology or histopathological findings of the specimen after biopsy or surgery. A ROC curve was constructed from the SUVmax values of each PET image, and the area under the curve (AUC) and cutoff values were calculated. RESULTS: The SUVmax for 18F-FDG PET/CT did not differ significantly for malignant tumors and benign lesions (10.9 ± 6.3 vs. 9.1 ± 2.7 P = 0.62), whereas SUVmax for 18F-FBPA PET/CT was significantly higher for malignant tumors (5.1 ± 3.0 vs. 2.9 ± 0.6, P < 0.001). The best SUVmax cutoffs for distinguishing malignant tumors from benign lesions were 11.16 for 18F-FDG PET/CT (sensitivity 0.909, specificity 0.390) and 3.24 for 18F-FBPA PET/CT (sensitivity 0.818, specificity 0.753). ROC analysis showed significantly different AUC values for 18F-FDG and 18F-FBPA PET/CT (0.547 vs. 0.834, p < 0.001). CONCLUSION: 18F-FBPA PET/CT showed superior diagnostic ability over 18F-FDG PET/CT in differential diagnosis of malignant tumors and benign lesions. The results of this study suggest that 18F-FBPA PET/CT diagnosis may reduce false-positive 18F-FDG PET/CT diagnoses.

Improvement in the neutron beam collimation for application in boron neutron capture therapy of the head and neck region.

In June 2020, the Japanese government approved boron neutron capture therapy for the treatment of head and neck cancer. The treatment is usually performed in a single fraction, with the neutron irradiation time being approximately 30-60 min. As neutrons scatter in air and loses its intensity, it is preferable to bring the patient as close to the beam port as possible to shorten the irradiation time. However, this can be a challenge, especially for patients with head and neck cancer, as the shoulders are an obstacle to a clean positioning. In this study, a novel neutron collimation system for an accelerator based neutron source was designed to allow for a more comfortable treatment, without compromising the irradiation time. Experimental measurements confirmed the simulation results and showed the new collimator can reduce the irradiation time by approximately 60% (under the same condition where the distance between the source and the patient surface was kept the same). The dose delivered to the surrounding healthy tissue was reduced with the new collimator, showing a 25% decrease in the D50 of the mucosal membrane. Overall, the use of the newly designed collimator will allow for a more comfortable treatment of the head and neck region, reduce the treatment time, and reduce the dose delivered to the surrounding healthy tissue.

Comprehensive evaluation of dosimetric impact against position errors in accelerator-based BNCT under different treatment parameter settings.

BACKGROUND: Patients who undergo accelerator-based (AB) boron neutron capture therapy (BNCT) for head and neck cancer in the sitting position are generally uncomfortably immobilized, and patient motion during this treatment may be greater than that in other radiotherapy techniques. Furthermore, the treatment time of BNCT is relatively long (up to approximately 1 h), which increases the possibility of patient movement during treatment. As most BNCT irradiations are performed in a single fraction, the dosimetric error due to patient motion is of greater consequence and needs to be evaluated and accounted for. Several treatment parameters are required for BNCT dose calculation. PURPOSE: To investigate the dosimetric impacts (DIs) against position errors using a simple cylindrical phantom for an AB-BNCT system under different treatment parameter settings. METHODS: The treatment plans were created in RayStation and the dose calculation was performed using the NeuCure® dose engine. A cylindrical phantom (16 cm diameter × 20 cm height) made of soft tissue was modeled. Dummy tumors in the form of a 3-cm-diameter sphere were arranged at depths of 2.5 and 6.5 cm (denoted by T2.5 and T6.5 , respectively). Reference plans were created by setting the following parameters: collimator size = 10, 12, or 15 cm in diameter, collimator-to-surface distance (CSD) = 4.0 or 8.0 cm, tumor-to-blood ratio (T/B ratio) using 18 F-fluoro-borono-phenylalanine = 2.5 or 5.0, and 10 B concentration in blood = 20, 25, or 30 ppm. The prescribed dose was D95%  ≥ 20 Gy-eq for both T2.5 and T6.5 . Based on the reference plans, phantom-shifted plans were created in 26 directions [all combinations of left-right (LR), anterior-posterior (AP), and superior-inferior (SI) directions) and three distances (1.0, 2.0, and 3.0 cm). The DIs were evaluated at D80% of the tumors. The shift direction dependency of the DI in the LR, AP, and SI directions was evaluated by conducting a multiple regression analysis (MRA) and other analyses where required. RESULTS: The coefficients of the MRA of the DIs for LR, AP, and SI shifts were -0.08, 2.16, and -0.04 (p-values = 0.084, <0.01, and 0.334) for T2.5 and -0.05, 2.08, and 0.15 (p-values = 0.526, <0.01, and 0.065) for T6.5 , respectively. The analysis of variance showed that DIs due to the AP shift were significantly greater for smaller collimator sizes on T2.5 and smaller CSD on T6.5 . Dose reduction due to SI or LR (lateral) shifts was significantly greater for smaller collimator sizes on both T2.5 and T6.5 and smaller CSD on T2.5 , according to the Student's t-test. There were no significant differences in the DIs against both the AP shift and the lateral shift between the different T/B ratios and 10 B concentrations. CONCLUSION: The DIs were largely affected by the shift in the AP direction and were influenced by the different treatment parameters.

Evaluation of internal margins for prostate for step and shoot intensity-modulated radiation therapy and volumetric modulated arc therapy using different margin formulas.

PURPOSE: This feasibility study evaluated the intra-fractional prostate motion using an ultrasound image-guided system during step and shoot intensity-modulated radiation therapy (SS-IMRT) and volumetric modulated arc therapy (VMAT). Moreover, the internal margins (IMs) using different margin formulas were calculated. METHODS: Fourteen consecutive patients with prostate cancer who underwent SS-IMRT (n = 5) or VMAT (n = 9) between March 2019 and April 2020 were considered. The intra-fractional prostate motion was observed in the superior-inferior (SI), anterior-posterior (AP), and left-right (LR) directions. The displacement of the prostate was defined as the displacement from the initial position at the scanning start time, which was evaluated using the mean ± standard deviation (SD). IMs were calculated using the van Herk and restricted maximum likelihood (REML) formulas for SS-IMRT and VMAT. RESULTS: For SS-IMRT, the maximum displacements of the prostate motion were 0.17 ± 0.18, 0.56 ± 0.86, and 0.18 ± 0.59 mm in the SI, AP, and LR directions, respectively. For VMAT, the maximum displacements of the prostate motion were 0.19 ± 0.64, 0.22 ± 0.35, and 0.14 ± 0.37 mm in the SI, AP, and LR directions, respectively. The IMs obtained for SS-IMRT and VMAT were within 2.3 mm and 1.2 mm using the van Herk formula and within 1.2 mm and 0.8 mm using the REML formula. CONCLUSIONS: This feasibility study confirmed that intra-fractional prostate motion was observed with SS-IMRT and VMAT using different margin formulas. The IMs should be determined according to each irradiation technique using the REML margin.

Dosiomic feature comparison between dose-calculation algorithms used for lung stereotactic body radiation therapy.

To evaluate the reproducibility of dose-based radiomic (dosiomic) features between dose-calculation algorithms for lung stereotactic body radiation therapy (SBRT). We analyzed 105 patients with early-stage non-small cell lung cancer who underwent lung SBRT between March 2011 and December 2017. Radiation doses of 48, 60, and 70 Gy were prescribed to the isocenter in 4-8 fractions. Dose calculations were performed using X-ray voxel Monte Carlo (XVMC) on the iPlan radiation treatment planning system (RTPS). Thereafter, the radiation doses were recalculated using the Acuros XB (AXB) and analytical anisotropic algorithm (AAA) on the Eclipse RTPS while maintaining the XVMC-calculated monitor units and beam arrangements. A total of 6808 dosiomic features were extracted without preprocessing (112 shape, 144 first-order, and 600 texture features) or with wavelet filters to eight decompositions (1152 first-order and 4800 texture features). Features with absolute pairwise concordance correlation coefficients-|CCcon|-values exceeding or equaling 0.85 were considered highly reproducible. Subgroup analyses were performed considering the wavelet filters and prescribed doses. The numbers of highly reproducible first-order and texture features were 34.8%, 26.9%, and 31.0% for the XVMC-AXB, XVMC-AAA, and AXB-AAA pairs, respectively. The maximum difference between the mean |CCcon| values was 0.70 and 0.11 for the subgroup analyses of wavelet filters and prescribed dose, respectively. The application of wavelet filter-based dosiomic analyses may be limited when using different types of dose-calculation algorithms for lung SBRT.

Vulnerabilities of radiomic features to respiratory motion on four-dimensional computed tomography-based average intensity projection images: A phantom study.

PURPOSE: To evaluate the influence of respiratory motion on the robustness of radiomic features on four-dimensional computed tomography (4DCT)-based average intensity projection (AIP) images by employing an anthropomorphic chest phantom. METHODS: Three spherical objects (φ30 mm), namely, acrylic (100 Hounsfield unit [HU], homogeneous), rubber (-140 HU, homogeneous), and cork (-630 HU, heterogeneous), were moved with motion amplitudes of 0, 1, 2.5, 4, 6, 8, and 10 mm in the phantom, and 4DCT scans were repeated at four different locations. Thereafter, the AIP images were generated considering the average of the 10 respiratory phases of the 4DCT images. Further, the targets were manually delineated on the AIP images in the lung window setting. A total of 851 radiomic features, including 107 unfiltered features and 744 wavelet filter-based features, were extracted from the region of interest for each material. The feature robustness among the different target motion amplitude (ε) was evaluated by normalizing the feature variability of the target motion relative to the variability of data from 573 patients with early-stage non-small cell lung cancer. The features with absolute ε values ≤0.5 were considered highly robust to target motions. RESULTS: The percentage of robust unfiltered and wavelet filter-based features with a motion amplitude of 1 mm was greater than 83.2% and 93.4%, respectively; however, the percentage decreased by more than 24.3% and 17.6%, respectively, for motion amplitudes greater than 2.5 mm. The movement of cork had a small effect on the feature robustness compared to that of acrylic and rubber, regardless of the target motion amplitudes. CONCLUSIONS: Our phantom study demonstrated that target motion amplitudes ≤1 mm led to the robustness of radiomic features on the 4DCT-based AIP images of thoracic regions. The frequency components and directions of the wavelet filters may be essential factors in 4DCT-based radiomic analysis.

Evaluation of a treatment planning system developed for clinical boron neutron capture therapy and validation against an independent Monte Carlo dose calculation system.

Boron neutron capture therapy (BNCT) for the treatment of unresectable, locally advanced, and recurrent carcinoma of the head and neck cancer has been approved by the Japanese government for reimbursement under the national health insurance as of June 2020. A new treatment planning system for clinical BNCT has been developed by Sumitomo Heavy Industries, Ltd. (Sumitomo), NeuCure® Dose Engine. To safely implement this system for clinical use, the simulated neutron flux and gamma ray dose rate inside a water phantom was compared against experimental measurements. Furthermore, to validate and verify the new planning system, the dose distribution inside an anthropomorphic head phantom was compared against a BNCT treatment planning system SERA and an in-house developed Monte Carlo dose calculation program. The simulated results closely matched the experimental results, within 5% for the thermal neutron flux and 10% for the gamma ray dose rate. The dose distribution inside the head phantom closely matched with SERA and the in-house developed dose calculation program, within 3% for the tumour and a difference of 0.3 Gyw for the brain.

Photoneutron-induced damage reduction for cardiac implantable electronic devices using neutron-shielding sheets in high-energy X-ray radiotherapy: A phantom study.

PURPOSE: To evaluate damage reduction in cardiac implantable electronic devices (CIEDs) caused by photoneutrons in high-energy X-ray radiotherapy using a neutron-shielding sheet (NSS). METHODS: The NSS consists of a bolus with a thickness of 1 or 2 cm (Bls1 or Bls2) as a moderator and several absorbers (20%, 50%, or 80% B4C silicone sheet [B4C20, B4C50, or B4C80] or a 40% LiF silicone sheet [LiF40]). First, a linear accelerator (LINAC) with a water-equivalent phantom was modeled in the simulation and measured experimentally. Several NSSs were placed on the phantom, a Eu:LiCaAlF6 scintillator was placed between the phantom and the NSS, and X-rays were irradiated. The relative counts (Cr = counts when placing the NSS or Bls2) were compared between the experiment and simulation. Second, CIED damage was evaluated in the simulation. The relative damage (Dr = damage when placing or not placing the NSS) was compared among all the NSSs. In addition, the γ-ray and leaking X-ray dose from B4C was measured using a dosimetric film. After determining the optimal NSS combination, Dr value analysis was performed by changing the length of one side and the thickness. RESULTS: The Cr values of the simulation and experiment agreed within a 30% percentage difference, except for Bare or LiF40-only. The Dr value was reduced by 43% when Bls2 + B4C80 was applied. The photon dose was less than 5 cGy/1500 MU. The Dr values were smaller for the smaller lengths of one side of B4C80 and decreased as the M-layer thickness increased. CONCLUSIONS: The CIED damage induced by photoneutrons generated by a LINAC was effectively reduced by applying the optimal NSS.

Multi-institutional dose-segmented dosiomic analysis for predicting radiation pneumonitis after lung stereotactic body radiation therapy.

PURPOSE: To predict radiation pneumonitis (RP) grade 2 or worse after lung stereotactic body radiation therapy (SBRT) using dose-based radiomic (dosiomic) features. METHODS: This multi-institutional study included 247 early-stage nonsmall cell lung cancer patients who underwent SBRT with a prescribed dose of 48-70 Gy at an isocenter between June 2009 and March 2016. Ten dose-volume indices (DVIs) were used, including the mean lung dose, internal target volume size, and percentage of entire lung excluding the internal target volume receiving greater than x Gy (x = 5, 10, 15, 20, 25, 30, 35, and 40). A total of 6,808 dose-segmented dosiomic features, such as shape, first order, and texture features, were extracted from the dose distribution. Patients were randomly partitioned into two groups: model training (70%) and test datasets (30%) over 100 times. Dosiomic features were converted to z-scores (standardized values) with a mean of zero and a standard deviation (SD) of one to put different variables on the same scale. The feature dimension was reduced using the following methods: interfeature correlation based on Spearman's correlation coefficients and feature importance based on a light gradient boosting machine (LightGBM) feature selection function. Three different models were developed using LightGBM as follows: (a) a model with ten DVIs (DVI model), (b) a model with the selected dosiomic features (dosiomic model), and (c) a model with ten DVIs and selected dosiomic features (hybrid model). Suitable hyperparameters were determined by searching the largest average area under the curve (AUC) value in the receiver operating characteristic curve (ROC-AUC) via stratified fivefold cross-validation. Each of the final three models with the closest the ROC-AUC value to the average ROC-AUC value was applied to the test datasets. The classification performance was evaluated by calculating the ROC-AUC, AUC in the precision-recall curve (PR-AUC), accuracy, precision, recall, and f1-score. The entire process was repeated 100 times with randomization, and 100 individual models were developed for each of the three models. Then the mean value and SD for the 100 random iterations were calculated for each performance metric. RESULTS: Thirty-seven (15.0%) patients developed RP after SBRT. The ROC-AUC and PR-AUC values in the DVI, dosiomic, and hybrid models were 0.660 ± 0.054 and 0.272 ± 0.052, 0.837 ± 0.054 and 0.510 ± 0.115, and 0.846 ± 0.049 and 0.531 ± 0.116, respectively. For each performance metric, the dosiomic and hybrid models outperformed the DVI models (P < 0.05). Texture-based dosiomic feature was confirmed as an effective indicator for predicting RP. CONCLUSIONS: Our dose-segmented dosiomic approach improved the prediction of the incidence of RP after SBRT.

Application and limitation of radiomics approach to prognostic prediction for lung stereotactic body radiotherapy using breath-hold CT images with random survival forest: A multi-institutional study.

PURPOSE: To predict local recurrence (LR) and distant metastasis (DM) in early stage non-small cell lung cancer (NSCLC) patients after stereotactic body radiotherapy (SBRT) in multiple institutions using breath-hold computed tomography (CT)-based radiomic features with random survival forest. METHODS: A total of 573 primary early stage NSCLC patients who underwent SBRT between January 2006 and March 2016 and met the eligibility criteria were included in this study. Patients were divided into two datasets: training (464 patients in 10 institutions) and test (109 patients in one institution) datasets. A total of 944 radiomic features were extracted from manually segmented gross tumor volumes (GTVs). Feature selection was performed by analyzing inter-segmentation reproducibility, GTV correlation, and inter-feature redundancy. Nine clinical factors, including histology and GTV size, were also used. Three prognostic models (clinical, radiomic, and combined) for LR and DM were constructed using random survival forest (RSF) to deal with total death as a competing risk in the training dataset. Robust models with optimal hyper-parameters were determined using fivefold cross-validation. The patients were dichotomized into two groups based on the median value of the patient-specific risk scores (high- and low-risk score groups). Gray's test was used to evaluate the statistical significance between the two risk score groups. The prognostic power was evaluated by the concordance index with the 95% confidence intervals (CI) via bootstrapping (2000 iterations). RESULTS: The concordance indices at 3 yr of clinical, radiomic, and combined models for LR were 0.57 [CI: 0.39-0.75], 0.55 [CI: 0.38-0.73], and 0.61 [CI: 0.43-0.78], respectively, whereas those for DM were 0.59 [CI: 0.54-0.79], 0.67 [CI: 0.54-0.79], and 0.68 [CI: 0.55-0.81], respectively, in the test dataset. The combined DM model significantly discriminated its cumulative incidence between high- and low-risk score groups (P < 0.05). The variable importance of RSF in the combined model for DM indicated that two radiomic features were more important than other clinical factors. The feature maps generated on the basis of the most important radiomic feature had visual difference between high- and low-risk score groups. CONCLUSIONS: The radiomics approach with RSF for competing risks using breath-hold CT-based radiomic features might predict DM in early stage NSCLC patients who underwent SBRT although that may not have potential to predict LR.

Comparison of radiomic features in diagnostic CT images with and without contrast enhancement in the delayed phase for NSCLC patients.

PURPOSE: To compare radiomic features extracted from diagnostic computed tomography (CT) images with and without contrast enhancement in delayed phase for non-small cell lung cancer (NSCLC) patients. METHODS: Diagnostic CT images from 269 tumors [non-contrast CT, 188 (dataset NE); contrast-enhanced CT, 81 (dataset CE)] were enrolled in this study. Eighteen first-order and seventy-five texture features were extracted by setting five bin width levels for CT values. Reproducible features were selected by the intraclass correlation coefficient (ICC). Radiomic features were compared between datasets NE and CE. Subgroup analyses were performed based on the CT acquisition period, exposure value, and patient characteristics. RESULTS: Eighty features were considered reproducible (0.5 ≤ ICC). Twelve of the sixteen first-order features, independent of the bin width levels, were statistically different between datasets NE and CE (p < 0.05), and the p-values of two first-order features depending on the bin width levels were reduced with narrower bin widths. Sixteen out of sixty-two features showed a significant difference, regardless of the bin width (p < 0.05). There were significant differences between datasets NE and CE with older age, lighter body weight, better performance status, being a smoker, larger gross tumor volume, and tumor location at central region. CONCLUSIONS: Contrast enhancement in the delayed phase of CT images for NSCLC patients affected some of the radiomic features and the variability of radiomic features due to contrast uptake may depend largely on the patient characteristics.