Radiomics-based hybrid model for predicting radiation pneumonitis: A systematic review and meta-analysis.

PURPOSE: This study reviewed and meta-analyzed evidence on radiomics-based hybrid models for predicting radiation pneumonitis (RP). These models are crucial for improving thoracic radiotherapy plans and mitigating RP, a common complication of thoracic radiotherapy. We examined and compared the RP prediction models developed in these studies with the radiomics features employed in RP models. METHODS: We systematically searched Google Scholar, Embase, PubMed, and MEDLINE for studies published up to April 19, 2024. Sixteen studies met the inclusion criteria. We compared the RP prediction models developed in these studies and the radiomics features employed. RESULTS: Radiomics, as a single-factor evaluation, achieved an area under the receiver operating characteristic curve (AUROC) of 0.73, accuracy of 0.69, sensitivity of 0.64, and specificity of 0.74. Dosiomics achieved an AUROC of 0.70. Clinical and dosimetric factors showed lower performance, with AUROCs of 0.59 and 0.58. Combining clinical and radiomic factors yielded an AUROC of 0.78, while combining dosiomic and radiomics factors produced an AUROC of 0.81. Triple combinations, including clinical, dosimetric, and radiomics factors, achieved an AUROC of 0.81. The study identifies key radiomics features, such as the Gray Level Co-occurrence Matrix (GLCM) and Gray Level Size Zone Matrix (GLSZM), which enhance the predictive accuracy of RP models. CONCLUSIONS: Radiomics-based hybrid models are highly effective in predicting RP. These models, combining traditional predictive factors with radiomic features, particularly GLCM and GLSZM, offer a clinically feasible approach for identifying patients at higher RP risk. This approach enhances clinical outcomes and improves patient quality of life. PROTOCOL REGISTRATION: The protocol of this study was registered on PROSPERO (CRD42023426565).

FLASH dose rate calculation based on log files in proton pencil beam scanning therapy.

BACKGROUND: In radiation therapy, irradiating healthy normal tissues in the beam trajectories is inevitable. This unnecessary dose means that patients undergoing treatment risk developing side effects. Recently, FLASH radiotherapy delivering ultra-high-dose-rate beams has been re-examined because of its normal-tissue-sparing effect. To confirm the mean and instantaneous dose rates of the FLASH beam, stable and accurate dosimetry is required. PURPOSE: Detailed verification of the FLASH effect requires dosimeters and a method to measure the average and instantaneous dose rate stably for 2- or 3-dimensional dose distributions. To verify the delivered FLASH beam, we utilized machine log files from the built-in monitor chamber to develop a dosimetry method to calculate the dose and average/instantaneous dose rate distributions in two or three dimensions in a phantom. METHODS: To create a spread-out Bragg peak (SOBP) and deliver a uniform dose in a target, a mini-ridge filter was created with a 3D printer. Proton pencil beam line scanning plans of 2 × 2 cm2 , 3 × 3 cm2 , 4 × 4 cm2 , and round shapes with 2.3 cm diameter patterns delivering 230 MeV energy protons were created. The absorbed dose in the solid water phantom of each plan was measured using a PPC05 ionization chamber (IBA Dosimetry, Virginia, USA) in the SOBP region, and the log files for each plan were exported from the treatment control system console. Using these log files, the delivered dose and average dose rate were calculated using two methods: a direct method and a Monte Carlo (MC) simulation method that uses log file information. The computed and average dose rates were compared with the ionization chamber measurements. Additionally, instantaneous dose rates in user-defined volumes were calculated using the MC simulation method with a temporal resolution of 5 ms. RESULTS: Compared to ionization chamber dosimetry, 10 of 12 cases using the direct calculation method and 9 of 11 cases using the MC method had a dose difference below ±3%. Nine of 12 cases using the direct calculation method and 8 of 11 cases using the MC method had dose rate differences below ±3%. The average and maximum dose differences for the direct calculation and MC method were-0.17, +0.72%, and -3.15, +3.32%, respectively. For the dose rate difference, the average and maximum for the direct calculation and MC method were +1.26, +1.12%, and +3.75, +3.15%, respectively. In the instantaneous dose rate calculation with the MC simulation, a large fluctuation with a maximum of 163 Gy/s and a minimum of 4.29 Gy/s instantaneous dose rate was observed in a specific position, whereas the mean dose rate was 62 Gy/s. CONCLUSIONS: We successfully developed methods in which machine log files are used to calculate the dose and the average and instantaneous dose rates for FLASH radiotherapy and demonstrated the feasibility of verifying the delivered FLASH beams.

Application of error classification model using indices based on dose distribution for characteristics evaluation of multileaf collimator position errors.

This study aims to evaluate the specific characteristics of various multileaf collimator (MLC) position errors that are correlated with the indices using dose distribution. The dose distribution was investigated using the gamma, structural similarity, and dosiomics indices. Cases from the American Association of Physicists in Medicine Task Group 119 were planned, and systematic and random MLC position errors were simulated. The indices were obtained from distribution maps and statistically significant indices were selected. The final model was determined when all values of the area under the curve, accuracy, precision, sensitivity, and specificity were higher than 0.8 (p < 0.05). The dose-volume histogram (DVH) relative percentage difference between the error-free and error datasets was examined to investigate clinical relations. Seven multivariate predictive models were finalized. The common significant dosiomics indices (GLCM Energy and GLRLM_LRHGE) can characterize the MLC position error. In addition, the finalized logistic regression model for MLC position error prediction showed excellent performance with AUC > 0.9. Furthermore, the results of the DVH were related to dosiomics analysis in that it reflects the characteristics of the MLC position error. It was also shown that dosiomics analysis could provide important information on localized dose-distribution differences in addition to DVH information.

Novel framework for determining TPS-calculated doses corresponding to detector locations using 3D camera inin vivosurface dosimetry.

Purpose. To address the shortcomings of current procedures for evaluating the measured-to-planned dose agreement inin vivodosimetry (IVD), this study aimed to develop an accurate and efficient novel framework to identify the detector location placed on a patient's skin surface using a 3D camera and determine the planned dose at the same anatomical position corresponding to the detector location.Methods. Breast cancer treatment was simulated using an anthropomorphic adult female phantom (ATOM 702D; CIRS, Norfolk, VA, USA). An optically stimulated luminescent dosimeter was used for surface dose measurements (MyOSLchip, RadPro International GmbH, Germany) at six IVD points. Three-dimensional surface imaging (3DSI) of the phantom with the detector was performed in the treatment position using a 3D camera. The developed framework, iSMART, was designed to import 3DSI and treatment planning data for determining the position of the IVD detectors in the 3D treatment planning DICOM image. The clinical usefulness of iSMART was evaluated in terms of accuracy and efficiency, for comparison with the results obtained using cone-beam computed tomography (CBCT) image guidance.Results. The relative dose difference between the planned doses determined using iSMART and CBCT images displayed similar accuracies (within approximately ±2.0%) at all detector locations. The relative dose differences between the planned and measured doses at the six detector locations ranged from -4.8% to 3.1% for the CBCT images and -3.5% to 2.1% for iSMART. The total time required to read the planned doses at six detector locations averaged at 8.1 and 0.8 min for the CBCT images and iSMART, respectively.Conclusions. The proposed framework can improve the robustness of IVD analyses and aid in accurate and efficient evaluations of the measured-to-planned dose agreement.

Monitor unit prediction model for wobbling proton therapy with ridge filters.

PURPOSE: We introduced an output factor (cGy/MU) prediction model for wobbling proton beams over the full range of proton energy, scatterer thickness, and the width of spread-out Bragg peak (SOBP). MATERIALS AND METHODS: From December 2015 to August 2020, 1990 wobbling proton fields were used to treat patients, where 1714 fields had a diameter smaller than 11 cm and 276 had a diameter between 11 and 16 cm, which were designated as small and middle wobbling radius cases, respectively. The output factor is defined as the ratio of proton absorbed dose at mid-depth of SOBP to monitor unit (MU). It depends dominantly on proton energy, scatterer thickness, and the width of SOBP. We established the prediction model using the polynomial fitting function and determined its coefficients for the small and middle wobbling radius cases. We evaluated the accuracy of our prediction model by calculating the difference between predicted and measured output factors. RESULTS: For the small wobbling radius cases, the mean value of the output factor difference was 0.22% with a standard deviation of 1.3%. For the middle wobbling radius cases, the mean value was 0.20% and with a standard deviation of 0.79%. The large deviation was especially observed for wobbling proton beams having small field size and small width of SOBP. CONCLUSIONS: We made a prediction model of output factor for wobbling proton beams, thereby determining MU of each beam. This included the dependency of the output factor on the proton energy between 70 and 230 MeV, scatterer thickness, and the width of SOBP. For 93.6% of the small and 95.5% of the middle wobbling radius cases, the deviation between predicted and measured output factor was below 3%. The cases with deviations of predicted and measured output factor above 3% had small field size and small width of SOBP. The accuracy of our prediction model would be improved by adopting the field size effect and measuring more cases of small field size and small SOBP width in the future.

A radiomics nomogram for predicting transcatheter arterial chemoembolization refractoriness of hepatocellular carcinoma without extrahepatic metastasis or macrovascular invasion.

OBJECTIVE: A radiomics nomogram for pretreatment prediction of TACE refractoriness was developed and validated for hepatocellular carcinoma (HCC) without extrahepatic metastasis or macrovascular invasion. MATERIALS AND METHODS: This study included 80 patients with HCC without extrahepatic metastasis or macrovascular involvement treated with TACE between July 2016 and November 2018. The datasets were divided into a training set (80%) and a test set (20%) for feature selection and tenfold cross-validation. Forty radiomic features were extracted from arterial-phase computed tomography (CT) using the Local Image Features Extraction software. The Lasso regression model was used for radiomics signature selection. The Lasso regression model was used for radiomics signature selection and the selected signatures were validated using the Mann-Whitney U-test. The radiomics nomogram was developed based on a multivariate logistic regression model incorporating the Rad-score, CT imaging factors, and clinical factors, and it was validated. RESULTS: The Rad-score, which consists of the Gray-Level Zone Length Matrix (GLZLM)-Long-Zone Low Gray-Level Emphasis (LZLGE) and GLZLM-Gray-Level Non-Uniformity (GLNU), T-stage, log α-fetoprotein (AFP), and bilobar distribution were significantly associated with TACE refractoriness (p < 0.05). Predictors in the radiomics nomogram were the Rad-score and T-stage (Rad-score + T-stage), Rad-score and bilobar distribution (Rad-score + bilobar distribution), or Rad-score and logAFP (Rad-score + logAFP). The multivariate logistic regression model showed a good predictive performance (Rad-score + T-stage, AUC, 0.95; Rad-score + bilobar distribution, AUC 0.91; and Rad-score + logAFP, AUC, 0.91). CONCLUSION: The radiomics nomogram could be used for the pretreatment prediction of TACE refractoriness.

Metastasis risk prediction model in osteosarcoma using metabolic imaging phenotypes: A multivariable radiomics model.

BACKGROUND: Osteosarcoma (OS) is the most common primary bone tumor affecting humans and it has extreme heterogeneity. Despite modern therapy, it recurs in approximately 30-40% of patients initially diagnosed with no metastatic disease, with the long-term survival rates of patients with recurrent OS being generally 20%. Thus, early prediction of metastases in OS management plans is crucial for better-adapted treatments and survival rates. In this study, a radiomics model for metastasis risk prediction in OS was developed and evaluated using metabolic imaging phenotypes. METHODS AND FINDINGS: The subjects were eighty-three patients with OS, and all were treated with surgery and chemotherapy for local control. All patients underwent a pretreatment 18F-FDG-PET scan. Forty-five features were extracted from the tumor region. The incorporation of features into multivariable models was performed using logistic regression. The multivariable modeling strategy involved cross validation in the following four steps leading to final prediction model construction: (1) feature set reduction and selection; (2) model coefficients computation through train and validation processing; and (3) prediction performance estimation. The multivariable logistic regression model was developed using two radiomics features, SUVmax and GLZLM-SZLGE. The trained and validated multivariable logistic model based on probability of endpoint (P) = 1/ (1+exp (-Z)) was Z = -1.23 + 1.53*SUVmax + 1.68*GLZLM-SZLGE with significant p-values (SUVmax: 0.0462 and GLZLM_SZLGE: 0.0154). The final multivariable logistic model achieved an area under the curve (AUC) receiver operating characteristics (ROC) curve of 0.80, a sensitivity of 0.66, and a specificity of 0.88 in cross validation. CONCLUSIONS: The SUVmax and GLZLM-SZLGE from metabolic imaging phenotypes are independent predictors of metastasis risk assessment. They show the association between 18F-FDG-PET and metastatic colonization knowledge. The multivariable model developed using them could improve patient outcomes by allowing aggressive treatment in patients with high metastasis risk.

Digital Imaging and Communications in Medicine (DICOM) information conversion procedure for SUV calculation of PET scanners with different DICOM header information.

PURPOSE: In nuclear medicine, the standardized uptake value (SUV) obtained using positron emission tomography with 2-deoxy-2-fluoro-D-glucose (FDG-PET) is widely used as a semi-quantitative diagnosis factor. We found that the header file of the Philips Allegro PET scanner using the Digital Imaging and Communications in Medicine (DICOM) standard was stored differently than with other scanners. Thus, the purpose of this study was to develop a DICOM header information conversion program to ensure compatibility between Allegro and other equipment. METHODS AND RESULTS: The NEMA IEC Body phantom was scanned using the Allegro PET scanner. We conducted measurements and performed calculations by using commercial software and the proposed self-developed program, respectively, to compare the SUVs by using conversion data. The program consists of three parts: an input part that can load data regardless of the number of DICOM images, and conversion and output parts that can be used to convert the DICOM header information and store it in the order of slices. The results of the calculation are in good agreement with the data measured at 12 circular regions of interest. The percent difference was lower than the 20%. CONCLUSION: In conclusion, this study suggested a simple and convenient method to solve the incompatibility through conversion of the DICOM header information. This study thus provides physicians more accurate information for diagnosis and treatment.

Development of a real-time internal and external marker tracking system for particle therapy: a phantom study using patient tumor trajectory data.

Target motion-induced uncertainty in particle therapy is more complicated than that in X-ray therapy, requiring more accurate motion management. Therefore, a hybrid motion-tracking system that can track internal tumor motion and as well as an external surrogate of tumor motion was developed. Recently, many correlation tests between internal and external markers in X-ray therapy have been developed; however, the accuracy of such internal/external marker tracking systems, especially in particle therapy, has not yet been sufficiently tested. In this article, the process of installing an in-house hybrid internal/external motion-tracking system is described and the accuracy level of tracking system was acquired. Our results demonstrated that the developed in-house external/internal combined tracking system has submillimeter accuracy, and can be clinically used as a particle therapy system as well as a simulation system for moving tumor treatment.