Multiomics-Based Outcome Prediction in Personalized Ultra-Fractionated Stereotactic Adaptive Radiotherapy (PULSAR).

Objectives: This retrospective study aims to develop a multiomics approach that integrates radiomics, dosiomics, and delta features to predict treatment responses in brain metastasis (BM) patients undergoing PULSAR. Methods: A retrospective study encompassing 39 BM patients with 69 lesions treated with PULSAR was undertaken. Radiomics, dosiomics, and delta features were extracted from both pre-treatment and intra-treatment MRI scans alongside dose distributions. Six individual models, alongside an ensemble feature selection (EFS) model, were evaluated. The classification task focused on distinguishing between two lesion groups based on whether they exhibited a volume reduction of more than 20% at follow-up. Performance metrics, including sensitivity, specificity, accuracy, precision, F1 score, and the area under the receiver operating characteristic (ROC) curve (AUC), were assessed. Results: The EFS model integrated the features from pre-treatment radiomics, pre-treatment dosiomics, intra-treatment radiomics, and delta radiomics. It outperformed six individual models, achieving an AUC of 0.979, accuracy of 0.917, and F1 score of 0.821. Among the top nine features of the EFS model, six features came from post-wavelet transformation and three from original images. Conclusions: The study demonstrated the feasibility of employing a data-driven multiomics approach to predict treatment outcomes in BM patients receiving PULSAR treatment. Integrating multiomics with intra-treatment decision support in PULSAR shows promise for optimizing patient management and reducing the risks of under- or over-treatment.

Development of a multi-feature-combined model: proof-of-concept with application to local failure prediction of post-SBRT or surgery early-stage NSCLC patients.

Objective: To develop a Multi-Feature-Combined (MFC) model for proof-of-concept in predicting local failure (LR) in NSCLC patients after surgery or SBRT using pre-treatment CT images. This MFC model combines handcrafted radiomic features, deep radiomic features, and patient demographic information in an integrated machine learning workflow. Methods: The MFC model comprised three key steps. (1) Extraction of 92 handcrafted radiomic features from the GTV segmented on pre-treatment CT images. (2) Extraction of 512 deep radiomic features from pre-trained U-Net encoder. (3) The extracted handcrafted radiomic features, deep radiomic features, along with 4 patient demographic information (i.e., gender, age, tumor volume, and Charlson comorbidity index), were concatenated as a multi-dimensional input to the classifiers for LR prediction. Two NSCLC patient cohorts from our institution were investigated: (1) the surgery cohort includes 83 patients with segmentectomy or wedge resection (7 LR), and (2) the SBRT cohort includes 84 patients with lung SBRT (9 LR). The MFC model was developed and evaluated independently for both cohorts, and was subsequently compared against the prediction models based on only handcrafted radiomic features (R models), patient demographic information (PI models), and deep learning modeling (DL models). ROC with AUC was adopted to evaluate model performance with leave-one-out cross-validation (LOOCV) and 100-fold Monte Carlo random validation (MCRV). The t-test was performed to identify the statistically significant differences. Results: In LOOCV, the AUC range (surgery/SBRT) of the MFC model was 0.858-0.895/0.868-0.913, which was higher than the three other models: 0.356-0.480/0.322-0.650 for PI models, 0.559-0.618/0.639-0.682 for R models, and 0.809/0.843 for DL models. In 100-fold MCRV, the MFC model again showed the highest AUC results (surgery/SBRT): 0.742-0.825/0.888-0.920, which were significantly higher than PI models: 0.464-0.564/0.538-0.628, R models: 0.557-0.652/0.551-0.732, and DL models: 0.702/0.791. Conclusion: We successfully developed an MFC model that combines feature information from multiple sources for proof-of-concept prediction of LR in patients with surgical and SBRT early-stage NSCLC. Initial results suggested that incorporating pre-treatment patient information from multiple sources improves the ability to predict the risk of local failure.

Design and fabrication of a personalized anthropomorphic phantom using 3D printing and tissue equivalent materials.

To fabricate an individualized anthropomorphic lung phantom with tissue-equivalent radiation attenuation properties using a cost-effective three-dimensional (3D) printing technique. Based on anonymized human chest CT images, the phantom contained a 3D-printed skin shell, filled with tissue equivalent materials with similar radiation attenuation characteristics. The filling materials were a mixture of CaCO3, MgO, agarose, NaCl, pearl powder and silica gel. The dose calculation accuracy of different treatment planning system (TPS) algorithms was validated and compared with the ion chamber measurements in the phantom, including tumor and surrounding normal tissues. The chest phantom was shown to represent a human's chest in terms of radiation attenuation property and human anatomy. The Hounsfield unit ranges were -60 to -100, 20 to 60, and 120 to 300 for fat, muscle, and bone, respectively. The actual measured values of the ionization chamber were 213.7 cGy for the tumor, 53.85 cGy for normal lung tissue, and 4.1 cGy for the spinal cord, compared to 214.1, 55.2, and 4.5 cGy, respectively, with use of the Monte Carlo algorithm in TPS. The application of 3D printing in anthropomorphic phantoms can improve personalized medical need and efficiency with reduce costs thus, can be used for radiation dose verification.

Fabrication of an anthropomorphic heterogeneous mouse phantom for multimodality medical imaging.

This work presents a comprehensive methodology for constructing a tissue equivalent mouse phantom using image modeling and 3D printing technology. The phantom can be used in multimodality imaging and irradiation experiments, quality control, and management. Computed tomography (CT) images of a mouse were acquired and imported into 3D modeling software. A skeleton and skin shell models were segmented in the modeling software and manufactured using 3D printing technology. The bone model was constructed with VERO-WHITE printing material with additional ingredients, including a photosensitive resin, polyurethane epoxy resin, and acrylate. Acrylonitrile butadiene styrene resin material was used to construct the skin shell. The skin shell was attached to the skeleton and filled with a specially formulated gel to act as a soft tissue substitute. The gel consisted of agarose, micro-pearl powder, sodium chloride, and magnevist solution (gadopentetate dimeglumine). A micro-container filled with 18F-fluorodeoxyglucose (18F-FDG) radioactive tracer was placed in the abdomen for micro and human positron emission tomography (PET)/CT imaging. The mouse phantom had tissue equivalency in dose attenuation with x-rays and relaxation times with magnetic resonance imaging (MRI). The CT Hounsfield Unit (HU) range for the gel soft tissue material was 31-36 HU. The 3D printed bone mimetic material had equivalent tissue/bone contrast compared with in vivo mouse measurements with a mean value of 130 ± 10 HU. At different magnetic field strengths, the T 1 relaxation time of the soft tissue was 382.75-506.48 ms, and T 2 was 51.11-70.76 ms. 18F-FDG tracer could be clearly observed in PET imaging. The 3D printed mouse phantom was successfully constructed with tissue-equivalent materials. Our model can be used for CT, MRI, and PET as a standard device for small-animal imaging and quality control.