Search

Background: To address the numerous unmeet clinical needs, in recent years several Machine Learning models applied to medical images and clinical data have been introduced and developed. Even when they achieve encouraging results, they lack evolutionary progression, thus perpetuating their status as autonomous entities. We postulated that different algorithms which have been proposed in the literature to address the same diagnostic task, can be aggregated to enhance classification performance. We suggested a proof of concept to define an ensemble approach useful for integrating different algorithms proposed to solve the same clinical task. Methods: The proposed approach was developed starting from a public database consisting of radiomic features extracted from CT images relating to 535 patients suffering from lung cancer. Seven algorithms were trained independently by participants in the AI4MP working group on Artificial Intelligence of the Italian Association of Physics in Medicine to discriminate metastatic from non-metastatic patients. The classification scores generated by these algorithms are used to train SVM classifier. The Explainable Artificial Intelligence approach is applied to the final model. The ensemble model was validated following an 80–20 hold-out and leave-one-out scheme on the training set. Results: Compared to individual algorithms, a more accurate result was achieved. On the independent test the ensemble model achieved an accuracy of 0.78, a F1-score of 0.57 and a log-loss of 0.49. Shapley values representing the contribution of each algorithm to the final classification result of the ensemble model were calculated. This information represents an added value for the end user useful for evaluating the appropriateness of the classification result on a particular case. It also allows us to evaluate on a global level which methodological approaches of the individual algorithms are likely to have the most impact. Conclusion: Our proposal represents an innovative approach useful for integrating different algorithms that populate the literature and which lays the foundations for future evaluations in broader application scenarios. [ABSTRACT FROM AUTHOR]

Background: Soft tissue sarcomas (STS) are a group of rare, heterogeneous tumors. Scarce terapeuthic options are currently available, given the limited efficacy of systemic chemotherapy. Patients with low-burden metastatic disease may benefit from adjunction of local ablative treatments. However, because of the presumed radioresistance of STS, stereotactic body radiotherapy (SBRT) has not been steadily considered as a viable treatment option. Methods: We retrospectively reviewed STS patients treated with SBRT in a single institution,, for oligometastatic and oligoprogressive metastatic disease. Local control (LC), disease-free survival (DFS), and overall survival (OS) were assessed. We also reported the incidence of early and late adverse events and their grade according to Common Terminology Criteria for Adverse Events (CTCAE) 5.0. Univariate and multivariate statistical analysis were carried out to investigate variables correlated with outcome and toxicity. Results: Forty patients were treated between 2012 and 2019 to 77 metastases with SBRT to a median biologic effective dose (BED5, assuming an α/β=5) of 105 Gy (range 66 to 305 Gy). 63% of patients had received 2 or more chemotherapy lines at the time of SBRT. LC, DFS and OS at two years were 67%, 23% and 40% respectively. At multivariate analysis, LC only significantly improved in patients receving a BED5>150 Gy (hazard ratio [HR] 3.9. 95% confidence interval [CI], 1.6-9.7; p=0.028), while an interval>24 months between primary tumor diagnosis and metastatic disease relapse was correlated with improved DFS and OS (HR, 0.46; 95% CI, 0.22-0.96; P = 0.01 and HR, 0.48; 95% CI, 0.23- 0.99; P = 0.03, respectively). Early toxicity was observed in 4 patients, late toxicity in 1 patient. No toxicity grade>2 was observed. Conclusions: SBRT is an effective therapeutic option in metastatic STS. A BED5>150 Gy5 is required to obtain better tumor control rates. Metastatic relapse>24 months after diagnosis is correlated with improved overall survival. Adverse events are scarce and mostly transient, meaning SBRT is safe and could be an option even in heavily pretreated patients. [ABSTRACT FROM AUTHOR]

Purpose: To present BrainTool, a new anthropomorphic phantom made to assess the accuracy of brain radiosurgery treatments. Materials and method: A synthetic human brain T1/ MRI study was segmented, simplified, and 3D printed. The printed objects were glued to two PMMA disks thus creating four compartments. Each compartment was filled with different substances to create realistic MRI and CT brain contrasts. A water-equivalent silicon rubber casing covers the phantom giving it a realistic head look. Eight markers were attached inside the BrainTool to test the accuracy of image registration procedures. A holder hosting an ionization chamber can be inserted into the BrainTool to assess the targeting accuracy. BrainTool can be placed on any CT, MRI or linac couch or fixed to a Lexell stereotactic frame. To exemplify a possible use of BrainTool, we measured the registration accuracy and the targeting accuracy of a GammaKnife system. Registration accuracy- CT and T1 MRI images of the BrainTool phantom were sent to GammaPlan (Elekta) system and registered with the provided tool. The distances between correspondent markers’ positions in CT and MRI registered studies were evaluated and mean and range values were calculated. Registrations were carried out in the entire image or in a smaller userdefined region of interest (ROI) inside the phantom. Targeting accuracy- It was assessed by comparing planned and measured target positions. The measured target position was obtained using an iterative cross scan approach aimed to localize the camera measuring point. Planned and measured target positions uncertainty were both assessed. Results: In Fig. 1, a picture of the BrainTool phantom, a CT slice with measured HU and axial coronal and sagittal slice of a T1 weighted MR images, are shown. Two NiCl2 solutions (1.2 mmol/l and 0.8 mmol/l) were used to fill the phantom and simulate grey and white matter T1 relaxation times (650 ms and 1200 ms). Several T1 or T2 MRI relaxation times, can be simulated by choosing different solutions or concentrations. Registration accuracy – The mean distances between correspondent markers in CT and MRI registered studies were 1.2± 0.5mm; range [1.0mm -1.4mm] and 0.3 ± 0.2mm; range [0.0mm -0.6mm] for the entire image and the selected ROI, respectively. Targeting accuracy - Targeting mismatches along x, y and z and cumulative (r) were x=-0.7±0.5 mm, y=0.3±0.4 mm z= -0.9± 0.4 mm and r=1.2±0.8 mm, respectively. Conclusions: We have developed a phantom and exemplified its use in brain radiosurgery treatments. [ABSTRACT FROM AUTHOR]

This paper presents a comprehensive study of hydrogenated amorphous silicon (a-Si)-based detectors, utilizing electrical characterization, Raman spectroscopy, photoemission, and inverse photoemission techniques. The unique properties of a-Si have sparked interest in its application for radiation detection in both physics and medicine. Although amorphous silicon (a-Si) is inherently a highly defective material, hydrogenation significantly reduces defect density, enabling its use in radiation detector devices. Spectroscopic measurements provide insights into the intricate relationship between the structure and electronic properties of a-Si, enhancing our understanding of how specific configurations, such as the choice of substrate, can markedly influence detector performance. In this study, we compare the performance of a-Si detectors deposited on two different substrates: crystalline silicon (c-Si) and flexible Kapton. Our findings suggest that detectors deposited on Kapton exhibit reduced sensitivity, despite having comparable noise and leakage current levels to those on crystalline silicon. We hypothesize that this discrepancy may be attributed to the substrate material, differences in film morphology, and/or the alignment of energy levels. Further measurements are planned to substantiate these hypotheses.