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]

Objective: Interstitial fibrosis/tubular atrophy (IFTA) is a common, irreversible, and progressive form of chronic kidney allograft injury, and it is considered a critical predictor of kidney allograft outcomes. The extent of IFTA is estimated through a graft biopsy, while a non-invasive test is lacking. The aim of this study was to evaluate the feasibility and accuracy of an MRI radiomic-based machine learning (ML) algorithm to estimate the degree of IFTA in a cohort of transplanted patients. Approach: Patients who underwent MRI and renal biopsy within a 6-month interval from 1 January 2012 to 1 March 2021 were included. Stable MRI sequences were selected, and renal parenchyma, renal cortex and medulla were segmented. After image filtering and pre-processing, we computed radiomic features that were subsequently selected through a LASSO algorithm for their highest correlation with the outcome and lowest intercorrelation. Selected features and relevant patients' clinical data were used to produce ML algorithms using 70% of the study cases for feature selection, model training and validation with a 10-fold cross-validation, and 30% for model testing. Performances were evaluated using AUC with 95% confidence interval. Main results: A total of 70 coupled tests (63 patients, 35.4% females, mean age 52.2 years) were included and subdivided into a wider cohort of 50 for training and a smaller cohort of 20 for testing. For IFTA ≥ 25%, the AUCs in test cohort were 0.60, 0.59, and 0.54 for radiomic features only, clinical variables only, and a combined radiomic–clinical model, respectively. For IFTA ≥ 50%, the AUCs in training cohort were 0.89, 0.84, and 0.96, and in the test cohort, they were 0.82, 0.83, and 0.86, for radiomic features only, clinical variables only, and the combined radiomic–clinical model, respectively. Significance: An ML-based MRI radiomic algorithm showed promising discrimination capacity for IFTA > 50%, especially when combined with clinical variables. These results need to be confirmed in larger cohorts. [ABSTRACT FROM AUTHOR]

In Internal Radiotherapy, radiopharmaceutical dosimetry provides an accurate estimation of absorbed radiation doses to organs at risk and tumours. In this paper Velocity Theranostics (Varian Medical Systems), is investigated. Its performances are compared to OLINDA 2.0 in both an anthropomorphic phantom and a group of patients. Velocity Theranostics was evaluated with a cohort of patients (15) treated with 177Lu radiolabelled peptides. The absorbed doses were calculated for the liver, spleen and kidneys, separately with OLINDA 2.0 and Velocity Theranostics using the same set of images. To reduce the contribution of Time-integrated activities (TIAs) on the results and to merely compare the dose calculation algorithms, the OLINDA 2.0 absorbed doses were calculated using the TIA values calculated in Velocity Theranostics. The absorbed doses from Velocity Theranostics were found to be correlated with the doses from OLINDA 2.0 with the TIAs from Theranostics (Lin's coefficient = 0.894 and R2 = 0.9531). Absorbed doses from Velocity Theranostics are reliable at least as reliable as those for OLINDA 2.0, with many advantages regarding accuracy of calculations and robustness. In conclusion, the personalisation of dosimetry may be totally fulfilled by computational systems for absorbed dose in internal radiotherapy, equipped with a complete workflow and borrowed from external radiotherapy. [ABSTRACT FROM AUTHOR]

Background: 18F-FDG PET/CT imaging allows to study oncological patients and their relative diagnosis through the standardised uptake value (SUV) evaluation. During radiopharmaceutical injection, an extravasation event may occur, making the SUV value less accurate and possibly leading to severe tissue damage. The study aimed to propose a new technique to monitor and manage these events, to provide an early evaluation and correction to the estimated SUV value through a SUV correction coefficient. Methods: A cohort of 70 patients undergoing 18F- FDG PET/CT examinations was enrolled. Two portable detectors were secured on the patients' arms. The dose-rate (DR) time curves on the injected DRin and contralateral DRcon arm were acquired during the first 10 min of injection. Such data were processed to calculate the parameters ΔpinNOR = (DRinmax- DRinmean)/DRinmax and ΔRt = (DRin(t) − DRcon(t)), where DRinmax is the maximum DR value, DRinmean is the average DR value in the injected arm. OLINDA software allowed dosimetric estimation of the dose in the extravasation region. The estimated residual activity in the extravasation site allowed the evaluation of the SUV's correction value and to define an SUV correction coefficient. Results: Four cases of extravasations were identified for which ΔRt [(390 ± 26) µSv/h], while ΔRt [(150 ± 22) µSv/h] for abnormal and ΔRt [(24 ± 11) µSv/h] for normal cases. The ΔpinNOR showed an average value of (0.44 ± 0.05) for extravasation cases and an average value of (0.91 ± 0.06) and (0.77 ± 0.23) in normal and abnormal classes, respectively. The percentage of SUV reduction (SUV%CR) ranges between 0.3% and 6%. The calculated self-tissue dose values range from 0.027 to 0.573 Gy, according to the segmentation modality. A similar correlation between the inverse of ΔpinNOR and the normalised ΔRt with the SUV correction coefficient was found. Conclusions: The proposed metrics allowed to characterised the extravasation events in the first few minutes after the injection, providing an early SUV correction when necessary. We also assume that the characterisation of the DR-time curve of the injection arm is sufficient for the detection of extravasation events. Further validation of these hypotheses and key metrics is recommended in larger cohorts. [ABSTRACT FROM AUTHOR]

[Display omitted] • AI4MP-Challenge is the first AIFM multicentric experience on machine learning. • The main objective is to improve knowledge and skills of medical physicists on machine learning. • Encountered pitfalls: violation of independence assumption, computation errors, data imbalance. • Providing both cross-validation and an independent test helps to detect implementation issue. • The exclusion of non-robust features does not allow to significantly increase model stability. A novel and unconventional approach to a machine learning challenge was designed to spread knowledge, identify robust methods and highlight potential pitfalls about machine learning within the Medical Physics community. A public dataset comprising 41 radiomic features and 535 patients was employed to assess the potential of radiomics in distinguishing between primary lung tumors and metastases. Each participant developed two classification models using: (i) all features (base model); (ii) only robust features (robust model). Both models were validated with cross-validation and on unseen data. The population stability index (PSI) was used as diagnostic metric for implementation issues. Performance was compared to reference. Base and robust models were compared in terms of performance and stability (coefficient of variation (CoV) of prediction probabilities). PSI detected potential implementation errors in 70 % of models. The dataset exhibited strong imbalance. The average Gmean (i.e. an appropriate metric for imbalance) among all participants was 0.67 ± 0.01, significantly higher than reference Gmean of 0.50 ± 0.04. Robust models performances were slightly worse than base models (p < 0.05). Regarding stability, robust models exhibited lower median CoV on training set only. AI4MP-Challenge models overperformed the reference, significantly improving the Gmean. Exclusion of less-robust features did not improve model robustness and it should be avoided when confounding effects are absent. Other methods, like harmonization or data augmentation, should be evaluated. This study demonstrated how the collaborative effort to foster knowledge on machine learning among medical physicists, through interactive sessions and exchange of information among participants, can result in improved models. [ABSTRACT FROM AUTHOR]

• Automatic vertex arranging in lattice radiotherapy. • optimal spatial distribution for organ at risk dose sparing. • avoid time-consuming manual arrangement. • standardization, promoting reproducibility for multicenter studies. Lattice radiotherapy (LRT) is a three dimensional (3D) implementation of spatially fractionated radiation therapy, based on regular spatial distribution of high dose spheres (vertices) inside the target. Due to tumour shape heterogeneity, finding the best lattice arrangement is not trivial. The aim of this study was to develop the LatticeOpt tool to generate the best lattice structures on clinical cases for treatment planning. Developed in MATLAB, LatticeOpt finds the 3D-spatial configurations that maximize the number of vertices within the gross target volume (GTV). If organs at risk (OARs) are considered, it chooses the solution that minimizes the overlapping volume histograms (OVH). Otherwise, the lattice structure with the minimum Hausdorff distance between vertices and GTV boundary is chosen to avoid unpopulated regions. Different lattice structures were created for 20 patients, with (OVHopt) and without (OVHunopt) OVH minimization. Imported into TPS (Eclipse, Varian), corresponding plans were generated and evaluated in terms of OAR mean and maximum doses, GTV vertex coverage and dose gradients, as well as pre-clinical plan dosimetry. Plans based on an optimized lattice structure (OVHopt, OVHunopt) had similar dose distributions in terms of vertex coverage and gradient index score. OAR sparing was observed in all patients, with a 4 % and 9 % difference for mean and max dose (both p-values <0.01), respectively. The best vertices dimensions and their relative distances were patient dependent. LatticeOpt was able to reduce the time-consuming procedures of LRT, as well as to achieve standardized and reproducible results, useful for multicentre studies. [ABSTRACT FROM AUTHOR]

On‐line in situ monitoring of radiation emitted by radiotracers during intravenous injection is of crucial importance for the safety of patients and personnel involved in nuclear medicine treatments and analysis. The aim of this work is to assess wearable perovskite ionizing radiation detectors for such application. With this purpose pixelated direct detector based on thin films of 2D layered perovskite PEA2PbBr4 are tested for the monitoring of gamma‐rays emitted by 18F radiopharmaceutical under different experimental conditions. The wearable detector is here demonstrated being able to detect in real‐time the gamma‐rays emitted by annihilation of 18F radiotracer. Its reliability is assessed by 18F half‐life measurement with a maximum accuracy deviation within 8%. The detector also successes in its validation for the foreseen application, demonstrating able to monitor in real‐time the whole injection process and extravasation event. [ABSTRACT FROM AUTHOR]

Anterior segment optical coherence tomography (AS-OCT) allows the explore not only the anterior chamber but also the front part of the vitreous cavity. Our cross-sectional single-centre study investigated whether AS-OCT can distinguish between vitreous involvement due to vitreoretinal lymphoma (VRL) and vitritis in uveitis. We studied AS-OCT images from 28 patients (11 with biopsy-proven VRL and 17 with differential diagnosis uveitis) using publicly available radiomics software written in MATLAB. Patients were divided into two balanced groups: training and testing. Overall, 3260/3705 (88%) AS-OCT images met our defined quality criteria, making them eligible for analysis. We studied five different sets of grey-level samplings (16, 32, 64, 128, and 256 levels), finding that 128 grey levels performed the best. We selected the five most effective radiomic features ranked by the ability to predict the class (VRL or uveitis). We built a classification model using the xgboost python function; through our model, 87% of eyes were correctly diagnosed as VRL or uveitis, regardless of exam technique or lens status. Areas under the receiver operating characteristic curves (AUC) in the 128 grey-level model were 0.95 [CI 0.94, 0.96] and 0.84 for training and testing datasets, respectively. This preliminary retrospective study highlights how AS-OCT can support ophthalmologists when there is clinical suspicion of VRL.