Laser ablation for spongiform thyroid nodules: identifying baseline nodule volume and energy to be delivered for successful treatment.

BACKGROUND: Laser ablation (LA) is a minimally invasive treatment. It has been widely used since the early 2000s to induce volume reduction of symptomatic benign thyroid nodules. Up to 40% of laser-treated nodules have been reported to achieve a volume reduction of <50% (technique inefficacy) at 12 months and tend to regrow over time. OBJECTIVE: This study aimed to assess the optimal baseline volume and energy to be delivered to minimize technique inefficacy. METHODS: This was a retrospective study. Data were collected, including baseline volume, energy delivered, and 12-month volume reduction ratio (VRR) of spongiform nodules (EU-TIRADS 2) treated with LA between 2010 and 2020. Based on these data, the optimal baseline volume and energy to be delivered were calculated to maximize the rate of nodules with technique efficacy (VRR ≥ 50% at 12-month follow-up). RESULTS: A total of 205 patients with spongiform nodules were included in this study. The energy delivered was positively associated with VRR. However, no association was observed between baseline volume and VRR. Delivering energy ≥500 J/mL to nodules with a mean baseline volume of 11.4 ± 4 mL resulted in technique efficacy in 83% of cases. CONCLUSION: Treating spongiform nodules with a baseline volume of ≤15 mL and delivering energy ≥500 J/mL are key factors for achieving a relevant rate of technique efficacy.

Risk prediction of clinical adverse outcomes with machine learning in a cohort of critically ill patients with atrial fibrillation.

Critically ill patients affected by atrial fibrillation are at high risk of adverse events: however, the actual risk stratification models for haemorrhagic and thrombotic events are not validated in a critical care setting. With this paper we aimed to identify, adopting topological data analysis, the risk factors for therapeutic failure (in-hospital death or intensive care unit transfer), the in-hospital occurrence of stroke/TIA and major bleeding in a cohort of critically ill patients with pre-existing atrial fibrillation admitted to a stepdown unit; to engineer newer prediction models based on machine learning in the same cohort. We selected all medical patients admitted for critical illness and a history of pre-existing atrial fibrillation in the timeframe 01/01/2002-03/08/2007. All data regarding patients' medical history, comorbidities, drugs adopted, vital parameters and outcomes (therapeutic failure, stroke/TIA and major bleeding) were acquired from electronic medical records. Risk factors for each outcome were analyzed adopting topological data analysis. Machine learning was used to generate three different predictive models. We were able to identify specific risk factors and to engineer dedicated clinical prediction models for therapeutic failure (AUC: 0.974, 95%CI: 0.934-0.975), stroke/TIA (AUC: 0.931, 95%CI: 0.896-0.940; Brier score: 0.13) and major bleeding (AUC: 0.930:0.911-0.939; Brier score: 0.09) in critically-ill patients, which were able to predict accurately their respective clinical outcomes. Topological data analysis and machine learning techniques represent a concrete viewpoint for the physician to predict the risk at the patients' level, aiding the selection of the best therapeutic strategy in critically ill patients affected by pre-existing atrial fibrillation.

Machine Learning Prediction of Radiofrequency Thermal Ablation Efficacy: A New Option to Optimize Thyroid Nodule Selection.

BACKGROUND: Radiofrequency (RF) is a therapeutic modality for reducing the volume of large benign thyroid nodules. If thermal therapies are interpreted as an alternative strategy to surgery, critical issues in their use are represented by the extent of nodule reduction and by the durability of nodule reduction over a long period of time. OBJECTIVE: To assess the ability of machine learning to discriminate nodules with volume reduction rate (VRR) < or ≥50% at 12 months following RF treatment. METHODS: A machine learning model was trained with a dataset of 402 cytologically benign thyroid nodules subjected to RF at six Italian Institutions. The model was trained with the following variables: baseline nodule volume, echostructure, macrocalcalcifications, vascularity, and 12-month VRR. RESULTS: After training, the model could distinguish between nodules having VRR <50% from those having VRR ≥50% in 85% of cases (accuracy: 0.85; 95% confidence interval [CI]: 0.80-0.90; sensitivity: 0.70; 95% CI: 0.62-0.75; specificity: 0.99; 95% CI: 0.98-1.0; positive predictive value: 0.95; 95% CI: 0.92-0.98; negative predictive value: 0.95; 95% CI: 0.92-0.98). CONCLUSIONS: This study demonstrates that a machine learning model can reliably identify those nodules that will have VRR < or ≥50% at 12 months after one RF treatment session. Predicting which nodules will be poor or good responders represents valuable data that may help physicians and patients decide on the best treatment option between thermal ablation and surgery or in predicting if more than one session might be necessary to obtain a significant volume reduction.

Twelve-Month Volume Reduction Ratio Predicts Regrowth and Time to Regrowth in Thyroid Nodules Submitted to Laser Ablation: A 5-Year Follow-Up Retrospective Study.

OBJECTIVE: Laser ablation is a therapeutic modality used to reduce the volume of large benign thyroid nodules. Unsatisfactory reduction and regrowth are observed in some treated nodules. The aim of the study was to evaluate the long-term outcomes of laser treatment for solid nodules during a 5-year follow-up period, the regrowth rate, and the predictive risk factors of nodule regrowth. MATERIALS AND METHODS: We retrospectively evaluated patients with benign, solid, cold thyroid nodules who underwent laser ablation and were followed-up for 5 years. According to the selection criteria, 104 patients were included (median baseline nodule volume, 12.5 mL [25.0-75.0%, 8-18 mL]; median energy delivered, 481.5 J/mL [25.0-75.0%, 370-620 J/mL]). Nodule volume, thyroid function test results, and ultrasound were evaluated at baseline and then annually after the procedure. RESULTS: Of 104 patients, 31 patients (29.8%) had a 12-month volume reduction ratio (VRR) < 50.0% and 39 (37.5%) experienced nodule regrowth. Of these 39 patients, 17 (43.6%) underwent surgery and 14 (35.9%) underwent a second laser treatment. The rate of nodule regrowth was inversely related to the 12-month VRR, i.e., the lower the 12-month VRR, the higher the risk of regrowth (p < 0.001). The mean time for nodule regrowth was 33.5 ± 16.6 months. The 12-month VRR was directly related to time to regrowth, i.e., the lower the 12-month VRR, the shorter the time to regrowth (p < 0.001; R² = 0.3516). Non-spongiform composition increased the risk of regrowth with an odds ratio of 4.3 (95% confidence interval [CI] 1.8-10.2; p < 0.001); 12-month VRR < 50.0% increased the risk of regrowth with an odds ratio of 11.7 (95% CI 4.2-32.2; p < 0.001). CONCLUSION: The VRR of thyroid nodules subjected to similar amounts of laser energy varies widely and depends on the nodule composition; non-spongiform nodules are reduced to a lesser extent and regrow more frequently than spongiform nodules. A 12-month VRR < 50.0% is a predictive risk factor for regrowth and correlates with the time to regrowth.

Topolnogical classifier for detecting the emergence of epileptic seizures.

OBJECTIVE: An innovative method based on topological data analysis is introduced for classifying EEG recordings of patients affected by epilepsy. We construct a topological space from a collection of EEGs signals using Persistent Homology; then, we analyse the space by Persistent entropy, a global topological feature, in order to classify healthy and epileptic signals. RESULTS: The performance of the resulting one-feature-based linear topological classifier is tested by analysing the Physionet dataset. The quality of classification is evaluated in terms of the Area Under Curve (AUC) of the receiver operating characteristic curve. It is shown that the linear topological classifier has an AUC equal to [Formula: see text] while the performance of a classifier based on Sample Entropy has an AUC equal to 62.0%.

Evaluation of low-grade glioma structural changes after chemotherapy using DTI-based histogram analysis and functional diffusion maps.

OBJECTIVES: To explore the role of diffusion tensor imaging (DTI)-based histogram analysis and functional diffusion maps (fDMs) in evaluating structural changes of low-grade gliomas (LGGs) receiving temozolomide (TMZ) chemotherapy. METHODS: Twenty-one LGG patients underwent 3T-MR examinations before and after three and six cycles of dose-dense TMZ, including 3D-fluid-attenuated inversion recovery (FLAIR) sequences and DTI (b = 1000 s/mm(2), 32 directions). Mean diffusivity (MD), fractional anisotropy (FA), and tensor-decomposition DTI maps (p and q) were obtained. Histogram and fDM analyses were performed on co-registered baseline and post-chemotherapy maps. DTI changes were compared with modifications of tumour area and volume [according to Response Assessment in Neuro-Oncology (RANO) criteria], and seizure response. RESULTS: After three cycles of TMZ, 20/21 patients were stable according to RANO criteria, but DTI changes were observed in all patients (Wilcoxon test, P ≤ 0.03). After six cycles, DTI changes were more pronounced (P ≤ 0.005). Seventy-five percent of patients had early seizure response with significant improvement of DTI values, maintaining stability on FLAIR. Early changes of the 25th percentiles of p and MD predicted final volume change (R(2) = 0.614 and 0.561, P < 0.0005, respectively). TMZ-related changes were located mainly at tumour borders on p and MD fDMs. CONCLUSIONS: DTI-based histogram and fDM analyses are useful techniques to evaluate the early effects of TMZ chemotherapy in LGG patients. KEY POINTS: • DTI helps to assess the efficacy of chemotherapy in low-grade gliomas. • Histogram analysis of DTI metrics quantifies structural changes in tumour tissue. • Functional diffusion maps (fDMs) spatially localize the changes of DTI metrics. • Changes in DTI histograms and fDMs precede changes in conventional MRI. • Early changes in DTI histograms and fDMs correlate with seizure response.

Neural hypernetwork approach for pulmonary embolism diagnosis.

BACKGROUND: Hypernetworks are based on topological simplicial complexes and generalize the concept of two-body relation to many-body relation. Furthermore, Hypernetworks provide a significant generalization of network theory, enabling the integration of relational structure, logic and analytic dynamics. A pulmonary embolism is a blockage of the main artery of the lung or one of its branches, frequently fatal. RESULTS: Our study uses data on 28 diagnostic features of 1427 people considered to be at risk of pulmonary embolism enrolled in the Department of Internal and Subintensive Medicine of an Italian National Hospital "Ospedali Riuniti di Ancona". Patients arrived in the department after a first screening executed by the emergency room. The resulting neural hypernetwork correctly recognized 94% of those developing pulmonary embolism. This is better than previous results obtained with other methods (statistical selection of features, partial least squares regression, topological data analysis in a metric space). CONCLUSION: In this work we successfully derived a new integrative approach for the analysis of partial and incomplete datasets that is based on Q-analysis with machine learning. The new approach, called Neural Hypernetwork, has been applied to a case study of pulmonary embolism diagnosis. The novelty of this method is that it does not use clinical parameters extracted by imaging analysis.