Data-based predictions of individual Cognitive Behavioral Therapy (CBT) treatment response are a fundamental step towards precision medicine. Past studies demonstrated only moderate prediction accuracy (i.e. ability to discriminate between responders and non-responders of a given treatment) when using clinical routine data such as demographic and questionnaire data, while neuroimaging data achieved superior prediction accuracy. However, these studies may be considerably biased due to very limited sample sizes and bias-prone methodology. Adequately powered and cross-validated samples are a prerequisite to evaluate predictive performance and to identify the most promising predictors. We therefore analyzed resting state functional magnet resonance imaging (rs-fMRI) data from two large clinical trials to test whether functional neuroimaging data continues to provide good prediction accuracy in much larger samples. Data came from two distinct German multicenter studies on exposure-based CBT for anxiety disorders, the Protect-AD and SpiderVR studies. We separately and independently preprocessed baseline rs-fMRI data from n = 220 patients (Protect-AD) and n = 190 patients (SpiderVR) and extracted a variety of features, including ROI-to-ROI and edge-functional connectivity, sliding-windows, and graph measures. Including these features in sophisticated machine learning pipelines, we found that predictions of individual outcomes never significantly differed from chance level, even when conducting a range of exploratory post-hoc analyses. Moreover, resting state data never provided prediction accuracy beyond the sociodemographic and clinical data. The analyses were independent of each other in terms of selecting methods to process resting state data for prediction input as well as in the used parameters of the machine learning pipelines, corroborating the external validity of the results. These similar findings in two independent studies, analyzed separately, urge caution regarding the interpretation of promising prediction results based on neuroimaging data from small samples and emphasizes that some of the prediction accuracies from previous studies may result from overestimation due to homogeneous data and weak cross-validation schemes. The promise of resting-state neuroimaging data to play an important role in the prediction of CBT treatment outcomes in patients with anxiety disorders remains yet to be delivered.

OBJECTIVE: A machine learning-based anterior cruciate ligament (ACL) revision prediction model has been developed using Norwegian Knee Ligament Register (NKLR) data, but lacks external validation outside Scandinavia. This study aimed to assess the external validity of the NKLR model (https://swastvedt.shinyapps.io/calculator_rev/) using the STABILITY 1 randomized clinical trial (RCT) data set. The hypothesis was that model performance would be similar.
METHODS: The NKLR Cox Lasso model was selected for external validation owing to its superior performance in the original study. STABILITY 1 patients with all five predictors required by the Cox Lasso model were included. The STABILITY 1 RCT was a prospective study which randomized patients to receive either a hamstring tendon autograft (HT) alone or HT plus a lateral extra-articular tenodesis (LET). Since all patients in the STABILITY 1 trial received HT ± LET, three configurations were tested: 1: all patients coded as HT, 2: HT + LET group coded as bone-patellar tendon-bone (BPTB) autograft, 3: HT + LET group coded as unknown/other graft choice. Model performance was assessed via concordance and calibration.
RESULTS: In total, 591/618 (95.6%) STABILITY 1 patients were eligible for inclusion, with 39 undergoing revisions within 2 years (6.6%). Model performance was best when patients receiving HT + LET were coded as BPTB. Concordance was similar to the original NKLR prediction model for 1- and 2-year revision prediction (STABILITY: 0.71; NKLR: 0.68-0.69). Concordance 95% confidence interval (CI) ranged from 0.63 to 0.79. The model was well calibrated for 1-year prediction while the 2-year prediction demonstrated evidence of miscalibration.
CONCLUSIONS: When patients in STABILITY 1 who received HT + LET were coded as BPTB in the NKLR prediction model, concordance was similar to the index study. However, due to a wide 95% CI, the true performance of the prediction model with this Canadian and European cohort is unclear and a larger data set is required to definitively determine the external validity. Further, better calibration for 1-year predictions aligns with general prediction modelling challenges over longer periods. While not a large enough sample size to elicit the true accuracy and external validity of the prediction model when applied to North American patients, this analysis provides more support for the notion that HT plus LET performs similarly to BPTB reconstruction. In addition, despite the wide confidence interval, this study suggests optimism regarding the accuracy of the model when applied outside of Scandinavia.
METHODS: Level 3, cohort study.

OBJECTIVE: This study aimed to predict the probability of grade ≥2 pneumonitis or dyspnea within 12 months of receiving conventionally fractionated or mildly hypofractionated proton beam therapy for locally advanced lung cancer using machine learning.
METHODS: Demographic and treatment characteristics were analyzed for 965 consecutive patients treated for lung cancer with conventionally fractionated or mildly hypofractionated (2.2-3 Gy/fraction) proton beam therapy across 12 institutions. Three machine learning models (gradient boosting, additive tree, and logistic regression with lasso regularization) were implemented to predict Common Terminology Criteria for Adverse Events version 4 grade ≥2 pulmonary toxicities using double 10-fold cross-validation for parameter hyper-tuning without leak of information. Balanced accuracy and area under the curve were calculated, and 95% confidence intervals were obtained using bootstrap sampling.
RESULTS: The median age of the patients was 70 years (range, 20-97), and they had predominantly stage IIIA or IIIB disease. They received a median dose of 60 Gy in 2 Gy/fraction, and 46.4% received concurrent chemotherapy. In total, 250 (25.9%) had grade ≥2 pulmonary toxicity. The probability of pulmonary toxicity was 0.08 for patients treated with pencil beam scanning and 0.34 for those treated with other techniques (P = 8.97e-13). Use of abdominal compression and breath hold were highly significant predictors of less toxicity (P = 2.88e-08). Higher total radiation delivered dose (P = .0182) and higher average dose to the ipsilateral lung (P = .0035) increased the likelihood of pulmonary toxicities. The gradient boosting model performed the best of the models tested, and when demographic and dosimetric features were combined, the area under the curve and balanced accuracy were 0.75 ± 0.02 and 0.67 ± 0.02, respectively. After analyzing performance versus the number of data points used for training, we observed that accuracy was limited by the number of observations.
CONCLUSIONS: In the largest analysis of prospectively enrolled patients with lung cancer assessing pulmonary toxicities from proton therapy to date, advanced machine learning methods revealed that pencil beam scanning, abdominal compression, and lower normal lung doses can lead to significantly lower probability of developing grade ≥2 pneumonitis or dyspnea.

BACKGROUND: Out-of-hospital cardiac arrest (OHCA) is a heterogeneous entity with multiple origins and prognoses. An early, reliable assessment of the prognosis is useful to adapt therapeutic strategy, tailor intensity of care, and inform relatives. We aimed primarily to undertake a prospective multicentric study to evaluate predictive performance of the Cardiac Arrest Prognosis (CAHP) Score as compare to historical dataset systematically collected after OHCA (Utstein style criteria). Our secondary aim was to evaluate other dedicated scores for predicting outcome after OHCA and to compare them to Utstein style criteria.
METHODS: We prospectively collected data from 24 French and Belgium Intensive Care Units (ICUs) between August 2020 and June 2022. All cases of non-traumatic OHCA (cardiac and non-cardiac causes) patients with stable return of spontaneous circulation (ROSC) and comatose at ICU admission (defined by Glasgow coma score ≤ 8) on ICU admission were included. The primary outcome was the modified Rankin scale (mRS) at day 90 after cardiac arrest, assessed by phone interviews. A wide range of developed scores (CAHP, OHCA, CREST, C-Graph, TTM, CAST, NULL-PLEASE, and MIRACLE2) were included, and their accuracies in predicting poor outcome at 90 days after OHCA (defined as mRS ≥ 4) were determined using the area under the receiving operating characteristic curve (AUROC) and the calibration belt.
RESULTS: During the study period, 907 patients were screened, and 658 were included in the study. Patients were predominantly male (72%), with a mean age of 61 ± 15, most having collapsed from a supposed cardiac cause (64%). The mortality rate at day 90 was 63% and unfavorable neurological outcomes were observed in 66%. The performance (AUROC) of Utstein criteria for poor outcome prediction was moderate at 0.79 [0.76-0.83], whereas AUROCs from other scores varied from 0.79 [0.75-0.83] to 0.88 [0.86-0.91]. For each score, the proportion of patients for whom individual values could not be calculated varied from 1.4% to 17.4%.
CONCLUSIONS: In patients admitted to ICUs after a successfully resuscitated OHCA, most of the scores available for the evaluation of the subsequent prognosis are more efficient than the usual Utstein criteria but calibration is unacceptable for some of them. Our results show that some scores (CAHP, sCAHP, mCAHP, OHCA, rCAST) have superior performance, and that their ease and speed of determination should encourage their use. Trial registration https://clinicaltrials.gov/ct2/show/NCT04167891.

UNASSIGNED: This study hypothesized that monitoring electrocardiogram (ECG) waveforms in patients with out-of-hospital cardiac arrest (OHCA) could have predictive value for survival or neurological outcomes. We aimed to establish a new prognostication model based on the single variable of monitoring ECG waveforms in patients with OHCA using machine learning (ML) techniques.
UNASSIGNED: This observational retrospective study included successfully resuscitated patients with OHCA aged ≥ 18 years admitted to an intensive care unit in Japan between April 2010 and April 2020. Waveforms from ECG monitoring for 1 h after admission were obtained from medical records and examined. Based on the open-access PTB-XL dataset, a large publicly available 12-lead ECG waveform dataset, we built an ML-supported premodel that transformed the II-lead waveforms of the monitoring ECG into diagnostic labels. The ECG diagnostic labels of the patients in this study were analyzed for prognosis using another model supported by ML. The endpoints were favorable neurological outcomes (cerebral performance category 1 or 2) and survival to hospital discharge.
UNASSIGNED: In total, 590 patients with OHCA were included in this study and randomly divided into 3 groups (training set, n = 283; validation set, n = 70; and test set, n = 237). In the test set, our ML model predicted neurological and survival outcomes, with the highest areas under the receiver operating characteristic curves of 0.688 (95% CI: 0.682-0.694) and 0.684 (95% CI: 0.680-0.689), respectively.
UNASSIGNED: Our ML predictive model showed that monitoring ECG waveforms soon after resuscitation could predict neurological and survival outcomes in patients with OHCA.

(1) Background: To predict clinical outcomes in patients with aneurysmal subarachnoid hemorrhage (aSAH) and delayed cerebral ischemia (DCI) by assessment of the cerebral perfusion using a 2D perfusion angiography (2DPA) time-contrast agent (CA) concentration model. (2) Methods: Digital subtraction angiography (DSA) data sets of n = 26 subjects were acquired and post-processed focusing on changes in contrast density using a time-concentration model at three time points: (i) initial presentation with SAH (T0); (ii) vasospasm-associated acute clinical impairment (T1); and (iii) directly after endovascular treatment (T2) of SAH-associated large vessel vasospasm (LVV), which resulted in n = 78 data sets. Maximum slope (MS in SI/ms), time-to-peak (TTP in ms), and maximum amplitude of a CA bolus (dSI) were measured in brain parenchyma using regions of interest (ROIs). First, acquired parameters were standardized to the arterial input function (AIF) and then statistically analyzed as mean values. Additionally, data were clustered into two subsets consisting of patients with regredient or with stable/progredient symptoms (or Doppler signals) after endovascular treatment (n = 10 vs. n = 16). (3) Results: Perfusion parameters (MS, TTP, and dSI) differed significantly between T0 and T1 (p = 0.003 each). Significant changes between T1 and T2 were only detectable for MS (0.041 ± 0.016 vs. 0.059 ± 0.026; p = 0.011) in patients with regredient symptoms at T2 (0.04 ± 0.012 vs. 0.066 ± 0.031; p = 0.004). For dSI, there were significant differences between T0 and T2 (5095.8 ± 2541.9 vs. 3012.3 ± 968.3; p = 0.001), especially for those with stable symptoms at T2 (5685.4 ± 2967.2 vs. 3102.8 ± 1033.2; p = 0.02). Multiple linear regression analysis revealed that a) the difference in MS between T1 and T2 and b) patient\'s age (R = 0.6; R2 = 0.34; p = 0.009) strongly predict the modified Rankin Scale (mRS) at discharge. (4) Conclusions: 2DPA allows the direct measurement of treatment effects in SAH associated DCI and may be used to predict outcomes in these critically ill patients.

UNASSIGNED: Cerebrospinal fluid rhinorrhoea (CSFR) is a common complication following endonasal skull base surgery, a technique that is fundamental to the treatment of pituitary adenomas and many other skull base tumours. The CRANIAL study explored CSFR incidence and related risk factors, particularly skull base repair techniques, via a multicentre prospective observational study. We sought to use machine learning to leverage this complex multicentre dataset for CSFR prediction and risk factor analysis.
UNASSIGNED: A dataset of 865 cases - 725 transsphenoidal approach (TSA) and 140 expanded endonasal approach (EEA) - with cerebrospinal fluid rhinorrhoea as the primary outcome, was used. Relevant variables were extracted from the data, and prediction variables were divided into two categories, preoperative risk factors; and repair techniques, with 6 and 11 variables respectively. Three types of machine learning models were developed in order to predict CSFR: logistic regression (LR); decision tree (DT); and neural network (NN). Models were validated using 5-fold cross-validation, compared via their area under the curve (AUC) evaluation metric, and key prediction variables were identified using their Shapley additive explanations (SHAP) score.
UNASSIGNED: CSFR rates were 3.9% (28/725) for the transsphenoidal approach and 7.1% (10/140) for the expanded endonasal approach. NNs outperformed LR and DT for CSFR prediction, with a mean AUC of 0.80 (0.70-0.90) for TSA and 0.78 (0.60-0.96) for EEA, when all risk factor and intraoperative repair data were integrated into the model. The presence of intraoperative CSF leak was the most prominent risk factor for CSFR. Elevated BMI and revision surgery were also associated with CSFR for the transsphenoidal approach. CSF diversion and gasket sealing appear to be strong predictors of the absence of CSFR for both approaches.
UNASSIGNED: Neural networks are effective at predicting CSFR and uncovering key CSFR predictors in patients following endonasal skull base surgery, outperforming traditional statistical methods. These models will be improved further with larger and more granular datasets, improved NN architecture, and external validation. In the future, such predictive models could be used to assist surgical decision-making and support more individualised patient counselling.

Virtual clinical trials (VCTs) can potentially simulate clinical trials on a computer, but their application with a limited number of past clinical cases is challenging due to the biased estimation of the statistical population. In this study, we developed ExMixup, a novel training technique based on machine learning, using iteratively redistributed extrapolated data. Information obtained from 100 patients with prostate cancer and 385 patients with oropharyngeal cancer was used to predict the recurrence after radiotherapy. Model performance was evaluated by developing outcome prediction models based on three types of training methods: training with original data (baseline), interpolation data (Mixup), and interpolation + extrapolation data (ExMixup). Two types of VCTs were conducted to predict the treatment response of patients with distinct characteristics compared to the training data obtained from patient cohorts categorized under risk classification or cancer stage. The prediction models developed with ExMixup yielded concordance indices (95% confidence intervals) of 0.751 (0.719-0.818) and 0.752 (0.734-0.785) for VCTs on the prostate and oropharyngeal cancer datasets, respectively, which significantly outperformed the baseline and Mixup models (P < 0.01). The proposed approach could enhance the ability of VCTs to predict treatment results in patients excluded from past clinical trials.

UNASSIGNED: Although carbon-ion radiotherapy (CIRT) may improve outcome for patients with locoregionally recurrent nasopharyngeal carcinoma (LR-NPC), local progression still remains one of the major failure patterns. This suggests an unmet need of markers for predicting disease control after re-irradiation and potentially guiding tailored treatment. The purpose of this study was to explore the predictive value of pre-treatment 3\'-deoxy-3\'-[18F]fluorothymidine (FLT)-positron emission tomography (PET) for patients with locally advanced LR-NPC.
UNASSIGNED: In this retrospective analysis, LR-NPC patients with locally advanced stage (stage III/IV) who received pre-treatment FLT-PET between June, 2015, and August, 2017, were retrospective reviewed and included in this study. OS and local progression-free survival (LPFS) were calculated using the Kaplan-Meier method. Univariable and multivariable Cox regression analyses of LPFS were performed. FLT-derived parameters, including SUVmax, metabolic tumor volume (MTV), and total lesion thymidine (TLT) were examined. The relationship between FLT-derived parameters and mucosal necrosis was tested by the Wilcoxon test.
UNASSIGNED: A total of 27 patients with a median follow-up of 31.3 months were included in this analysis. The 2-year OS and LPFS rates were 85.2% and 47.9%, respectively. In multivariable analysis, except for TLT-40% (P=0.059), all pre-treatment MTVs (P=0.040 for MTV-40%; P=0.021 for MTV-50%; P=0.026 for MTV-60%) and TLTs (P=0.043 for TLT-50%; P=0.048 for TLT-60%) were significantly related to LPFS. Moreover, MTVs and TLTs with various boundaries (except for MTV-40%) were also associated with the development of mucosal necrosis after CIRT.
UNASSIGNED: In the current study, a significant association between pre-treatment FLT-PET and LPFS was observed in patients with locally advanced LR-NPC. Further investigations are warranted to confirm the predictive role of FLT-PET.

BACKGROUND: Disorders of consciousness (DoC) are severe neurological conditions in which consciousness is impaired to various degrees. They are caused by injury or malfunction of neural systems regulating arousal and awareness. Over the last decades, major efforts in improving and individualizing diagnostic and prognostic accuracy for patients affected by DoC have been made, mainly focusing on introducing multimodal assessments to complement behavioral examination. The present EU-funded multicentric research project \"PerBrain\" is aimed at developing an individualized diagnostic hierarchical pathway guided by both behavior and multimodal neurodiagnostics for DoC patients.
METHODS: In this project, each enrolled patient undergoes repetitive behavioral, clinical, and neurodiagnostic assessments according to a patient-tailored multi-layer workflow. Multimodal diagnostic acquisitions using state-of-the-art techniques at different stages of the patients\' clinical evolution are performed. The techniques applied comprise well-established behavioral scales, innovative neurophysiological techniques (such as quantitative electroencephalography and transcranial magnetic stimulation combined with electroencephalography), structural and resting-state functional magnetic resonance imaging, and measurements of physiological activity (i.e. nasal airflow respiration). In addition, the well-being and treatment decision attitudes of patients\' informal caregivers (primarily family members) are investigated. Patient and caregiver assessments are performed at multiple time points within one year after acquired brain injury, starting at the acute disease phase.
CONCLUSIONS: Accurate classification and outcome prediction of DoC are of crucial importance for affected patients as well as their caregivers, as individual rehabilitation strategies and treatment decisions are critically dependent on the latter. The PerBrain project aims at optimizing individual DoC diagnosis and accuracy of outcome prediction by integrating data from the suggested multimodal examination methods into a personalized hierarchical diagnosis and prognosis procedure. Using the parallel tracking of both patients\' neurological status and their caregivers\' mental situation, well-being, and treatment decision attitudes from the acute to the chronic phase of the disease and across different countries, this project aims at significantly contributing to the current clinical routine of DoC patients and their family members.
BACKGROUND: ClinicalTrials.gov, NCT04798456 . Registered 15 March 2021 - Retrospectively registered.