Machine Learning Models for Prediction of Posttreatment Recurrence in Early-Stage Hepatocellular Carcinoma Using Pretreatment Clinical and MRI Features: A Proof-of-Concept Study.

BACKGROUND. Posttreatment recurrence is an unpredictable complication after liver transplant for hepatocellular carcinoma (HCC) that is associated with poor survival. Biomarkers are needed to estimate recurrence risk before organ allocation. OBJECTIVE. This proof-of-concept study evaluated the use of machine learning (ML) to predict recurrence from pretreatment laboratory, clinical, and MRI data in patients with early-stage HCC initially eligible for liver transplant. METHODS. This retrospective study included 120 patients (88 men, 32 women; median age, 60.0 years) with early-stage HCC diagnosed who were initially eligible for liver transplant and underwent treatment by transplant, resection, or thermal ablation between June 2005 and March 2018. Patients underwent pretreatment MRI and posttreatment imaging surveillance. Imaging features were extracted from postcontrast phases of pretreatment MRI examinations using a pretrained convolutional neural network. Pretreatment clinical characteristics (including laboratory data) and extracted imaging features were integrated to develop three ML models (clinical model, imaging model, combined model) for predicting recurrence within six time frames ranging from 1 through 6 years after treatment. Kaplan-Meier analysis with time to recurrence as the endpoint was used to assess the clinical relevance of model predictions. RESULTS. Tumor recurred in 44 of 120 (36.7%) patients during follow-up. The three models predicted recurrence with AUCs across the six time frames of 0.60-0.78 (clinical model), 0.71-0.85 (imaging model), and 0.62-0.86 (combined model). The mean AUC was higher for the imaging model than the clinical model (0.76 vs 0.68, respectively; p = .03), but the mean AUC was not significantly different between the clinical and combined models or between the imaging and combined models (p > .05). Kaplan-Meier curves were significantly different between patients predicted to be at low risk and those predicted to be at high risk by all three models for the 2-, 3-, 4-, 5-, and 6-year time frames (p < .05). CONCLUSION. The findings suggest that ML-based models can predict recurrence before therapy allocation in patients with early-stage HCC initially eligible for liver transplant. Adding MRI data as model input improved predictive performance over clinical parameters alone. The combined model did not surpass the imaging model's performance. CLINICAL IMPACT. ML-based models applied to currently underutilized imaging features may help design more reliable criteria for organ allocation and liver transplant eligibility.

Quantitative Automated Segmentation of Lipiodol Deposits on Cone-Beam CT Imaging Acquired during Transarterial Chemoembolization for Liver Tumors: A Deep Learning Approach.

PURPOSE: To show that a deep learning (DL)-based, automated model for Lipiodol (Guerbet Pharmaceuticals, Paris, France) segmentation on cone-beam computed tomography (CT) after conventional transarterial chemoembolization performs closer to the "ground truth segmentation" than a conventional thresholding-based model. MATERIALS AND METHODS: This post hoc analysis included 36 patients with a diagnosis of hepatocellular carcinoma or other solid liver tumors who underwent conventional transarterial chemoembolization with an intraprocedural cone-beam CT. Semiautomatic segmentation of Lipiodol was obtained. Subsequently, a convolutional U-net model was used to output a binary mask that predicted Lipiodol deposition. A threshold value of signal intensity on cone-beam CT was used to obtain a Lipiodol mask for comparison. The dice similarity coefficient (DSC), mean squared error (MSE), center of mass (CM), and fractional volume ratios for both masks were obtained by comparing them to the ground truth (radiologist-segmented Lipiodol deposits) to obtain accuracy metrics for the 2 masks. These results were used to compare the model versus the threshold technique. RESULTS: For all metrics, the U-net outperformed the threshold technique: DSC (0.65 ± 0.17 vs 0.45 ± 0.22, P < .001) and MSE (125.53 ± 107.36 vs 185.98 ± 93.82, P = .005). The difference between the CM predicted and the actual CM was 15.31 mm ± 14.63 versus 31.34 mm ± 30.24 (P < .001), with lesser distance indicating higher accuracy. The fraction of volume present ([predicted Lipiodol volume]/[ground truth Lipiodol volume]) was 1.22 ± 0.84 versus 2.58 ± 3.52 (P = .048) for the current model's prediction and threshold technique, respectively. CONCLUSIONS: This study showed that a DL framework could detect Lipiodol in cone-beam CT imaging and was capable of outperforming the conventionally used thresholding technique over several metrics. Further optimization will allow for more accurate, quantitative predictions of Lipiodol depositions intraprocedurally.

MR Imaging Biomarkers for the Prediction of Outcome after Radiofrequency Ablation of Hepatocellular Carcinoma: Qualitative and Quantitative Assessments of the Liver Imaging Reporting and Data System and Radiomic Features.

PURPOSE: To assess the Liver Imaging Reporting and Data System (LI-RADS) and radiomic features in pretreatment magnetic resonance (MR) imaging for predicting progression-free survival (PFS) in patients with nodular hepatocellular carcinoma (HCC) treated with radiofrequency (RF) ablation. MATERIAL AND METHODS: Sixty-five therapy-naïve patients with 85 nodular HCC tumors <5 cm in size were included in this Health Insurance Portability and Accountability Act-compliant, institutional review board-approved, retrospective study. All patients underwent RF ablation as first-line treatment and demonstrated complete response on the first follow-up imaging. Gadolinium-enhanced MR imaging biomarkers were analyzed for LI-RADS features by 2 board-certified radiologists or by analysis of nodular and perinodular radiomic features from 3-dimensional segmentations. A radiomic signature was calculated with the most informative features of a least absolute shrinkage and selection operator Cox regression model using leave-one-out cross-validation. The association between both LI-RADS features and radiomic signatures with PFS was assessed via the Kaplan-Meier analysis and a weighted log-rank test. RESULTS: The median PFS was 19 months (95% confidence interval, 16.1-19.4) for a follow-up period of 24 months. Multifocality (P = .033); the appearance of capsular continuity, compared with an absent or discontinuous capsule (P = .012); and a higher radiomic signature based on nodular and perinodular features (P = .030) were associated with poorer PFS in early-stage HCC. The observation size, presence of arterial hyperenhancement, nonperipheral washout, and appearance of an enhancing "capsule" were not associated with PFS (P > .05). CONCLUSIONS: Although multifocal HCC clearly indicates a more aggressive phenotype even in early-stage disease, the continuity of an enhancing capsule and a higher radiomic signature may add value as MR imaging biomarkers for poor PFS in HCC treated with RF ablation.

Reliable prediction of survival in advanced-stage hepatocellular carcinoma treated with sorafenib: comparing 1D and 3D quantitative tumor response criteria on MRI.

OBJECTIVES: To compare 1D and 3D quantitative tumor response criteria applied to DCE-MRI in patients with advanced-stage HCC undergoing sorafenib therapy to predict overall survival (OS) early during treatment. METHODS: This retrospective analysis included 29 patients with advanced-stage HCC who received sorafenib for at least 60 days. All patients underwent baseline and follow-up DCE-MRI at 81.5 ± 29.3 days (range 35-140 days). Response to sorafenib was assessed in 46 target lesions using 1D criteria RECIST1.1 and mRECIST. In addition, a segmentation-based 3D quantification of absolute enhancing lesion volume (vqEASL) was performed on the arterial phase MRI, and the enhancement fraction of total tumor volume (%qEASL) was calculated. Accordingly, patients were stratified into groups of disease control (DC) and disease progression (DP). OS was evaluated using Kaplan-Meier curves with log-rank test and Cox proportional hazards regression model. RESULTS: The Kaplan-Meier analysis revealed that stratification of patients in DC vs. DP according to mRECIST (p = 0.0371) and vqEASL (p = 0.0118) successfully captured response and stratified OS, while stratification according to RECIST and %qEASL did not correlate with OS (p = 0.6273 and p = 0.7474, respectively). Multivariable Cox regression identified tumor progression according to mRECIST and qEASL as independent risk factors of decreased OS (p = 0.039 and p = 0.006, respectively). CONCLUSIONS: The study identified enhancement-based vqEASL and mRECIST as reliable predictors of patient survival early after initiation of treatment with sorafenib. This data provides evidence for potential advantages 3D quantitative, enhancement-based tumor response analysis over conventional techniques regarding early identification of treatment success or failure. KEY POINTS: • Tumor response criteria on MRI can be used to predict survival benefit of sorafenib therapy in patients with advanced HCC. • Stratification into DC and DP using mRECIST and vqEASL significantly correlates with OS (p = 0.0371 and p = 0.0118, respectively) early after initiation of sorafenib, while stratification according to RECIST and %qEASL did not correlate with OS (p = 0.6273 and p = 0.7474, respectively). • mRECIST (HR = 0.325, p = 0.039. 95%CI 0.112-0.946) and qEASL (HR = 0.183, p = 0.006, 95%CI 0.055-0.613) are independent prognostic factors of survival in HCC patients undergoing sorafenib therapy.

Admission computed tomography radiomic signatures outperform hematoma volume in predicting baseline clinical severity and functional outcome in the ATACH-2 trial intracerebral hemorrhage population.

BACKGROUND AND PURPOSE: Radiomics provides a framework for automated extraction of high-dimensional feature sets from medical images. We aimed to determine radiomics signature correlates of admission clinical severity and medium-term outcome from intracerebral hemorrhage (ICH) lesions on baseline head computed tomography (CT). METHODS: We used the ATACH-2 (Antihypertensive Treatment of Acute Cerebral Hemorrhage II) trial dataset. Patients included in this analysis (n = 895) were randomly allocated to discovery (n = 448) and independent validation (n = 447) cohorts. We extracted 1130 radiomics features from hematoma lesions on baseline noncontrast head CT scans and generated radiomics signatures associated with admission Glasgow Coma Scale (GCS), admission National Institutes of Health Stroke Scale (NIHSS), and 3-month modified Rankin Scale (mRS) scores. Spearman's correlation between radiomics signatures and corresponding target variables was compared with hematoma volume. RESULTS: In the discovery cohort, radiomics signatures, compared to ICH volume, had a significantly stronger association with admission GCS (0.47 vs. 0.44, p = 0.008), admission NIHSS (0.69 vs. 0.57, p < 0.001), and 3-month mRS scores (0.44 vs. 0.32, p < 0.001). Similarly, in independent validation, radiomics signatures, compared to ICH volume, had a significantly stronger association with admission GCS (0.43 vs. 0.41, p = 0.02), NIHSS (0.64 vs. 0.56, p < 0.001), and 3-month mRS scores (0.43 vs. 0.33, p < 0.001). In multiple regression analysis adjusted for known predictors of ICH outcome, the radiomics signature was an independent predictor of 3-month mRS in both cohorts. CONCLUSIONS: Limited by the enrollment criteria of the ATACH-2 trial, we showed that radiomics features quantifying hematoma texture, density, and shape on baseline CT can provide imaging correlates for clinical presentation and 3-month outcome. These findings couldtrigger a paradigm shift where imaging biomarkers may improve current modelsfor prognostication, risk-stratification, and treatment triage of ICH patients.

PET/CT radiomics signature of human papilloma virus association in oropharyngeal squamous cell carcinoma.

PURPOSE: To devise, validate, and externally test PET/CT radiomics signatures for human papillomavirus (HPV) association in primary tumors and metastatic cervical lymph nodes of oropharyngeal squamous cell carcinoma (OPSCC). METHODS: We analyzed 435 primary tumors (326 for training, 109 for validation) and 741 metastatic cervical lymph nodes (518 for training, 223 for validation) using FDG-PET and non-contrast CT from a multi-institutional and multi-national cohort. Utilizing 1037 radiomics features per imaging modality and per lesion, we trained, optimized, and independently validated machine-learning classifiers for prediction of HPV association in primary tumors, lymph nodes, and combined "virtual" volumes of interest (VOI). PET-based models were additionally validated in an external cohort. RESULTS: Single-modality PET and CT final models yielded similar classification performance without significant difference in independent validation; however, models combining PET and CT features outperformed single-modality PET- or CT-based models, with receiver operating characteristic area under the curve (AUC) of 0.78, and 0.77 for prediction of HPV association using primary tumor lesion features, in cross-validation and independent validation, respectively. In the external PET-only validation dataset, final models achieved an AUC of 0.83 for a virtual VOI combining primary tumor and lymph nodes, and an AUC of 0.73 for a virtual VOI combining all lymph nodes. CONCLUSION: We found that PET-based radiomics signatures yielded similar classification performance to CT-based models, with potential added value from combining PET- and CT-based radiomics for prediction of HPV status. While our results are promising, radiomics signatures may not yet substitute tissue sampling for clinical decision-making.

Effects of ambient conditions on the risk of pressure injuries in bedridden patients-multi-physics modelling of microclimate.

Scientific evidence regarding microclimate and its effects on the risk of pressure ulcers (PU) remains sparse. It is known that elevated skin temperatures and moisture may affect metabolic demand as well as the mechanical behaviour of the tissue. In this study, we incorporated these microclimate factors into a novel, 3-dimensional multi-physics coupled model of the human buttocks, which simultaneously determines the biothermal and biomechanical behaviours of the buttocks in supine lying on different support surfaces. We compared 3 simulated thermally controlled mattresses with 2 reference foam mattresses. A tissue damage score was numerically calculated in a relevant volume of the model, and the cooling effect of each 1°C decrease of tissue temperature was deduced. Damage scores of tissues were substantially lower for the non-foam mattresses compared with the foams. The percentage tissue volume at risk within the volume of interest was found to grow exponentially as the average tissue temperature increased. The resultant average sacral skin temperature was concluded to be a good predictor for an increased risk of PU/injuries. Each 1°C increase contributes approximately 14 times as much to the risk with respect to an increase of 1 mmHg of pressure. These findings highlight the advantages of using thermally controlled support surfaces as well as the need to further assess the potential damage that may be caused by uncontrolled microclimate conditions on inadequate support surfaces in at-risk patients.

Reliable Prostate Cancer Risk Mapping From MRI Using Targeted and Systematic Core Needle Biopsy Histopathology.

OBJECTIVE: To compute a dense prostate cancer risk map for the individual patient post-biopsy from magnetic resonance imaging (MRI) and to provide a more reliable evaluation of its fitness in prostate regions that were not identified as suspicious for cancer by a human-reader in pre- and intra-biopsy imaging analysis. METHODS: Low-level pre-biopsy MRI biomarkers from targeted and non-targeted biopsy locations were extracted and statistically tested for representativeness against biomarkers from non-biopsied prostate regions. A probabilistic machine learning classifier was optimized to map biomarkers to their core-level pathology, followed by extrapolation of pathology scores to non-biopsied prostate regions. Goodness-of-fit was assessed at targeted and non-targeted biopsy locations for the post-biopsy individual patient. RESULTS: Our experiments showed high predictability of imaging biomarkers in differentiating histopathology scores in thousands of non-targeted core-biopsy locations (ROC-AUCs: 0.85-0.88), but also high variability between patients (Median ROC-AUC [IQR]: 0.81-0.89 [0.29-0.40]). CONCLUSION: The sparseness of prostate biopsy data makes the validation of a whole gland risk mapping a non-trivial task. Previous studies i) focused on targeted-biopsy locations although biopsy-specimens drawn from systematically scattered locations across the prostate constitute a more representative sample to non-biopsied regions, and ii) estimated prediction-power across predicted instances (e.g., biopsy specimens) with no patient distinction, which may lead to unreliable estimation of model fitness to the individual patient due to variation between patients in instance count, imaging characteristics, and pathologies. SIGNIFICANCE: This study proposes a personalized whole-gland prostate cancer risk mapping post-biopsy to allow clinicians to better stage and personalize focal therapy treatment plans.

Impact of 18F-FDG PET Intensity Normalization on Radiomic Features of Oropharyngeal Squamous Cell Carcinomas and Machine Learning-Generated Biomarkers.

We aimed to investigate the effects of 18F-FDG PET voxel intensity normalization on radiomic features of oropharyngeal squamous cell carcinoma (OPSCC) and machine learning-generated radiomic biomarkers. Methods: We extracted 1,037 18F-FDG PET radiomic features quantifying the shape, intensity, and texture of 430 OPSCC primary tumors. The reproducibility of individual features across 3 intensity-normalized images (body-weight SUV, reference tissue activity ratio to lentiform nucleus of brain and cerebellum) and the raw PET data was assessed using an intraclass correlation coefficient (ICC). We investigated the effects of intensity normalization on the features' utility in predicting the human papillomavirus (HPV) status of OPSCCs in univariate logistic regression, receiver-operating-characteristic analysis, and extreme-gradient-boosting (XGBoost) machine-learning classifiers. Results: Of 1,037 features, a high (ICC ≥ 0.90), medium (0.90 > ICC ≥ 0.75), and low (ICC < 0.75) degree of reproducibility across normalization methods was attained in 356 (34.3%), 608 (58.6%), and 73 (7%) features, respectively. In univariate analysis, features from the PET normalized to the lentiform nucleus had the strongest association with HPV status, with 865 of 1,037 (83.4%) significant features after multiple testing corrections and a median area under the receiver-operating-characteristic curve (AUC) of 0.65 (interquartile range, 0.62-0.68). Similar tendencies were observed in XGBoost models, with the lentiform nucleus-normalized model achieving the numerically highest average AUC of 0.72 (SD, 0.07) in the cross validation within the training cohort. The model generalized well to the validation cohorts, attaining an AUC of 0.73 (95% CI, 0.60-0.85) in independent validation and 0.76 (95% CI, 0.58-0.95) in external validation. The AUCs of the XGBoost models were not significantly different. Conclusion: Only one third of the features demonstrated a high degree of reproducibility across intensity-normalization techniques, making uniform normalization a prerequisite for interindividual comparability of radiomic markers. The choice of normalization technique may affect the radiomic features' predictive value with respect to HPV. Our results show trends that normalization to the lentiform nucleus may improve model performance, although more evidence is needed to draw a firm conclusion.

Peri-hematomal edema shape features related to 3-month outcome in acute supratentorial intracerebral hemorrhage.

INTRODUCTION: Perihematomal edema (PHE) represents secondary brain injury and a potential treatment target in intracerebral hemorrhage (ICH). However, studies differ on optimal PHE volume metrics as prognostic factor(s) after spontaneous, non-traumatic ICH. This study examines associations of baseline and 24-h PHE shape features with 3-month outcomes. PATIENTS AND METHODS: We included 796 patients from a multicentric trial dataset and manually segmented ICH and PHE on baseline and follow-up CTs, extracting 14 shape features. We explored the association of baseline, follow-up, difference (baseline/follow-up) and temporal rate (difference/time gap) of PHE shape changes with 3-month modified Rankin Score (mRS) - using Spearman correlation. Then, using multivariable analysis, we determined if PHE shape features independently predict outcome adjusting for patients' age, sex, NIH stroke scale (NIHSS), Glasgow Coma Scale (GCS), and hematoma volume. RESULTS: Baseline PHE maximum diameters across various planes, main axes, volume, surface, and sphericity correlated with 3-month mRS adjusting for multiple comparisons. The 24-h difference and temporal change rates of these features had significant association with outcome - but not the 24-h absolute values. In multivariable regression, baseline PHE shape sphericity (OR = 2.04, CI = 1.71-2.43) and volume (OR = 0.99, CI = 0. 98-1.0), alongside admission NIHSS (OR = 0.86, CI = 0.83-0.88), hematoma volume (OR = 0.99, CI = 0. 99-1.0), and age (OR = 0.96, CI = 0.95-0.97) were independent predictors of favorable outcomes. CONCLUSION: In acute ICH patients, PHE shape sphericity at baseline emerged as an independent prognostic factor, with a less spherical (more irregular) shape associated with worse outcome. The PHE shape features absolute values over the first 24 h provide no added prognostic value to baseline metrics.

Uncertainty-aware deep-learning model for prediction of supratentorial hematoma expansion from admission non-contrast head computed tomography scan.

Hematoma expansion (HE) is a modifiable risk factor and a potential treatment target in patients with intracerebral hemorrhage (ICH). We aimed to train and validate deep-learning models for high-confidence prediction of supratentorial ICH expansion, based on admission non-contrast head Computed Tomography (CT). Applying Monte Carlo dropout and entropy of deep-learning model predictions, we estimated the model uncertainty and identified patients at high risk of HE with high confidence. Using the receiver operating characteristics area under the curve (AUC), we compared the deep-learning model prediction performance with multivariable models based on visual markers of HE determined by expert reviewers. We randomly split a multicentric dataset of patients (4-to-1) into training/cross-validation (n = 634) versus test (n = 159) cohorts. We trained and tested separate models for prediction of ≥6 mL and ≥3 mL ICH expansion. The deep-learning models achieved an AUC = 0.81 for high-confidence prediction of HE≥6 mL and AUC = 0.80 for prediction of HE≥3 mL, which were higher than visual maker models AUC = 0.69 for HE≥6 mL (p = 0.036) and AUC = 0.68 for HE≥3 mL (p = 0.043). Our results show that fully automated deep-learning models can identify patients at risk of supratentorial ICH expansion based on admission non-contrast head CT, with high confidence, and more accurately than benchmark visual markers.

Radiomics-Based Prediction of Collateral Status from CT Angiography of Patients Following a Large Vessel Occlusion Stroke.

BACKGROUND: A major driver of individual variation in long-term outcomes following a large vessel occlusion (LVO) stroke is the degree of collateral arterial circulation. We aimed to develop and evaluate machine-learning models that quantify LVO collateral status using admission computed tomography angiography (CTA) radiomics. METHODS: We extracted 1116 radiomic features from the anterior circulation territories from admission CTAs of 600 patients experiencing an acute LVO stroke. We trained and validated multiple machine-learning models for the prediction of collateral status based on consensus from two neuroradiologists as ground truth. Models were first trained to predict (1) good vs. intermediate or poor, or (2) good vs. intermediate or poor collateral status. Then, model predictions were combined to determine a three-tier collateral score (good, intermediate, or poor). We used the receiver operating characteristics area under the curve (AUC) to evaluate prediction accuracy. RESULTS: We included 499 patients in training and 101 in an independent test cohort. The best-performing models achieved an averaged cross-validation AUC of 0.80 ± 0.05 for poor vs. intermediate/good collateral and 0.69 ± 0.05 for good vs. intermediate/poor, and AUC = 0.77 (0.67-0.87) and AUC = 0.78 (0.70-0.90) in the independent test cohort, respectively. The collateral scores predicted by the radiomics model were correlated with (rho = 0.45, p = 0.002) and were independent predictors of 3-month clinical outcome (p = 0.018) in the independent test cohort. CONCLUSIONS: Automated tools for the assessment of collateral status from admission CTA-such as the radiomics models described here-can generate clinically relevant and reproducible collateral scores to facilitate a timely treatment triage in patients experiencing an acute LVO stroke.

Radiomic markers of intracerebral hemorrhage expansion on non-contrast CT: independent validation and comparison with visual markers.

Objective: To devise and validate radiomic signatures of impending hematoma expansion (HE) based on admission non-contrast head computed tomography (CT) of patients with intracerebral hemorrhage (ICH). Methods: Utilizing a large multicentric clinical trial dataset of hypertensive patients with spontaneous supratentorial ICH, we developed signatures predictive of HE in a discovery cohort (n = 449) and confirmed their performance in an independent validation cohort (n = 448). In addition to n = 1,130 radiomic features, n = 6 clinical variables associated with HE, n = 8 previously defined visual markers of HE, the BAT score, and combinations thereof served as candidate variable sets for signatures. The area under the receiver operating characteristic curve (AUC) quantified signatures' performance. Results: A signature combining select radiomic features and clinical variables attained the highest AUC (95% confidence interval) of 0.67 (0.61-0.72) and 0.64 (0.59-0.70) in the discovery and independent validation cohort, respectively, significantly outperforming the clinical (pdiscovery = 0.02, pvalidation = 0.01) and visual signature (pdiscovery = 0.03, pvalidation = 0.01) as well as the BAT score (pdiscovery < 0.001, pvalidation < 0.001). Adding visual markers to radiomic features failed to improve prediction performance. All signatures were significantly (p < 0.001) correlated with functional outcome at 3-months, underlining their prognostic relevance. Conclusion: Radiomic features of ICH on admission non-contrast head CT can predict impending HE with stable generalizability; and combining radiomic with clinical predictors yielded the highest predictive value. By enabling selective anti-expansion treatment of patients at elevated risk of HE in future clinical trials, the proposed markers may increase therapeutic efficacy, and ultimately improve outcomes.

Predicting tumor recurrence on baseline MR imaging in patients with early-stage hepatocellular carcinoma using deep machine learning.

Tumor recurrence affects up to 70% of early-stage hepatocellular carcinoma (HCC) patients, depending on treatment option. Deep learning algorithms allow in-depth exploration of imaging data to discover imaging features that may be predictive of recurrence. This study explored the use of convolutional neural networks (CNN) to predict HCC recurrence in patients with early-stage HCC from pre-treatment magnetic resonance (MR) images. This retrospective study included 120 patients with early-stage HCC. Pre-treatment MR images were fed into a machine learning pipeline (VGG16 and XGBoost) to predict recurrence within six different time frames (range 1-6 years). Model performance was evaluated with the area under the receiver operating characteristic curves (AUC-ROC). After prediction, the model's clinical relevance was evaluated using Kaplan-Meier analysis with recurrence-free survival (RFS) as the endpoint. Of 120 patients, 44 had disease recurrence after therapy. Six different models performed with AUC values between 0.71 to 0.85. In Kaplan-Meier analysis, five of six models obtained statistical significance when predicting RFS (log-rank p < 0.05). Our proof-of-concept study indicates that deep learning algorithms can be utilized to predict early-stage HCC recurrence. Successful identification of high-risk recurrence candidates may help optimize follow-up imaging and improve long-term outcomes post-treatment.

Machine Learning Tools for Image-Based Glioma Grading and the Quality of Their Reporting: Challenges and Opportunities.

Technological innovation has enabled the development of machine learning (ML) tools that aim to improve the practice of radiologists. In the last decade, ML applications to neuro-oncology have expanded significantly, with the pre-operative prediction of glioma grade using medical imaging as a specific area of interest. We introduce the subject of ML models for glioma grade prediction by remarking upon the models reported in the literature as well as by describing their characteristic developmental workflow and widely used classifier algorithms. The challenges facing these models-including data sources, external validation, and glioma grade classification methods -are highlighted. We also discuss the quality of how these models are reported, explore the present and future of reporting guidelines and risk of bias tools, and provide suggestions for the reporting of prospective works. Finally, this review offers insights into next steps that the field of ML glioma grade prediction can take to facilitate clinical implementation.

Prediction of Adverse Pathology at Radical Prostatectomy in Grade Group 2 and 3 Prostate Biopsies Using Machine Learning.

PURPOSE: There is ongoing clinical need to improve estimates of disease outcome in prostate cancer. Machine learning (ML) approaches to pathologic diagnosis and prognosis are a promising and increasingly used strategy. In this study, we use an ML algorithm for prediction of adverse outcomes at radical prostatectomy (RP) using whole-slide images (WSIs) of prostate biopsies with Grade Group (GG) 2 or 3 disease. METHODS: We performed a retrospective review of prostate biopsies collected at our institution which had corresponding RP, GG 2 or 3 disease one or more cores, and no biopsies with higher than GG 3 disease. A hematoxylin and eosin-stained core needle biopsy from each site with GG 2 or 3 disease was scanned and used as the sole input for the algorithm. The ML pipeline had three phases: image preprocessing, feature extraction, and adverse outcome prediction. First, patches were extracted from each biopsy scan. Subsequently, the pre-trained Visual Geometry Group-16 convolutional neural network was used for feature extraction. A representative feature vector was then used as input to an Extreme Gradient Boosting classifier for predicting the binary adverse outcome. We subsequently assessed patient clinical risk using CAPRA score for comparison with the ML pipeline results. RESULTS: The data set included 361 WSIs from 107 patients (56 with adverse pathology at RP). The area under the receiver operating characteristic curves for the ML classification were 0.72 (95% CI, 0.62 to 0.81), 0.65 (95% CI, 0.53 to 0.79) and 0.89 (95% CI, 0.79 to 1.00) for the entire cohort, and GG 2 and GG 3 patients, respectively, similar to the performance of the CAPRA clinical risk assessment. CONCLUSION: We provide evidence for the potential of ML algorithms to use WSIs of needle core prostate biopsies to estimate clinically relevant prostate cancer outcomes.

Machine Learning Models for Classifying High- and Low-Grade Gliomas: A Systematic Review and Quality of Reporting Analysis.

Objectives: To systematically review, assess the reporting quality of, and discuss improvement opportunities for studies describing machine learning (ML) models for glioma grade prediction. Methods: This study followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses of Diagnostic Test Accuracy (PRISMA-DTA) statement. A systematic search was performed in September 2020, and repeated in January 2021, on four databases: Embase, Medline, CENTRAL, and Web of Science Core Collection. Publications were screened in Covidence, and reporting quality was measured against the Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis (TRIPOD) Statement. Descriptive statistics were calculated using GraphPad Prism 9. Results: The search identified 11,727 candidate articles with 1,135 articles undergoing full text review and 85 included in analysis. 67 (79%) articles were published between 2018-2021. The mean prediction accuracy of the best performing model in each study was 0.89 ± 0.09. The most common algorithm for conventional machine learning studies was Support Vector Machine (mean accuracy: 0.90 ± 0.07) and for deep learning studies was Convolutional Neural Network (mean accuracy: 0.91 ± 0.10). Only one study used both a large training dataset (n>200) and external validation (accuracy: 0.72) for their model. The mean adherence rate to TRIPOD was 44.5% ± 11.1%, with poor reporting adherence for model performance (0%), abstracts (0%), and titles (0%). Conclusions: The application of ML to glioma grade prediction has grown substantially, with ML model studies reporting high predictive accuracies but lacking essential metrics and characteristics for assessing model performance. Several domains, including generalizability and reproducibility, warrant further attention to enable translation into clinical practice. Systematic Review Registration: PROSPERO, identifier CRD42020209938.

CT angiographic radiomics signature for risk stratification in anterior large vessel occlusion stroke.

BACKGROUND AND PURPOSE: As "time is brain" in acute stroke triage, the need for automated prognostication tools continues to increase, particularly in rapidly expanding tele-stroke settings. We aimed to create an automated prognostication tool for anterior circulation large vessel occlusion (LVO) stroke based on admission CTA radiomics. METHODS: We automatically extracted 1116 radiomics features from the anterior circulation territory on admission CTAs of 829 acute LVO stroke patients who underwent mechanical thrombectomy in two academic centers. We trained, optimized, validated, and compared different machine-learning models to predict favorable outcome (modified Rankin Scale ≤ 2) at discharge and 3-month follow-up using four different input sets: "Radiomics", "Radiomics + Treatment" (radiomics, post-thrombectomy reperfusion grade, and intravenous thrombolysis), "Clinical + Treatment" (baseline clinical variables and treatment), and "Combined" (radiomics, treatment, and baseline clinical variables). RESULTS: For discharge outcome prediction, models were optimized/trained on n = 494 and tested on an independent cohort of n = 100 patients from Yale. Receiver operating characteristic analysis of the independent cohort showed no significant difference between best-performing Combined input models (area under the curve, AUC = 0.77) versus Radiomics + Treatment (AUC = 0.78, p = 0.78), Radiomics (AUC = 0.78, p = 0.55), or Clinical + Treatment (AUC = 0.77, p = 0.87) models. For 3-month outcome prediction, models were optimized/trained on n = 373 and tested on an independent cohort from Yale (n = 72), and an external cohort from Geisinger Medical Center (n = 232). In the independent cohort, there was no significant difference between Combined input models (AUC = 0.76) versus Radiomics + Treatment (AUC = 0.72, p = 0.39), Radiomics (AUC = 0.72, p = 0.39), or Clinical + Treatment (AUC = 76, p = 0.90) models; however, in the external cohort, the Combined model (AUC = 0.74) outperformed Radiomics + Treatment (AUC = 0.66, p < 0.001) and Radiomics (AUC = 0.68, p = 0.005) models for 3-month prediction. CONCLUSION: Machine-learning signatures of admission CTA radiomics can provide prognostic information in acute LVO stroke candidates for mechanical thrombectomy. Such objective and time-sensitive risk stratification can guide treatment decisions and facilitate tele-stroke assessment of patients. Particularly in the absence of reliable clinical information at the time of admission, models solely using radiomics features can provide a useful prognostication tool.

Machine Learning Applications for Differentiation of Glioma from Brain Metastasis-A Systematic Review.

Glioma and brain metastasis can be difficult to distinguish on conventional magnetic resonance imaging (MRI) due to the similarity of imaging features in specific clinical circumstances. Multiple studies have investigated the use of machine learning (ML) models for non-invasive differentiation of glioma from brain metastasis. Many of the studies report promising classification results, however, to date, none have been implemented into clinical practice. After a screening of 12,470 studies, we included 29 eligible studies in our systematic review. From each study, we aggregated data on model design, development, and best classifiers, as well as quality of reporting according to the TRIPOD statement. In a subset of eligible studies, we conducted a meta-analysis of the reported AUC. It was found that data predominantly originated from single-center institutions (n = 25/29) and only two studies performed external validation. The median TRIPOD adherence was 0.48, indicating insufficient quality of reporting among surveyed studies. Our findings illustrate that despite promising classification results, reliable model assessment is limited by poor reporting of study design and lack of algorithm validation and generalizability. Therefore, adherence to quality guidelines and validation on outside datasets is critical for the clinical translation of ML for the differentiation of glioma and brain metastasis.

Dataset on acute stroke risk stratification from CT angiographic radiomics.

With advances in high-throughput image processing technologies and increasing availability of medical mega-data, the growing field of radiomics opened the door for quantitative analysis of medical images for prediction of clinically relevant information. One clinical area in which radiomics have proven useful is stroke neuroimaging, where rapid treatment triage is vital for patient outcomes and automated decision assistance tools have potential for significant clinical impact. Recent research, for example, has applied radiomics features extracted from CT angiography (CTA) images and a machine learning framework to facilitate risk-stratification in acute stroke. We here provide methodological guidelines and radiomics data supporting the referenced article "CT angiographic radiomics signature for risk-stratification in anterior large vessel occlusion stroke." The data were extracted from the stroke center registry at Yale New Haven Hospital between 1/1/2014 and 10/31/2020; and Geisinger Medical Center between 1/1/2016 and 12/31/2019. It includes detailed radiomics features of the anterior circulation territories on admission CTA scans in stroke patients with large vessel occlusion stroke who underwent thrombectomy. We also provide the methodological details of the analysis framework utilized for training, optimization, validation and external testing of the machine learning and feature selection algorithms. With the goal of advancing the feasibility and quality of radiomics-based analyses to improve patient care within and beyond the field of stroke, the provided data and methodological support can serve as a baseline for future studies applying radiomics algorithms to machine-learning frameworks, and allow for analysis and utilization of radiomics features extracted in this study.