Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): the TRIPOD statement

Development and external validation of a head and neck cancer risk prediction model

BACKGROUND

Head and neck cancer (HNC) incidence is on the rise, often diagnosed at late stage and associated with poor prognoses. Risk prediction tools have a potential role in prevention and early detection.

METHODS

The IARC-ARCAGE European case-control study was used as the model development dataset. A clinical HNC risk prediction model using behavioral and demographic predictors was developed via multivariable logistic regression analyses. The model was then externally validated in the UK Biobank cohort. Model performance was tested using discrimination and calibration metrics.

RESULTS

1926 HNC cases and 2043 controls were used for the development of the model. The development dataset model including sociodemographic, smoking, and alcohol variables had moderate discrimination, with an area under curve (AUC) value of 0.75 (95% CI, 0.74-0.77); the calibration slope (0.75) and tests were suggestive of good calibration. 384 616 UK Biobank participants (with 1177 HNC cases) were available for external validation of the model. Upon external validation, the model had an AUC of 0.62 (95% CI, 0.61-0.64).

CONCLUSION

We developed and externally validated a HNC risk prediction model using the ARCAGE and UK Biobank studies, respectively. This model had moderate performance in the development population and acceptable performance in the validation dataset. Demographics and risk behaviors are strong predictors of HNC, and this model may be a helpful tool in primary dental care settings to promote prevention and determine recall intervals for dental examination. Future addition of HPV serology or genetic factors could further enhance individual risk prediction.

ICGA-GPT: report generation and question answering for indocyanine green angiography images

BACKGROUND

Indocyanine green angiography (ICGA) is vital for diagnosing chorioretinal diseases, but its interpretation and patient communication require extensive expertise and time-consuming efforts. We aim to develop a bilingual ICGA report generation and question-answering (QA) system.

METHODS

Our dataset comprised 213 129 ICGA images from 2919 participants. The system comprised two stages: image-text alignment for report generation by a multimodal transformer architecture, and large language model (LLM)-based QA with ICGA text reports and human-input questions. Performance was assessed using both qualitative metrics (including Bilingual Evaluation Understudy (BLEU), Consensus-based Image Description Evaluation (CIDEr), Recall-Oriented Understudy for Gisting Evaluation-Longest Common Subsequence (ROUGE-L), Semantic Propositional Image Caption Evaluation (SPICE), accuracy, sensitivity, specificity, precision and F1 score) and subjective evaluation by three experienced ophthalmologists using 5-point scales (5 refers to high quality).

RESULTS

We produced 8757 ICGA reports covering 39 disease-related conditions after bilingual translation (66.7% English, 33.3% Chinese). The ICGA-GPT model's report generation performance was evaluated with BLEU scores (1-4) of 0.48, 0.44, 0.40 and 0.37; CIDEr of 0.82; ROUGE of 0.41 and SPICE of 0.18. For disease-based metrics, the average specificity, accuracy, precision, sensitivity and F1 score were 0.98, 0.94, 0.70, 0.68 and 0.64, respectively. Assessing the quality of 50 images (100 reports), three ophthalmologists achieved substantial agreement (kappa=0.723 for completeness, kappa=0.738 for accuracy), yielding scores from 3.20 to 3.55. In an interactive QA scenario involving 100 generated answers, the ophthalmologists provided scores of 4.24, 4.22 and 4.10, displaying good consistency (kappa=0.779).

CONCLUSION

This pioneering study introduces the ICGA-GPT model for report generation and interactive QA for the first time, underscoring the potential of LLMs in assisting with automated ICGA image interpretation.

Prediction model and assessment of malnutrition in patients with stable chronic obstructive pulmonary disease

Chronic obstructive pulmonary disease (COPD) combined with malnutrition results in decreased exercise capacity and a worse quality of life. We aimed to develop an observational case-control study to explore the effective and convenient method to identify potential individuals is lacking. This study included data from 251 patients with COPD and 85 participants in the control group. Parameters and body composition were compared between groups, and among patients with varied severity. The LASSO approach was employed to select the features for fitting a logistic model to predict the risk of malnutrition in patients with stable COPD. Patients with COPD exhibited significantly lower 6-min walk distance (6MWD), handgrip strength, fat-free mass index (FFMI), skeletal muscle mass (SMM) and protein. The significant predictors identified following LASSO selection included 6MWD, waist-to-hip ratio (WHR), GOLD grades, the COPD Assessment Test (CAT) score, and the prevalence of acute exacerbations. The risk score model yielded good accuracy (C-index, 0.866 [95% CI 0.824-0.909]) and calibration (Brier score = 0.150). After internal validation, the adjusted C-index and Brier score were 0.849, and 0.165, respectively. This model may provide primary physicians with a simple scoring system to identify malnourished patients with COPD and develop appropriate rehabilitation interventions.

Prediction of symptomatic anastomotic leak after rectal cancer surgery: A machine learning approach

INTRODUCTION

Anastomotic leakage (AL) remains the most dreaded and unpredictable major complication after low anterior resection for mid-low rectal cancer. The aim of this study is to identify patients with high risk for AL based on the machine learning method.

METHODS

Patients with mid-low rectal cancer undergoing low anterior resection were enrolled from West China Hospital between January 2008 and October 2019 and were split by time into training cohort and validation cohort. The least absolute shrinkage and selection operator (LASSO) method and stepwise method were applied for variable selection and predictive model building in the training cohort. The area under the receiver operating characteristic curve (AUC) and calibration curves were used to evaluate the performance of the models.

RESULTS

The rate of AL was 5.8% (38/652) and 7.2% (15/208) in the training cohort and validation cohort, respectively. The LASSO-logistic model selected almost the same variables (hypertension, operating time, cT4, tumor location, intraoperative blood loss) compared to the stepwise logistic model except for tumor size (the LASSO-logistic model) and American Society of Anesthesiologists score (the stepwise logistic model). The predictive performance of the LASSO-logistics model was better than the stepwise-logistics model (AUC: 0.790 vs. 0.759). Calibration curves showed mean absolute error of 0.006 and 0.013 for the LASSO-logistics model and stepwise-logistics model, respectively.

CONCLUSION

Our study developed a feasible predictive model with a machine-learning algorithm to classify patients with a high risk of AL, which would assist surgical decision-making and reduce unnecessary stoma diversion. The involved machine learning algorithms provide clinicians with an innovative alternative to enhance clinical management.

Developing and Validating a Multivariable Prognostic-Predictive Classifier for Treatment Escalation of Oropharyngeal Squamous Cell Carcinoma: The PREDICTR-OPC Study

PURPOSE

While there are several prognostic classifiers, to date, there are no validated predictive models that inform treatment selection for oropharyngeal squamous cell carcinoma (OPSCC).Our aim was to develop clinical and/or biomarker predictive models for patient outcome and treatment escalation for OPSCC.

EXPERIMENTAL DESIGN

We retrospectively collated clinical data and samples from a consecutive cohort of OPSCC cases treated with curative intent at ten secondary care centers in United Kingdom and Poland between 1999 and 2012. We constructed tissue microarrays, which were stained and scored for 10 biomarkers. We then undertook multivariable regression of eight clinical parameters and 10 biomarkers on a development cohort of 600 patients. Models were validated on an independent, retrospectively collected, 385-patient cohort.

RESULTS

A total of 985 subjects (median follow-up 5.03 years, range: 4.73-5.21 years) were included. The final biomarker classifier, comprising p16 and survivin immunohistochemistry, high-risk human papillomavirus (HPV) DNA in situ hybridization, and tumor-infiltrating lymphocytes, predicted benefit from combined surgery + adjuvant chemo/radiotherapy over primary chemoradiotherapy in the high-risk group [3-year overall survival (OS) 63.1% vs. 41.1%, respectively, HR = 0.32; 95% confidence interval (CI), 0.16-0.65; P = 0.002], but not in the low-risk group (HR = 0.4; 95% CI, 0.14-1.24; P = 0.114). On further adjustment by propensity scores, the adjusted HR in the high-risk group was 0.34, 95% CI = 0.17-0.67, P = 0.002, and in the low-risk group HR was 0.5, 95% CI = 0.1-2.38, P = 0.384. The concordance index was 0.73.

CONCLUSIONS

We have developed a prognostic classifier, which also appears to demonstrate moderate predictive ability. External validation in a prospective setting is now underway to confirm this and prepare for clinical adoption.

An updated PREDICT breast cancer prognostic model including the benefits and harms of radiotherapy

PREDICT Breast ( www.breast .predict.nhs.uk ) is a prognostication tool for early invasive breast cancer. The current version was based on cases diagnosed in 1999-2003 and did not incorporate the benefits of radiotherapy or the harms associated with therapy. Since then, there has been a substantial improvement in the outcomes for breast cancer cases. The aim of this study was to update PREDICT Breast to ensure that the underlying model is appropriate for contemporary patients. Data from the England National Cancer Registration and Advisory Service for invasive breast cancer cases diagnosed 2000-17 were used for model development and validation. Model development was based on 35,474 cases diagnosed and registered by the Eastern Cancer Registry. A Cox model was used to estimate the prognostic effects of the year of diagnosis, age at diagnosis, tumour size, tumour grade and number of positive nodes. Separate models were developed for ER-positive and ER-negative disease. Data on 32,408 cases from the West Midlands Cancer Registry and 100,551 cases from other cancer registries were used for validation. The new model was well-calibrated; predicted breast cancer deaths at 5-, 10- and 15-year were within 10 per cent of the observed validation data. Discrimination was also good: The AUC for 15-year breast cancer survival was 0.809 in the West Midlands data set and 0.846 in the data set for the other registries. The new PREDICT Breast model outperformed the current model and will be implemented in the online tool which should lead to more accurate absolute treatment benefit predictions for individual patients.

Machine learning and deep learning predictive models for long-term prognosis in patients with chronic obstructive pulmonary disease: a systematic review and meta-analysis

BACKGROUND

Machine learning and deep learning models have been increasingly used to predict long-term disease progression in patients with chronic obstructive pulmonary disease (COPD). We aimed to summarise the performance of such prognostic models for COPD, compare their relative performances, and identify key research gaps.

METHODS

We conducted a systematic review and meta-analysis to compare the performance of machine learning and deep learning prognostic models and identify pathways for future research. We searched PubMed, Embase, the Cochrane Library, ProQuest, Scopus, and Web of Science from database inception to April 6, 2023, for studies in English using machine learning or deep learning to predict patient outcomes at least 6 months after initial clinical presentation in those with COPD. We included studies comprising human adults aged 18-90 years and allowed for any input modalities. We reported area under the receiver operator characteristic curve (AUC) with 95% CI for predictions of mortality, exacerbation, and decline in forced expiratory volume in 1 s (FEV1). We reported the degree of interstudy heterogeneity using Cochran's Q test (significant heterogeneity was defined as p≤0·10 or I2>50%). Reporting quality was assessed using the TRIPOD checklist and a risk-of-bias assessment was done using the PROBAST checklist. This study was registered with PROSPERO (CRD42022323052).

FINDINGS

We identified 3620 studies in the initial search. 18 studies were eligible, and, of these, 12 used conventional machine learning and six used deep learning models. Seven models analysed exacerbation risk, with only six reporting AUC and 95% CI on internal validation datasets (pooled AUC 0·77 [95% CI 0·69-0·85]) and there was significant heterogeneity (I2 97%, p<0·0001). 11 models analysed mortality risk, with only six reporting AUC and 95% CI on internal validation datasets (pooled AUC 0·77 [95% CI 0·74-0·80]) with significant degrees of heterogeneity (I2 60%, p=0·027). Two studies assessed decline in lung function and were unable to be pooled. Machine learning and deep learning models did not show significant improvement over pre-existing disease severity scores in predicting exacerbations (p=0·24). Three studies directly compared machine learning models against pre-existing severity scores for predicting mortality and pooled performance did not differ (p=0·57). Of the five studies that performed external validation, performance was worse than or equal to regression models. Incorrect handling of missing data, not reporting model uncertainty, and use of datasets that were too small relative to the number of predictive features included provided the largest risks of bias.

INTERPRETATION

There is limited evidence that conventional machine learning and deep learning prognostic models demonstrate superior performance to pre-existing disease severity scores. More rigorous adherence to reporting guidelines would reduce the risk of bias in future studies and aid study reproducibility.

FUNDING

None.

Development of a population-level prediction model for intensive care unit (ICU) survivorship and mortality in older adults: A population-based cohort study

Background and Aims

Given the growing utilization of critical care services by an aging population, development of population-level risk models which predict intensive care unit (ICU) survivorship and mortality may offer advantages for researchers and health systems. Our objective was to develop a risk model for ICU survivorship and mortality among community dwelling older adults.

Methods

This was a population-based cohort study of 48,127 patients who were 50 years and older with at least one primary care visit between January 1, 2017, and December 31, 2017. We used electronic health record (EHR) data to identify variables predictive of ICU survivorship.

Results

ICU admission and mortality within 2 years after index primary care visit date were used to divide patients into three groups of "alive without ICU admission", "ICU survivors," and "death." Multinomial logistic regression was used to identify EHR predictive variables for the three patient outcomes. Cross-validation by randomly splitting the data into derivation and validation data sets (60:40 split) was used to identify predictor variables and validate model performance using area under the receiver operating characteristics (AUC) curve. In our overall sample, 92.2% of patients were alive without ICU admission, 6.2% were admitted to the ICU at least once and survived, and 1.6% died. Greater deciles of age over 50 years, diagnoses of chronic obstructive pulmonary disorder or chronic heart failure, and laboratory abnormalities in alkaline phosphatase, hematocrit, and albumin contributed highest risk score weights for mortality. Risk scores derived from the model discriminated between patients that died versus remained alive without ICU admission (AUC = 0.858), and between ICU survivors versus alive without ICU admission (AUC = 0.765).

Conclusion

Our risk scores provide a feasible and scalable tool for researchers and health systems to identify patient cohorts at increased risk for ICU admission and survivorship. Further studies are needed to prospectively validate the risk scores in other patient populations.

Prediction of toxicity outcomes following radiotherapy using deep learning-based models: A systematic review

PURPOSE

This study aims to perform a comprehensive systematic review of deep learning (DL) models in predicting RT-induced toxicity.

MATERIALS AND METHODS

A literature review was performed following the PRISMA guidelines. Systematic searches were performed in PubMed, Scopus, Cochrane and Embase databases from the earliest record up to September 2022. Related studies on deep learning models for radiotherapy toxicity prediction were selected based on predefined PICOS criteria.

RESULTS

Fourteen studies of radiotherapy-treated patients on different types of cancer [prostate (n=2), HNC (n=4), liver (n=2), lung (n=4), cervical (n=1), and oesophagus (n=1)] were eligible for inclusion in the systematic review. Information regarding patient characteristics and model development was summarized. Several approaches, such as ensemble learning, data augmentation, and transfer learning, that were utilized by selected studies were discussed.

CONCLUSION

Deep learning techniques are able to produce a consistent performance for toxicity prediction. Future research using large and diverse datasets and standardization of the study methodologies are required to improve the consistency of the research output.

Assessment of bleeding risk in cancer patients treated with anticoagulants for venous thromboembolic events

Introduction

Anticoagulant is the cornerstone of the management of VTE at the cost of a non-negligible risk of bleeding. Reliable and validated clinical tools to predict thromboembolic and hemorrhagic events are crucial for individualized decision-making for the type and duration of anticoagulant treatment. We evaluate the available risk models in real life cancer patients with VTE. The objectives of the study were to describe the bleeding of cancer patients with VTE and to evaluate the performance of the different bleeding models to predict the risk of bleeding during a 6-month follow-up.

Materials and Methods

VTE-diagnosed patient's demographic and clinical characteristics, treatment regimens and outcomes for up to 6 months were collected. The primary endpoint was the occurrence of a major bleeding (MB) or a clinically relevant non major bleeding (CRNMB) event, categorized according to the ISTH criteria.

Results

During the 6-months follow-up period, 26 out of 110 included patients (26.7%) experienced a bleeding event, with 3 recurrences of bleeding. Out of the 29 bleeding events, 19 events were CRNMB and 10 MB. One patient died because of a MB. Bleeding occurred in 27 % of the patients treated with DOACs and 22% of the patients treated with LMWH. Most of the bleedings were gastrointestinal (9 events, 31%); 26.9% of the bleedings occurred in patient with colorectal cancer and 19.6% in patients with lung cancer. In our cohort, none of the 10 RAMs used in our study were able to distinguish cancer patients with a low risk of bleeding, from all bleeding or non-bleeding patients. The Nieto et al. RAM had the best overall performance (C-statistic = 0.730, 95% CI (0.619-0.840)). However, it classified 1 out of 5 patients with major bleeding in the low risk of bleeding group. The rest of the RAMs showed a suboptimal result, with a range of C-statistic between 0.489, 95%CI (0.360-0.617)) and 0.532, 95%CI (0.406-0.658)).

Conclusions

The management of CAT patients is challenging due to a higher risk of both recurrent VTE and bleeding events, as compared with non-cancer patients with VTE. None of the existing RAMs was able to consistently identify patients with risk of anticoagulant associated bleeding events.

A self-supervised vision transformer to predict survival from histopathology in renal cell carcinoma

PURPOSE

To develop and validate an interpretable deep learning model to predict overall and disease-specific survival (OS/DSS) in clear cell renal cell carcinoma (ccRCC).

METHODS

Digitised haematoxylin and eosin-stained slides from The Cancer Genome Atlas were used as a training set for a vision transformer (ViT) to extract image features with a self-supervised model called DINO (self-distillation with no labels). Extracted features were used in Cox regression models to prognosticate OS and DSS. Kaplan-Meier for univariable evaluation and Cox regression analyses for multivariable evaluation of the DINO-ViT risk groups were performed for prediction of OS and DSS. For validation, a cohort from a tertiary care centre was used.

RESULTS

A significant risk stratification was achieved in univariable analysis for OS and DSS in the training (n = 443, log rank test, p < 0.01) and validation set (n = 266, p < 0.01). In multivariable analysis, including age, metastatic status, tumour size and grading, the DINO-ViT risk stratification was a significant predictor for OS (hazard ratio [HR] 3.03; 95%-confidence interval [95%-CI] 2.11-4.35; p < 0.01) and DSS (HR 4.90; 95%-CI 2.78-8.64; p < 0.01) in the training set but only for DSS in the validation set (HR 2.31; 95%-CI 1.15-4.65; p = 0.02). DINO-ViT visualisation showed that features were mainly extracted from nuclei, cytoplasm, and peritumoural stroma, demonstrating good interpretability.

CONCLUSION

The DINO-ViT can identify high-risk patients using histological images of ccRCC. This model might improve individual risk-adapted renal cancer therapy in the future.

Radiomics could predict surgery at 10 years in Crohn's disease

BACKGROUND

Predicting clinical outcomes represents a major challenge in Crohn's disease (CD). Radiomics provides a method to extract quantitative features from medical images and may successfully predict clinical course.

AIMS

The aim of this pilot study is to evaluate the use of radiomics to predict 10-year surgery for CD patients.

METHODS

We selected a cohort of CD patients with CT scan enterographies and a 10-year follow up. The R library Moddicom was used to extract radiomic features from each lesion of CD, segmented in the CT scans. A logistic regression model based on selected radiomic features was developed to predict 10-year surgery. The model was evaluated by computing the area under the curve (AUC) of the receiver operating characteristic curve, sensitivity, specificity, positive and negative predictive values (PPV, NPV).

RESULTS

We enroled 30 patients, with 44 CT scans and 93 lesions. We extracted 217 radiomic features from each lesion. The developed model was based on two radiomic features and presented an AUC (95% CI) of 0.83 (0.73-0.91) in predicting 10-year surgery. Sensitivity, specificity, PPV, NPV of the radiomic model were equal to 0.72, 0.90, 0.79, 0.86, respectively.

CONCLUSION

Radiomics could be a helpful tool to identify patients with high risk for surgery and needing a stricter monitoring.

Development and validation of a clinical prediction model for return to work after arthroscopic anterior crucial ligament reconstruction

BACKGROUND

Occupational reintegration after anterior cruciate ligament (ACL) rupture is an important clinical issue including economic and health-related perspectives. This study aims to develop and validate a clinical prediction model of return to work in patients with ACL reconstruction surgery considering evidence-based clinical, anthropometric and occupational factors.

METHODS

Data of 562 patients with an ACL rupture receiving an arthroscopic ACL reconstruction were used for analysis. A model for the binary outcome of experiencing an inability to work period of less or more than 14 days (model 1), and a model for finding predictor variables that are linearly associated with a continuous longer inability to work period of over 14 days (model 2) was calculated. Pre-operative determinants including patient characteristics and peri-operative factors were used as predictors for both models.

RESULTS

For model 1, the highest increase in odds was observed for the occupational type of work, followed by injury of the medial collateral ligament together with partial weight bearing. Small protective effects were observed for female sex, meniscal suture and work with light occupational strain. The type of occupational work, together with revision surgery, a longer duration of limited range of motion and the presence of cartilage therapy were risk factors for longer inability to work. Discrimination and calibration statistics were satisfactory in internal validation.

CONCLUSION

Within the framework of clinical consideration, these prediction models will serve as an estimator for patients, their treating physicians and the socioeconomic partners to forecast the individual cost and benefit of ACL injury.

Five-year follow-up mortality prognostic index for colorectal patients

PURPOSE

To identify 5-year survival prognostic variables in patients with colorectal cancer (CRC) and to propose a survival prognostic score that also takes into account changes over time in the patient's health-related quality of life (HRQoL) status.

METHODS

Prospective observational cohort study of CRC patients. We collected data from their diagnosis, intervention, and at 1, 2, 3, and 5 years following the index intervention, also collecting HRQoL data using the EuroQol-5D-5L (EQ-5D-5L), European Organization for Research and Treatment of Cancer's Quality of Life Questionnaire-Core 30 (EORTC-QLQ-C30), and Hospital Anxiety and Depression Scale (HADS) questionnaires. Multivariate Cox proportional models were used.

RESULTS

We found predictors of mortality over the 5-year follow-up to be being older; being male; having a higher TNM stage; having a higher lymph node ratio; having a result of CRC surgery classified as R1 or R2; invasion of neighboring organs; having a higher score on the Charlson comorbidity index; having an ASA IV; and having worse scores, worse quality of life, on the EORTC and EQ-5D questionnaires, as compared to those with higher scores in each of those questionnaires respectively.

CONCLUSIONS

These results allow preventive and controlling measures to be established on long-term follow-up of these patients, based on a few easily measurable variables.

IMPLICATIONS FOR CANCER SURVIVORS

Patients with colorectal cancer should be monitored more closely depending on the severity of their disease and comorbidities as well as the perceived health-related quality of life, and preventive measures should be established to prevent adverse outcomes and therefore to ensure that better treatment is received.

TRIAL REGISTRATION

ClinicalTrials.gov identifier: NCT02488161.

A Prediction Nomogram for Recurrent Retinal Detachment

Purpose

Recurrent retinal detachment (re-RD) is one of the complications in rhegmatogenous retinal detachment patients who underwent surgical treatment. We investigated the risk factors for re-RD and developed a nomogram for estimating clinical risk.

Methods

Univariate and multivariable logistic regression models were performed to determine the association between variables and re-RD, and a nomogram was then developed for re-RD. The nomogram performance was assessed based on its discrimination, calibration, and clinical usefulness.

Results

This study analyzed 15 potential variables of re-RD in 403 rhegmatogenous retinal detachment patients who underwent initial surgical treatment. Axial length, inferior breaks, retinal break diameter, and surgical methods were independent risk factors for re-RD. A clinical nomogram incorporating these four independent risk factors was constructed. The diagnostic performance of the nomogram was excellent (area under the curve = 0.892, 95% CI: 0.831-0.953). Our study further validated this nomogram by bootstrapping for 500 repetitions. The area under the curve of the bootstrap model was 0.797 (95% CI: 0.712-0.881). This model showed good calibration curve fitting and a positive net benefit in decision curve analysis.

Conclusion

Axial length, inferior breaks, retinal break diameter, and surgical methods could be risk factors for re-RD. We have developed a prediction nomogram of re-RD for rhegmatogenous retinal detachment following initial surgical treatment.

The risk assessment tool for intensive care unit readmission: A systematic review and meta-analysis

OBJECTIVE

To review and evaluate existing risk assessment tools for intensive care unitreadmission.

METHODS

Nine electronic databases (Medline, CINAHL, Web of Science, Cochrane Library, Embase, Sino Med, CNKI, VIP, and Wan fang) were systematically searched from their inception to September 2022. Two authors independently extracted data from the literature included. Meta-analysis was performed under the bivariate modeling and summary receiver operating characteristic curve method.

RESULTS

A total of 29 studies were included in this review, among which 11 were quantitatively Meta-analyzed. The results showed Stability and Workload Index for Transfer: Sensitivity = 0.55, Specificity = 0.65, Area under curve = 0.63. And Early warning score: Sensitivity = 0.78, Specificity = 0.83, Area under curve = 0.88. The remaining tools included scores, nomograms, machine learning models, and deep learning models. These studies, with varying reports on thresholds, case selection, data preprocessing, and model performance, have a high risk of bias.

CONCLUSION

We cannot identify a tool that can be used directly in intensive care unit readmission risk assessment. Scores based on early warning score are moderately accurate in predicting readmission, but there is heterogeneity and publication bias that requires model adjustment for local factors such as resources, demographics, and case mix. Machine learning models present a promising modeling technique but have a high methodological bias and require further validation.

IMPLICATIONS FOR CLINICAL PRACTICE

Using reliable risk assessment tools is essential for the early identification of unplanned intensive care unit readmission risk in critically ill patients.  A reliable risk assessment tool must be developed, which is the focus of further research.

Explainable ensemble learning model improves identification of candidates for oral cancer screening

OBJECTIVES

Artificial intelligence could enhance the use of disparate risk factors (crude method) for better stratification of patients to be screened for oral cancer. This study aims to construct a meta-classifier that considers diverse risk factors to identify patients at risk of oral cancer and other suspicious oral diseases for targeted screening.

MATERIALS AND METHODS

A retrospective dataset from a community oral cancer screening program was used to construct and train the novel voting meta-classifier. Comprehensive risk factor information from this dataset was used as input features for eleven supervised learning algorithms which served as base learners and provided predicted probabilities that are weighted and aggregated by the meta-classifier. Training dataset was augmented using SMOTE-ENN. Additionally, Shapley additive explanations (SHAP) values were generated to implement the explainability of the model and display the important risk factors.

RESULTS

Our meta-classifier had an internal validation recall, specificity, and AUROC of 0.83, 0.86, and 0.85 for identifying the risk of oral cancer and 0.92, 0.60, and 0.76 for identifying suspicious oral mucosal disease respectively. Upon external validation, the meta-classifier had a significantly higher AUROC than the crude/current method used for identifying the risk of oral cancer (0.78 vs 0.46; p = 0.001) Also, the meta-classifier had better recall than the crude method for predicting the risk of suspicious oral mucosal diseases (0.78 vs 0.47).

CONCLUSION

Overall, these findings showcase that our approach optimizes the use of risk factors in identifying patients for oral screening which suggests potential clinical application.

Development and validation of a nomogram to predict poor short-term response to recombinant human growth hormone treatment in children with growth disorders

PURPOSE

The purpose of this study was to develop and validate a clinical predictive model for predicting the likelihood of a poor therapeutic response during the first year of recombinant human growth hormone (rhGH) treatment in children with growth disorders.

METHODS

A total of 627 pediatric patients with growth disorders (GHD, ISS, TS, SGA) from The LG Growth Study cohort were evaluated. Restricted cubic splines (RCS) were utilized to investigate the association between predictors and the risk of poor rhGH response. Variables were selected using LASSO regression, and multivariate logistics regression models were established. Receiver operating characteristic (ROC) curves, calibration curves, decision curve analysis (DCA), and clinical impact curves (CIC) were used to assess the predictive model's accuracy and clinical value. The predictive accuracy of the model was validated on the testing set.

RESULTS

Two predictive models containing 8 baseline predictors (diagnosis, age, height SDS, bone age minus chronological age, rhGH dosage, distance from mid-parental height in SDS, weight SDS, IGF-1 SDS) and 1 post-treatment predictor (height SDS gain at 6 months) were constructed by multivariate logistic regression analyses. The nomogram was built based on the multivariate predictive model and showed good discrimination and model fit effects in both the training set and the testing set. DCA and CIC analyses presented good clinical usability.

CONCLUSION

The clinical predictive model for predicting the probability of poor short-term response of rhGH treatment in pediatric patients with growth disorders is useful and can assist physicians in making clinical decisions.

Nomogram Based on Clinical and Radiomics Data for Predicting Radiation-induced Temporal Lobe Injury in Patients with Non-metastatic Stage T4 Nasopharyngeal Carcinoma

AIMS

To use pre-treatment magnetic resonance imaging-based radiomics data with clinical data to predict radiation-induced temporal lobe injury (RTLI) in nasopharyngeal carcinoma (NPC) patients with stage T4/N0-3/M0 within 5 years after radiotherapy.

MATERIALS AND METHODS

This study retrospectively examined 98 patients (198 temporal lobes) with stage T4/N0-3/M0 NPC. Participants were enrolled into a training cohort or a validation cohort in a ratio of 7:3. Radiomics features were extracted from pre-treatment magnetic resonance imaging that were T1-and T2-weighted. Spearman rank correlation, the t-test and the least absolute shrinkage and selection operator (LASSO) algorithm were used to select significant radiomics features; machine-learning models were used to generate radiomics signatures (Rad-Scores). Rad-Scores and clinical factors were integrated into a nomogram for prediction of RTLI. Nomogram discrimination was evaluated using receiver operating characteristic analysis and clinical benefits were evaluated using decision curve analysis.

RESULTS

Participants were enrolled into a training cohort (n = 139) or a validation cohort (n = 59). In total, 3568 radiomics features were initially extracted from T1-and T2-weighted images. Age, D_max, D_1cc and 16 stable radiomics features (six from T1-weighted and 10 from T2-weighted images) were identified as independent predictive factors. A greater Rad-Score was associated with a greater risk of RTLI. The nomogram showed good discrimination, with a C-index of 0.85 (95% confidence interval 0.79-0.92) in the training cohort and 0.82 (95% confidence interval 0.71-0.92) in the validation cohort.

CONCLUSION

We developed models for the prediction of RTLI in patients with stage T4/N0-3/M0 NPC using pre-treatment radiomics data and clinical data. Nomograms from these pre-treatment data improved the prediction of RTLI. These results may allow the selection of patients for earlier clinical interventions.

Artificial intelligence 101 for veterinary diagnostic imaging

The prevalence and pervasiveness of artificial intelligence (AI) with medical images in veterinary and human medicine is rapidly increasing. This article provides essential definitions of AI with medical images with a focus on veterinary radiology. Machine learning methods common in medical image analysis are compared, and a detailed description of convolutional neural networks commonly used in deep learning classification and regression models is provided. A brief introduction to natural language processing (NLP) and its utility in machine learning is also provided. NLP can economize the creation of "truth-data" needed when training AI systems for both diagnostic radiology and radiation oncology applications. The goal of this publication is to provide veterinarians, veterinary radiologists, and radiation oncologists the necessary background needed to understand and comprehend AI-focused research projects and publications.

The role of artificial intelligence in veterinary radiation oncology

Veterinary radiation oncology regularly deploys sophisticated contouring, image registration, and treatment planning optimization software for patient care. Over the past decade, advances in computing power and the rapid development of neural networks, open-source software packages, and data science have been realized and resulted in new research and clinical applications of artificial intelligent (AI) systems in radiation oncology. These technologies differ from conventional software in their level of complexity and ability to learn from representative and local data. We provide clinical and research application examples of AI in human radiation oncology and their potential applications in veterinary medicine throughout the patient's care-path: from treatment simulation, deformable registration, auto-segmentation, automated treatment planning and plan selection, quality assurance, adaptive radiotherapy, and outcomes modeling. These technologies have the potential to offer significant time and cost savings in the veterinary setting; however, since the range of usefulness of these technologies have not been well studied nor understood, care must be taken if adopting AI technologies in clinical practice. Over the next several years, some practical and realizable applications of AI in veterinary radiation oncology include automated segmentation of normal tissues and tumor volumes, deformable registration, multi-criteria plan optimization, and adaptive radiotherapy. Keys in achieving success in adopting AI in veterinary radiation oncology include: establishing "truth-data"; data harmonization; multi-institutional data and collaborations; standardized dose reporting and taxonomy; adopting an open access philosophy, data collection and curation; open-source algorithm development; and transparent and platform-independent code development.

Construction of machine learning-based models for cancer outcomes in low and lower-middle income countries: A scoping review

Background

The impact and utility of machine learning (ML)-based prediction tools for cancer outcomes including assistive diagnosis, risk stratification, and adjunctive decision-making have been largely described and realized in the high income and upper-middle-income countries. However, statistical projections have estimated higher cancer incidence and mortality risks in low and lower-middle-income countries (LLMICs). Therefore, this review aimed to evaluate the utilization, model construction methods, and degree of implementation of ML-based models for cancer outcomes in LLMICs.

Methods

PubMed/Medline, Scopus, and Web of Science databases were searched and articles describing the use of ML-based models for cancer among local populations in LLMICs between 2002 and 2022 were included. A total of 140 articles from 22,516 citations that met the eligibility criteria were included in this study.

Results

ML-based models from LLMICs were often based on traditional ML algorithms than deep or deep hybrid learning. We found that the construction of ML-based models was skewed to particular LLMICs such as India, Iran, Pakistan, and Egypt with a paucity of applications in sub-Saharan Africa. Moreover, models for breast, head and neck, and brain cancer outcomes were frequently explored. Many models were deemed suboptimal according to the Prediction model Risk of Bias Assessment tool (PROBAST) due to sample size constraints and technical flaws in ML modeling even though their performance accuracy ranged from 0.65 to 1.00. While the development and internal validation were described for all models included (n=137), only 4.4% (6/137) have been validated in independent cohorts and 0.7% (1/137) have been assessed for clinical impact and efficacy.

Conclusion

Overall, the application of ML for modeling cancer outcomes in LLMICs is increasing. However, model development is largely unsatisfactory. We recommend model retraining using larger sample sizes, intensified external validation practices, and increased impact assessment studies using randomized controlled trial designs

Systematic review registration

https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=308345, identifier CRD42022308345.

A microRNA signature for clinical outcomes of pediatric ALL patients treated with TPOG protocols

MicroRNA (miRNA) expression is reportedly associated with clinical outcomes in childhood acute lymphoblastic leukemia (ALL). Here, we aimed at investigating whether miRNA expression is associated with clinical outcomes in pediatric ALL patients treated with the Taiwan Pediatric Oncology Group (TPOG) protocols. The expression of 397 miRNAs was measured using stem-loop quantitative real-time polymerase chain reaction miRNA arrays in 60 pediatric ALL patients treated with TPOG-ALL-93 or TPOG-ALL-97 VHR (very high-risk) protocols. In order to identify prognosis-related miRNAs, original cohort was randomly split into the training and testing cohort in a 2:1 ratio, and univariate Cox proportional hazards regression was applied to identify associations between event-free survival (EFS) and expressions of miRNAs. Four prognosis-related miRNAs were selected and validated in another independent cohort composed of 103 patients treated with the TPOG-ALL-2002 protocol. Risk score, including the impact of four prognosis-related miRNAs, was calculated for each patients, followed by grouping patients into the high or low risk-score groups. Irrespective of the training, testing, or validation cohort, risk-score group was significantly associated with EFS and overall survival (OS). Risk-score group combining with clinical characteristics including the age onset (≥10 years), white blood cell counts (≥100 × 10⁹/L), cell type (T- or B-cell), sex, and risk groups of the treatment protocols were used as predictors of EFS using the multivariate Cox proportional hazards regression. Results showed that the risk-score group was the strongest predictor. In the validation cohort, hazard ratios (HRs) of the risk-score group were 7.06 (95% CI=1.93-25.84, p-value =0.003) and 14.03 (95% CI=3.34-59.04, p-value =0.003) for EFS and OS, respectively. High risk-score group had higher risk of having poor prognosis and risk of death than that in the low risk group. Accuracy of the prediction model for 5-year EFS could reach 0.76. For the prediction of 5-year OS, accuracy was 0.75. In conclusion, a miRNA signature was associated with clinical outcomes in childhood ALL patients treated with TPOG protocols and might be a suitable prognostic biomarker.

Deep learning can predict survival directly from histology in clear cell renal cell carcinoma

For clear cell renal cell carcinoma (ccRCC) risk-dependent diagnostic and therapeutic algorithms are routinely implemented in clinical practice. Artificial intelligence-based image analysis has the potential to improve outcome prediction and thereby risk stratification. Thus, we investigated whether a convolutional neural network (CNN) can extract relevant image features from a representative hematoxylin and eosin-stained slide to predict 5-year overall survival (5y-OS) in ccRCC. The CNN was trained to predict 5y-OS in a binary manner using slides from TCGA and validated using an independent in-house cohort. Multivariable logistic regression was used to combine of the CNNs prediction and clinicopathological parameters. A mean balanced accuracy of 72.0% (standard deviation [SD] = 7.9%), sensitivity of 72.4% (SD = 10.6%), specificity of 71.7% (SD = 11.9%) and area under receiver operating characteristics curve (AUROC) of 0.75 (SD = 0.07) was achieved on the TCGA training set (n = 254 patients / WSIs) using 10-fold cross-validation. On the external validation cohort (n = 99 patients / WSIs), mean accuracy, sensitivity, specificity and AUROC were 65.5% (95%-confidence interval [CI]: 62.9–68.1%), 86.2% (95%-CI: 81.8–90.5%), 44.9% (95%-CI: 40.2–49.6%), and 0.70 (95%-CI: 0.69–0.71). A multivariable model including age, tumor stage and metastasis yielded an AUROC of 0.75 on the TCGA cohort. The inclusion of the CNN-based classification (Odds ratio = 4.86, 95%-CI: 2.70–8.75, p < 0.01) raised the AUROC to 0.81. On the validation cohort, both models showed an AUROC of 0.88. In univariable Cox regression, the CNN showed a hazard ratio of 3.69 (95%-CI: 2.60–5.23, p < 0.01) on TCGA and 2.13 (95%-CI: 0.92–4.94, p = 0.08) on external validation. The results demonstrate that the CNN’s image-based prediction of survival is promising and thus this widely applicable technique should be further investigated with the aim of improving existing risk stratification in ccRCC.

Open-source distributed learning validation for a larynx cancer survival model following radiotherapy

INTRODUCTION

Prediction models are useful to design personalised treatment. However, safe and effective implementation relies on external validation. Retrospective data are available in many institutions, but sharing between institutions can be challenging due to patient data sensitivity and governance or legal barriers. This study validates a larynx cancer survival model performed using distributed learning without any sensitive data leaving the institution.

METHODS

Open-source distributed learning software based on a stratified Cox proportional hazard model was developed and used to validate the Egelmeer et al. MAASTRO survival model across two hospitals in two countries. The validation optimised a single scaling parameter multiplied by the original predicted prognostic index. All analyses and figures were based on the distributed system, ensuring no information leakage from the individual centres. All applied software is provided as freeware to facilitate distributed learning in other institutions.

RESULTS

1745 patients received radiotherapy for larynx cancer in the two centres from Jan 2005 to Dec 2018. Limiting to a maximum of one missing value in the parameters of the survival model reduced the cohort to 1095 patients. The Harrell C-index was 0.74 (CI95%, 0.71-0.76) and 0.70 (0.66-0.75) for the two centres. However, the model needed a scaling update. In addition, it was found that survival predictions of patients undergoing hypofractionation were less precise.

CONCLUSION

Open-source distributed learning software was able to validate, and suggest a minor update to the original survival model without central access to patient sensitive information. Even without the update, the original MAASTRO survival model of Egelmeer et al. performed reasonably well, providing similar results in this validation as in its original validation.

Bias and Reporting Quality of Clinical Prognostic Models for Idiopathic Pulmonary Fibrosis: A Cross-Sectional Study

Objective

This study aims to evaluate the risk of bias (ROB) and reporting quality of idiopathic pulmonary fibrosis (IPF) prediction models by assessing characteristics of these models.

Methods

The development and/or validation of IPF prognostic models were identified via an electronic search of PubMed, Embase, and Web of Science (from inception to 12 August, 2021). Two researchers independently assessed the risk of bias (ROB) and reporting quality of IPF prediction models based on the Prediction model Risk Of Bias Assessment Tool (PROBAST) and Transparent Reporting of a multivariable prognostic model for Individual Prognosis or Diagnosis (TRIPOD) checklist.

Results

Twenty prognostic model studies for IPF were included, including 7 (35%) model development and external validation studies, 8 (40%) development studies, and 5 (25%) external validation studies. According to PROBAST, all studies were appraised with high ROB, because of deficient reporting in the domains of participants (45.0%) and analysis (67.3%), and at least 55% studies were susceptible to 4 of 20 sources of bias. For the reporting quality, none of them completely adhered to the TRIPOD checklist, with the lowest mean reporting score for the methods and results domains (46.6% and 44.7%). For specific items, eight sub-items had a reporting rate ≥80% and adhered to the TRIPOD checklist, and nine sub-items had a very poor reporting rate, less than 30%.

Conclusion

Studies adhering to PROBAST and TRIPOD checklists are recommended in the future. The reproducibility and transparency can be improved when studies completely adhere to PROBAST and TRIPOD checklists.

Performance of a simplified scoring system for risk stratification in oral cancer and oral potentially malignant disorders screening

BACKGROUND

Impact and efficiency of oral cancer and oral potentially malignant disorders screening are most realized in "at-risk" individuals. However, tools that can provide essential knowledge on individuals' risks are not applied in risk-based screening. This study aims to optimize a simplified risk scoring system for risk stratification in organized oral cancer and oral potentially malignant disorders screening.

METHODS

Participants were invited to attend a community-based oral cancer and oral potentially malignant disorders screening program in Hong Kong. Visual oral examination was performed for all attendees and information on sociodemographic characteristics as well as habitual, lifestyle, familial, and comorbidity risk factors were obtained. Individuals' status of those found to have suspicious lesions following biopsy and histopathology were classified as positive/negative and this outcome was used in a multiple logistic regression analysis with variables collected during screening. Odds ratio weightings were then used to develop a simplified risk scoring system which was validated in an external cohort.

RESULTS

Of 979 participants, 4.5% had positive status following confirmatory diagnosis. A 12-variable simplified risk scoring system with weightings was generated with an AUC, sensitivity, and specificity of 0.82, 0.71, and 0.78 for delineating high-risk cases. Further optimization on the validation cohort of 491 participants yielded a sensitivity and specificity of 0.75 and 0.87 respectively.

CONCLUSIONS

The simplified risk scoring system was able to stratify oral cancer and oral potentially malignant disorders risk with satisfactory sensitivity and specificity and can be applied in risk-based disease screening.

Simulation CT-based radiomics for prediction of response after neoadjuvant chemo-radiotherapy in patients with locally advanced rectal cancer

Background

To report on the discriminative ability of a simulation Computed Tomography (CT)-based radiomics signature for predicting response to treatment in patients undergoing neoadjuvant chemo-radiation for locally advanced adenocarcinoma of the rectum.

Methods

Consecutive patients treated at the Universities of Tübingen (from 1/1/07 to 31/12/10, explorative cohort) and Florence (from 1/1/11 to 31/12/17, external validation cohort) were considered in our dual-institution, retrospective analysis. Long-course neoadjuvant chemo-radiation was performed according to local policy. On simulation CT, the rectal Gross Tumor Volume was manually segmented. A feature selection process was performed yielding mineable data through an in-house developed software (written in Python 3.6). Model selection and hyper-parametrization of the model was performed using a fivefold cross validation approach. The main outcome measure of the study was the rate of pathologic good response, defined as the sum of Tumor regression grade (TRG) 3 and 4 according to Dworak’s classification.

Results

Two-hundred and one patients were included in our analysis, of whom 126 (62.7%) and 75 (37.3%) cases represented the explorative and external validation cohorts, respectively. Patient characteristics were well balanced between the two groups. A similar rate of good response to neoadjuvant treatment was obtained in in both cohorts (46% and 54.7%, respectively; p = 0.247). A total of 1150 features were extracted from the planning scans. A 5-metafeature complex consisting of Principal component analysis (PCA)-clusters (whose main components are LHL Grey-Level-Size-Zone: Large Zone Emphasis, Elongation, HHH Intensity Histogram Mean, HLL Run-Length: Run Level Variance and HHH Co-occurence: Cluster Tendency) in combination with 5-nearest neighbour model was the most robust signature. When applied to the explorative cohort, the prediction of good response corresponded to an average Area under the curve (AUC) value of 0.65 ± 0.02. When the model was tested on the external validation cohort, it ensured a similar accuracy, with a slightly lower predictive ability (AUC of 0.63).

Conclusions

Radiomics-based, data-mining from simulation CT scans was shown to be feasible and reproducible in two independent cohorts, yielding fair accuracy in the prediction of response to neoadjuvant chemo-radiation.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13014-022-02053-y.

Pulse perfusion index for predicting intrapartum fever during epidural analgesia

STUDY OBJECTIVE

To assess whether pulse perfusion index (PI) values could be employed to predict intrapartum fever and to provide a cut-off PI value for predicting intrapartum fever occurrence.

DESIGN

We conducted a single-center, prospective, observational study.

SETTING

Delivery room at the Department of Obstetrics, Affiliated Hospital of Jiangsu University.

PATIENTS

117 parturients who intended to have a vaginal delivery.

INTERVENTIONS

Each parturient received epidural analgesia.

MEASUREMENTS

We checked each parturient's tympanic temperature before analgesia (T₀), at 1 h (T₁) and 2 h (T₂) after analgesia, immediately at the end of the second (T₃) and third (T₄) stages of labor, and at 1 h postpartum (T₅). A temperature of ≥38°C was defined as fever. PI, measured on the right second toe, was recorded before analgesia (PI₀) and at 10 min (PI₁₀), 20 min (PI₂₀), and 30 min (PI₃₀) after analgesia. The PI change rate was calculated as the incremental change in PI₃₀ from PI₀, divided by the PI₀. Receiver operating characteristic (ROC) curves were used to verify the utility of the PI₃₀ and PI change rate values for predicting intrapartum fever.

MAIN RESULTS

We found that peak temperature (T_P) occurred at the end of the second or the third stage of labor. Within 30 min after analgesia, the PI showed a significant increase over time and there was a linear correlation between PI₃₀ and T_P values (P < 0.001, r = 0.544). The PI₁₀, PI₂₀, PI₃₀ and PI change rate in febrile parturients were higher than those in afebrile parturients (P < 0.001). The area under the ROC (AUROC) for PI₃₀ was 0.818 (P < 0.001) with a cut-off of 9.30. The AUROC of the PI change rate was 0.738 (P < 0.001) with a cut-off of 3.45.

CONCLUSIONS

PI₃₀ and PI change rate values could be used to predict intrapartum fever in parturients after epidural analgesia.

Marginal versus conditional odds ratios when updating risk prediction models

Risk prediction models often need to be updated when applied to new settings. A simple updating method involves fixed odds ratio transformation of predicted risks to adjust the model for outcome prevalence in the new setting. When a sample from the target population is available, the gold standard is to use a logistic regression model to estimate this odds ratio. A simpler method has been proposed that calculates this odds ratio from the prevalence estimates in the original and new samples. We show that the marginal odds ratio estimated in this way is generally closer to one than the correct (conditional) odds ratio; thus, the simpler method should be avoided when individual-level data are available. When prevalence estimates are the only information at hand, we suggest an approximate method for recovering the conditional odds ratio from the variance of predicted risks in the development sample. Brief simulations and examples show that this approach reduces undercorrection, often substantially.

Prognostic significance of tumor budding thresholds in oral tongue squamous cell carcinoma

Oral tongue squamous cell carcinoma (OTSCC) represents the most common malignancy of the oral cavity. Tumor budding (TB) is a reliable prognostic factor in OTSCC; however, a standardized scoring system is not still validated. The study aims to evaluate the prognostic role of TB in 211 OTSCC patients treated between 1997-2018. TB was evaluated on haematoxylin and eosin-stained sections in the hotspot area of the infiltrative front (×200-magnification). It was scored using a two-tier, a three-tier system, and according to BD-model and revised-Grading system. Univariate and multivariate Cox regression analyses of disease-specific survival (DSS) and disease-free survival (DFS) were performed. A p-values<0.05 was considered as statistically significant. The two-tier and three-tier system resulted an independent prognostic factor of DSS. High-risk patients had a 2.21 and 3.08 times-increased probability of poor DSS compared to low-risk group. It is significantly increased even for intermediate-risk group. No significant differences emerged classifying patients according to BD-model and revised-Grading. These data confirm the prognostic value of TB in predicting DSS in OTSCC. Classifying patients in two groups using the 5-buds cut-off significantly discriminates their outcomes. Since the established role of DOI and the poor prognostic value of grading, TB could be considered an independent prognostic marker.

Artificial intelligence in liver diseases: improving diagnostics, prognostics and response prediction

Clinical routine in hepatology involves the diagnosis and treatment of a wide spectrum of metabolic, infectious, autoimmune and neoplastic diseases. Clinicians integrate qualitative and quantitative information from multiple data sources to make a diagnosis, prognosticate the disease course, and recommend a treatment. In the last 5 years, advances in artificial intelligence (AI), particularly in deep learning, have made it possible to extract clinically relevant information from complex and diverse clinical datasets. In particular, histopathology and radiology image data contain diagnostic, prognostic and predictive information which AI can extract. Ultimately, such AI systems could be implemented in clinical routine as decision support tools. However, in the context of hepatology, this requires further large-scale clinical validation and regulatory approval. Herein, we summarise the state of the art in AI in hepatology with a particular focus on histopathology and radiology data. We present a roadmap for the further development of novel biomarkers in hepatology and outline critical obstacles which need to be overcome.

Computed tomography-based radiomic model predicts radiological response following stereotactic body radiation therapy in early-stage non-small-cell lung cancer and pulmonary oligo-metastases

Purpose

Radiomic models elaborate geometric and texture features of tumors extracted from imaging to develop predictors for clinical outcomes. Stereotactic body radiation therapy (SBRT) has been increasingly applied in the ablative treatment of thoracic tumors. This study aims to identify predictors of treatment responses in patients affected by early stage non-small cell lung cancer (NSCLC) or pulmonary oligo-metastases treated with SBRT and to develop an accurate machine learning model to predict radiological response to SBRT.

Materials and Methods

Computed tomography (CT) images of 85 tumors (stage I–II NSCLC and pulmonary oligo-metastases) from 69 patients treated with SBRT were analyzed. Gross tumor volumes (GTV) were contoured on CT images. Patients that achieved complete response (CR) or partial response (PR) were defined as responders. One hundred ten radiomic features were extracted using PyRadiomics module based on the GTV. The association of features with response to SBRT was evaluated. A model using support vector machine (SVM) was then trained to predict response based solely on the extracted radiomics features. Receiver operating characteristic curves were constructed to evaluate model performance of the identified radiomic predictors.

Results

Sixty-nine patients receiving thoracic SBRT from 2008 to 2018 were retrospectively enrolled. Skewness and root mean squared were identified as radiomic predictors of response to SBRT. The SVM machine learning model developed had an accuracy of 74.8%. The area under curves for CR, PR, and non-responder prediction were 0.86 (95% confidence interval [CI], 0.794–0.921), 0.946 (95% CI, 0.873–0.978), and 0.857 (95% CI, 0.789–0.915), respectively.

Conclusion

Radiomic analysis of pre-treatment CT scan is a promising tool that can predict tumor response to SBRT.

A prognostic model in patients with advanced biliary tract cancer receiving first-line chemotherapy

BACKGROUND

Standard treatment of advanced biliary tract cancer (aBTC) is represented by first-line chemotherapy (CT1). However, some patients do not gain any benefit from CT1, contributing to the overall dismal prognosis of aBTC. The present study aimed to devise a prognostic model in aBTC patients receiving CT1.

METHODS

A large panel of clinical, laboratory, and pathology variables, available before the start of CT1, were retrospectively assessed in a multi-centric cohort to determine their prognostic value on univariate and multivariate regression analysis. The variables that showed a significant correlation with overall survival (OS) were computed in a three-tier prognostic score. External validation of the prognostication performance was carried out.

RESULTS

Clinical histories of 935 patients (median OS 10.3 months), with diagnosis dates ranging from 2001 to 2017, were retrieved from 14 institutions. According to multivariate analysis, Eastern Cooperative Oncology Group performance status, carbohydrate antigen 19.9, albumin levels, and neutrophil/lymphocyte ratio were strongly associated with OS (p <0.01). The prognostic score could generate a highly significant stratification (all between-group p values ≤0.001) into groups of favorable (comprising 51.5% of the sample), intermediate (39.2%), and poor prognosis (9.3%): median OS was 12.7 (CI_95% 11.0-14.4), 7.1 (CI_95% 5.8-8.4), and 3.2 months (CI_95% 1.7-4.7), respectively. This OS gradient was replicated in the validation set (129 patients), with median OS of 12.7 (CI_95% 11.0-14.3), 7.5 (CI_95% 6.1-8.9), and 1.4 months (CI_95% 0.1-2.7), respectively (all between-group p values ≤0.05).

CONCLUSION

A prognostic score, derived from a limited set of easily-retrievable variables, efficiently stratified a large population of unselected aBTC patients undergoing CT1. This tool could be useful to clinicians, to ascertain the potential benefit from CT1 at the start of treatment.

Radiomics in PET Imaging:: A Practical Guide for Newcomers

Radiomics has undergone considerable development in recent years. In PET imaging, very promising results concerning the ability of handcrafted features to predict the biological characteristics of lesions and to assess patient prognosis or response to treatment have been reported in the literature. This article presents a checklist for designing a reliable radiomic study, gives an overview of the steps of the pipeline, and outlines approaches for data harmonization. Tips are provided for critical reading of the content of articles. The advantages and limitations of handcrafted radiomics compared with deep-learning approaches for the characterization of PET images are also discussed.

Phase I prognostic online (PIPO): A web tool to improve patient selection for oncology early phase clinical trials

PURPOSE

Patient selection in phase 1 clinical trials (Ph1t) continues to be a challenge. The aim of this study was to develop a user-friendly prognostic calculator for predicting overall survival (OS) outcomes in patients to be included in Ph1t with immune checkpoint inhibitors (ICIs) or targeted agents (TAs) based on clinical parameters assessed at baseline.

METHODS

Using a training cohort with consecutive patients from the VHIO phase 1 unit, we constructed a prognostic model to predict median OS (mOS) as a primary endpoint and 3-month (3m) OS rate as a secondary endpoint. The model was validated in an internal cohort after temporal data splitting and represented as a web application.

RESULTS

We recruited 799 patients (training and validation sets, 558 and 241, respectively). Median follow-up was 21.2 months (m), mOS was 10.2 m (95% CI, 9.3-12.7) for ICIs cohort and 7.7 m (95% CI, 6.6-8.6) for TAs cohort. In the multivariable analysis, six prognostic variables were independently associated with OS - ECOG, number of metastatic sites, presence of liver metastases, derived neutrophils/(leukocytes minus neutrophils) ratio [dNLR], albumin and lactate dehydrogenase (LDH) levels. The phase 1 prognostic online (PIPO) calculator showed adequate discrimination and calibration performance for OS, with C-statistics of 0.71 (95% CI 0.64-0.78) in the validation set. The overall accuracy of the model for 3m OS prediction was 87.2% (95% CI 85%-90%).

CONCLUSIONS

PIPO is a user-friendly objective and interactive tool to calculate specific survival probabilities for each patient before enrolment in a Ph1t. The tool is available at https://pipo.vhio.net/.

Prognostic assessment of different methods for eosinophils detection in oral tongue cancer

BACKGROUND

TATE has been proposed as a prognostic factor in oral cancer staging; however, the controversial literature data limit its application in the routine diagnosis. The aim of this study was to evaluate the prognostic value of TATE in patients with oral tongue cancer. The second aim was to identify any difference in the methods of eosinophil quantification or in the cutoff values reported in literature.

METHODS

Clinic-pathological data of 204 patients treated at "Ospedali Riuniti" Hospital, Ancona, Italy, were collected. Evaluation of TATE was performed on hematoxylin-and-eosin-stained slides and correlation with survival outcomes was evaluated. The number of eosinophils per square millimeter was evaluated by using two methods, namely density (TATE-1) and classical (TATE-2) methods. For each of the 2 methods tested, patients were stratified into two or three groups, according to the most used cutoff values reported in literature.

RESULTS

Regardless of the method of eosinophil quantification or the cutoff values used, patients with high TATE had a significantly better disease-specific survival. The density method (TATE-1) showed a better predictive performance, in particular when applying a single cutoff of 67 eosinophils/mm² , two cutoffs of 10 and 100 eosinophils/mm² , or two cutoffs of 50 and 120 eosinophils/mm² .

CONCLUSION

The evaluation of TATE is simple, cost-effective, and easy to implement in daily practice with the aim of improving risk stratification of patients affected by oral tongue cancer. Results of prognostic performance analysis suggest using density (TATE-1) method as the standard approach to evaluate TATE in future studies, enhancing replicability.

Cardiac magnetic resonance radiomics: basic principles and clinical perspectives

Radiomics is a novel image analysis technique, whereby voxel-level information is extracted from digital images and used to derive multiple numerical quantifiers of shape and tissue character. Cardiac magnetic resonance (CMR) is the reference imaging modality for assessment of cardiac structure and function. Conventional analysis of CMR scans is mostly reliant on qualitative image analysis and basic geometric quantifiers. Small proof-of-concept studies have demonstrated the feasibility and superior diagnostic accuracy of CMR radiomics analysis over conventional reporting. CMR radiomics has the potential to transform our approach to defining image phenotypes and, through this, improve diagnostic accuracy, treatment selection, and prognostication. The purpose of this article is to provide an overview of radiomics concepts for clinicians, with particular consideration of application to CMR. We will also review existing literature on CMR radiomics, discuss challenges, and consider directions for future work.

A novel HBsAg-based model for predicting significant liver fibrosis among Chinese patients with immune-tolerant phase chronic hepatitis B: a multicenter retrospective study

BACKGROUND

Hepatitis B e antigen (HBeAg)-positive chronic hepatitis B (CHB) in the immune-tolerant (IT) phase is significantly associated with high risk for hepatocellular carcinoma, suggesting requirement for antiviral therapy, particularly for those with histological liver injury. This study aimed to establish a non-invasive panel to assess significant liver fibrosis in IT chronic hepatitis B.

PATIENTS AND METHODS

One hundred and thirteen IT-phase CHB patients were retrospectively recruited and divided into two histopathological groups according to their histological profiles: necroinflammatory score <4 (N <4)/fibrosis score ⩽1 (F0-1), and necroinflammatory score ⩾4 (N ⩾4)/fibrosis score ⩾2 (F2-4). Multivariate analysis was conducted to assess the predictive value of the non-invasive model for significant liver fibrosis.

RESULTS

IT-phase CHB patients with N <4/F0-1 had significantly higher HBsAg levels than those with N ⩾4/F2-4. The optimal HBsAg level of log 4.44 IU/mL for significant liver fibrosis (F ⩾2) gave an area under the curve (AUC) of 0.83, sensitivity of 81.1%, specificity of 81.6%, positive predictive value (PPV) of 68.2%, and negative predictive value (NPV) of 89.9%. An IT model with HBsAg and gamma glutamyl transpeptidase (GGT) in combination was established, and it had an AUC of 0.86, sensitivity of 86.5%, specificity of 81.6%, PPV of 69.6, NPV of 92.5, and accuracy of 83.2% to predict F ⩾2 in the IT-phase CHB patients. Notably, the IT model exhibited higher predictive value than the existing aspartate aminotransferase-to-platelet ratio index, Fibrosis-4 score, and GGT to platelet ratio.

CONCLUSION

The established IT model combining HBsAg and GGT has good performance in predicting significant liver fibrosis in IT-phase CHB patients.

Deciphering the glioblastoma phenotype by computed tomography radiomics

INTRODUCTION

Glioblastoma (GBM) is the most common malignant primary brain tumour which has, despite extensive treatment, a median overall survival of 15 months. Radiomics is the high-throughput extraction of large amounts of image features from radiographic images, which allows capturing the tumour phenotype in 3D and in a non-invasive way. In this study we assess the prognostic value of CT radiomics for overall survival in patients with a GBM.

MATERIALS AND METHODS

Clinical data and pre-treatment CT images were obtained from 218 patients diagnosed with a GBM via biopsy who underwent radiotherapy +/- temozolomide between 2004 and 2015 treated at three independent institutes (n = 93, 62 and 63). A clinical prognostic score (CPS), a simple radiomics model consisting of volume based score (VPS), a complex radiomics prognostic score (RPS) and a combined clinical and radiomics (C + R)PS model were developed. The population was divided into three risk groups for each prognostic score and respective Kaplan-Meier curves were generated.

RESULTS

Patient characteristics were broadly comparable. Clinically significant differences were observed with regards to radiation dose, tumour volume and performance status between datasets. Image acquisition parameters differed between institutes. The cross-validated c-indices were moderately discriminative and for the CPS ranged from 0.63 to 0.65; the VPS c-indices ranged between 0.52 and 0.61; the RPS c-indices ranged from 0.57 to 0.64 and the combined clinical and radiomics model resulted in c-indices of 0.59-0.71.

CONCLUSION

In this study clinical and CT radiomics features were used to predict OS in GBM. Discrimination between low-, middle- and high-risk patients based on the combined clinical and radiomics model was comparable to previous MRI-based models.

Robust prediction of mortality of COVID-19 patients based on quantitative, operator-independent, lung CT densitometry

PURPOSE

To train and validate a predictive model of mortality for hospitalized COVID-19 patients based on lung densitometry.

METHODS

Two-hundred-fifty-one patients with respiratory symptoms underwent CT few days after hospitalization. "Aerated" (AV), "consolidated" (CV) and "intermediate" (IV) lung sub-volumes were quantified by an operator-independent method based on individual HU maximum gradient recognition. AV, CV, IV, CV/AV, IV/AV, and HU of the first peak position were extracted. Relevant clinical parameters were prospectively collected. The population was composed by training (n = 166) and validation (n = 85) consecutive cohorts, and backward multi-variate logistic regression was applied on the training group to build a CT_model. Similarly, models including only clinical parameters (CLIN_model) and both CT/clinical parameters (COMB_model) were developed. Model's performances were assessed by goodness-of-fit (H&L-test), calibration and discrimination. Model's performances were tested in the validation group.

RESULTS

Forty-three patients died (25/18 in training/validation). CT_model included AVmax (i.e. maximum AV between lungs), CV and CV/AE, while CLIN_model included random glycemia, C-reactive protein and biological drugs (protective). Goodness-of-fit and discrimination were similar (H&L:0.70 vs 0.80; AUC:0.80 vs 0.80). COMB_model including AVmax, CV, CV/AE, random glycemia, biological drugs and active cancer, outperformed both models (H&L:0.91; AUC:0.89, 95%CI:0.82-0.93). All models showed good calibration (R²:0.77-0.97). Despite several patient's characteristics were different between training and validation cohorts, performances in the validation cohort confirmed good calibration (R²:0-70-0.81) and discrimination for CT_model/COMB_model (AUC:0.72/0.76), while CLIN_model performed worse (AUC:0.64).

CONCLUSIONS

Few automatically extracted densitometry parameters with clear functional meaning predicted mortality of COVID-19 patients. Combined with clinical features, the resulting predictive model showed higher discrimination/calibration.

Clinical Implications of Combined Lymphocyte and Neutrophil Count in Locally Advanced Rectal Cancer After Preoperative Chemoradiotherapy

BACKGROUND

There are controversies about the ability of neutrophil to lymphocyte ratio to predict the recurrence and survival in patients with locally advanced rectal cancer (LARC) treated with neoadjuvant chemoradiation. The objective of this study is to investigate the prognostic potential of combined lymphocyte count (LC) and neutrophil count (NC) in LARC patients treated with chemoradiotherapy (CRT) followed by curative surgery.

METHODS

Patients with LARC who underwent surgical resection between January 2010 and December 2017 were reviewed retrospectively. We divided the patients into three groups: high LC and low NC, low LC and high NC, and the remaining patients. The cut-off values of LC and NC were determined by receiver operating characteristic curve analysis and log-rank test statistics. We compared the disease-free survival (DFS) rate between the groups.

RESULTS

A total of 176 consecutive patients were included in this study. The 5 year DFS rate was significantly different among the three groups in pathologic node (pN)+ patients (73.2% vs. 61.9% vs. 14.2%; P = 0.025). Cox multivariate analysis for pN+ patients demonstrated that combination of low LC and high NC (hazard ratio, 3.630; 95% confidence interval [CI], 1.306-10.093; P = 0.013) was significantly correlated with decreased DFS.

CONCLUSIONS

This study showed that the combination of LC and NC is a powerful predictive factor for disease recurrence in pN+ LARC patients who underwent CRT.

Radiomic Analysis to Predict Outcome in Recurrent Glioblastoma Based on Multi-Center MR Imaging From the Prospective DIRECTOR Trial

Background

Based on promising results from radiomic approaches to predict O⁶-methylguanine DNA methyltransferase promoter methylation status (MGMT status) and clinical outcome in patients with newly diagnosed glioblastoma, the current study aimed to evaluate radiomics in recurrent glioblastoma patients.

Methods

Pre-treatment MR-imaging data of 69 patients enrolled into the DIRECTOR trial in recurrent glioblastoma served as a training cohort, and 49 independent patients formed an external validation cohort. Contrast-enhancing tumor and peritumoral volumes were segmented on MR images. 180 radiomic features were extracted after application of two MR intensity normalization techniques: fixed number of bins and linear rescaling. Radiomic feature selection was performed via principal component analysis, and multivariable models were trained to predict MGMT status, progression-free survival from first salvage therapy, referred to herein as PFS₂, and overall survival (OS). The prognostic power of models was quantified with concordance index (CI) for survival data and area under receiver operating characteristic curve (AUC) for the MGMT status.

Results

We established and validated a radiomic model to predict MGMT status using linear intensity interpolation and considering features extracted from gadolinium-enhanced T1-weighted MRI (training AUC = 0.670, validation AUC = 0.673). Additionally, models predicting PFS₂ and OS were found for the training cohort but were not confirmed in our validation cohort.

Conclusions

A radiomic model for prediction of MGMT promoter methylation status from tumor texture features in patients with recurrent glioblastoma was successfully established, providing a non-invasive approach to anticipate patient's response to chemotherapy if biopsy cannot be performed. The radiomic approach to predict PFS₂ and OS failed.

Cytogenetic characteristics of and prognosis for acute myeloid leukemia in 107 children

Background

Patients diagnosed with acute myeloid leukemia (AML) in childhood have a poor prognosis. A better understanding of prognostic factors will assist patients and clinicians in making difficult treatment decisions.

Objectives

To understand further the cytogenetic characteristics of and reassess the prognostic value of cytogenetic abnormalities in childhood AML.

Methods

Conventional karyotypes of 107 children with AML were analyzed retrospectively. The cases were divided into 4 groups based on genetic rearrangements; namely patients with: t(15;17)/PML-RARA; t(8;21)/RUNX1-RUNX1T1 or inv(16)(p13;q22) and t(16;16)/CBFB-MYH11; -7 or complex karyotypes; normal karyotypes or other cytogenetic changes. Differences in age, sex, leukocyte count, event-free survival (EFS), and overall survival (OS) were analyzed.

Results

All French-American-British (FAB) subtypes of AML were detected in 107 patients. We successfully cultured 81 of 107 bone marrow specimens, of which 60 cases had abnormal karyotypes. The most common abnormal karyotypes were t(8;21) (17/81 cases), followed by t(15;17) (13/81 cases), -X/Y (10/81 cases). There were no significant differences (P > 0.05) in age, sex, or leukocyte counts between the 4 groups. The differences in 3-year EFS and OS between each pair were significant, except for groups of patients with t(8;21)/RUNX1-RUNX1T1 and patients with normal karyotypes or other cytogenetic changes (P = 0.054).

Conclusions

Chromosomal abnormalities may provide important prognostic factors for AML in children and be helpful for risk stratification and individual treatment.

Development and validation of a predictive model of in-hospital mortality in COVID-19 patients

We retrospectively evaluated 2879 hospitalized COVID-19 patients from four hospitals to evaluate the ability of demographic data, medical history, and on-admission laboratory parameters to predict in-hospital mortality. Association of previously published risk factors (age, gender, arterial hypertension, diabetes mellitus, smoking habit, obesity, renal failure, cardiovascular/ pulmonary diseases, serum ferritin, lymphocyte count, APTT, PT, fibrinogen, D-dimer, and platelet count) with death was tested by a multivariate logistic regression, and a predictive model was created, with further validation in an independent sample. A total of 2070 hospitalized COVID-19 patients were finally included in the multivariable analysis. Age 61–70 years (p<0.001; OR: 7.69; 95%CI: 2.93 to 20.14), age 71–80 years (p<0.001; OR: 14.99; 95%CI: 5.88 to 38.22), age >80 years (p<0.001; OR: 36.78; 95%CI: 14.42 to 93.85), male gender (p<0.001; OR: 1.84; 95%CI: 1.31 to 2.58), D-dimer levels >2 ULN (p = 0.003; OR: 1.79; 95%CI: 1.22 to 2.62), and prolonged PT (p<0.001; OR: 2.18; 95%CI: 1.49 to 3.18) were independently associated with increased in-hospital mortality. A predictive model performed with these parameters showed an AUC of 0.81 in the development cohort (n = 1270) [sensitivity of 95.83%, specificity of 41.46%, negative predictive value of 98.01%, and positive predictive value of 24.85%]. These results were then validated in an independent data sample (n = 800). Our predictive model of in-hospital mortality of COVID-19 patients has been developed, calibrated and validated. The model (MRS-COVID) included age, male gender, and on-admission coagulopathy markers as positively correlated factors with fatal outcome.

A Position Statement on the Utility of Interval Imaging in Standard of Care Brain Tumour Management: Defining the Evidence Gap and Opportunities for Future Research

Objectiv e

To summarise current evidence for the utility of interval imaging in monitoring disease in adult brain tumours, and to develop a position for future evidence gathering while incorporating the application of data science and health economics.

Methods

Experts in 'interval imaging' (imaging at pre-planned time-points to assess tumour status); data science; health economics, trial management of adult brain tumours, and patient representatives convened in London, UK. The current evidence on the use of interval imaging for monitoring brain tumours was reviewed. To improve the evidence that interval imaging has a role in disease management, we discussed specific themes of data science, health economics, statistical considerations, patient and carer perspectives, and multi-centre study design. Suggestions for future studies aimed at filling knowledge gaps were discussed.

Results

Meningioma and glioma were identified as priorities for interval imaging utility analysis. The "monitoring biomarkers" most commonly used in adult brain tumour patients were standard structural MRI features. Interval imaging was commonly scheduled to provide reported imaging prior to planned, regular clinic visits. There is limited evidence relating interval imaging in the absence of clinical deterioration to management change that alters morbidity, mortality, quality of life, or resource use. Progression-free survival is confounded as an outcome measure when using structural MRI in glioma. Uncertainty from imaging causes distress for some patients and their caregivers, while for others it provides an important indicator of disease activity. Any study design that changes imaging regimens should consider the potential for influencing current or planned therapeutic trials, ensure that opportunity costs are measured, and capture indirect benefits and added value.

Conclusion

Evidence for the value, and therefore utility, of regular interval imaging is currently lacking. Ongoing collaborative efforts will improve trial design and generate the evidence to optimise monitoring imaging biomarkers in standard of care brain tumour management.

Clinical Evidence for Locoregional Surgery of the Primary Tumor in Patients with De Novo Stage IV Breast Cancer

BACKGROUND

Whether primary tumor surgery is better than no surgery in patients with de novo stage IV breast cancer remains controversial.

METHODS

This study combined prospective clinical trials and a multicenter cohort to evaluate the impact of locoregional surgery in de novo stage IV breast cancer. The GRADE approach was used to assess the quality of evidence in meta-analysis, and propensity score matching analysis was used in the cohort study. This study was registered with PROSPERO CRD42016043766 and ClinicalTrials.gov NCT04456855.

RESULTS

A total of 1110 patients from six trials and 353 patients from the cohort study were included. The meta-analysis showed that compared with no surgery, locoregional surgery did not prolong overall survival (hazard ratio [HR] = 0.90, P = 0.40; moderate-quality) but had a significantly longer locoregional progression-free survival (HR = 0.23, P < 0.001; moderate-quality). The subgroup analysis of solitary bone-only metastasis (HR = 0.47, P = 0.04; high-quality) resulted in prolonged overall survival. In the cohort study, locoregional surgery showed a survival benefit (HR = 0.63, P = 0.041) before matching, but not (HR = 0.84, P = 0.579) after matching. Patients with bone-only metastasis showed a survival advantage in surgery compared with no surgery before matching (HR = 0.36, P = 0.034) as well as after matching (HR = 0.18, P = 0.017).

CONCLUSIONS

This study indicated that locoregional surgery had a significantly longer locoregional progression-free survival than no surgery in de novo stage IV breast cancer, and patients with bone-only metastasis tended to show an overall survival benefit from surgery.