Utilization of model-agnostic explainable artificial intelligence frameworks in oncology: a narrative review

Using XBGoost, an interpretable machine learning model, for diagnosing prostate cancer in patients with PSA < 20 ng/ml based on the PSAMR indicator

To create a diagnostic tool before biopsy for patients with prostate-specific antigen (PSA) levels < 20 ng/ml to minimize prostate biopsy-related discomfort and risks. Data from 655 patients who underwent transperineal prostate biopsy at the First Affiliated Hospital of Wannan Medical College from July 2021 to January 2023 were collected and analyzed. After applying the Synthetic Minority Over-sampling TEchnique class balancing on the training set, multiple machine learning models were constructed by using the Least Absolute Shrinkage and Selection Operator (LASSO) feature selection to identify the significant variables. The best-performing model was selected and evaluated through tenfold cross-validation to ensure interpretability. Finally, the performance was assessed using the test set data for validation. The age, prostate-specific antigen mass ratio (PSAMR), Prostate Imaging-Reporting and Data System, and prostate volume were selected as the variables for model construction based on the LASSO regression. The receiver operating characteristic (ROC) results for multiple models in the validation set were as follows: XGBoost: 0.93 (0.88-0.97); logistic: 0.89 (0.83-0.95); LightGBM: 0.87 (0.80-0.93); AdaBoost: 0.90 (0.85-0.96); GNB: 0.88 (0.82-0.95); CNB: 0.79 (0.71-0.87); MLP: 0.78 (0.69-0.86); and Support Vector Machine: 0.81 (0.73-0.89). XGBoost was selected as the best model and reconstructed with tenfold cross-validation on the training data, resulting in the following ROC scores: training set 0.995 (0.991-0.999), validation set 0.945 (0.885-0.997 ), and test set 0.920 (0.868-0.972). The Kolmogorov-Smirnov curve, calibration curve and learning curve yielded positive results; The decision curve demonstrates that patients with threshold probabilities ranging from 10 to 95% can benefit from this model. We developed an XGBoost machine learning model based on the PSAMR indicator and interpreted it using the SHapley Additive exPlanations method. The model offered a high-performance non-invasive technique to diagnose prostate cancer in patients with PSA levels < 20 ng/ml.

Multi-Class Brain Tumor Grades Classification Using a Deep Learning-Based Majority Voting Algorithm and Its Validation Using Explainable-AI

Biopsy is considered the gold standard for diagnosing brain tumors, but its invasive nature can pose risks to patients. Additionally, tissue analysis can be cumbersome and inconsistent among observers. This research aims to develop a cost-effective, non-invasive, MRI-based computer-aided diagnosis tool that can reliably, accurately and swiftly identify brain tumor grades. Our system employs ensemble deep learning (EDL) within an MRI multiclass framework that includes five datasets: two-class (C2), three-class (C3), four-class (C4), five-class (C5) and six-class (C6). The EDL utilizes a majority voting algorithm to classify brain tumors by combining seven renowned deep learning (DL) models-EfficientNet, VGG16, ResNet18, GoogleNet, ResNet50, Inception-V3 and DarkNet-and seven machine learning (ML) models, including support vector machine, K-nearest neighbour, Naïve Bayes, decision tree, linear discriminant analysis, artificial neural network and random forest. Additionally, local interpretable model-agnostic explanations (LIME) are employed as an explainable AI algorithm, providing a visual representation of the CNN's internal workings to enhance the credibility of the results. Through extensive five-fold cross-validation experiments, the DL-based majority voting algorithm outperformed the ML-based majority voting algorithm, achieving the highest average accuracies of 100 ± 0.00%, 98.55 ± 0.35%, 98.47 ± 0.63%, 95.34 ± 1.17% and 96.61 ± 0.85% for the C2, C3, C4, C5 and C6 datasets, respectively. Majority voting algorithms typically yield consistent results across different folds of the brain tumor data and enhance performance compared to any individual deep learning and machine learning models.

Cherry on Top or Real Need? A Review of Explainable Machine Learning in Kidney Transplantation

Research on solid organ transplantation has taken advantage of the substantial acquisition of medical data and the use of artificial intelligence (AI) and machine learning (ML) to answer diagnostic, prognostic, and therapeutic questions for many years. Nevertheless, despite the question of whether AI models add value to traditional modeling approaches, such as regression models, their "black box" nature is one of the factors that have hindered the translation from research to clinical practice. Several techniques that make such models understandable to humans were developed with the promise of increasing transparency in the support of medical decision-making. These techniques should help AI to close the gap between theory and practice by yielding trust in the model by doctors and patients, allowing model auditing, and facilitating compliance with emergent AI regulations. But is this also happening in the field of kidney transplantation? This review reports the use and explanation of "black box" models to diagnose and predict kidney allograft rejection, delayed graft function, graft failure, and other related outcomes after kidney transplantation. In particular, we emphasize the discussion on the need (or not) to explain ML models for biological discovery and clinical implementation in kidney transplantation. We also discuss promising future research paths for these computational tools.

AI-Enhanced Detection of Clinically Relevant Structural and Functional Anomalies in MRI: Traversing the Landscape of Conventional to Explainable Approaches

Anomaly detection in medical imaging, particularly within the realm of magnetic resonance imaging (MRI), stands as a vital area of research with far-reaching implications across various medical fields. This review meticulously examines the integration of artificial intelligence (AI) in anomaly detection for MR images, spotlighting its transformative impact on medical diagnostics. We delve into the forefront of AI applications in MRI, exploring advanced machine learning (ML) and deep learning (DL) methodologies that are pivotal in enhancing the precision of diagnostic processes. The review provides a detailed analysis of preprocessing, feature extraction, classification, and segmentation techniques, alongside a comprehensive evaluation of commonly used metrics. Further, this paper explores the latest developments in ensemble methods and explainable AI, offering insights into future directions and potential breakthroughs. This review synthesizes current insights, offering a valuable guide for researchers, clinicians, and medical imaging experts. It highlights AI's crucial role in improving the precision and speed of detecting key structural and functional irregularities in MRI. Our exploration of innovative techniques and trends furthers MRI technology development, aiming to refine diagnostics, tailor treatments, and elevate patient care outcomes. LEVEL OF EVIDENCE: 5 TECHNICAL EFFICACY: Stage 1.

Machine learning and explainable artificial intelligence to predict pathologic stage in men with localized prostate cancer

BACKGROUND

Though several nomograms exist, machine learning (ML) approaches might improve prediction of pathologic stage in patients with prostate cancer. To develop ML models to predict pathologic stage that outperform existing nomograms that use readily available clinicopathologic variables.

METHODS

Patients with prostate adenocarcinoma who underwent surgery were identified in the National Cancer Database. Seven ML models were trained to predict organ-confined (OC) disease, extracapsular extension, seminal vesicle invasion (SVI), and lymph node involvement (LNI). Model performance was measured using area under the curve (AUC) on a holdout testing data set. Clinical utility was evaluated using decision curve analysis (DCA). Performance metrics were confirmed on an external validation data set.

RESULTS

The ML-based extreme gradient boosted trees model achieved the best performance with an AUC of 0.744, 0.749, 0.816, 0.811 for the OC, ECE, SVI, and LNI models, respectively. The MSK nomograms achieved an AUC of 0.708, 0.742, 0.806, 0.802 for the OC, ECE, SVI, and LNI models, respectively. These models also performed the best on DCA. Findings were consistent on both a holdout internal validation data set as well as an external validation data set.

CONCLUSIONS

Our ML models better predicted pathologic stage relative to existing nomograms at predicting pathologic stage. Accurate prediction of pathologic stage can help oncologists and patients determine optimal definitive treatment options for patients with prostate cancer.

Deep learning algorithms for melanoma detection using dermoscopic images: A systematic review and meta-analysis

BACKGROUND

Melanoma is a serious risk to human health and early identification is vital for treatment success. Deep learning (DL) has the potential to detect cancer using imaging technologies and many studies provide evidence that DL algorithms can achieve high accuracy in melanoma diagnostics.

OBJECTIVES

To critically assess different DL performances in diagnosing melanoma using dermatoscopic images and discuss the relationship between dermatologists and DL.

METHODS

Ovid-Medline, Embase, IEEE Xplore, and the Cochrane Library were systematically searched from inception until 7th December 2021. Studies that reported diagnostic DL model performances in detecting melanoma using dermatoscopic images were included if they had specific outcomes and histopathologic confirmation. Binary diagnostic accuracy data and contingency tables were extracted to analyze outcomes of interest, which included sensitivity (SEN), specificity (SPE), and area under the curve (AUC). Subgroup analyses were performed according to human-machine comparison and cooperation. The study was registered in PROSPERO, CRD42022367824.

RESULTS

2309 records were initially retrieved, of which 37 studies met our inclusion criteria, and 27 provided sufficient data for meta-analytical synthesis. The pooled SEN was 82 % (range 77-86), SPE was 87 % (range 84-90), with an AUC of 0.92 (range 0.89-0.94). Human-machine comparison had pooled AUCs of 0.87 (0.84-0.90) and 0.83 (0.79-0.86) for DL and dermatologists, respectively. Pooled AUCs were 0.90 (0.87-0.93), 0.80 (0.76-0.83), and 0.88 (0.85-0.91) for DL, and junior and senior dermatologists, respectively. Analyses of human-machine cooperation were 0.88 (0.85-0.91) for DL, 0.76 (0.72-0.79) for unassisted, and 0.87 (0.84-0.90) for DL-assisted dermatologists.

CONCLUSIONS

Evidence suggests that DL algorithms are as accurate as senior dermatologists in melanoma diagnostics. Therefore, DL could be used to support dermatologists in diagnostic decision-making. Although, further high-quality, large-scale multicenter studies are required to address the specific challenges associated with medical AI-based diagnostics.

Artificial intelligence in cardiac surgery: A systematic review

BACKGROUND

Artificial intelligence (AI) has emerged as a tool to potentially increase the efficiency and efficacy of cardiovascular care and improve clinical outcomes. This study aims to provide an overview of applications of AI in cardiac surgery.

METHODS

A systematic literature search on AI applications in cardiac surgery from inception to February 2024 was conducted. Articles were then filtered based on the inclusion and exclusion criteria and the risk of bias was assessed. Key findings were then summarized.

RESULTS

A total of 81 studies were found that reported on AI applications in cardiac surgery. There is a rapid rise in studies since 2020. The most popular machine learning technique was random forest (n = 48), followed by support vector machine (n = 33), logistic regression (n = 32), and eXtreme Gradient Boosting (n = 31). Most of the studies were on adult patients, conducted in China, and involved procedures such as valvular surgery (24.7%), heart transplant (9.4%), coronary revascularization (11.8%), congenital heart disease surgery (3.5%), and aortic dissection repair (2.4%). Regarding evaluation outcomes, 35 studies examined the performance, 26 studies examined clinician outcomes, and 20 studies examined patient outcomes.

CONCLUSION

AI was mainly used to predict complications following cardiac surgeries and improve clinicians' decision-making by providing better preoperative risk assessment, stratification, and prognostication. While the application of AI in cardiac surgery has greatly progressed in the last decade, further studies need to be conducted to verify accuracy and ensure safety before use in clinical practice.

Predicting progression-free survival in patients with epithelial ovarian cancer using an interpretable random forest model

Prognostic models play a crucial role in providing personalised risk assessment, guiding treatment decisions, and facilitating the counselling of patients with cancer. However, previous imaging-based artificial intelligence models of epithelial ovarian cancer lacked interpretability. In this study, we aimed to develop an interpretable machine-learning model to predict progression-free survival in patients with epithelial ovarian cancer using clinical variables and radiomics features. A total of 102 patients with epithelial ovarian cancer who underwent contrast-enhanced computed tomography scans were enrolled in this retrospective study. Pre-surgery clinical data, including age, performance status, body mass index, tumour stage, venous blood cancer antigen-125 (CA125) level, white blood cell count, neutrophil count, red blood cell count, haemoglobin level, and platelet count, were obtained from medical records. The volume of interest for each tumour was manually delineated slice-by-slice along the boundary. A total of 2074 radiomic features were extracted from the pre- and post-contrast computed tomography images. Optimal radiomic features were selected using the Least Absolute Shrinkage and Selection Operator logistic regression. Multivariate Cox analysis was performed to identify independent predictors of three-year progression-free survival. The random forest algorithm developed radiomic and combined models using four-fold cross-validation. Finally, the Shapley additive explanation algorithm was applied to interpret the predictions of the combined model. Multivariate Cox analysis identified CA-125 levels (P = 0.015), tumour stage (P = 0.019), and Radscore (P < 0.001) as independent predictors of progression-free survival. The combined model based on these factors achieved an area under the curve of 0.812 (95 % confidence interval: 0.802-0.822) in the training cohort and 0.772 (95 % confidence interval: 0.727-0.817) in the validation cohort. The most impactful features on the model output were Radscore, followed by tumour stage and CA-125. In conclusion, the Shapley additive explanation-based interpretation of the prognostic model enables clinicians to understand the reasoning behind predictions better.

Development and Validation of an Interpretable Machine Learning Model for Early Prognosis Prediction in ICU Patients with Malignant Tumors and Hyperkalemia

This study aims to develop and validate a machine learning (ML) predictive model for assessing mortality in patients with malignant tumors and hyperkalemia (MTH). We extracted data on patients with MTH from the Medical Information Mart for Intensive Care-IV, version 2.2 (MIMIC-IV v2.2) database. The dataset was split into a training set (75%) and a validation set (25%). We used the Least Absolute Shrinkage and Selection Operator (LASSO) regression to identify potential predictors, which included clinical laboratory indicators and vital signs. Pearson correlation analysis tested the correlation between predictors. In-hospital death was the prediction target. The Area Under the Curve (AUC) and accuracy of the training and validation sets of 7 ML algorithms were compared, and the optimal 1 was selected to develop the model. The calibration curve was used to evaluate the prediction accuracy of the model further. SHapley Additive exPlanations (SHAP) and Local Interpretable Model-agnostic Explanations (LIME) enhanced model interpretability. 496 patients with MTH in the Intensive Care Unit (ICU) were included. After screening, 17 clinical features were included in the construction of the ML model, and the Pearson correlation coefficient was <0.8, indicating that the correlation between the clinical features was small. eXtreme Gradient Boosting (XGBoost) outperformed other algorithms, achieving perfect scores in the training set (accuracy: 1.000, AUC: 1.000) and high scores in the validation set (accuracy: 0.734, AUC: 0.733). The calibration curves indicated good predictive calibration of the model. SHAP analysis identified the top 8 predictive factors: urine output, mean heart rate, maximum urea nitrogen, minimum oxygen saturation, minimum mean blood pressure, maximum total bilirubin, mean respiratory rate, and minimum pH. In addition, SHAP and LIME performed in-depth individual case analyses. This study demonstrates the effectiveness of ML methods in predicting mortality risk in ICU patients with MTH. It highlights the importance of predictors like urine output and mean heart rate. SHAP and LIME significantly enhanced the model's interpretability.

Predicting axillary lymph node metastasis in breast cancer patients: A radiomics-based multicenter approach with interpretability analysis

PURPOSE

To develop a MRI-based radiomics model, integrating the intratumoral and peritumoral imaging information to predict axillary lymph node metastasis (ALNM) in patients with breast cancer and to elucidate the model's decision-making process via interpretable algorithms.

METHODS

This study included 376 patients from three institutions who underwent contrast-enhanced breast MRI between 2021 and 2023. We used multiple machine learning algorithms to combine peritumoral, intratumoral, and radiological characteristics with the building of radiological, radiomics, and combined models. The model's performance was compared based on the area under the curve (AUC) obtained from the receiver operating characteristic analysis and interpretable machine learning techniques to analyze the operating mechanism of the model.

RESULTS

The radiomics model, incorporating features from both intratumoral tissue and the 3 mm peritumoral region and utilizing the backpropagation neural network (BPNN) algorithm, demonstrated superior diagnostic efficacy, achieving an AUC of 0.820. The AUC of the combination of the RAD score, clinical T stage, and spiculated margin was as high as 0.855. Furthermore, we conducted SHapley Additive exPlanations (SHAP) analysis to evaluate the contributions of RAD score, clinical T stage, and spiculated margin in ALNM status prediction.

CONCLUSIONS

The interpretable radiomics model we propose can better predict the ALNM status of breast cancer and help inform clinical treatment decisions.

An explainable AI-assisted web application in cancer drug value prediction

In recent years, there has been an increase in the interest in adopting Explainable Artificial Intelligence (XAI) for healthcare. The proposed system includes•An XAI model for cancer drug value prediction. The model provides data that is easy to understand and explain, which is critical for medical decision-making. It also produces accurate projections.•A model outperformed existing models due to extensive training and evaluation on a large cancer medication chemical compounds dataset.•Insights into the causation and correlation between the dependent and independent actors in the chemical composition of the cancer cell. While the model is evaluated on Lung Cancer data, the architecture offered in the proposed solution is cancer agnostic. It may be scaled out to other cancer cell data if the properties are similar. The work presents a viable route for customizing treatments and improving patient outcomes in oncology by combining XAI with a large dataset. This research attempts to create a framework where a user can upload a test case and receive forecasts with explanations, all in a portable PDF report.

Explainable artificial intelligence analysis of brachytherapy boost receipt in cervical cancer during the COVID-19 era

PURPOSE

Brachytherapy is a critical component of the standard-of-care curative radiotherapy regimen for women with locally advanced cervical cancer (LACC). However, existing literature suggests that many patients will not receive the brachytherapy boost. We used machine learning (ML) and explainable artificial intelligence to characterize this disparity.

MATERIALS AND METHODS

Patients with LACC diagnosed from 2004 to 2020 who received definitive radiation were identified in the National Cancer Database. Five ML models were trained to predict if a patient received a brachytherapy boost. The best-performing model was explained using SHapley Additive exPlanation (SHAP) values. To identify trends that may be attributable to the coronavirus disease 2019 (COVID-19) pandemic, the previous analysis was repeated and limited to 2019 to 2020.

RESULTS

A total of 37,564 patients with LACC were identified; 5799 were diagnosed from 2019 to 2020 (COVID cohort). Of these patients, 59.3% received a brachytherapy boost, with 76.4% of patients diagnosed in 2019 to 2020 receiving a boost. The random forest model achieved the best performance for both the overall and COVID cohorts. In the overall cohort, the most important predictive features were the year of diagnosis, stage, age, and insurance status. In the COVID cohort, the most important predictive features were FIGO stage, age, insurance status, and hospital type. Of the 26 patients who tested positive for COVID-19 during their course of radiotherapy, 19 (73.1%) received a brachytherapy boost.

CONCLUSIONS

A gradual increase in brachytherapy boost utilization has been noted, which did not seem to be significantly impacted by the onset of the COVID-19 pandemic. ML could be considered to identify patient populations where brachytherapy is underutilized, which can provide actionable feedback for improving access.

The Promise and Perils of Artificial Intelligence in Health Professions Education Practice and Scholarship

ABSTRACT

Artificial intelligence (AI) methods, especially machine learning and natural language processing, are increasingly affecting health professions education (HPE), including the medical school application and selection processes, assessment, and scholarship production. The rise of large language models over the past 18 months, such as ChatGPT, has raised questions about how best to incorporate these methods into HPE. The lack of training in AI among most HPE faculty and scholars poses an important challenge in facilitating such discussions. In this commentary, the authors provide a primer on the AI methods most often used in the practice and scholarship of HPE, discuss the most pressing challenges and opportunities these tools afford, and underscore that these methods should be understood as part of the larger set of statistical tools available.Despite their ability to process huge amounts of data and their high performance completing some tasks, AI methods are only as good as the data on which they are trained. Of particular importance is that these models can perpetuate the biases that are present in those training datasets, and they can be applied in a biased manner by human users. A minimum set of expectations for the application of AI methods in HPE practice and scholarship is discussed in this commentary, including the interpretability of the models developed and the transparency needed into the use and characteristics of such methods.The rise of AI methods is affecting multiple aspects of HPE including raising questions about how best to incorporate these models into HPE practice and scholarship. In this commentary, we provide a primer on the AI methods most often used in HPE and discuss the most pressing challenges and opportunities these tools afford.

Insights from Explainable Artificial Intelligence of Pollution and Socioeconomic Influences for Respiratory Cancer Mortality in Italy

Respiratory malignancies, encompassing cancers affecting the lungs, the trachea, and the bronchi, pose a significant and dynamic public health challenge. Given that air pollution stands as a significant contributor to the onset of these ailments, discerning the most detrimental agents becomes imperative for crafting policies aimed at mitigating exposure. This study advocates for the utilization of explainable artificial intelligence (XAI) methodologies, leveraging remote sensing data, to ascertain the primary influencers on the prediction of standard mortality rates (SMRs) attributable to respiratory cancer across Italian provinces, utilizing both environmental and socioeconomic data. By scrutinizing thirteen distinct machine learning algorithms, we endeavor to pinpoint the most accurate model for categorizing Italian provinces as either above or below the national average SMR value for respiratory cancer. Furthermore, employing XAI techniques, we delineate the salient factors crucial in predicting the two classes of SMR. Through our machine learning scrutiny, we illuminate the environmental and socioeconomic factors pertinent to mortality in this disease category, thereby offering a roadmap for prioritizing interventions aimed at mitigating risk factors.

From pixels to prognosis: unveiling radiomics models with SHAP and LIME for enhanced interpretability

Radiomics-based prediction models have shown promise in predicting Radiation Pneumonitis (RP), a common adverse outcome of chest irradiation. Τhis study looks into more than just RP: it also investigates a bigger shift in the way radiomics-based models work. By integrating multi-modal radiomic data, which includes a wide range of variables collected from medical images including cutting-edge PET/CT imaging, we have developed predictive models that capture the intricate nature of illness progression. Radiomic features were extracted using PyRadiomics, encompassing intensity, texture, and shape measures. The high-dimensional dataset formed the basis for our predictive models, primarily Gradient Boosting Machines (GBM)-XGBoost, LightGBM, and CatBoost. Performance evaluation metrics, including Multi-Modal AUC-ROC, Sensitivity, Specificity, and F1-Score, underscore the superiority of the Deep Neural Network (DNN) model. The DNN achieved a remarkable Multi-Modal AUC-ROC of 0.90, indicating superior discriminatory power. Sensitivity and specificity values of 0.85 and 0.91, respectively, highlight its effectiveness in detecting positive occurrences while accurately identifying negatives. External validation datasets, comprising retrospective patient data and a heterogeneous patient population, validate the robustness and generalizability of our models. The focus of our study is the application of sophisticated model interpretability methods, namely SHAP (SHapley Additive exPlanations) and LIME (Local Interpretable Model-Agnostic Explanations), to improve the clarity and understanding of predictions. These methods allow clinicians to visualize the effects of features and provide localized explanations for every prediction, enhancing the comprehensibility of the model. This strengthens trust and collaboration between computational technologies and medical competence. The integration of data-driven analytics and medical domain expertise represents a significant shift in the profession, advancing us from analyzing pixel-level information to gaining valuable prognostic insights.

SurvIAE: Survival prediction with Interpretable Autoencoders from Diffuse Large B-Cells Lymphoma gene expression data

BACKGROUND

In Diffuse Large B-Cell Lymphoma (DLBCL), several methodologies are emerging to derive novel biomarkers to be incorporated in the risk assessment. We realized a pipeline that relies on autoencoders (AE) and Explainable Artificial Intelligence (XAI) to stratify prognosis and derive a gene-based signature.

METHODS

AE was exploited to learn an unsupervised representation of the gene expression (GE) from three publicly available datasets, each with its own technology. Multi-layer perceptron (MLP) was used to classify prognosis from latent representation. GE data were preprocessed as normalized, scaled, and standardized. Four different AE architectures (Large, Medium, Small and Extra Small) were compared to find the most suitable for GE data. The joint AE-MLP classified patients on six different outcomes: overall survival at 12, 36, 60 months and progression-free survival (PFS) at 12, 36, 60 months. XAI techniques were used to derive a gene-based signature aimed at refining the Revised International Prognostic Index (R-IPI) risk, which was validated in a fourth independent publicly available dataset. We named our tool SurvIAE: Survival prediction with Interpretable AE.

RESULTS

From the latent space of AEs, we observed that scaled and standardized data reduced the batch effect. SurvIAE models outperformed R-IPI with Matthews Correlation Coefficient up to 0.42 vs. 0.18 for the validation-set (PFS36) and to 0.30 vs. 0.19 for the test-set (PFS60). We selected the SurvIAE-Small-PFS36 as the best model and, from its gene signature, we stratified patients in three risk groups: R-IPI Poor patients with High levels of GAB1, R-IPI Poor patients with Low levels of GAB1 or R-IPI Good/Very Good patients with Low levels of GPR132, and R-IPI Good/Very Good patients with High levels of GPR132.

CONCLUSIONS

SurvIAE showed the potential to derive a gene signature with translational purpose in DLBCL. The pipeline was made publicly available and can be reused for other pathologies.

A critical moment in machine learning in medicine: on reproducible and interpretable learning

Over the past two decades, advances in computational power and data availability combined with increased accessibility to pre-trained models have led to an exponential rise in machine learning (ML) publications. While ML may have the potential to transform healthcare, this sharp increase in ML research output without focus on methodological rigor and standard reporting guidelines has fueled a reproducibility crisis. In addition, the rapidly growing complexity of these models compromises their interpretability, which currently impedes their successful and widespread clinical adoption. In medicine, where failure of such models may have severe implications for patients' health, the high requirements for accuracy, robustness, and interpretability confront ML researchers with a unique set of challenges. In this review, we discuss the semantics of reproducibility and interpretability, as well as related issues and challenges, and outline possible solutions to counteracting the "black box". To foster reproducibility, standard reporting guidelines need to be further developed and data or code sharing encouraged. Editors and reviewers may equally play a critical role by establishing high methodological standards and thus preventing the dissemination of low-quality ML publications. To foster interpretable learning, the use of simpler models more suitable for medical data can inform the clinician how results are generated based on input data. Model-agnostic explanation tools, sensitivity analysis, and hidden layer representations constitute further promising approaches to increase interpretability. Balancing model performance and interpretability are important to ensure clinical applicability. We have now reached a critical moment for ML in medicine, where addressing these issues and implementing appropriate solutions will be vital for the future evolution of the field.

Interpretable CT radiomics model for invasiveness prediction in patients with ground-glass nodules

AIM

To evaluate the performance of an interpretable computed tomography (CT) radiomic model in predicting the invasiveness of ground-glass nodules (GGNs).

MATERIALS AND METHODS

The study was conducted retrospectively from 1 August 2017 to 1 August 2022, at three different centres. Two hundred and thirty patients with GGNs were enrolled at centre I as a training cohort. Centres II (n=157) and III (n=156) formed two external validation cohorts. Radiomics features extracted based on CT were reduced by a coarse-fine feature screening strategy. A radiomic model was developed through the use of the LASSO (least absolute shrinkage and selection operator) and XGBoost algorithms. Then, a radiological model was established through multivariate logistic regression analysis. Finally, the interpretability of the model was explored using SHapley Additive exPlanations (SHAP).

RESULTS

The radiomic XGBoost model outperformed the radiomic logistic model and radiological model in assessing the invasiveness of GGNs. The area under the curve (AUC) values for the radiomic XGBoost model were 0.885 (95% confidence interval [CI] 0.836-0.923), 0.853 (95% CI 0.790-0.906), and 0.838 (95% CI 0.773-0.902) in the training and the two external validation cohorts, respectively. The SHAP method allowed for both a quantitative and visual representation of how decisions were made using a given model for each individual patient. This can provide a deeper understanding of the decision-making mechanisms within the model and the factors that contribute to its prediction effectiveness.

CONCLUSIONS

The present interpretable CT radiomics model has the potential to preoperatively evaluate the invasiveness of GGNs. Furthermore, it can provide personalised, image-based clinical-decision support.

Graph Neural Networks in Cancer and Oncology Research: Emerging and Future Trends

Next-generation cancer and oncology research needs to take full advantage of the multimodal structured, or graph, information, with the graph data types ranging from molecular structures to spatially resolved imaging and digital pathology, biological networks, and knowledge graphs. Graph Neural Networks (GNNs) efficiently combine the graph structure representations with the high predictive performance of deep learning, especially on large multimodal datasets. In this review article, we survey the landscape of recent (2020-present) GNN applications in the context of cancer and oncology research, and delineate six currently predominant research areas. We then identify the most promising directions for future research. We compare GNNs with graphical models and "non-structured" deep learning, and devise guidelines for cancer and oncology researchers or physician-scientists, asking the question of whether they should adopt the GNN methodology in their research pipelines.

Automated machine learning-based model for predicting benign anastomotic strictures in patients with rectal cancer who have received anterior resection

BACKGROUND

Benign anastomotic strictures (BAS) significantly impact patients' quality of life and long-term prognosis. However, the current clinical practice lacks accurate tools for predicting BAS. This study aimed to develop a machine-learning model to predict BAS in patients with rectal cancer who have undergone anterior resection.

METHODS

Data from 1973 patients who underwent anterior resection for rectal cancer were collected. Multiple machine learning classification models were integrated to analyze the data and identify the optimal model. Model performance was evaluated using receiver operator characteristic (ROC) curves, decision curve analysis (DCA), and calibration curves. The Shapley Additive exPlanation (SHAP) algorithm was utilized to assess the impact of various clinical characteristics on the optimal model to enhance the interpretability of the model results.

RESULTS

A total of 10 clinical features were considered in constructing the machine learning model. The model evaluation results indicated that the random forest (RF)model was optimal, with the area under the test set curve (AUC: 0.888, 95% CI: 0.810-0.965), accuracy: 0.792, sensitivity: 0.846, specificity: 0.791. The SHAP algorithm analysis identified prophylactic ileostomy, operative time, and anastomotic leakage as significant contributing factors influencing the predictions of the RF model.

CONCLUSION

We developed a robust machine-learning model and user-friendly online prediction tool for predicting BAS following anterior resection of rectal cancer. This tool offers a potential foundation for BAS prevention and aids clinical practice by enabling more efficient disease management and precise medical interventions.

Artificial intelligence and allied subsets in early detection and preclusion of gynecological cancers

Gynecological cancers including breast, cervical, ovarian, uterine, and vaginal, pose the greatest threat to world health, with early identification being crucial to patient outcomes and survival rates. The application of machine learning (ML) and artificial intelligence (AI) approaches to the study of gynecological cancer has shown potential to revolutionize cancer detection and diagnosis. The current review outlines the significant advancements, obstacles, and prospects brought about by AI and ML technologies in the timely identification and accurate diagnosis of different types of gynecological cancers. The AI-powered technologies can use genomic data to discover genetic alterations and biomarkers linked to a particular form of gynecologic cancer, assisting in the creation of targeted treatments. Furthermore, it has been shown that the potential benefits of AI and ML technologies in gynecologic tumors can greatly increase the accuracy and efficacy of cancer diagnosis, reduce diagnostic delays, and possibly eliminate the need for needless invasive operations. In conclusion, the review focused on the integrative part of AI and ML based tools and techniques in the early detection and exclusion of various cancer types; together with a collaborative coordination between research clinicians, data scientists, and regulatory authorities, which is suggested to realize the full potential of AI and ML in gynecologic cancer care.

A deep learning-based interpretable decision tool for predicting high risk of chemotherapy-induced nausea and vomiting in cancer patients prescribed highly emetogenic chemotherapy

OBJECTIVE

This study aims to develop a risk prediction model for chemotherapy-induced nausea and vomiting (CINV) in cancer patients receiving highly emetogenic chemotherapy (HEC) and identify the variables that have the most significant impact on prediction.

METHODS

Data from Tianjin Medical University General Hospital were collected and subjected to stepwise data preprocessing. Deep learning algorithms, including deep forest, and typical machine learning algorithms such as support vector machine (SVM), categorical boosting (CatBoost), random forest, decision tree, and neural network were used to develop the prediction model. After training the model and conducting hyperparameter optimization (HPO) through cross-validation in the training set, the performance was evaluated using the test set. Shapley additive explanations (SHAP), partial dependence plot (PDP), and Local Interpretable Model-Agnostic Explanations (LIME) techniques were employed to explain the optimal model. Model performance was assessed using AUC, F1 score, accuracy, specificity, sensitivity, and Brier score.

RESULTS

The deep forest model exhibited good discrimination, outperforming typical machine learning models, with an AUC of 0.850 (95%CI, 0.780-0.919), an F1 score of 0.757, an accuracy of 0.852, a specificity of 0.863, a sensitivity of 0.784, and a Brier score of 0.082. The top five important features in the model were creatinine clearance (Ccr), age, gender, anticipatory nausea and vomiting, and antiemetic regimen. Among these, Ccr had the most significant predictive value. The risk of CINV decreased with increased Ccr and age, while it was higher in the presence of anticipatory nausea and vomiting, female gender, and non-standard antiemetic regimen.

CONCLUSION

The deep forest model demonstrated good discrimination in predicting the risk of CINV in cancer patients prescribed HEC. Kidney function, as represented by Ccr, played a crucial role in the model's prediction. The clinical application of this predictive tool can help assess individual risks and improve patient care by proactively optimizing the use of antiemetics in cancer patients receiving HEC.

Assessment of brain cancer atlas maps with multimodal imaging features

BACKGROUND

Glioblastoma Multiforme (GBM) is a fast-growing and highly aggressive brain tumor that invades the nearby brain tissue and presents secondary nodular lesions across the whole brain but generally does not spread to distant organs. Without treatment, GBM can result in death in about 6 months. The challenges are known to depend on multiple factors: brain localization, resistance to conventional therapy, disrupted tumor blood supply inhibiting effective drug delivery, complications from peritumoral edema, intracranial hypertension, seizures, and neurotoxicity.

MAIN TEXT

Imaging techniques are routinely used to obtain accurate detections of lesions that localize brain tumors. Especially magnetic resonance imaging (MRI) delivers multimodal images both before and after the administration of contrast, which results in displaying enhancement and describing physiological features as hemodynamic processes. This review considers one possible extension of the use of radiomics in GBM studies, one that recalibrates the analysis of targeted segmentations to the whole organ scale. After identifying critical areas of research, the focus is on illustrating the potential utility of an integrated approach with multimodal imaging, radiomic data processing and brain atlases as the main components. The templates associated with the outcome of straightforward analyses represent promising inference tools able to spatio-temporally inform on the GBM evolution while being generalizable also to other cancers.

CONCLUSIONS

The focus on novel inference strategies applicable to complex cancer systems and based on building radiomic models from multimodal imaging data can be well supported by machine learning and other computational tools potentially able to translate suitably processed information into more accurate patient stratifications and evaluations of treatment efficacy.

Integration of artificial intelligence in lung cancer: Rise of the machine

The goal of oncology is to provide the longest possible survival outcomes with the therapeutics that are currently available without sacrificing patients' quality of life. In lung cancer, several data points over a patient's diagnostic and treatment course are relevant to optimizing outcomes in the form of precision medicine, and artificial intelligence (AI) provides the opportunity to use available data from molecular information to radiomics, in combination with patient and tumor characteristics, to help clinicians provide individualized care. In doing so, AI can help create models to identify cancer early in diagnosis and deliver tailored therapy on the basis of available information, both at the time of diagnosis and in real time as they are undergoing treatment. The purpose of this review is to summarize the current literature in AI specific to lung cancer and how it applies to the multidisciplinary team taking care of these complex patients.

Stratification of Length of Stay Prediction following Surgical Cytoreduction in Advanced High-Grade Serous Ovarian Cancer Patients Using Artificial Intelligence; the Leeds L-AI-OS Score

(1) Background: Length of stay (LOS) has been suggested as a marker of the effectiveness of short-term care. Artificial Intelligence (AI) technologies could help monitor hospital stays. We developed an AI-based novel predictive LOS score for advanced-stage high-grade serous ovarian cancer (HGSOC) patients following cytoreductive surgery and refined factors significantly affecting LOS. (2) Methods: Machine learning and deep learning methods using artificial neural networks (ANN) were used together with conventional logistic regression to predict continuous and binary LOS outcomes for HGSOC patients. The models were evaluated in a post-hoc internal validation set and a Graphical User Interface (GUI) was developed to demonstrate the clinical feasibility of sophisticated LOS predictions. (3) Results: For binary LOS predictions at differential time points, the accuracy ranged between 70-98%. Feature selection identified surgical complexity, pre-surgery albumin, blood loss, operative time, bowel resection with stoma formation, and severe postoperative complications (CD3-5) as independent LOS predictors. For the GUI numerical LOS score, the ANN model was a good estimator for the standard deviation of the LOS distribution by ± two days. (4) Conclusions: We demonstrated the development and application of both quantitative and qualitative AI models to predict LOS in advanced-stage EOC patients following their cytoreduction. Accurate identification of potentially modifiable factors delaying hospital discharge can further inform services performing root cause analysis of LOS.