Clinical decision support and individualized prediction of survival in colon cancer: bayesian belief network model

Predicting Colorectal Cancer Survival Using Time-to-Event Machine Learning: Retrospective Cohort Study

BACKGROUND

Machine learning (ML) methods have shown great potential in predicting colorectal cancer (CRC) survival. However, the ML models introduced thus far have mainly focused on binary outcomes and have not considered the time-to-event nature of this type of modeling.

OBJECTIVE

This study aims to evaluate the performance of ML approaches for modeling time-to-event survival data and develop transparent models for predicting CRC-specific survival.

METHODS

The data set used in this retrospective cohort study contains information on patients who were newly diagnosed with CRC between December 28, 2012, and December 27, 2019, at West China Hospital, Sichuan University. We assessed the performance of 6 representative ML models, including random survival forest (RSF), gradient boosting machine (GBM), DeepSurv, DeepHit, neural net-extended time-dependent Cox (or Cox-Time), and neural multitask logistic regression (N-MTLR) in predicting CRC-specific survival. Multiple imputation by chained equations method was applied to handle missing values in variables. Multivariable analysis and clinical experience were used to select significant features associated with CRC survival. Model performance was evaluated in stratified 5-fold cross-validation repeated 5 times by using the time-dependent concordance index, integrated Brier score, calibration curves, and decision curves. The SHapley Additive exPlanations method was applied to calculate feature importance.

RESULTS

A total of 2157 patients with CRC were included in this study. Among the 6 time-to-event ML models, the DeepHit model exhibited the best discriminative ability (time-dependent concordance index 0.789, 95% CI 0.779-0.799) and the RSF model produced better-calibrated survival estimates (integrated Brier score 0.096, 95% CI 0.094-0.099), but these are not statistically significant. Additionally, the RSF, GBM, DeepSurv, Cox-Time, and N-MTLR models have comparable predictive accuracy to the Cox Proportional Hazards model in terms of discrimination and calibration. The calibration curves showed that all the ML models exhibited good 5-year survival calibration. The decision curves for CRC-specific survival at 5 years showed that all the ML models, especially RSF, had higher net benefits than default strategies of treating all or no patients at a range of clinically reasonable risk thresholds. The SHapley Additive exPlanations method revealed that R0 resection, tumor-node-metastasis staging, and the number of positive lymph nodes were important factors for 5-year CRC-specific survival.

CONCLUSIONS

This study showed the potential of applying time-to-event ML predictive algorithms to help predict CRC-specific survival. The RSF, GBM, Cox-Time, and N-MTLR algorithms could provide nonparametric alternatives to the Cox Proportional Hazards model in estimating the survival probability of patients with CRC. The transparent time-to-event ML models help clinicians to more accurately predict the survival rate for these patients and improve patient outcomes by enabling personalized treatment plans that are informed by explainable ML models.

Prediction model of gleason score upgrading after radical prostatectomy based on a bayesian network

OBJECTIVE

To explore the clinical value of the Gleason score upgrading (GSU) prediction model after radical prostatectomy (RP) based on a Bayesian network.

METHODS

The data of 356 patients who underwent prostate biopsy and RP in our hospital from January 2018 to May 2021 were retrospectively analysed. Fourteen risk factors, including age, body mass index (BMI), total prostate-specific antigen (tPSA), prostate volume, total prostate-specific antigen density (PSAD), the number and proportion of positive biopsy cores, PI-RADS score, clinical stage and postoperative pathological characteristics, were included in the analysis. Data were used to establish a prediction model for Gleason score elevation based on the tree augmented naive (TAN) Bayesian algorithm. Moreover, the Bayesia Lab validation function was used to calculate the importance of polymorphic Birnbaum according to the results of the posterior analysis and to obtain the importance of each risk factor.

RESULTS

In the overall cohort, 110 patients (30.89%) had GSU. Based on all of the risk factors that were included in this study, the AUC of the model was 81.06%, and the accuracy was 76.64%. The importance ranking results showed that lymphatic metastasis, the number of positive biopsy cores, ISUP stage and PI-RADS score were the top four influencing factors for GSU after RP.

CONCLUSIONS

The prediction model of GSU after RP based on a Bayesian network has high accuracy and can more accurately evaluate the Gleason score of prostate biopsy specimens and guide treatment decisions.

Establishment of a Bayesian network model to predict the survival of malignant peritoneal mesothelioma patients after cytoreductive surgery plus hyperthermic intraperitoneal chemotherapy

OBJECTIVES

To establish a Bayesian network (BN) model to predict the survival of patients with malignant peritoneal mesothelioma (MPM) treated with cytoreductive surgery (CRS) plus hyperthermic intraperitoneal chemotherapy (HIPEC).

METHODS

The clinicopathological data of 154 MPM patients treated with CRS + HIPEC at our hospital from April 2015 to November 2022 were retrospectively analyzed. They were randomly divided into two groups in a 7:3 ratio. Survival analysis was conducted on the training set and a BN model was established. The accuracy of the model was validated using a confusion matrix of the testing set. The receiver operating characteristic (ROC) curve and area under the curve were used to evaluate the overall performance of the BN model.

RESULTS

Survival analysis of 107 patients (69.5%) in the training set found ten factors affecting patient prognosis: age, Karnofsky performance score, surgical history, ascites volume, peritoneal cancer index, organ resections, red blood cell transfusion, pathological types, lymphatic metastasis, and Ki-67 index (all p < 0.05). The BN model was successfully established after the above factors were included, and the BN model structure was adjusted according to previous research and clinical experience. The results of confusion matrix obtained by internal validation of 47 cases in the testing set showed that the accuracy of BN model was 72.7%, and the area under ROC was 0.74.

CONCLUSIONS

The BN model was established successfully with good overall performance and can be used as a clinical decision reference.

Qualitative Evaluation of an Artificial Intelligence–Based Clinical Decision Support System to Guide Rhythm Management of Atrial Fibrillation: Survey Study

Background

Despite the numerous studies evaluating various rhythm control strategies for atrial fibrillation (AF), determination of the optimal strategy in a single patient is often based on trial and error, with no one-size-fits-all approach based on international guidelines/recommendations. The decision, therefore, remains personal and lends itself well to help from a clinical decision support system, specifically one guided by artificial intelligence (AI). QRhythm utilizes a 2-stage machine learning (ML) model to identify the optimal rhythm management strategy in a given patient based on a set of clinical factors, in which the model first uses supervised learning to predict the actions of an expert clinician and identifies the best strategy through reinforcement learning to obtain the best clinical outcome—a composite of symptomatic recurrence, hospitalization, and stroke.

Objective

We qualitatively evaluated a novel, AI-based, clinical decision support system (CDSS) for AF rhythm management, called QRhythm, which uses both supervised and reinforcement learning to recommend either a rate control or one of 3 types of rhythm control strategies—external cardioversion, antiarrhythmic medication, or ablation—based on individual patient characteristics.

Methods

Thirty-three clinicians, including cardiology attendings and fellows and internal medicine attendings and residents, performed an assessment of QRhythm, followed by a survey to assess relative comfort with automated CDSS in rhythm management and to examine areas for future development.

Results

The 33 providers were surveyed with training levels ranging from resident to fellow to attending. Of the characteristics of the app surveyed, safety was most important to providers, with an average importance rating of 4.7 out of 5 (SD 0.72). This priority was followed by clinical integrity (a desire for the advice provided to make clinical sense; importance rating 4.5, SD 0.9), backward interpretability (transparency in the population used to create the algorithm; importance rating 4.3, SD 0.65), transparency of the algorithm (reasoning underlying the decisions made; importance rating 4.3, SD 0.88), and provider autonomy (the ability to challenge the decisions made by the model; importance rating 3.85, SD 0.83). Providers who used the app ranked the integrity of recommendations as their highest concern with ongoing clinical use of the model, followed by efficacy of the application and patient data security. Trust in the app varied; 1 (17%) provider responded that they somewhat disagreed with the statement, “I trust the recommendations provided by the QRhythm app,” 2 (33%) providers responded with neutrality to the statement, and 3 (50%) somewhat agreed with the statement.

Conclusions

Safety of ML applications was the highest priority of the providers surveyed, and trust of such models remains varied. Widespread clinical acceptance of ML in health care is dependent on how much providers trust the algorithms. Building this trust involves ensuring transparency and interpretability of the model.

Analysis of Prognostic Factors of Rectal Cancer and Construction of a Prognostic Prediction Model Based on Bayesian Network

Background

The existing prognostic models of rectal cancer after radical resection ignored the relationships among prognostic factors and their mutual effects on prognosis. Thus, a new modeling method is required to remedy this defect. The present study aimed to construct a new prognostic prediction model based on the Bayesian network (BN), a machine learning tool for data mining, clinical decision-making, and prognostic prediction.

Methods

From January 2015 to December 2017, the clinical data of 705 patients with rectal cancer who underwent radical resection were analyzed. The entire cohort was divided into training and testing datasets. A new prognostic prediction model based on BN was constructed and compared with a nomogram.

Results

A univariate analysis showed that age, Carcinoembryonic antigen (CEA), Carbohydrate antigen19-9 (CA19-9), Carbohydrate antigen 125 (CA125), preoperative chemotherapy, macropathology type, tumor size, differentiation status, T stage, N stage, vascular invasion, KRAS mutation, and postoperative chemotherapy were associated with overall survival (OS) of the training dataset. Based on the above-mentioned variables, a 3-year OS prognostic prediction BN model of the training dataset was constructed using the Tree Augmented Naïve Bayes method. In addition, age, CEA, CA19-9, CA125, differentiation status, T stage, N stage, KRAS mutation, and postoperative chemotherapy were identified as independent prognostic factors of the training dataset through multivariate Cox regression and were used to construct a nomogram. Then, based on the testing dataset, the two models were evaluated using the receiver operating characteristic (ROC) curve. The results showed that the area under the curve (AUC) of ROC of the BN model and nomogram was 80.11 and 74.23%, respectively.

Conclusion

The present study established a BN model for prognostic prediction of rectal cancer for the first time, which was demonstrated to be more accurate than a nomogram.

Comparisons of Forecasting for Survival Outcome for Head and Neck Squamous Cell Carcinoma by using Machine Learning Models based on Multi-omics

Background: Machine learning methods showed excellent predictive ability in a wide range of fields. For the survival of head and neck squamous cell carcinoma (HNSC), its multi-omics influence is crucial. This study attempts to establish a variety of machine learning multi-omics models to predict the survival of HNSC and find the most suitable machine learning prediction method. Methods: The HNSC clinical data and multi-omics data were downloaded from the TCGA database. The important variables were screened by the LASSO algorithm. We used a total of 12 supervised machine learning models to predict the outcome of HNSC survival and compared the results. In vitro qPCR was performed to verify core genes predicted by the random forest algorithm. Results: For omics of HNSC, the results of the twelve models showed that the performance of multi-omics was better than each single-omic alone. Results were presented, which showed that the Bayesian network(BN) model (area under the curve [AUC] 0.8250, F1 score=0.7917) and random forest(RF) model (area under the curve [AUC] 0.8002,F1 score=0.7839) played good prediction performance in HNSC multi-omics data. The results of in vitro qPCR were consistent with the RF algorithm. Conclusion: Machine learning methods could better forecast the survival outcome of HNSC. Meanwhile, this study found that the BN model and the RF model were the most superior. Moreover, the forecast result of multi-omics was better than single-omic alone in HNSC.

Data mining and mathematical models in cancer prognosis and prediction

Cancer is a fetal and complex disease. Individual differences of the same cancer type or the same patient at different stages of cancer development may require distinct treatments. Pathological differences are reflected in tissues, cells and gene levels etc. The interactions between the cancer cells and nearby microenvironments can also influence the cancer progression and metastasis. It is a huge challenge to understand all of these mechanistically and quantitatively. Researchers applied pattern recognition algorithms such as machine learning or data mining to predict cancer types or classifications. With the rapidly growing and available computing powers, researchers begin to integrate huge data sets, multi-dimensional data types and information. The cells are controlled by the gene expressions determined by the promoter sequences and transcription regulators. For example, the changes in the gene expression through these underlying mechanisms can modify cell progressing in the cell-cycle. Such molecular activities can be governed by the gene regulations through the underlying gene regulatory networks, which are essential for cancer study when the information and gene regulations are clear and available. In this review, we briefly introduce several machine learning methods of cancer prediction and classification which include Artificial Neural Networks (ANNs), Decision Trees (DTs), Support Vector Machine (SVM) and naive Bayes. Then we describe a few typical models for building up gene regulatory networks such as Correlation, Regression and Bayes methods based on available data. These methods can help on cancer diagnosis such as susceptibility, recurrence, survival etc. At last, we summarize and compare the modeling methods to analyze the development and progression of cancer through gene regulatory networks. These models can provide possible physical strategies to analyze cancer progression in a systematic and quantitative way.

Predicting Colorectal Cancer Recurrence and Patient Survival Using Supervised Machine Learning Approach: A South African Population-Based Study

Background: South Africa (SA) has the highest incidence of colorectal cancer (CRC) in Sub-Saharan Africa (SSA). However, there is limited research on CRC recurrence and survival in SA. CRC recurrence and overall survival are highly variable across studies. Accurate prediction of patients at risk can enhance clinical expectations and decisions within the South African CRC patients population. We explored the feasibility of integrating statistical and machine learning (ML) algorithms to achieve higher predictive performance and interpretability in findings.

Methods: We selected and compared six algorithms:- logistic regression (LR), naïve Bayes (NB), C5.0, random forest (RF), support vector machine (SVM) and artificial neural network (ANN). Commonly selected features based on OneR and information gain, within 10-fold cross-validation, were used for model development. The validity and stability of the predictive models were further assessed using simulated datasets.

Results: The six algorithms achieved high discriminative accuracies (AUC-ROC). ANN achieved the highest AUC-ROC for recurrence (87.0%) and survival (82.0%), and other models showed comparable performance with ANN. We observed no statistical difference in the performance of the models. Features including radiological stage and patient's age, histology, and race are risk factors of CRC recurrence and patient survival, respectively.

Conclusions: Based on other studies and what is known in the field, we have affirmed important predictive factors for recurrence and survival using rigorous procedures. Outcomes of this study can be generalised to CRC patient population elsewhere in SA and other SSA countries with similar patient profiles.

Machine Learning Model for Predicting Postoperative Survival of Patients with Colorectal Cancer

Purpose

Machine learning (ML) is a strong candidate for making accurate predictions, as we can use large amount of data with powerful computational algorithms. We developed a ML based model to predict survival of patients with colorectal cancer (CRC) using data from two independent datasets.

Materials and Methods

A total of 364,316 and 1,572 CRC patients were included from the Surveillance, Epidemiology, and End Results (SEER) and a Korean dataset, respectively. As SEER combines data from 18 cancer registries, internal validation was done using 18-Fold-Cross-Validation then external validation was performed by testing the trained model on the Korean dataset. Performance was evaluated using area under the receiver operating characteristic curve (AUROC), sensitivity and positive predictive values.

Results

Clinicopathological characteristics were significantly different between the two datasets and the SEER showed a significant lower 5-year survival rate compared to the Korean dataset (60.1% vs. 75.3%, p < 0.001). The ML-based model using the Light gradient boosting algorithm achieved a better performance in predicting 5-year-survival compared to American Joint Committee on Cancer stage (AUROC, 0.804 vs. 0.736; p < 0.001). The most important features which influenced model performance were age, number of examined lymph nodes, and tumor size. Sensitivity and positive predictive values of predicting 5-year-survival for classes including dead or alive were reported as 68.14%, 77.51% and 49.88%, 88.1% respectively in the validation set. Survival probability can be checked using the web-based survival predictor (http://colorectalcancer.pythonanywhere.com).

Conclusion

ML-based model achieved a much better performance compared to staging in individualized estimation of survival of patients with CRC.

A Prognostic Nomogram of Colon Cancer With Liver Metastasis: A Study of the US SEER Database and a Chinese Cohort

Background

Among colon cancer patients, liver metastasis is a commonly deadly phenomenon, but there are few prognostic models for these patients.

Methods

The clinicopathologic data of colon cancer with liver metastasis (CCLM) patients were downloaded from the Surveillance, Epidemiology and End Results (SEER) database. All patients were randomly divided into training and internal validation sets based on the ratio of 7:3. A prognostic nomogram was established with Cox analysis in the training set, which was validated by two independent validation sets.

Results

A total of 5,700 CCLM patients were included. Age, race, tumor size, tumor site, histological type, grade, AJCC N status, carcinoembryonic antigen (CEA), lung metastasis, bone metastasis, surgery, and chemotherapy were independently associated with the overall survival (OS) of CCLM in the training set, which were used to establish a nomogram. The AUCs of 1-, 2- and 3-year were higher than or equal to 0.700 in the training, internal validation, and external validation sets, indicating the favorable effects of our nomogram. Besides, whether in overall or subgroup analysis, the risk score calculated by this nomogram can divide CCLM patients into high-, middle- and low-risk groups, which suggested that the nomogram can significantly determine patients with different prognosis and is suitable for different patients.

Conclusion

Higher age, the race of black, larger tumor size, higher grade, histological type of mucinous adenocarcinoma and signet ring cell carcinoma, higher N stage, RCC, lung metastasis, bone metastasis, without surgery, without chemotherapy, and elevated CEA were independently associated with poor prognosis of CCLM patients. A nomogram incorporating the above variables could accurately predict the prognosis of CCLM.

PI Prob: A risk prediction and clinical guidance system for evaluating patients with recurrent infections

Background

Primary immunodeficiency diseases represent an expanding set of heterogeneous conditions which are difficult to recognize clinically. Diagnostic rates outside of the newborn period have not changed appreciably. This concern underscores a need for novel methods of disease detection.

Objective

We built a Bayesian network to provide real-time risk assessment about primary immunodeficiency and to facilitate prescriptive analytics for initiating the most appropriate diagnostic work up. Our goal is to improve diagnostic rates for primary immunodeficiency and shorten time to diagnosis. We aimed to use readily available health record data and a small training dataset to prove utility in diagnosing patients with relatively rare features.

Methods

We extracted data from the Texas Children’s Hospital electronic health record on a large population of primary immunodeficiency patients (n = 1762) and appropriately-matched set of controls (n = 1698). From the cohorts, clinically relevant prior probabilities were calculated enabling construction of a Bayesian network probabilistic model(PI Prob). Our model was constructed with clinical-immunology domain expertise, trained on a balanced cohort of 100 cases-controls and validated on an unseen balanced cohort of 150 cases-controls. Performance was measured by area under the receiver operator characteristic curve (AUROC). We also compared our network performance to classic machine learning model performance on the same dataset.

Results

PI Prob was accurate in classifying immunodeficiency patients from controls (AUROC = 0.945; p<0.0001) at a risk threshold of ≥6%. Additionally, the model was 89% accurate for categorizing validation cohort members into appropriate International Union of Immunological Societies diagnostic categories. Our network outperformed 3 other machine learning models and provides superior transparency with a prescriptive output element.

Conclusion

Artificial intelligence methods can classify risk for primary immunodeficiency and guide management. PI Prob enables accurate, objective decision making about risk and guides the user towards the appropriate diagnostic evaluation for patients with recurrent infections. Probabilistic models can be trained with small datasets underscoring their utility for rare disease detection given appropriate domain expertise for feature selection and network construction.

Survival prediction for gallbladder carcinoma after curative resection: Comparison of nomogram and Bayesian network models

BACKGROUND

In this study, we developed a nomogram and a Bayesian network (BN) model for prediction of survival in gallbladder carcinoma (GBC) patients following surgery and compared the performance of the two models.

METHODS

Survival prediction models were established and validated using data from 698 patients with GBC who underwent curative-intent resection between 2008 and 2017 at one of six Chinese tertiary hospitals. Model construction and internal validation were performed using data from 381 patients at one hepatobiliary center, and external validation was then performed using data from 317 patients at the other five centers. A BN model and a nomogram model were constructed based on the independent prognostic variables. Performance of the BN and nomogram models was compared based on area under receiver operating characteristic curves (AUC), model accuracy, and a confusion matrix.

RESULTS

Independent prognostic variables included age, pathological grade, liver infiltration, T stage, N stage, and margin. In internal validation, AUC was 84.14% and 78.22% for the BN and nomogram, respectively, and model accuracy was 75.65% and 72.17%, respectively. In external validation, AUC was 76.46% and 70.19% for the BN and nomogram, respectively, with model accuracy of 66.88% and 60.25%, respectively. Based on the confusion matrix, the nomogram had a higher true positive rate but a substantially lower true negative rate compared to the BN.

CONCLUSION

A BN model was more accurate than a Cox regression-based nomogram for prediction of survival in GBC patients undergoing curative-intent resection.

Time-dependent Estimates of Recurrence and Survival in Colon Cancer: Clinical Decision Support System Tool Development for Adjuvant Therapy and Oncological Outcome Assessment

Unanswered questions remain in determining which high-risk node-negative colon cancer (CC) cohorts benefit from adjuvant therapy and how it may differ in an equal access population. Machine-learned Bayesian Belief Networks (ml-BBNs) accurately estimate outcomes in CC, providing clinicians with Clinical Decision Support System (CDSS) tools to facilitate treatment planning. We evaluated ml-BBNs ability to estimate survival and recurrence in CC. We performed a retrospective analysis of registry data of patients with CC to train–test–crossvalidate ml-BBNs using the Department of Defense Automated Central Tumor Registry (January 1993 to December 2004). Cases with events or follow-up that passed quality control were stratified into 1-, 2-, 3-, and 5-year survival cohorts. ml-BBNs were trained using machine-learning algorithms and k-fold crossvalidation and receiver operating characteristic curve analysis used for validation. BBNs were comprised of 5301 patients and areas under the curve ranged from 0.85 to 0.90. Positive predictive values for recurrence and mortality ranged from 78 to 84 per cent and negative predictive values from 74 to 90 per cent by survival cohort. In the 12-month model alone, 1,132,462,080 unique rule sets allow physicians to predict individual recurrence/mortality estimates. Patients with Stage II (N0M0) CC benefit from chemotherapy at different rates. At one year, all patients older than 73 years of age with T2–4 tumors and abnormal carcinoembryonic antigen levels benefited, whereas at five years, all had relative reduction in mortality with the largest benefit amongst elderly, highest T-stage patients. ml-BBN can readily predict which high-risk patients benefit from adjuvant therapy. CDSS tools yield individualized, clinically relevant estimates of outcomes to assist clinicians in treatment planning.

Clinical and Statistical Validation of a Probabilistic Prediction Tool of Total Knee Arthroplasty Outcome

BACKGROUND

Predicting patients at risk of a poor outcome would be useful in patient selection for total knee arthroplasty (TKA). Existing models to predict outcome have seen limited functional implementation. This study aims to validate a model and shared decision-making tool for both clinical utility and predictive accuracy.

METHODS

A Bayesian belief network statistical model was developed using data from the Osteoarthritis Initiative. A consecutive series of consultations for osteoarthritis before and after introduction of the tool was used to evaluate the clinical impact of the tool. A data audit of postoperative outcomes of TKA patients exposed to the tool was used to evaluate the accuracy of predictions.

RESULTS

The tool changed consultation outcomes and identified patients at risk of limited improvement. After introduction of the tool, patients booked for surgery reported worse Knee Osteoarthritis and Injury Outcome Score pain scores (difference, 15.2; P < .001) than those not booked, with no significant difference prior. There was a 27% chance of not improving if predicted at risk, and a 1.4% chance if predicted to improve. This gives a risk ratio of 19× (P < .001) for patients not improving if predicted at risk.

CONCLUSION

For a prediction tool to be clinically useful, it needs to provide a better understanding of the likely clinical outcome of an intervention than existed without its use when the clinical decisions are made. The tool presented here has the potential to direct patients to surgical or nonsurgical pathways on a patient-specific basis, ensuring patients who will benefit most from TKA surgery are selected.

Estimating survival benefit of adjuvant therapy based on a Bayesian network prediction model in curatively resected advanced gallbladder adenocarcinoma

BACKGROUND

The factors affecting the prognosis and role of adjuvant therapy in advanced gallbladder carcinoma (GBC) after curative resection remain unclear.

AIM

To provide a survival prediction model to patients with GBC as well as to identify the role of adjuvant therapy.

METHODS

Patients with curatively resected advanced gallbladder adenocarcinoma (T3 and T4) were selected from the Surveillance, Epidemiology, and End Results database between 2004 and 2015. A survival prediction model based on Bayesian network (BN) was constructed using the tree-augmented naïve Bayes algorithm, and composite importance measures were applied to rank the influence of factors on survival. The dataset was divided into a training dataset to establish the BN model and a testing dataset to test the model randomly at a ratio of 7:3. The confusion matrix and receiver operating characteristic curve were used to evaluate the model accuracy.

RESULTS

A total of 818 patients met the inclusion criteria. The median survival time was 9.0 mo. The accuracy of BN model was 69.67%, and the area under the curve value for the testing dataset was 77.72%. Adjuvant radiation, adjuvant chemotherapy (CTx), T stage, scope of regional lymph node surgery, and radiation sequence were ranked as the top five prognostic factors. A survival prediction table was established based on T stage, N stage, adjuvant radiotherapy (XRT), and CTx. The distribution of the survival time (>9.0 mo) was affected by different treatments with the order of adjuvant chemoradiotherapy (cXRT) > adjuvant radiation > adjuvant chemotherapy > surgery alone. For patients with node-positive disease, the larger benefit predicted by the model is adjuvant chemoradiotherapy. The survival analysis showed that there was a significant difference among the different adjuvant therapy groups (log rank, surgery alone vs CTx, P < 0.001; surgery alone vs XRT, P = 0.014; surgery alone vs cXRT, P < 0.001).

CONCLUSION

The BN-based survival prediction model can be used as a decision-making support tool for advanced GBC patients. Adjuvant chemoradiotherapy is expected to improve the survival significantly for patients with node-positive disease.

Balancing accuracy and interpretability of machine learning approaches for radiation treatment outcomes modeling

Radiation outcomes prediction (ROP) plays an important role in personalized prescription and adaptive radiotherapy. A clinical decision may not only depend on an accurate radiation outcomes’ prediction, but also needs to be made based on an informed understanding of the relationship among patients’ characteristics, radiation response and treatment plans. As more patients’ biophysical information become available, machine learning (ML) techniques will have a great potential for improving ROP. Creating explainable ML methods is an ultimate task for clinical practice but remains a challenging one. Towards complete explainability, the interpretability of ML approaches needs to be first explored. Hence, this review focuses on the application of ML techniques for clinical adoption in radiation oncology by balancing accuracy with interpretability of the predictive model of interest. An ML algorithm can be generally classified into an interpretable (IP) or non-interpretable (NIP) (“black box”) technique. While the former may provide a clearer explanation to aid clinical decision-making, its prediction performance is generally outperformed by the latter. Therefore, great efforts and resources have been dedicated towards balancing the accuracy and the interpretability of ML approaches in ROP, but more still needs to be done. In this review, current progress to increase the accuracy for IP ML approaches is introduced, and major trends to improve the interpretability and alleviate the “black box” stigma of ML in radiation outcomes modeling are summarized. Efforts to integrate IP and NIP ML approaches to produce predictive models with higher accuracy and interpretability for ROP are also discussed.

Mining patterns of comorbidity evolution in patients with multiple chronic conditions using unsupervised multi-level temporal Bayesian network

Over the past few decades, the rise of multiple chronic conditions has become a major concern for clinicians. However, it is still not known precisely how multiple chronic conditions emerge among patients. We propose an unsupervised multi-level temporal Bayesian network to provide a compact representation of the relationship among emergence of multiple chronic conditions and patient level risk factors over time. To improve the efficiency of the learning process, we use an extension of maximum weight spanning tree algorithm and greedy search algorithm to study the structure of the proposed network in three stages, starting with learning the inter-relationship of comorbidities within each year, followed by learning the intra-relationship of comorbidity emergence between consecutive years, and finally learning the hierarchical relationship of comorbidities and patient level risk factors. We also use a longest path algorithm to identify the most likely sequence of comorbidities emerging from and/or leading to specific chronic conditions. Using a de-identified dataset of more than 250,000 patients receiving care from the U.S. Department of Veterans Affairs for a period of five years, we compare the performance of the proposed unsupervised Bayesian network in comparison with those of Bayesian networks developed based on supervised and semi-supervised learning approaches, as well as multivariate probit regression, multinomial logistic regression, and latent regression Markov mixture clustering focusing on traumatic brain injury (TBI), post-traumatic stress disorder (PTSD), depression (Depr), substance abuse (SuAb), and back pain (BaPa). Our findings show that the unsupervised approach has noticeably accurate predictive performance that is comparable to the best performing semi-supervised and the second-best performing supervised approaches. These findings also revealed that the unsupervised approach has improved performance over multivariate probit regression, multinomial logistic regression, and latent regression Markov mixture clustering.

Predictors of one and two years' mortality in patients with colon cancer: A prospective cohort study

Background

Tools to aid in the prognosis assessment of colon cancer patients in terms of risk of mortality are needed. Goals of this study are to develop and validate clinical prediction rules for 1- and 2-year mortality in these patients.

Methods

This is a prospective cohort study of patients diagnosed with colon cancer who underwent surgery at 22 hospitals. The main outcomes were mortality at 1 and 2 years after surgery. Background, clinical parameters, and diagnostic tests findings were evaluated as possible predictors. Multivariable multilevel logistic regression and survival models were used in the analyses to create the clinical prediction rules. Models developed in the derivation sample were validated in another sample of the study.

Results

American Society of Anesthesiologists Physical Status Classification System (ASA), Charlson comorbidity index (> = 4), age (>75 years), residual tumor (R2), TNM stage IV and log of lymph nodes ratio (> = -0.53) were predictors of 1-year mortality (C-index (95% CI): 0.865 (0.792–0.938)). Adjuvant chemotherapy was an additional predictor. Again ASA, Charlson Index (> = 4), age (>75 years), log of lymph nodes ratio (> = -0.53), TNM, and residual tumor were predictors of 2-year mortality (C-index:0.821 (0.766–0.876). Chemotherapy was also an additional predictor.

Conclusions

These clinical prediction rules show very good predictive abilities of one and two years survival and provide clinicians and patients with an easy and quick-to-use decision tool for use in the clinical decision process while the patient is still in the index admission.

Development of a Fully Cross-Validated Bayesian Network Approach for Local Control Prediction in Lung Cancer

The purpose of this study is to demonstrate that a Bayesian network (BN) approach can explore hierarchical biophysical relationships that influence tumor response and predict tumor local control (LC) in non-small-cell lung cancer (NSCLC) patients before and during radiotherapy from a large-scale dataset. Our BN building approach has two steps. First, relevant biophysical predictors influencing LC before and during the treatment are selected through an extended Markov blanket (eMB) method. From this eMB process, the most robust BN structure for LC prediction was found via a wrapper-based approach. Sixty-eight patients with complete feature information were used to identify a full BN model for LC prediction before and during the treatment. Fifty more recent patients with some missing information were reserved for independent testing of the developed pre- and during-therapy BNs. A nested cross-validation (N-CV) was developed to evaluate the performance of the two-step BN approach. An ensemble BN model is generated from the N-CV sampling process to assess its similarity with the corresponding full BN model, and thus evaluate the sensitivity of our BN approach. Our results show that the proposed BN development approach is a stable and robust approach to identify hierarchical relationships among biophysical features for LC prediction. Furthermore, BN predictions can be improved by incorporating during treatment information.

Combining statistical techniques to predict postsurgical risk of 1-year mortality for patients with colon cancer

Introduction

Colorectal cancer is one of the most frequently diagnosed malignancies and a common cause of cancer-related mortality. The aim of this study was to develop and validate a clinical predictive model for 1-year mortality among patients with colon cancer who survive for at least 30 days after surgery.

Methods

Patients diagnosed with colon cancer who had surgery for the first time and who survived 30 days after the surgery were selected prospectively. The outcome was mortality within 1 year. Random forest, genetic algorithms and classification and regression trees were combined in order to identify the variables and partition points that optimally classify patients by risk of mortality. The resulting decision tree was categorized into four risk categories. Split-sample and bootstrap validation were performed. ClinicalTrials.gov Identifier: NCT02488161.

Results

A total of 1945 patients were enrolled in the study. The variables identified as the main predictors of 1-year mortality were presence of residual tumor, American Society of Anesthesiologists Physical Status Classification System risk score, pathologic tumor staging, Charlson Comorbidity Index, intraoperative complications, adjuvant chemotherapy and recurrence of tumor. The model was internally validated; area under the receiver operating characteristic curve (AUC) was 0.896 in the derivation sample and 0.835 in the validation sample. Risk categorization leads to AUC values of 0.875 and 0.832 in the derivation and validation samples, respectively. Optimal cut-off point of estimated risk had a sensitivity of 0.889 and a specificity of 0.758.

Conclusion

The decision tree was a simple, interpretable, valid and accurate prediction rule of 1-year mortality among colon cancer patients who survived for at least 30 days after surgery.

Survivability prediction of colon cancer patients using neural networks

We utilize deep neural networks to develop prediction models for patient survival and conditional survival of colon cancer. Our models are trained and validated on data obtained from the Surveillance, Epidemiology, and End Results Program. We provide an online outcome calculator for 1, 2, and 5 years survival periods. We experimented with multiple neural network structures and found that a network with five hidden layers produces the best results for these data. Moreover, the online outcome calculator provides conditional survival of 1, 2, and 5 years after surviving the mentioned survival periods. In this article, we report an approximate 0.87 area under the receiver operating characteristic curve measurements, higher than the 0.85 reported by Stojadinovic et al.

Comprehensible knowledge model creation for cancer treatment decision making

BACKGROUND

A wealth of clinical data exists in clinical documents in the form of electronic health records (EHRs). This data can be used for developing knowledge-based recommendation systems that can assist clinicians in clinical decision making and education. One of the big hurdles in developing such systems is the lack of automated mechanisms for knowledge acquisition to enable and educate clinicians in informed decision making.

MATERIALS AND METHODS

An automated knowledge acquisition methodology with a comprehensible knowledge model for cancer treatment (CKM-CT) is proposed. With the CKM-CT, clinical data are acquired automatically from documents. Quality of data is ensured by correcting errors and transforming various formats into a standard data format. Data preprocessing involves dimensionality reduction and missing value imputation. Predictive algorithm selection is performed on the basis of the ranking score of the weighted sum model. The knowledge builder prepares knowledge for knowledge-based services: clinical decisions and education support.

RESULTS

Data is acquired from 13,788 head and neck cancer (HNC) documents for 3447 patients, including 1526 patients of the oral cavity site. In the data quality task, 160 staging values are corrected. In the preprocessing task, 20 attributes and 106 records are eliminated from the dataset. The Classification and Regression Trees (CRT) algorithm is selected and provides 69.0% classification accuracy in predicting HNC treatment plans, consisting of 11 decision paths that yield 11 decision rules.

CONCLUSION

Our proposed methodology, CKM-CT, is helpful to find hidden knowledge in clinical documents. In CKM-CT, the prediction models are developed to assist and educate clinicians for informed decision making. The proposed methodology is generalizable to apply to data of other domains such as breast cancer with a similar objective to assist clinicians in decision making and education.

Prediction of lung cancer incidence on the low-dose computed tomography arm of the National Lung Screening Trial: A dynamic Bayesian network

INTRODUCTION

Identifying high-risk lung cancer individuals at an early disease stage is the most effective way of improving survival. The landmark National Lung Screening Trial (NLST) demonstrated the utility of low-dose computed tomography (LDCT) imaging to reduce mortality (relative to X-ray screening). As a result of the NLST and other studies, imaging-based lung cancer screening programs are now being implemented. However, LDCT interpretation results in a high number of false positives. A set of dynamic Bayesian networks (DBN) were designed and evaluated to provide insight into how longitudinal data can be used to help inform lung cancer screening decisions.

METHODS

The LDCT arm of the NLST dataset was used to build and explore five DBNs for high-risk individuals. Three of these DBNs were built using a backward construction process, and two using structure learning methods. All models employ demographics, smoking status, cancer history, family lung cancer history, exposure risk factors, comorbidities related to lung cancer, and LDCT screening outcome information. Given the uncertainty arising from lung cancer screening, a cancer state-space model based on lung cancer staging was utilized to characterize the cancer status of an individual over time. The models were evaluated on balanced training and test sets of cancer and non-cancer cases to deal with data imbalance and overfitting.

RESULTS

Results were comparable to expert decisions. The average area under the curve (AUC) of the receiver operating characteristic (ROC) for the three intervention points of the NLST trial was higher than 0.75 for all models. Evaluation of the models on the complete LDCT arm of the NLST dataset (N=25,486) demonstrated satisfactory generalization. Consensus of predictions over similar cases is reported in concordance statistics between the models' and the physicians' predictions. The models' predictive ability with respect to missing data was also evaluated with the sample of cases that missed the second screening exam of the trial (N=417). The DBNs outperformed comparison models such as logistic regression and naïve Bayes.

CONCLUSION

The lung cancer screening DBNs demonstrated high discrimination and predictive power with the majority of cancer and non-cancer cases.

Mining causal relationships among clinical variables for cancer diagnosis based on Bayesian analysis

BACKGROUND

Cancer is the second leading cause of death around the world after cardiovascular diseases. Over the past decades, various data mining studies have tried to predict the outcome of cancer. However, only a few reports describe the causal relationships among clinical variables or attributes, which may provide theoretical guidance for cancer diagnosis and therapy. Different restricted Bayesian classifiers have been used to discover information from numerous domains. This research work designed a novel Bayesian learning strategy to predict cause-specific death classes and proposed a graphical structure of key attributes to clarify the implicit relationships implicated in the data set.

RESULTS

The working mechanisms of 3 classical restricted Bayesian classifiers, namely, NB, TAN and KDB, were analysed and summarised. To retain the properties of global optimisation and high-order dependency representation, the proposed learning algorithm, i.e., flexible K-dependence Bayesian network (FKBN), applies the greedy search of conditional mutual information space to identify the globally optimal ordering of the attributes and to allow the classifiers to be constructed at arbitrary points (values of K) along the attribute dependence spectrum. This method represents the relationships between different attributes by using a directed acyclic graph (DAG) model. A total of 12 data sets were selected from the SEER database and KRBM repository by 10-fold cross-validation for evaluation purposes. The findings revealed that the FKBN model outperformed NB, TAN and KDB.

CONCLUSIONS

A Bayesian classifier can graphically describe the conditional dependency among attributes. The proposed algorithm offers a trade-off between probability estimation and network structure complexity. The direct and indirect relationships between the predictive attributes and class variable should be considered simultaneously to achieve global optimisation and high-order dependency representation. By analysing the DAG inferred from the breast cancer data set of the SEER database we divided the attributes into two subgroups, namely, key attributes that should be considered first for cancer diagnosis and those that are independent of each other but are closely related to key attributes. The statistical analysis results clarify some of the causal relationships implicated in the DAG.

Lung Cancer Assistant: a hybrid clinical decision support application for lung cancer care

Multidisciplinary team (MDT) meetings are becoming the model of care for cancer patients worldwide. While MDTs have improved the quality of cancer care, the meetings impose substantial time pressure on the members, who generally attend several such MDTs. We describe Lung Cancer Assistant (LCA), a clinical decision support (CDS) prototype designed to assist the experts in the treatment selection decisions in the lung cancer MDTs. A novel feature of LCA is its ability to provide rule-based and probabilistic decision support within a single platform. The guideline-based CDS is based on clinical guideline rules, while the probabilistic CDS is based on a Bayesian network trained on the English Lung Cancer Audit Database (LUCADA). We assess rule-based and probabilistic recommendations based on their concordances with the treatments recorded in LUCADA. Our results reveal that the guideline rule-based recommendations perform well in simulating the recorded treatments with exact and partial concordance rates of 0.57 and 0.79, respectively. On the other hand, the exact and partial concordance rates achieved with probabilistic results are relatively poorer with 0.27 and 0.76. However, probabilistic decision support fulfils a complementary role in providing accurate survival estimations. Compared to recorded treatments, both CDS approaches promote higher resection rates and multimodality treatments.

Computer-assisted abdominal surgery: new technologies

BACKGROUND

Computer-assisted surgery is a wide field of technologies with the potential to enable the surgeon to improve efficiency and efficacy of diagnosis, treatment, and clinical management.

PURPOSE

This review provides an overview of the most important new technologies and their applications.

METHODS

A MEDLINE database search was performed revealing a total of 1702 references. All references were considered for information on six main topics, namely image guidance and navigation, robot-assisted surgery, human-machine interface, surgical processes and clinical pathways, computer-assisted surgical training, and clinical decision support. Further references were obtained through cross-referencing the bibliography cited in each work. Based on their respective field of expertise, the authors chose 64 publications relevant for the purpose of this review.

CONCLUSION

Computer-assisted systems are increasingly used not only in experimental studies but also in clinical studies. Although computer-assisted abdominal surgery is still in its infancy, the number of studies is constantly increasing, and clinical studies start showing the benefits of computers used not only as tools of documentation and accounting but also for directly assisting surgeons during diagnosis and treatment of patients. Further developments in the field of clinical decision support even have the potential of causing a paradigm shift in how patients are diagnosed and treated.

Future directions for the early detection of colorectal cancer recurrence

Surgical resection remains a mainstay of treatment and is highly effective for localized colorectal cancer. However, ~30-40% of patients develop recurrence following surgery and 40-50% of recurrences are apparent within the first few years after initial surgical resection. Several variables factor into the ultimate outcome of these patients, including the extent of disease, tumor biology, and patient co-morbidities. Additionally, the time from initial treatment to the development of recurrence is strongly associated with overall survival, particularly in patients who recur within one year of their surgical resection. Current post-resection surveillance strategies involve physical examination, laboratory, endoscopic and imaging studies utilizing various high and low-intensity protocols. Ultimately, the goal is to detect recurrence as early as possible, and ideally in the asymptomatic localized phase, to allow initiation of treatment that may still result in cure. While current strategies have been effective, several efforts are evolving to improve our ability to identify recurrent disease at its earliest phase. Our aim with this article is to briefly review the options available and, more importantly, examine emerging and future options to assist in the early detection of colon and rectal cancer recurrence.

Bayesian networks for clinical decision support in lung cancer care

Survival prediction and treatment selection in lung cancer care are characterised by high levels of uncertainty. Bayesian Networks (BNs), which naturally reason with uncertain domain knowledge, can be applied to aid lung cancer experts by providing personalised survival estimates and treatment selection recommendations. Based on the English Lung Cancer Database (LUCADA), we evaluate the feasibility of BNs for these two tasks, while comparing the performances of various causal discovery approaches to uncover the most feasible network structure from expert knowledge and data. We show first that the BN structure elicited from clinicians achieves a disappointing area under the ROC curve of 0.75 (± 0.03), whereas a structure learned by the CAMML hybrid causal discovery algorithm, which adheres with the temporal restrictions, achieves 0.81 (± 0.03). Second, our causal intervention results reveal that BN treatment recommendations, based on prescribing the treatment plan that maximises survival, can only predict the recorded treatment plan 29% of the time. However, this percentage rises to 76% when partial matches are included.

Patients at risk for peritoneal surface malignancy of colorectal cancer origin: the role of second look laparotomy

Peritoneal surface malignancy (PSM) is a frequent occurrence in the natural history of colorectal cancer (CRC). Although significant advances have been made in screening of CRC, similar progress has yet to be made in the early detection of PSM of colorectal cancer origin. The fact that advanced CRC can be confined to the peritoneal surface without distant dissemination forms the basis for aggressive multi-modality therapy consisting of cytoreductive surgery (CRS) plus hyperthermic intra-peritoneal chemotherapy (HIPEC), and neoadjuvant and/or adjuvant systemic therapy. Reported overall survival with complete CRS+HIPEC exceeds that of systemic therapy alone for the treatment of PSM from CRC, underscoring the advantage of this multi-modality therapeutic approach. Patients with limited peritoneal disease from CRC can undergo complete cytoreduction, which is associated with the best reported outcomes. As early or limited peritoneal carcinomatosis is undetectable by conventional imaging modalities, second look laparotomy is an important means to identify disease in high-risk patients at a stage most amenable to complete cytoreduction. This review focuses on the identification of patients at risk for PSM from CRC and discusses the role of second look laparotomy.

Evidence-based Guidelines for Precision Risk Stratification-Based Screening (PRSBS) for Colorectal Cancer: Lessons learned from the US Armed Forces: Consensus and Future Directions

Colorectal cancer (CRC) is the third most common cause of cancer-related death in the United States (U.S.), with estimates of 143,460 new cases and 51,690 deaths for the year 2012. Numerous organizations have published guidelines for CRC screening; however, these numerical estimates of incidence and disease-specific mortality have remained stable from years prior. Technological, genetic profiling, molecular and surgical advances in our modern era should allow us to improve risk stratification of patients with CRC and identify those who may benefit from preventive measures, early aggressive treatment, alternative treatment strategies, and/or frequent surveillance for the early detection of disease recurrence. To better negotiate future economic constraints and enhance patient outcomes, ultimately, we propose to apply the principals of personalized and precise cancer care to risk-stratify patients for CRC screening (Precision Risk Stratification-Based Screening, PRSBS). We believe that genetic, molecular, ethnic and socioeconomic disparities impact oncological outcomes in general, those related to CRC, in particular. This document highlights evidence-based screening recommendations and risk stratification methods in response to our CRC working group private-public consensus meeting held in March 2012. Our aim was to address how we could improve CRC risk stratification-based screening, and to provide a vision for the future to achieving superior survival rates for patients diagnosed with CRC.