A Novel Prognostic Scoring System of Intrahepatic Cholangiocarcinoma With Machine Learning Basing on Real-World Data

Application of AI on cholangiocarcinoma

Cholangiocarcinoma, classified as intrahepatic, perihilar, and extrahepatic, is considered a deadly malignancy of the hepatobiliary system. Most cases of cholangiocarcinoma are asymptomatic. Therefore, early detection of cholangiocarcinoma is significant but still challenging. The routine screening of a tumor lacks specificity and accuracy. With the application of AI, high-risk patients can be easily found by analyzing their clinical characteristics, serum biomarkers, and medical images. Moreover, AI can be used to predict the prognosis including recurrence risk and metastasis. Although they have some limitations, AI algorithms will still significantly improve many aspects of cholangiocarcinoma in the medical field with the development of computing power and technology.

Deep Learning for Automatic Diagnosis and Morphologic Characterization of Malignant Biliary Strictures Using Digital Cholangioscopy: A Multicentric Study

Digital single-operator cholangioscopy (D-SOC) has enhanced the ability to diagnose indeterminate biliary strictures (BSs). Pilot studies using artificial intelligence (AI) models in D-SOC demonstrated promising results. Our group aimed to develop a convolutional neural network (CNN) for the identification and morphological characterization of malignant BSs in D-SOC. A total of 84,994 images from 129 D-SOC exams in two centers (Portugal and Spain) were used for developing the CNN. Each image was categorized as either a normal/benign finding or as malignant lesion (the latter dependent on histopathological results). Additionally, the CNN was evaluated for the detection of morphologic features, including tumor vessels and papillary projections. The complete dataset was divided into training and validation datasets. The model was evaluated through its sensitivity, specificity, positive and negative predictive values, accuracy and area under the receiver-operating characteristic and precision-recall curves (AUROC and AUPRC, respectively). The model achieved a 82.9% overall accuracy, 83.5% sensitivity and 82.4% specificity, with an AUROC and AUPRC of 0.92 and 0.93, respectively. The developed CNN successfully distinguished benign findings from malignant BSs. The development and application of AI tools to D-SOC has the potential to significantly augment the diagnostic yield of this exam for identifying malignant strictures.

Development and Validation of a Machine-Learning Model to Predict Early Recurrence of Intrahepatic Cholangiocarcinoma

BACKGROUND

The high incidence of early recurrence after hepatectomy for intrahepatic cholangiocarcinoma (ICC) has a detrimental effect on overall survival (OS). Machine-learning models may improve the accuracy of outcome prediction for malignancies.

METHODS

Patients who underwent curative-intent hepatectomy for ICC were identified using an international database. Three machine-learning models were trained to predict early recurrence (< 12 months after hepatectomy) using 14 clinicopathologic characteristics. The area under the receiver operating curve (AUC) was used to assess their discrimination ability.

RESULTS

In this study, 536 patients were randomly assigned to training (n = 376, 70.1%) and testing (n = 160, 29.9%) cohorts. Overall, 270 (50.4%) patients experienced early recurrence (training: n = 150 [50.3%] vs testing: n = 81 [50.6%]), with a median tumor burden score (TBS) of 5.6 (training: 5.8 [interquartile range {IQR}, 4.1-8.1] vs testing: 5.5 [IQR, 3.7-7.9]) and metastatic/undetermined nodes (N1/NX) in the majority of the patients (training: n = 282 [75.0%] vs testing n = 118 [73.8%]). Among the three different machine-learning algorithms, random forest (RF) demonstrated the highest discrimination in the training/testing cohorts (RF [AUC, 0.904/0.779] vs support vector machine [AUC, 0.671/0.746] vs logistic regression [AUC, 0.668/0.745]). The five most influential variables in the final model were TBS, perineural invasion, microvascular invasion, CA 19-9 lower than 200 U/mL, and N1/NX disease. The RF model successfully stratified OS relative to the risk of early recurrence.

CONCLUSIONS

Machine-learning prediction of early recurrence after ICC resection may inform tailored counseling, treatment, and recommendations. An easy-to-use calculator based on the RF model was developed and made available online.

Call to Improve the Quality of Prediction Tools for Intrahepatic Cholangiocarcinoma Resection: A Critical Appraisal, Systematic Review, and External Validation Study

Objective

To conduct a systematic review, critical appraisal, and external validation of survival prediction tools for patients undergoing intrahepatic cholangiocarcinoma (iCCA) resection.

Summary background data

Despite the development of several survival prediction tools in recent years for patients undergoing iCCA resections, there is a lack of critical appraisal and external validation of these models.

Methods

We conducted a systematic review and critical appraisal of survival and recurrence prediction models for patients undergoing curative-intent iCCA resections. Studies were evaluated based on their model design, risk of bias, reporting, performance, and validation results. We identified the best model and externally validated it using our institution's data.

Results

This review included a total of 31 studies, consisting of 26 studies with original prediction tools and 5 studies that only conducted external validations. Among the 26, 54% of the studies conducted internal validations, 46% conducted external validations, and only 1 study scored a low risk of bias. Harrell's C-statistics ranged from 0.67 to 0.76 for internal validation and from 0.64 to 0.75 for external validation. Only 81% of the studies reported model calibration. Our external validation of the best model (Intrahepatic Cholangiocarcinoma [ICC]-Metroticket) estimated Harrell's and Uno's C-statistics of 0.67 (95% CI: 0.56-0.77) and Uno's time-dependent area under the receiver operating characteristic curve (AUC) of 0.71 (95% CI: 0.53-0.88), with a Brier score of 0.20 (95% CI: 0.15-0.26) and good calibration plots.

Conclusions

Many prediction models have been published in recent years, but their quality remains poor, and minimal methodological quality improvement has been observed. The ICC-Metroticket was selected as the best model (Uno's time-dependent AUC of 0.71) for 5-year overall survival prediction in patients undergoing curative-intent iCCA resection.

Machine learning‑based prediction of survival prognosis in esophageal squamous cell carcinoma

The current prognostic tools for esophageal squamous cell carcinoma (ESCC) lack the necessary accuracy to facilitate individualized patient management strategies. To address this issue, this study was conducted to develop a machine learning (ML) prediction model for ESCC patients' survival management. Six ML approaches, including Rpart, Elastic Net, GBM, Random Forest, GLMboost, and the machine learning-extended CoxPH method, were employed to develop risk prediction models. The model was trained on a dataset of 1954 ESCC patients with 27 clinical features and validated on a dataset of 487 ESCC patients. The discriminative performance of the models was assessed using the concordance index (C-index). The best performing model was used for risk stratification and clinical evaluation. The study found that N stage, T stage, surgical margin, tumor grade, tumor length, sex, MPV, AST, FIB, and Mg are the important feature for ESCC patients' survival. The machine learning-extended CoxPH model, Elastic Net, and Random Forest had similar performance in predicting the mortality risk of ESCC patients, and outperformed GBM, GLMboost, and Rpart. The risk scores derived from the CoxPH model effectively stratified ESCC patients into low-, intermediate-, and high-risk groups with distinctly different 3-year overall survival (OS) probabilities of 80.8%, 58.2%, and 29.5%, respectively. This risk stratification was also observed in the validation cohort. Furthermore, the risk model demonstrated greater discriminative ability and net benefit than the AJCC8th stage, suggesting its potential as a prognostic tool for predicting survival events and guiding clinical decision-making. The classical algorithm of the CoxPH method was also found to be sufficiently good for interpretive studies.

Dynamic evaluation of postoperative survival in intrahepatic cholangiocarcinoma patients who did not undergo lymphadenectomy: a multicenter study

BACKGROUND

The prognosis of Intrahepatic cholangiocarcinoma (ICC) patients who did not undergo lymphadenectomy is difficult to assess. This study aims to have a dynamic evaluation on the postoperative survival of ICC patients by calculating conditional survival.

METHODS

Relevant data were from patients treated in 12 large-scale hospitals from December 2011 to December 2017. The influence of relevant clinical baseline data on the prognosis of ICC patients was analyzed by Cox regression. Conditional survival (CS) is a method that may predict the prognostic probability dynamically. For a patient with x years of survival, the 1-year CS (CS1) may be calculated as CS1= OS(x + 1)/OS(x).

RESULT

A total of 361 patients who met the criteria were included in the study. Conditional survival (CS) means that the patients' prognosis varies with survival time, meanwhile, relevant factors affecting the prognosis have a time-varying effect. The probability of survival assessed by CS1 increased year by year and the 1,2,3-year survival improved from 68.4% to 87.8%, while the postoperative actuarial OS decreased from 69.4% at 1 years to 36.9% at 3 years.

CONCLUSIONS

In terms of CS, the estimated survival for ICC varies with the increase of survival time after excision. Patients who live longer were likely to live longer. At the same time, with the passage of time, the role of the original adverse factors of the tumor would gradually decrease. Conditional survival allows a more accurate assessment of ICC patients who did not undergo lymphadenectomy.

Predicting Lapatinib Dose Regimen Using Machine Learning and Deep Learning Techniques Based on a Real-World Study

Lapatinib is used for the treatment of metastatic HER2(+) breast cancer. We aim to establish a prediction model for lapatinib dose using machine learning and deep learning techniques based on a real-world study. There were 149 breast cancer patients enrolled from July 2016 to June 2017 at Fudan University Shanghai Cancer Center. The sequential forward selection algorithm based on random forest was applied for variable selection. Twelve machine learning and deep learning algorithms were compared in terms of their predictive abilities (logistic regression, SVM, random forest, Adaboost, XGBoost, GBDT, LightGBM, CatBoost, TabNet, ANN, Super TML, and Wide&Deep). As a result, TabNet was chosen to construct the prediction model with the best performance (accuracy = 0.82 and AUC = 0.83). Afterward, four variables that strongly correlated with lapatinib dose were ranked via importance score as follows: treatment protocols, weight, number of chemotherapy treatments, and number of metastases. Finally, the confusion matrix was used to validate the model for a dose regimen of 1,250 mg lapatinib (precision = 81% and recall = 95%), and for a dose regimen of 1,000 mg lapatinib (precision = 87% and recall = 64%). To conclude, we established a deep learning model to predict lapatinib dose based on important influencing variables selected from real-world evidence, to achieve an optimal individualized dose regimen with good predictive performance.

Artificial intelligence and cholangiocarcinoma: Updates and prospects

Artificial intelligence (AI) is the timeliest field of computer science and attempts to mimic cognitive function of humans to solve problems. In the era of “Big data”, there is an ever-increasing need for AI in all aspects of medicine. Cholangiocarcinoma (CCA) is the second most common primary malignancy of liver that has shown an increase in incidence in the last years. CCA has high mortality as it is diagnosed in later stages that decreases effect of surgery, chemotherapy, and other modalities. With technological advancement there is an immense amount of clinicopathologic, genetic, serologic, histologic, and radiologic data that can be assimilated together by modern AI tools for diagnosis, treatment, and prognosis of CCA. The literature shows that in almost all cases AI models have the capacity to increase accuracy in diagnosis, treatment, and prognosis of CCA. Most studies however are retrospective, and one study failed to show AI benefit in practice. There is immense potential for AI in diagnosis, treatment, and prognosis of CCA however limitations such as relative lack of studies in use by human operators in improvement of survival remains to be seen.

Survival Prediction in Intrahepatic Cholangiocarcinoma: A Proof of Concept Study Using Artificial Intelligence for Risk Assessment

Several scoring systems have been devised to objectively predict survival for patients with intrahepatic cholangiocellular carcinoma (ICC) and support treatment stratification, but they have failed external validation. The aim of the present study was to improve prognostication using an artificial intelligence-based approach. We retrospectively identified 417 patients with ICC who were referred to our tertiary care center between 1997 and 2018. Of these, 293 met the inclusion criteria. Established risk factors served as input nodes for an artificial neural network (ANN). We compared the performance of the trained model to the most widely used conventional scoring system, the Fudan score. Predicting 1-year survival, the ANN reached an area under the ROC curve (AUC) of 0.89 for the training set and 0.80 for the validation set. The AUC of the Fudan score was significantly lower in the validation set (0.77, p < 0.001). In the training set, the Fudan score yielded a lower AUC (0.74) without reaching significance (p = 0.24). Thus, ANNs incorporating a multitude of known risk factors can outperform conventional risk scores, which typically consist of a limited number of parameters. In the future, such artificial intelligence-based approaches have the potential to improve treatment stratification when models trained on large multicenter data are openly available.