Survival Prediction in Gallbladder Cancer Using CT Based Machine Learning

Contrast-Enhanced CT-Based Deep Learning Radiomics Nomogram for the Survival Prediction in Gallbladder Cancer

RATIONALE AND OBJECTIVES

An accurate prognostic model is essential for the development of treatment strategies for gallbladder cancer (GBC). This study proposes an integrated model using clinical features, radiomics, and deep learning based on contrast-enhanced computed tomography (CT) images for survival prediction in patients with GBC after surgical resection.

METHODS

A total of 167 patients with GBC who underwent surgical resection at two medical institutions were retrospectively enrolled. After obtaining the pre-treatment CT images, the tumor lesions were manually segmented, and handcrafted radiomics features were extracted. A clinical prognostic signature and radiomics signature were built using machine learning algorithms based on the optimal clinical features or handcrafted radiomics features, respectively. Subsequently, a DenseNet121 model was employed for transfer learning on the radiomics image data and as the basis for the deep learning signature. Finally, we used logistic regression on the three signatures to obtain the unified multimodal model for comprehensive interpretation and analysis.

RESULTS

The integrated model performed better than the other models, exhibiting the highest area under the curve (AUC) of 0.870 in the test set, and the highest concordance index (C-index) of 0.736 in predicting patient survival rates. A Kaplan-Meier analysis demonstrated that patients in high-risk group had a lower survival probability compared to those in low-risk group (log-rank p < 0.05).

CONCLUSION

The nomogram is useful for predicting the survival of patients with GBC after surgical resection, helping in the identification of high-risk patients with poor prognosis and ultimately facilitating individualized management of patients with GBC.

Applications of artificial intelligence in biliary tract cancers

Biliary tract cancers are malignant neoplasms arising from bile duct epithelial cells. They include cholangiocarcinomas and gallbladder cancer. Gallbladder cancer has a marked geographical preference and is one of the most common cancers in women in northern India. Biliary tract cancers are usually diagnosed at an advanced, unresectable stage. Hence, the prognosis is extremely dismal. The five-year survival rate in advanced gallbladder cancer is < 5%. Hence, early detection and radical surgery are critical to improving biliary tract cancer prognoses. Radiological imaging plays an essential role in diagnosing and managing biliary tract cancers. However, the diagnosis is challenging because the biliary tract is affected by many diseases that may have radiological appearances similar to cancer. Artificial intelligence (AI) can improve radiologists' performance in various tasks. Deep learning (DL)-based approaches are increasingly incorporated into medical imaging to improve diagnostic performance. This paper reviews the AI-based strategies in biliary tract cancers to improve the diagnosis and prognosis.

The Many Hidden Faces of Gallbladder Carcinoma on CT and MRI Imaging-From A to Z

Gallbladder carcinoma represents the most aggressive biliary tract cancer and the sixth most common gastrointestinal malignancy. The diagnosis is a challenging clinical task due to its clinical presentation, which is often non-specific, mimicking a heterogeneous group of diseases, as well as benign processes such as complicated cholecystitis, xanthogranulomatous cholecystitis, adenomyomatosis, porcelain gallbladder or metastasis to the gallbladder (most frequently derived from melanoma, renal cell carcinoma). Risk factors include gallstones, carcinogen exposure, porcelain gallbladder, typhoid carrier state, gallbladder polyps and abnormal pancreaticobiliary ductal junction. Typical imaging features on CT or MRI reveal three major patterns: asymmetric focal or diffuse wall-thickening of the gallbladder, a solid mass that replaces the gallbladder and invades the adjacent organs or as an intraluminal enhancement mass arising predominantly from the gallbladder fundus. The tumor can spread to the liver, the adjacent internal organs and lymph nodes. Depending on the disease stage, surgical resection is the curative treatment option in early stages and adjuvant combination chemotherapy at advanced stages. The purpose of this scientific paper is to fully illustrate and evaluate, through multimodality imaging findings (CT and MRI), different presentations and imaging scenarios of gallbladder cancer in six patients and thoroughly analyze the risk factors, patterns of spread and differential diagnosis regarding each particular case.

Radiomics-based machine learning and deep learning to predict serosal involvement in gallbladder cancer

OBJECTIVE

Our study aimed to determine whether radiomics models based on contrast-enhanced computed tomography (CECT) have considerable ability to predict serosal involvement in gallbladder cancer (GBC) patients.

MATERIALS AND METHODS

A total of 152 patients diagnosed with GBC were retrospectively enrolled and divided into the serosal involvement group and no serosal involvement group according to paraffin pathology results. The regions of interest (ROIs) in the lesion on all CT images were drawn by two radiologists using ITK-SNAP software (version 3.8.0). A total of 412 features were extracted from the CT images of each patient. The Mann‒Whitney U test was applied to identify features with significant differences between groups. Seven machine learning algorithms and a deep learning model based on fully connected neural networks (f-CNNs) were used for radiomics model construction. The prediction efficacy of the models was evaluated using receiver operating characteristic (ROC) curve analysis.

RESULTS

Through the Mann‒Whitney U test, 75 of the 412 features extracted from the CT images of patients were significantly different between groups (P < 0.05). Among all the algorithms, logistic regression achieved the highest performance with an area under the curve (AUC) of 0.944 (sensitivity 0.889, specificity 0.8); the f-CNN deep learning model had an AUC of 0.916, and the model showed high predictive power for serosal involvement, with a sensitivity of 0.733 and a specificity of 0.801.

CONCLUSION

Radiomics models based on features derived from CECT showed convincing performances in predicting serosal involvement in GBC.

CT-based radiomics nomogram for overall survival prediction in patients with cervical cancer treated with concurrent chemoradiotherapy

Background and purpose

To establish and validate a hybrid radiomics model to predict overall survival in cervical cancer patients receiving concurrent chemoradiotherapy (CCRT).

Methods

We retrospectively collected 367 cervical cancer patients receiving chemoradiotherapy from the First Affiliated Hospital of Soochow University in China and divided them into a training set and a test set in a ratio of 7:3. Handcrafted and deep learning (DL)-based radiomics features were extracted from the contrast-enhanced computed tomography (CT), and the two types of radiomics signatures were calculated based on the features selected using the least absolute shrinkage and selection operator (LASSO) Cox regression. A hybrid radiomics nomogram was constructed by integrating independent clinical risk factors, handcrafted radiomics signature, and DL-based radiomics signature in the training set and was validated in the test set.

Results

The hybrid radiomics nomogram exhibited favorable performance in predicting overall survival, with areas under the receiver operating characteristic curve (AUCs) for 1, 3, and 5 years in the training set of 0.833, 0.777, and 0.871, respectively, and in the test set of 0.811, 0.713, and 0.730, respectively. Furthermore, the hybrid radiomics nomogram outperformed the single clinical model, handcrafted radiomics signature, and DL-based radiomics signature in both the training (C-index: 0.793) and test sets (C-index: 0.721). The calibration curves and decision curve analysis (DCA) indicated that our hybrid nomogram had good calibration and clinical benefits. Finally, our hybrid nomogram demonstrated value in stratifying patients into high- and low-risk groups (cutoff value: 5.6).

Conclusion

A high-performance hybrid radiomics model based on pre-radiotherapy CT was established, presenting strengths in risk stratification.

Effectivity of Palliative Care Bundle on Advanced Gallbladder Cancer: A Randomised Controlled Trial

Objectives

The main aim of this study was to develop, test, and compare palliative care bundles to improve functional recovery, resilience, and quality of life among advanced gallbladder cancer patient with their routine palliative care.

Material and Methods

This study was to test a palliative care bundle, a single-center, and two-arm randomised controlled trial done on a total of 116 participants (58 in each arm) from July 2019 to December 2021 at All India Institute of Medical Sciences, Rishikesh.

Results

By the end of 4th month, the recruitment rate was 96.7%, retention rate acceptance rate was 95%, and adherence rate was 85%. The palliative care bundle showed that a significant difference in trial outcome index score (P = 0.014*) indicates the effectiveness of the palliative care bundle related to improvement in physical mobility, resilience, and quality of life of patients and reduced caregiver burden. Reported barriers faced by participants were physical exhaustion (65%), psychological factors (25%), social factors (15%) and unfamiliar surroundings (5%). Caregivers reported barriers that their job (40%), physical fatigue related to the care of their patient (40%), their education (10%), and lack of support for their other family members (10%) were some reasons forcing them not to practice palliative care bundle.

Conclusion

The palliative care bundle did not interfere with the palliative treatment plan of any patients and significantly improved physical mobility, resilience, quality of life of patients, and reduced caregiver burden. Hence, a palliative care bundle can be considered in the palliative care of advanced cancer patients during their palliative treatment to provide holistic care.

Optimal radiological gallbladder lesion characterization by combining visual assessment with CT-based radiomics

OBJECTIVES

Differentiating benign gallbladder diseases from gallbladder cancer (GBC) remains a radiological challenge because they can appear very similar on imaging. This study aimed at investigating whether CT-based radiomic features of suspicious gallbladder lesions analyzed by machine learning algorithms could adequately discriminate benign gallbladder disease from GBC. In addition, the added value of machine learning models to radiological visual CT-scan interpretation was assessed.

METHODS

Patients were retrospectively selected based on confirmed histopathological diagnosis and available contrast-enhanced portal venous phase CT-scan. The radiomic features were extracted from the entire gallbladder, then further analyzed by machine learning classifiers based on Lasso regression, Ridge regression, and XG Boosting. The results of the best-performing classifier were combined with radiological visual CT diagnosis and then compared with radiological visual CT assessment alone.

RESULTS

In total, 127 patients were included: 83 patients with benign gallbladder lesions and 44 patients with GBC. Among all machine learning classifiers, XG boosting achieved the best AUC of 0.81 (95% CI 0.72-0.91) and the highest accuracy rate of 73% (95% CI 65-80%). When combining radiological visual interpretation and predictions of the XG boosting classifier, the highest diagnostic performance was achieved with an AUC of 0.98 (95% CI 0.96-1.00), a sensitivity of 91% (95% CI 86-100%), a specificity of 93% (95% CI 90-100%), and an accuracy of 92% (95% CI 90-100%).

CONCLUSIONS

Machine learning analysis of CT-based radiomic features shows promising results in discriminating benign from malignant gallbladder disease. Combining CT-based radiomic analysis and radiological visual interpretation provided the most optimal strategy for GBC and benign gallbladder disease differentiation.

KEY POINTS

Radiomic-based machine learning algorithms are able to differentiate benign gallbladder disease from gallbladder cancer. Combining machine learning algorithms with a radiological visual interpretation of gallbladder lesions at CT increases the specificity, compared to visual interpretation alone, from 73 to 93% and the accuracy from 85 to 92%. Combined use of machine learning algorithms and radiological visual assessment seems the most optimal strategy for GBC and benign gallbladder disease differentiation.

Differential diagnosis of gallbladder neoplastic polyps and cholesterol polyps with radiomics of dual modal ultrasound: a pilot study

PURPOSE

To verify whether radiomics techniques based on dual-modality ultrasound consisting of B-mode and superb microvascular imaging (SMI) can improve the accuracy of the differentiation between gallbladder neoplastic polyps and cholesterol polyps.

METHODS

A total of 100 patients with 100 pathologically proven gallbladder polypoid lesions were enrolled in this retrospective study. Radiomics features on B-mode ultrasound and SMI of each lesion were extracted. Support vector machine was used to classify adenomas and cholesterol polyps of gallbladder for B-mode, SMI and dual-modality ultrasound, respectively, and the classification results were compared among the three groups.

RESULTS

Six, eight and nine features were extracted for each lesion at B-mode ultrasound, SMI and dual-modality ultrasound, respectively. In dual-modality ultrasound model, the area under the receiver operating characteristic curve (AUC), classification accuracy, sensitivity, specificity, and Youden's index were 0.850 ± 0.090, 0.828 ± 0.097, 0.892 ± 0.144, 0.803 ± 0.149 and 0.695 ± 0.157, respectively. The AUC and Youden's index of the dual-modality model were higher than those of the B-mode model (p < 0.05). The AUC, accuracy, specificity and Youden's index of the dual-modality model were higher than those of the SMI model (p < 0.05).

CONCLUSIONS

Radiomics analysis of the dual-modality ultrasound composed of B-mode and SMI can improve the accuracy of classification between gallbladder neoplastic polyps and cholesterol polyps.

Development and Validation of an Ultrasound-Based Radiomics Nomogram for Identifying HER2 Status in Patients with Breast Carcinoma

(1) Objective: To evaluate the performance of ultrasound-based radiomics in the preoperative prediction of human epidermal growth factor receptor 2-positive (HER2+) and HER2- breast carcinoma. (2) Methods: Ultrasound images from 309 patients (86 HER2+ cases and 223 HER2- cases) were retrospectively analyzed, of which 216 patients belonged to the training set and 93 patients assigned to the time-independent validation set. The region of interest of the tumors was delineated, and the radiomics features were extracted. Radiomics features underwent dimensionality reduction analyses using the intra-class correlation coefficient (ICC), Mann-Whitney U test, and the least absolute shrinkage and selection operator (LASSO) algorithm. The radiomics score (Rad-score) for each patient was calculated through a linear combination of the nonzero coefficient features. The support vector machine (SVM), K nearest neighbors (KNN), logistic regression (LR), decision tree (DT), random forest (RF), naive Bayes (NB) and XGBoost (XGB) machine learning classifiers were trained to establish prediction models based on the Rad-score. A clinical model based on significant clinical features was also established. In addition, the logistic regression method was used to integrate Rad-score and clinical features to generate the nomogram model. The leave-one-out cross validation (LOOCV) method was used to validate the reliability and stability of the model. (3) Results: Among the seven classifier models, the LR achieved the best performance in the validation set, with an area under the receiver operating characteristic curve (AUC) of 0.786, and was obtained as the Rad-score model, while the RF performed the worst. Tumor size showed a statistical difference between the HER2+ and HER2- groups (p = 0.028). The nomogram model had a slightly higher AUC than the Rad-score model (AUC, 0.788 vs. 0.786), but no statistical difference (Delong test, p = 0.919). The LOOCV method yielded a high median AUC of 0.790 in the validation set. (4) Conclusion: The Rad-score model performs best among the seven classifiers. The nomogram model based on Rad-score and tumor size has slightly better predictive performance than the Rad-score model, and it has the potential to be utilized as a routine modality for preoperatively determining HER2 status in BC patients non-invasively.

Radiomics-based nomogram as predictive model for prognosis of hepatocellular carcinoma with portal vein tumor thrombosis receiving radiotherapy

Background

This study aims to establish and validate a predictive model based on radiomics features, clinical features, and radiation therapy (RT) dosimetric parameters for overall survival (OS) in hepatocellular carcinoma (HCC) patients treated with RT for portal vein tumor thrombosis (PVTT).

Methods

We retrospectively reviewed 131 patients. Patients were randomly divided into the training (n = 105) and validation (n = 26) cohorts. The clinical target volume was contoured on pre-RT computed tomography images and 48 textural features were extracted. The least absolute shrinkage and selection operator regression was used to determine the radiomics score (rad-score). A nomogram based on rad-score, clinical features, and dosimetric parameters was developed using the results of multivariate regression analysis. The predictive nomogram was evaluated using Harrell’s concordance index (C-index), area under the curve (AUC), and calibration curve.

Results

Two radiomics features were extracted to calculate the rad-score for the prediction of OS. The radiomics-based nomogram had better performance than the clinical nomogram for the prediction of OS, with a C-index of 0.73 (95% CI, 0.67–0.79) and an AUC of 0.71 (95% CI, 0.62–0.79). The predictive accuracy was assessed by a calibration curve.

Conclusion

The radiomics-based predictive model significantly improved OS prediction in HCC patients treated with RT for PVTT.

Contrast-enhanced CT radiomics for prediction of recurrence-free survival in gallbladder carcinoma after surgical resection

OBJECTIVES

Gallbladder carcinoma (GBC) is the most common and aggressive biliary tract malignancy with high postoperative recurrence rates. This single-center study aimed to develop and validate a radiomics signature to estimate GBC recurrence-free survival (RFS).

METHODS

This study retrospectively included 204 consecutive patients with pathologically diagnosed GBC and were randomly divided into development (n = 142) and validation (n = 62) cohorts (7:3). The radiomics features of tumor were extracted from preoperative contrast-enhanced CT imaging for each patient. In the development cohort, the least absolute shrinkage and selection operator (LASSO) Cox regression was employed to develop a radiomics signature for RFS prediction. The patients were stratified into high-score or low-score groups according to their median value of radiomics score. A nomogram was established using multivariable Cox regression by incorporating significant pathological predictors and radiomics signatures.

RESULTS

The radiomics signature based on 12 features could discriminate high-risk patients with poor RFS. Multivariate Cox analysis revealed that pT3/4 stage (hazard ratio, [HR] = 2.691), pN2 stage (HR = 3.60), poor differentiation grade (HR = 2.651), and high radiomics score (HR = 1.482) were independent risk variables associated with worse RFS and were incorporated to construct a nomogram. The nomogram displayed good prediction performance in estimating RFS with AUC values of 0.895, 0.935, and 0.907 at 1, 3, and 5 years, respectively.

CONCLUSIONS

The radiomics signature and combined nomogram may assist in predicting RFS in GBC patients.

KEY POINTS

• A radiomics signature extracted from preoperative contrast-enhanced CT can be a useful tool to preoperatively predict RFS of GBC. • T3/T4 stage, N2, poor tumor differentiation, and high radiomics score were positively associated with postoperative recurrence.

FDG PET/CT to Predict Recurrence of Early Breast Invasive Ductal Carcinoma

This study investigated the prognostic value of FDG PET/CT radiomic features for predicting recurrence in patients with early breast invasive ductal carcinoma (IDC). The medical records of consecutive patients who were newly diagnosed with primary breast IDC after curative surgery were reviewed. Patients who received any neoadjuvant treatment before surgery were not included. FDG PET/CT radiomic features, such as a maximum standardized uptake value (SUVmax), metabolic tumor volume (MTV), total lesion glycolysis (TLG), skewness, kurtosis, entropy, and uniformity, were measured for the primary breast tumor using LIFEx software to evaluate recurrence-free survival (RFS). A total of 124 patients with early breast IDC were evaluated. Eleven patients had a recurrence (8.9%). Univariate survival analysis identified large tumor size (>2 cm, p = 0.045), high Ki-67 expression (≥30%, p = 0.017), high AJCC prognostic stage (≥II, p = 0.044), high SUVmax (≥5.0, p = 0.002), high MTV (≥3.25 mL, p = 0.044), high TLG (≥10.5, p = 0.004), and high entropy (≥3.15, p = 0.003) as significant predictors of poor RFS. After multivariate survival analysis, only high MTV (p = 0.045) was an independent prognostic predictor. Evaluation of the MTV of the primary tumor by FDG PET/CT in patients with early breast IDC provides useful prognostic information regarding recurrence.

Radiomics of Contrast-Enhanced Computed Tomography: A Potential Biomarker for Pretreatment Prediction of the Response to Bacillus Calmette-Guerin Immunotherapy in Non-Muscle-Invasive Bladder Cancer

Background:Bacillus Calmette-Guerin (BCG) instillation is recommended postoperatively after transurethral resection of bladder cancer (TURBT) in patients with high-risk non-muscle-invasive bladder cancer (NMIBC). An accurate prediction model for the BCG response can help identify patients with NMIBC who may benefit from alternative therapy.

Objective: To investigate the value of computed tomography (CT) radiomics features in predicting the response to BCG instillation among patients with primary high-risk NMIBC.

Methods: Patients with pathologically confirmed high-risk NMIBC were retrospectively reviewed. Patients who underwent contrast-enhanced CT examination within one to 2 weeks before TURBT and received ≥5 BCG instillation treatments in two independent hospitals were enrolled. Patients with a routine follow-up of at least 1 year at the outpatient department were included in the final cohort. Radiomics features based on CT images were extracted from the tumor and its periphery in the training cohort, and a radiomics signature was built with recursive feature elimination. Selected features further underwent an unsupervised radiomics analysis using the newly introduced method, non-negative matrix factorization (NMF), to compute factor factorization decompositions of the radiomics matrix. Finally, a robust component, which was most associated with BCG failure in 1 year, was selected. The performance of the selected component was assessed and tested in an external validation cohort.

Results: Overall, 128 patients (training cohort, n = 104; external validation cohort, n = 24) were included, including 12 BCG failures in the training cohort and 11 failures in the validation cohort each. NMF revealed five components, of which component 3 was selected for the best discrimination of BCG failure; it had an area under the curve (AUC) of .79, sensitivity of .79, and specificity of .65 in the training set. In the external validation cohort, it achieved an AUC of .68, sensitivity of .73, and specificity of .69. Survival analysis showed that patients with higher component scores had poor recurrence-free survival (RFS) in both cohorts (C-index: training cohort, .69; validation cohort, .68).

Conclusion: The study suggested that radiomics components based on NMF might be a potential biomarker to predict BCG response and RFS after BCG treatment in patients with high-risk NMIBC.

Prediction of lymph node metastasis in early colorectal cancer based on histologic images by artificial intelligence

Risk evaluation of lymph node metastasis (LNM) for endoscopically resected submucosal invasive (T1) colorectal cancers (CRC) is critical for determining therapeutic strategies, but interobserver variability for histologic evaluation remains a major problem. To address this issue, we developed a machine-learning model for predicting LNM of T1 CRC without histologic assessment. A total of 783 consecutive T1 CRC cases were randomly split into 548 training and 235 validation cases. First, we trained convolutional neural networks (CNN) to extract cancer tile images from whole-slide images, then re-labeled these cancer tiles with LNM status for re-training. Statistical parameters of the tile images based on the probability of primary endpoints were assembled to predict LNM in cases with a random forest algorithm, and defined its predictive value as random forest score. We evaluated the performance of case-based prediction models for both training and validation datasets with area under the receiver operating characteristic curves (AUC). The accuracy for classifying cancer tiles was 0.980. Among cancer tiles, the accuracy for classifying tiles that were LNM-positive or LNM-negative was 0.740. The AUCs of the prediction models in the training and validation sets were 0.971 and 0.760, respectively. CNN judged the LNM probability by considering histologic tumor grade.

Machine Learning and Radiomic Features to Predict Overall Survival Time for Glioblastoma Patients

Glioblastoma is an aggressive brain tumor with a low survival rate. Understanding tumor behavior by predicting prognosis outcomes is a crucial factor in deciding a proper treatment plan. In this paper, an automatic overall survival time prediction system (OST) for glioblastoma patients is developed on the basis of radiomic features and machine learning (ML). This system is designed to predict prognosis outcomes by classifying a glioblastoma patient into one of three survival groups: short-term, mid-term, and long-term. To develop the prediction system, a medical dataset based on imaging information from magnetic resonance imaging (MRI) and non-imaging information is used. A novel radiomic feature extraction method is proposed and developed on the basis of volumetric and location information of brain tumor subregions extracted from MRI scans. This method is based on calculating the volumetric features from two brain sub-volumes obtained from the whole brain volume in MRI images using brain sectional planes (sagittal, coronal, and horizontal). Many experiments are conducted on the basis of various ML methods and combinations of feature extraction methods to develop the best OST system. In addition, the feature fusions of both radiomic and non-imaging features are examined to improve the accuracy of the prediction system. The best performance was achieved by the neural network and feature fusions.