Molecular markers for bladder cancer: the road to a multimarker approach

Automatic image segmentation and online survival prediction model of medulloblastoma based on machine learning

OBJECTIVES

To develop a dynamic nomogram containing radiomics signature and clinical features for estimating the overall survival (OS) of patients with medulloblastoma (MB) and design an automatic image segmentation model to reduce labor and time costs.

METHODS

Data from 217 medulloblastoma (MB) patients over the past 4 years were collected and separated into a training set and a test set. Intraclass correlation coefficient (ICC), random survival forest (RSF), and least absolute shrinkage and selection operator (LASSO) regression methods were employed to select variables in the training set. Univariate and multivariate Cox proportional hazard models, as well as Kaplan-Meier analysis, were utilized to determine the relationship among the radiomics signature, clinical features, and overall survival. A dynamic nomogram was developed. Additionally, a 3D-Unet deep learning model was used to train the automatic tumor delineation model.

RESULTS

Higher Rad-scores were significantly associated with worse OS in both the training and validation sets (p < 0.001 and p = 0.047, respectively). The Cox model combined clinical and radiomics signatures ([IBS = 0.079], [C-index = 0.747, SE = 0.045]) outperformed either radiomics signatures alone ([IBS = 0.081], [C-index = 0.738, SE = 0.041]) or clinical features alone ([IBS = 0.085], [C-index = 0.565, SE = 0.041]). The segmentation model had mean Dice coefficients of 0.80, 0.82, and 0.78 in the training, validation, and test sets respectively. A deep learning-based tumor segmentation model was built with Dice coefficients of 0.8372, 0.8017, and 0.7673 on the training set, validation set, and test set, respectively.

CONCLUSIONS

A combination of radiomics features and clinical characteristics enhances the accuracy of OS prediction in medulloblastoma patients. Additionally, building an MRI image automatic segmentation model reduces labor and time costs.

CLINICAL RELEVANCE STATEMENT

A survival prognosis model based on radiomics and clinical characteristics could improve the accuracy of prognosis estimation for medulloblastoma patients, and an MRI-based automatic tumor segmentation model could reduce the cost of time.

KEY POINTS

• A model that combines radiomics and clinical features can predict the survival prognosis of patients with medulloblastoma. • Online nomogram and image automatic segmentation model can help doctors better judge the prognosis of medulloblastoma and save working time. • The developed AI system can help doctors judge the prognosis of diseases and promote the development of precision medicine.

Nomogram based on multimodal magnetic resonance combined with B7-H3mRNA for preoperative lymph node prediction in esophagus cancer

BACKGROUND

Accurate preoperative prediction of lymph node metastasis (LNM) in esophageal cancer (EC) patients is of crucial clinical significance for treatment planning and prognosis.

AIM

To develop a clinical radiomics nomogram that can predict the preoperative lymph node (LN) status in EC patients.

METHODS

A total of 32 EC patients confirmed by clinical pathology (who underwent surgical treatment) were included. Real-time fluorescent quantitative reverse transcription-polymerase chain reaction was used to detect the expression of B7-H3 mRNA in EC tissue obtained during preoperative gastroscopy, and its correlation with LNM was analyzed. Radiomics features were extracted from multi-modal magnetic resonance imaging of EC using Pyradiomics in Python. Feature extraction, data dimensionality reduction, and feature selection were performed using XGBoost model and leave-one-out cross-validation. Multivariable logistic regression analysis was used to establish the prediction model, which included radiomics features, LN status from computed tomography (CT) reports, and B7-H3 mRNA expression, represented by a radiomics nomogram. Receiver operating characteristic area under the curve (AUC) and decision curve analysis (DCA) were used to evaluate the predictive performance and clinical application value of the model.

RESULTS

The relative expression of B7-H3 mRNA in EC patients with LNM was higher than in those without metastasis, and the difference was statistically significant (P < 0.05). The AUC value in the receiver operating characteristic (ROC) curve was 0.718 (95%CI: 0.528-0.907), with a sensitivity of 0.733 and specificity of 0.706, indicating good diagnostic performance. The individualized clinical prediction nomogram included radiomics features, LN status from CT reports, and B7-H3 mRNA expression. The ROC curve demonstrated good diagnostic value, with an AUC value of 0.765 (95%CI: 0.598-0.931), sensitivity of 0.800, and specificity of 0.706. DCA indicated the practical value of the radiomics nomogram in clinical practice.

CONCLUSION

This study developed a radiomics nomogram that includes radiomics features, LN status from CT reports, and B7-H3 mRNA expression, enabling convenient preoperative individualized prediction of LNM in EC patients.

Development and validation of a novel radiomics nomogram for prediction of early recurrence in colorectal cancer

BACKGROUND

Early recurrence (ER) is a significant concern following curative resection of advanced colorectal cancer (CRC) and is linked to poor long-term survival. Reliable prediction of ER is challenging, necessitating the development of a novel radiomics-based nomogram for CRC patients.

METHODS

We enrolled 405 patients, with 298 in the training set and 107 in the external test set. Radiomic features were extracted from preoperative venous-phase computed tomography (CT) images. A radiomics signature was created using univariate logistic regression analyses and the least absolute shrinkage and selection operator algorithm. Clinical factors were integrated into the analyses to develop a comprehensive predictive tool in a multivariate logistic regression model, resulting in a radiomics nomogram. Subsequently, the calibration, discrimination, and clinical usefulness of the nomogram were evaluated.

RESULTS

The radiomics signature, consisting of four selected CT features, was significantly associated with ER in both the training and test datasets (P < 0.05). Independent predictors of ER included TNM stage, carcinoembryonic antigen level and differentiation grade were identified. The radiomics nomogram, incorporating all these predictors, exhibited good predictive ability in both the training set with an area under the curve (AUC) of 0.82 (95 % confidence interval (CI), 0.74-0.90) and the test set with an AUC of 0.85 (95 % CI, 0.72-0.99), surpassing the performance of any single candidate factor alone. Furthermore, additional analysis demonstrated that the nomogram was clinically useful.

CONCLUSIONS

We have developed a radiomics-based nomogram that effectively predicts early recurrence in CRC patients, enhancing the potential for timely intervention and improved outcomes.

A tumor microenvironment preoperative nomogram for prediction of lymph node metastasis in bladder cancer

Background

Growing evidence suggests that tumor metastasis necessitates multi-step microenvironmental regulation. Lymph node metastasis (LNM) influences both pre- and post-operative bladder cancer (BLCA) treatment strategies. Given that current LNM diagnosis methods are still insufficient, we intend to investigate the microenvironmental changes in BLCA with and without LNM and develop a prediction model to confirm LNM status.

Method

"Estimation of Stromal and Immune cells in Malignant Tumors using Expression data" (ESTIMATE) algorithm was used to characterize the tumor microenvironment pattern of TCGA-BLCA cohort, and dimension reduction, feature selection, and StrLNM signature construction were accomplished using least absolute shrinkage and selection operator (LASSO) regression. StrLNM signature was combined with the genomic mutation to establish an LNM nomogram by using multivariable logistic regression. The performance of the nomogram was evaluated in terms of calibration, discrimination, and clinical utility. The testing set from the TCGA-BLCA cohort was used for internal validation. Moreover, three independent cohorts were used for external validation, and BLCA patients from our cohort were also used for further validation.

Results

The StrLNM signature, consisting of 22 selected features, could accurately predict LNM status in the TCGA-BLCA cohort and several independent cohorts. The nomogram performed well in discriminating LNM status, with the area under curve (AUC) of 75.1% and 65.4% in training and testing datasets from the TCGA-BLCA cohort. Furthermore, the StrLNM nomogram demonstrated good calibration with p >0.05 in the Hosmer-Lemeshow goodness of fit test. Decision curve analysis (DCA) revealed that the StrLNM nomogram had a high potential for clinical utility. Additionally, 14 of 22 stably expressed genes were identified by survival analysis and confirmed by qPCR in BLCA patient samples in our cohort.

Conclusion

In summary, we developed a nomogram that included an StrLNM signature and facilitated the preoperative prediction of LNM status in BLCA patients.

Nomograms for prediction of breast cancer in breast imaging reporting and data system (BI-RADS) ultrasound category 4 or 5 lesions: A single-center retrospective study based on radiomics features

Purpose

To develop nomograms for predicting breast malignancy in BI-RADS ultrasound (US) category 4 or 5 lesions based on radiomics features.

Methods

Between January 2020 and January 2022, we prospectively collected and retrospectively analyzed the medical records of 496 patients pathologically proven breast lesions in our hospital. The data set was divided into model training group and validation testing group with a 75/25 split. Radiomics features were obtained using the PyRadiomics package, and the radiomics score was established by least absolute shrinkage and selection operator regression. A nomogram was developed for BI-RADS US category 4 or 5 lesions according to the results of multivariate regression analysis from the training group.

Result

The AUCs of radiomics score consisting of 31 US features was 0.886. The AUC of the model constructed with radiomics score, patient age, lesion diameter identified by US and BI-RADS category involved was 0.956 (95% CI, 0.910–0.972) for the training group and 0.937 (95% CI, 0.893–0.965) for the validation cohort. The calibration curves showed good agreement between the predictions and observations.

Conclusions

Both nomogram and radiomics score can be used as methods to assist radiologists and clinicians in predicting breast malignancy in BI-RADS US category 4 or 5 lesions.

Progress in the development of tissue-based biomarkers for urothelial cancer

INTRODUCTION

As the understanding of molecular mechanisms of bladder cancer advances, molecularly-guided precision medicine becomes increasingly relevant. Biomarkers play a critical role in this setting, predicting treatment response and identifying candidates for targeted therapies.

AREAS COVERED

Current literature on biomarkers in their role in disease prognosis, and response to neoadjuvant and adjuvant therapies. In non-muscle invasive bladder cancer, particular focus is on markers of disease progression, and response to intravesical therapy. In muscle invasive and advanced bladder cancer, particular emphasis is on markers associated with neoadjuvant chemotherapy, as well as systemic immunotherapy. We discuss current shortcomings and pitfalls in contemporary markers, and future avenues of prospective research.

EXPERT OPINION

The focus on biomarkers has moved from immunohistochemical analysis and tumor-related phenotypic changes to examining genetic alterations. Single marker analysis has been shown to be insufficient in predicting both disease course and response to therapy, and studies have shifted towards examining marker combinations and genetic classifiers. Ultimately, significant progress in implementing biomarkers into clinical guidelines remains elusive, largely due to lack of prospective studies in well-defined patient cohorts and with clinically-meaningful endpoints. Until then, despite their promising value, tissue markers should be limited to experimental settings and clinical trials.

Radiomics Nomogram for Predicting Locoregional Failure in Locally Advanced Non-small Cell Lung Cancer Treated with Definitive Chemoradiotherapy

RATIONALE AND OBJECTIVES

To develop and validate a computed tomography (CT)-based radiomics nomogram for predicting locoregional failure (LRF) in patients with locally advanced non-small cell lung cancer (NSCLC) treated with definitive chemoradiotherapy (CRT).

MATERIALS AND METHODS

A total of 141 patients with locally advanced NSCLC treated with definitive CRT from January 2014 to December 2017 were included and divided into testing cohort (n = 100) and validation (n = 41) cohort. Radiomics features were extracted from pretreatment contrast enhanced CT. The least absolute shrinkage and selection operator logistic regression was processed to select predictive features from the testing cohort and constructed a radiomics signature. Clinical characteristics and the radiomics signature were analyzed using univariable and multivariate Cox regression. The radiomics nomogram was established with the radiomics signature and independent clinical factors. Harrell's C-index, calibration curves and decision curves were used to assess the performance of the radiomics nomogram.

RESULTS

The radiomics signature, which consisted of eight selected features, was an independent factor of LRF. The clinical predictors of LRF were the histologic type and clinical stage. The radiomics nomogram combined with the radiomics signature and clinical prognostic factors showed good performance with C-indexes of 0.796 (95% confidence interval [CI]: 0.709-0.883) and 0.756 (95% CI: 0.674-0.838) in the testing and validation cohorts respectively. Additionally, the combined nomogram resulted in better performance (p < 0.001) for the estimation of LRF than the nomograms with the radiomics signature (C-index: 0.776; 95% CI: 0.686-0.866) or clinical predictors (C-index: 0.641; 95% CI: 0.542-0.740) alone.

CONCLUSION

The radiomics nomogram provided the best performance for LRF prediction in patients with locally advanced NSCLC, which may help optimize individual treatments.

Elaboration of Multiparametric MRI-Based Radiomics Signature for the Preoperative Quantitative Identification of the Histological Grade in Patients With Non-Small-Cell Lung Cancer

BACKGROUND

The histological grading plays an essential role in the treatment decision of lung cancer. Detected tumors are usually biopsied to confirm histologic grade. How to use MRI extracted radiomics features for accurately grading lung cancer is still challenging.

PURPOSE

To examine the diagnostic utility of multiparametric MRI radiomics and clinical factors for grading non-small-cell lung cancer (NSCLC).

STUDY TYPE

Retrospective.

POPULATION

A total of 148 patients (25.7% female) with postoperative pathologically confirmed NSCLC and divided into the training cohort (N = 110) and the validation cohort (N = 38).

FIELD STRENGTH/SEQUENCE

A 1.5 T; single-shot turbo spin-echo (TSE), T2-weighted imaging (T2WI), and integrated shimming-echo planar imaging (ISHIM-EPI) diffusion-weighted imaging (DWI).

ASSESSMENT

A total of 2775 radiomics features were extracted from carcinomatous regions of interest on T2WI, DWI, and the apparent diffusion coefficient (ADC) maps. The five optimal features were selected by using the Student' s t-test, the least absolute shrinkage and selection operator (LASSO) and stepwise regression. The Radscore combined with clinical factors, which selected by univariate and multivariate analyses, to develop a radiomics-clinical nomogram. Its performance was evaluated in the training cohort and the validation cohort. The potential clinical usefulness was analyzed by the receiver operating characteristic curve (ROC), area under the curve (AUC), and the Hosmer-Lemeshow test.

STATISTICAL TESTS

Student's t-test, univariate analyses, multivariate analyses, LASSO, ROC, AUC, and the Hosmer-Lemeshow test. P < 0.05 was considered statistically significant.

RESULTS

Favorable discrimination performance was obtained for five optimal features (out of the 2775 features), using the training cohorts (AUC 0.761) and validation cohorts (AUC 0.753). In addition, the radiomics-clinical nomogram significantly improved the ability to identify histological grades in the training cohort (AUC 0.814) and the validation cohort (AUC 0.767).

DATA CONCLUSIONS

The radiomics-clinical nomogram based on multiparametric MRI might have the potential to distinguish the histological grade of NSCLC.

EVIDENCE LEVEL

3 TECHNICAL EFFICACY: Stage 2.

Development of a prognostic gene signature for hepatocellular carcinoma

Accurate prediction of overall survival is important for prognosis and the assignment of appropriate personalized clinical treatment in hepatocellular carcinoma (HCC) patients. The aim of the present study was to establish an optimal gene model for the independent prediction of prognosis associated with common clinical patterns. Gene expression profiles and the corresponding clinical information of the LIHC cohort were obtained from The Cancer Genome Atlas. Differentially expressed genes were found using the R package "limma". Subsequently, a prognostic gene signature was developed using the LASSO Cox regression model. Kaplan-Meier, log-rank, and receiver operating characteristic (ROC) analyses were performed to verify the predictive accuracy of the prognostic model. Finally, a nomogram and calibration plot were created using the "rms" package. Differentially expressed genes were screened with threshold criteria (FDR < 0.01 and |log FC|>3) and 563 differentially expressed genes were obtained, including 448 downregulated and 115 upregulated genes. Using the LASSO Cox regression model, a prognostic gene signature was developed based on nine genes, IQGAP3, BIRC5, PTTG1, STC2, CDKN3, PBK, EXO1, NEIL3, and HOXD9, the expression levels of which were quantitated using RT-qPCR. According to the risk scores, patients were separated into high-risk and low-risk groups. In conclusion, the prognostic gene signature can be used as a combined biomarker for the independent prediction of overall survival in HCC patients. Moreover, we created a nomogram that can be used to infer prognosis and aid individualized decisions regarding treatment and surveillance.

An interactive nomogram based on clinical and molecular signatures to predict prognosis in multiple myeloma patients

Although novel drugs and treatments have been developed and improved, multiple myeloma (MM) is still recurrent and difficult to cure. In the present study, the magenta module containing 400 hub genes was determined from the training dataset of GSE24080 through weighted gene co-expression network analysis (WGCNA). Then, using the least absolute shrinkage and selection operator (Lasso) analysis, a fifteen-gene signature was firstly selected and the predictive performance for overall survival (OS) was favorable, which was identified by Receiver Operating Characteristic (ROC) curves. The risk score model was constructed based on survival-associated fifteen genes from the Lasso model, which classified MM patients into high-risk and low-risk groups. Areas under the curve (AUC) of ROC curve and log-rank test showed that the high-risk group was correlated to the dismal survival outcome of MM patients, which was also identified in testing dataset of GSE9782. The calibration plot, the AUC value of the ROC curve and Concordance-index showed that the interactive nomogram with risk score could favorably predict the probability of multi-year OS of MM patients. Therefore, it may help clinicians make a precise therapeutic decision based on the easy-to-use tool of the nomogram.

Preoperative prediction of perineural invasion with multi-modality radiomics in rectal cancer

Perineural invasion (PNI) as a grossly underreported independent risk predictor in rectal cancer is hard to identify preoperatively. We aim to predict PNI status in rectal cancer using multi-modality radiomics. In total, 396 radiomics features were extracted from T2-weighted images (T2WIs), diffusion-weighted images (DWIs), and portal venous phase of contrast-enhanced CT (CE-CT) respectively of 94 consecutive patients with histologically confirmed rectal cancer. T2WI score, DWI score, and CT score were calculated via the radiomics features selection and optimization. Discrimination, calibration, and clinical benefit ability were used to evaluate the performance of the radiomics scores in both training and testing datasets. CT score and T2WI score were independent risk predictors [CT score, OR (95% CI) = 4.218 (1.070-16.620); T2WI score, OR (95% CI) = 105.721 (3.091-3615.790)]. The concise score which combined CT score and T2WI score, showed the best performance [training dataset, AUC (95% CI) = 0.906 (0.833-0.979); testing dataset, AUC (95% CI) = 0.884 (0.761-1.000)] and good calibration (P > 0.05 in the Hosmer-Lemeshow test for the training and testing datasets). Decision curve analysis showed that the multi-modality radiomics nomogram had a higher clinical net benefit. The multi-modality radiomics score could be used to preoperatively assess PNI status in rectal cancer.

Individualized Prediction of Colorectal Cancer Metastasis Using a Radiogenomics Approach

Objectives: To evaluate whether incorporating the radiomics, genomics, and clinical features allows prediction of metastasis in colorectal cancer (CRC) and to develop a preoperative nomogram for predicting metastasis.

Methods: We retrospectively analyzed radiomics features of computed tomography (CT) images in 134 patients (62 in the primary cohort, 28 in the validation cohort, and 44 in the independent-test cohort) clinicopathologically diagnosed with CRC at Dazhou Central Hospital from February 2018 to October 2019. Tumor tissues were collected from all patients for RNA sequencing, and clinical data were obtained from medical records. A total of 854 radiomics features were extracted from enhanced venous-phase CT of CRC. Least absolute shrinkage and selection operator regression analysis was utilized for data dimension reduction, feature screen, and radiomics signature development. Multivariable logistic regression analysis was performed to build a multiscale predicting model incorporating the radiomics, genomics, and clinical features. The receiver operating characteristic curve, calibration curve, and decision curve were conducted to evaluate the performance of the nomogram.

Results: The radiomics signature based on 16 selected radiomics features showed good performance in metastasis assessment in both primary [area under the curve (AUC) = 0.945, 95% confidence interval (CI) 0.892–0.998] and validation cohorts (AUC = 0.754, 95% CI 0.570–0.938). The multiscale nomogram model contained radiomics features signatures, four-gene expression related to cell cycle pathway, and CA 19-9 level. The multiscale model showed good discrimination performance in the primary cohort (AUC = 0.981, 95% CI 0.953–1.000), the validation cohort (AUC = 0.822, 95% CI 0.635–1.000), and the independent-test cohort (AUC = 0.752, 95% CI 0.608–0.896) and good calibration. Decision curve analysis confirmed the clinical application value of the multiscale model.

Conclusion: This study presented a multiscale model that incorporated the radiological eigenvalues, genomics features, and CA 19-9, which could be conveniently utilized to facilitate the individualized preoperatively assessing metastasis in CRC patients.

A Novel Signature for Predicting Prognosis of Smoking-Related Squamous Cell Carcinoma

Tobacco smoking is an established risk factor for squamous cell carcinoma (SCC). We obtained smoking-related SCC, including cervical SCC (CSCC), esophageal SCC (ESCC), head and neck SCC (HNSC), and lung SCC (LUSC), from The Cancer Genome Atlas (TCGA) database to investigate the association between smoking status (reformed and current smoking) and prognosis. We found that reformed smokers had a better prognosis than current smokers in CSCC (p = 0.003), HNSC (p = 0.019), and LUSC (p < 0.01) cohorts. Then, we selected LUSC cohorts as the training cohort and other SCC cohorts as the test cohorts. Function analysis revealed that homologous recombination (HR) was the most significant pathway involved in smoking-induced LUSC. Moreover, the effect of cross-talk between the smoking status and HR deficiency (HRD) on the prognosis was further evaluated, revealing that quitting smoking with high HRD scores could significantly improve patients’ prognosis (p < 0.01). To improve prognosis prediction and more effectively screen suitable populations for platinum drugs and poly-ADP-ribose polymerase (PARP) inhibitors, we constructed a risk score model using smoking- and HRD-related genes in LUSC. The risk score model had high power for predicting 2-, 3-, and 5-year survival (p < 0.01, AUC = 0.67, 0.66, and 0.66). In addition, the risk scores were an independent risk factor for LUSC (HR = 2.34, 95%CI = 1.70–3.23). The practical nomogram was also built using the risk score, smoking status, and other clinical information with a good c-index (0.72, 95%CI = 0.70–0.74). Finally, we used other TCGA SCC cohorts to confirm the reliability and validity of the risk score model (p < 0.01 and AUC > 0.6 at 2, 3, and 5 years in CSCC and HNSC cohorts). In conclusion, the present study suggested that smoking cessation should be a part of smoking-related SCC treatment, and also provided a risk score model to predict prognosis and improve the effectiveness of screening the platinum/PARP population.

Development of Machine Learning Models for Predicting Postoperative Delayed Remission in Patients With Cushing's Disease

CONTEXT

Postoperative hypercortisolemia mandates further therapy in patients with Cushing's disease (CD). Delayed remission (DR) is defined as not achieving postoperative immediate remission (IR), but having spontaneous remission during long-term follow-up.

OBJECTIVE

We aimed to develop and validate machine learning (ML) models for predicting DR in non-IR patients with CD.

METHODS

We enrolled 201 CD patients, and randomly divided them into training and test datasets. We then used the recursive feature elimination (RFE) algorithm to select features and applied 5 ML algorithms to construct DR prediction models. We used permutation importance and local interpretable model-agnostic explanation (LIME) algorithms to determine the importance of the selected features and interpret the ML models.

RESULTS

Eighty-eight (43.8%) of the 201 CD patients met the criteria for DR. Overall, patients who were younger, had a low body mass index, a Knosp grade of III-IV, and a tumor not found by pathological examination tended to achieve a lower rate of DR. After RFE feature selection, the Adaboost model, which comprised 18 features, had the greatest discriminatory ability, and its predictive ability was significantly better than using Knosp grading and postoperative immediate morning serum cortisol (PoC). The results obtained from permutation importance and LIME algorithms showed that preoperative 24-hour urine free cortisol, PoC, and age were the most important features, and showed the reliability and clinical practicability of the Adaboost model in DC prediction.

CONCLUSIONS

Machine learning-based models could serve as an effective noninvasive approach to predicting DR, and could aid in determining individual treatment and follow-up strategies for CD patients.

Radiomics Nomogram of DCE-MRI for the Prediction of Axillary Lymph Node Metastasis in Breast Cancer

Purpose

This study aimed to establish and validate a radiomics nomogram based on dynamic contrast-enhanced (DCE)-MRI for predicting axillary lymph node (ALN) metastasis in breast cancer.

Method

This retrospective study included 296 patients with breast cancer who underwent DCE-MRI examinations between July 2017 and June 2018. A total of 396 radiomics features were extracted from primary tumor. In addition, the least absolute shrinkage and selection operator (LASSO) algorithm was used to select the features. Radiomics signature and independent risk factors were incorporated to build a radiomics nomogram model. Calibration and receiver operator characteristic (ROC) curves were used to confirm the performance of the nomogram in the training and validation sets. The clinical usefulness of the nomogram was evaluated by decision curve analysis (DCA).

Results

The radiomics signature consisted of three ALN-status-related features, and the nomogram model included the radiomics signature and the MR-reported lymph node (LN) status. The model showed good calibration and discrimination with areas under the ROC curve (AUC) of 0.92 [95% confidence interval (CI), 0.87-0.97] in the training set and 0.90 (95% CI, 0.85-0.95) in the validation set. In the MR-reported LN-negative (cN0) subgroup, the nomogram model also exhibited favorable discriminatory ability (AUC, 0.79; 95% CI, 0.70-0.87). DCA findings indicated that the nomogram model was clinically useful.

Conclusions

The MRI-based radiomics nomogram model could be used to preoperatively predict the ALN metastasis of breast cancer.

An Epithelial-Mesenchymal Transition (EMT) Preoperative Nomogram for Prediction of Lymph Node Metastasis in Bladder Cancer (BLCA)

Lymph node (LN) metastasis is a lethal independent risk factor for patients with bladder cancer (BLCA). Accurate evaluation of LN metastasis is of vital importance for disease staging, treatment selection, and prognosis prediction. Several histopathologic parameters are available to predict LN metastasis postoperatively. To date, medical imaging techniques have made a great contribution to preoperatively diagnosis of LN metastasis, but it also exhibits substantial false positives. Therefore, a reliable and robust method to preoperatively predict LN metastasis is urgently needed. Here, we selected 19 candidate genes related to epithelial-mesenchymal transition (EMT) across the LN metastasis samples, which was previously reported to be responsible for the subtype transition and correlation with malignancy and prognosis of BLCA, to establish an EMT-LN signature through LASSO logistic regression analysis. The EMT-LN signature could significantly predict LN metastasis with high accuracy in the TCGA-BLCA cohort, as well as several independent cohorts. As integrating with C3orf70 mutation, we developed an individualized prediction nomogram based on the EMT-LN signature. The nomogram exhibited good discrimination on LN metastasis status, with AUC of 71.7% and 75.9% in training and testing datasets of the TCGA-BLCA cohort. Moreover, the EMT-LN nomogram displayed good calibration with p > 0.05 in the Hosmer-Lemeshow goodness of fit test. Decision curve analysis (DCA) revealed that the EMT-LN nomogram was of high potential for clinical utility. In summary, we established an EMT-LN nomogram integrating an EMT-LN signature and C3orf70 mutation status, which acted as an easy-to-use tool to facilitate preoperative prediction of LN metastasis in BLCA individuals.

Using texture analysis as a predictive factor of subtype, grade and stage of renal cell carcinoma

OBJECTIVE

The aim of this study was to evaluate the correlation between the tissue texture analysis and the histological subtypes, grade and stage of the disease in patients with renal cell carcinoma (RCC).

MATERIALS AND METHODS

Seventy-seven patients who underwent computed tomography due to renal mass and diagnosed with RCC as a result of pathological examination were retrospectively analyzed. In these analyses, the demographic characteristics, pathological and radiological findings of the patients were evaluated. The masses were introduced to the Radiomics extension of the software and the first- and second-order texture analysis parameters were obtained. The correlation of these parameters with histological subtype, Fuhrman grade and TNM stage was investigated.

RESULTS

In the comparison of the Radiomics values by stages, "minimum", "Long Run Low Gray-level Emphasis" values were higher in the stage 1-2 group, while "Energy", "Total energy", "Range", "Joint Average", "Sum Average", "Gray-Level Non-Uniformity", "Short-Run High Gray-level Emphasis ", "Run Length Non-Uniformity "and "High Gray-Level Run Emphasis "values were higher in the stage 3-4 group. Of these parameters, only "Gray-Level Non-Uniformity" and "Run Length Non-Uniformity'' values were significantly lower in tumors with low Fuhrman grade (1-2) and low TNM stage (1-2). There was no statistically significant correlation between the parameters found to be significant in histological subtype differentiation and Fuhrman grade and TNM stage.

CONCLUSION

This study demonstrates that "Gray-Level Non-Uniformity" and "Run Length Non-Uniformity "parameters in the texture analysis method can be used to evaluate the prognosis in patients with RCC.

Deep Learning Features Improve the Performance of a Radiomics Signature for Predicting KRAS Status in Patients with Colorectal Cancer

RATIONALE AND OBJECTIVES

We assess the performance of a model combining a deep convolutional neural network and a hand-crafted radiomics signature for predicting KRAS status in patients with colorectal cancer (CRC).

MATERIALS AND METHODS

The primary cohort consisted of 279 patients with clinicopathologically confirmed CRC between April 2011 and April 2015. Portal venous phase computed tomographic images were analyzed to extract traditional hand-crafted radiomics features as well as deep learning features. A Wilcoxon rank sum test, the minimum redundancy maximum relevance algorithm, and multivariable logistic regression analysis were used to select features and build a radiomics signature. A combined model was then developed using multivariable logistic regression analysis. An independent validation cohort of 119 patients from May 2015 to April 2016 was used to confirm the combined model's predictive performance.

RESULTS

The C-index of hand-crafted radiomics signature's discriminative ability was 0.719 (95% confidence interval, CI: 0.658-0.776) for the primary cohort and 0.720 (95% CI: 0.625-0.813) for the validation cohort. The C-index of the deep radiomics signature's discriminative ability was 0.754 (95% CI: 0.696-0.813) for the primary cohort and 0.786 (95% CI: 0.702-0.863) for the validation cohort. The combined model, which merged the hand-crafted radiomics features and deep radiomics features, achieve a C-index of 0.815 (95% CI: 0.766-0.868) for the primary cohort and 0.832 (95% CI: 0.762-0.905) for the validation cohort.

CONCLUSION

This study presents a model that incorporates the hand-crafted and deep radiomics signature, which can be used for individualized preoperative prediction of KRAS mutations in patients with CRC.

Preoperative Prediction Nomogram Based on Integrated Profiling for Glioblastoma Multiforme in Glioma Patients

Introduction: Traditional classification that divided gliomas into glioblastoma multiformes (GBM) and lower grade gliomas (LGG) based on pathological morphology has been challenged over the past decade by improvements in molecular stratification, however, the reproducibility and diagnostic accuracy of glioma classification still remains poor. This study aimed to establish and validate a novel nomogram for the preoperative diagnosis of GBM by using integrated data combined with feasible baseline characteristics and preoperative tests. Material and method: The models were established in a primary cohort that included 259 glioma patients who had undergone surgical resection and were pathologically diagnosed from March 2014 to May 2016 in the First Affiliated Hospital of Xi'an Jiaotong University. The preoperative data were used to construct three models by the best subset regression, the forward stepwise regression, and the least absolute shrinkage and selection operator, and to furthermore establish the nomogram among those models. The assessment of nomogram was carried out by the discrimination and calibration in internal cohorts and external cohorts. Results and discussion: Out of all three models, model 2 contained eight clinical-related variables, which exhibited the minimum Akaike Information Criterion (173.71) and maximum concordance index (0.894). Compared with the other two models, the integrated discrimination index for model 2 was significantly improved, indicating that the nomogram obtained from model 2 was the most appropriate model. Likewise, the nomogram showed great calibration and significant clinical benefit according to calibration curves and the decision curve analysis. Conclusion: In conclusion, our study showed a novel preoperative model that incorporated clinically relevant variables and imaging features with laboratory data that could be used for preoperative prediction in glioma patients, thus providing more reliable evidence for surgical decision-making.

Preoperative prediction of peritoneal metastasis in colorectal cancer using a clinical-radiomics model

PURPOSE

To establish and validate a combined clinical-radiomics model for preoperative prediction of synchronous peritoneal metastasis (PM) in patients with colorectal cancer (CRC).

METHOD

We enrolled 779 patients (585 in the training set: 553 with nonmetastasis (NM) and 32 with PM; 194 in the validation set: 184 with NM and 10 with PM) with clinicopathologically confirmed CRC. The significant clinical risk factors were used to build the clinical model; the least absolute shrinkage and selection operator (LASSO) algorithm was adopted to construct a radiomics signature, which included imaging features of the primary lesion and the largest peripheral lymph node, and stepwise logistic regression was applied to select the significant variables to develop the clinical-radiomics model. We used the Akaike information criterion (AIC) and receiver operating characteristic analysis to compare the goodness of fit and the prediction performance of the three models respectively. An independent validation cohort, containing 139 consecutive patients from February to September 2018, was used to evaluate the performance of the optimal model.

RESULTS

Among the three models, the clinical-radiomics model (AUC = 0.855; AIC = 1043.2) was identified as the optimal model, with the maximum AUC value and the minimum AIC value (the clinical-only model: AUC = 0.771, AIC = 1277.7; the radiomics-only model: AUC = 0.764, AIC = 1280.5). The clinical-radiomics model also showed good discrimination in both the validation cohort (AUC = 0.793) and the independent validation cohort (AUC = 0.781).

CONCLUSIONS

The present study proposes a clinical-radiomics model created with the CT-based radiomics signature and key clinical features that can potentially be applied in the individual preoperative prediction of synchronous PM for CRC patients.

An immune risk score with potential implications in prognosis and immunotherapy of metastatic melanoma

Melanoma is a highly aggressive cancer with a poor prognosis. We found that immune response played important roles in melanoma metastasis by GSEA analysis. Therefore, we constructed the immune risk score (IRS) by the LASSO-COX analysis in the sequencing metastatic samples from the TCGA database. Then, initial diagnosis patients with metastasis were selected as the test cohort. Importantly, we adopted overall survival (OS) as the survival outcome for initial diagnosis patients, while adopting the observed survival interval (OBS) as the survival outcome for sequencing samples which could avoid biologically meaningless associations. We found that the IRS had high power for predicting 2, 3 and 5-year survival in training (AUC = 0.70, 0.69 and 0.68) and test cohorts (AUC = 0.72, 0.70 and 0.65). The IRS was significantly associated with prognosis both in the metastatic samples (HR = 1.60, 95% CI = 1.16-2.19) and patients with metastasis (HR = 2.89, 95% CI = 1.69-4.53). we further used other independent melanoma cohorts from the GEO databases to confirm the reliability and validity of the IRS (P < 0.01 in all cohorts). The practical nomogram was also built using the IRS and clinical information with high c-index both in training (0.76, 95%CI = 0.72-0.80) and test cohorts (0.72, 95%CI = 0.65-0.79). Finally, IRS showed the predictive value of survival outcome and response of immunotherapy patients, and increased the predictive ability of current immune checkpoint gene markers. In conclusion, the IRS can serve as a potential biomarker for prognosis and responsiveness to immune checkpoint blockade immunotherapy in metastatic melanoma patients.

MR imaging of thymomas: a combined radiomics nomogram to predict histologic subtypes

OBJECTIVES

Accurately predicting the WHO classification of thymomas is urgently needed to optimize individualized therapeutic strategies. We aimed to develop and validate a combined radiomics nomogram for personalized prediction of histologic subtypes in patients with thymomas.

METHODS

A total of 182 thymoma patients were divided into training (n = 128) and test (n = 54) cohorts. Radiomics features were extracted from T2-weighted, T2-weighted fat suppression, and diffusion-weighted images to establish a radiomics signature in the training cohort. Multivariate logistic regression analysis was used to develop a combined radiomics nomogram that incorporated clinical, conventional MR imaging variables, apparent diffusion coefficient (ADC) value, and radiomics signature. The efficacy of clinical, conventional MR imaging, or ADC model was also evaluated respectively. The performances of different models were compared by receiver operating characteristic analysis and Delong test. The discrimination, calibration, and clinical usefulness of the combined radiomics nomogram were assessed.

RESULTS

The radiomics signature, consisting of 14 features, achieved favorable predictive efficacy in differentiating low-risk from high-risk thymomas, outperforming clinical, conventional MR imaging, and ADC models. The combined radiomics nomogram incorporating tumor shape, ADC value, and radiomics signature yielded the best performance (training cohort: area under the curve [AUC] = 0.946, test cohort: AUC = 0.878). The calibration curve and decision curve analysis indicated the clinical utility of the combined radiomics nomogram.

CONCLUSIONS

The radiomics signature is a useful tool that can be used to predict histologic subtypes of thymomas. The combined radiomics nomogram improved the individualized subtype prediction in patients with thymomas.

KEY POINTS

• Fourteen robust features were selected to develop a radiomics signature for preoperative prediction of thymoma subtype. • MRI-based radiomics signature can differentiate low-risk thymomas from high-risk thymomas with favorable predictive efficacy compared with clinical, conventional MR imaging, and ADC models. • Combined radiomics nomogram based on tumor shape, ADC value, and radiomics signature could improve the individualized subtype prediction in patients with thymomas.

Prediction of Malignant Acute Middle Cerebral Artery Infarction via Computed Tomography Radiomics

Malignant middle cerebral artery infarction (mMCAi) is a serious complication of cerebral infarction usually associated with poor patient prognosis. In this retrospective study, we analyzed clinical information as well as non-contrast computed tomography (NCCT) and computed tomography angiography (CTA) data from patients with cerebral infarction in the middle cerebral artery (MCA) territory acquired within 24 h from symptoms onset. Then, we aimed to develop a model based on the radiomics signature to predict the development of mMCAi in cerebral infarction patients. Patients were divided randomly into training (n = 87) and validation (n = 39) sets. A total of 396 texture features were extracted from each NCCT image from the 126 patients. The least absolute shrinkage and selection operator regression analysis was used to reduce the feature dimension and construct an accurate radiomics signature based on the remaining texture features. Subsequently, we developed a model based on the radiomics signature and Alberta Stroke Program Early CT Score (ASPECTS) based on NCCT to predict mMCAi. Our prediction model showed a good predictive performance with an AUC of 0.917 [95% confidence interval (CI), 0.863-0.972] and 0.913 [95% CI, 0.795-1] in the training and validation sets, respectively. Additionally, the decision curve analysis (DCA) validated the clinical efficacy of the combined risk factors of radiomics signature and ASPECTS based on NCCT in the prediction of mMCAi development in patients with acute stroke across a wide range of threshold probabilities. Our research indicates that radiomics signature can be an instrumental tool to predict the risk of mMCAi.

Nomogram-based prediction of pre-eclampsia in the first trimester of gestation

OBJECTIVES

Pre-eclampsia is a specific disease during gestation without a clear etiology. The lack of effective forecasting and prevention methods threatens the safety of maternal and infant lives.

STUDY DESIGN

We prospectively evaluated 356 pregnancies in their first trimester at the First Affiliated Hospital of Xi'an Jiaotong University and followed up the whole gestation.

MAIN OUTCOME MEASURES

Pre-eclampsia correlation parameters were subjected to dimensionality reduction analysis using the least absolute shrinkage and selection operator (LASSO) method.

RESULTS

Of the 356 pregnancies, 25 developed pre-eclampsia during late gestation. Sonographic parameters of the placenta, uterine artery, and umbilical artery were calculated using LASSO regression analysis. Five factors (vascularization and blood flow index of the placenta, peak systolic velocity and peak systolic to end-diastolic artery ratio of the left uterine artery, and pulse index of the umbilical artery) were applied in a final nomogram model. The fitting curve was closely correlated with the actual situation, with a C-index of 0.877.

CONCLUSIONS

The nomogram described here could be used to predict the risk of pre-eclampsia in pregnant women and provide strong support for early intervention.

Preoperative Prediction of Extramural Venous Invasion in Rectal Cancer: Comparison of the Diagnostic Efficacy of Radiomics Models and Quantitative Dynamic Contrast-Enhanced Magnetic Resonance Imaging

Background: To compare the diagnostic performance of radiomics models with that of dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) perfusion parameters for the preoperative prediction of extramural venous invasion (EMVI) in rectal cancer patients and to develop a preoperative nomogram for predicting the EMVI status. Methods: In total, 106 rectal cancer patients were enrolled in our study. All patients under went preoperative rectal high-resolution MRI and DCE-MRI. We built five models based on the perfusion parameters of DCE-MRI (quantitative model), the radiomics of T₂-weighted (T₂W) CUBE imaging (R₁ model), DCE-MRI (R₂ model), clinical features (clinical model), and clinical-radiomics features. The predictive efficacy of the radiomics signature was assessed and internally verified. The area under the receiver operating curve (AUC) was used to compare the diagnostic performance of different radiomics models and DCE-MRI quantitative parameters. The radiomics score and clinical-pathologic risk factors were incorporated into an easy-to-use nomogram. Results: The quantitative parameters K ^trans and Ve were significantly higher in the EMVI-positive group than in the EMVI-negative group (both P =0.02). K ^trans combined with Ve showed a fair degree of accuracy (AUC 0.680 in the training cohort and AUC 0.715 in the validation cohort) compared with K ^trans or Ve alone. The AUCs of the R₁ and R₂ models were 0.826, 0.715 and 0.872, 0.812 in the training and validation cohorts, respectively. In addition, the R₂-C model yielded an AUC of 0.904 in the training cohort and 0.812 in the validation cohort. The nomogram was presented based on the clinical-radiomics model. The calibration curves showed good agreement. Conclusion: The radiomics nomogram that incorporates the radiomics score, histopathological grade and T stage demonstrated better diagnostic accuracy than the DCE-MRI quantitative parameters and may have significant clinical implications for the preoperative individualized prediction of EMVI in rectal cancer patients.

Predicting the Risk of Melanoma Metastasis Using an Immune Risk Score in the Melanoma Cohort

Melanoma is a highly aggressive cancer, attracting increasing attention worldwide. The 5-year survival rate of patients with metastatic melanoma is low. Therefore, it is critical to identify potential effective biomarkers for diagnosis of melanoma metastasis. In the present study, the melanoma cohort and immune genes were obtained from the Cancer Genome Atlas (TCGA) database and the ImmPort database, respectively. Then, we constructed the immune risk score (IRS) using univariate and multivariate logistic analysis. The area under the curve (AUC) of IRS in sequencing samples and the initial diagnosis patients was 0.90 and 0.80, respectively. Besides, IRS could add benefits for metastasis diagnosis. For sequencing samples, IRS (OR = 16.35, 95% CI = 8.74–30.59) increased the odds for melanoma metastasis. Similar results were obtained in the initial diagnosis patients (OR = 8.93, 95% CI = 3.53–22.61). A composite nomogram was built based on IRS and clinical information with well-fitted calibration curves. We further used other independent melanoma cohorts from Gene Expression Omnibus (GEO) databases to confirm the reliability and validity of the IRS (AUC > 0.75, OR > 1.04, and P value < 0.01 in all cohorts). In conclusion, IRS is significantly associated with melanoma metastasis and can be a novel effective signature for predicting the metastasis risk.

Multiparametric MRI-Based Radiomics Nomogram for Predicting Lymph Node Metastasis in Early-Stage Cervical Cancer

BACKGROUND

Lymph node metastasis (LNM) is a critical risk factor affecting treatment strategy and prognosis in patients with early-stage cervical cancer.

PURPOSE

To establish a multiparametric MRI (mpMRI)-based radiomics nomogram for preoperatively predicting LNM status.

STUDY TYPE

Retrospective.

POPULATION

Among 233 consecutive patients, 155 patients were randomly allocated to the primary cohort and 78 patients to the validation cohort.

FIELD STRENGTH

Radiomic features were extracted from a 1.5T mpMRI scan (T₁ -weighted imaging [T₁ WI], fat-saturated T₂ -weighted imaging [FS-T₂ WI], contrast-enhanced [CE], diffusion-weighted imaging [DWI], and apparent diffusion coefficient [ADC] maps).

ASSESSMENT

The performance of the nomogram was assessed with respect to its calibration, discrimination, and clinical usefulness. The area under the receiver operating characteristics curve (ROC AUC), accuracy, sensitivity, and specificity were also calculated.

STATISTICAL TESTS

The least absolute shrinkage and selection operator (LASSO) method was used for dimension reduction, feature selection, and radiomics signature building. Multivariable logistic regression analysis was used to develop the radiomics nomogram. An independent sample t-test and chi-squared test were used to compare the differences in continuous and categorical variables, respectively.

RESULTS

The radiomic signature allowed a good discrimination between the LNM and non-LNM groups, with a C-index of 0.856 (95% confidence interval [CI], 0.794-0.918) in the primary cohort and 0.883 (95% CI, 0.809-0.957) in the validation cohort. Additionally, the radiomics nomogram also had a good discriminating performance and yielded good calibration both in the primary and validation cohorts (C-index, 0.882 [95% CI, 0.827-0.937], C-index, 0.893 [95% CI, 0.822-0.964], respectively). Decision curve analysis demonstrated that the radiomics nomogram was clinically useful.

DATA CONCLUSION

A radiomics nomogram was developed by incorporating the radiomics signature with the MRI-reported LN status and FIGO stage. This nomogram might be used to facilitate the individualized prediction of LNM in patients with early-stage cervical cancer.

LEVEL OF EVIDENCE

3 TECHNICAL EFFICACY STAGE: 2 J. Magn. Reson. Imaging 2020;52:885-896.

Predicting response to electroconvulsive therapy combined with antipsychotics in schizophrenia using multi-parametric magnetic resonance imaging

Electroconvulsive therapy (ECT) has been shown to be effective in schizophrenia, particularly when rapid symptom reduction is needed or in cases of resistance to drug treatment. However, there are no markers available to predict response to ECT. Here, we examine whether multi-parametric magnetic resonance imaging (MRI)-based radiomic features can predict response to ECT for individual patients. A total of 57 treatment-resistant schizophrenia patients, or schizophrenia patients with an acute episode or suicide attempts were randomly divided into primary (42 patients) and test (15 patients) cohorts. We collected T1-weighted structural MRI and diffusion MRI for 57 patients before receiving ECT and extracted 600 radiomic features for feature selection and prediction. To predict a continuous improvement in symptoms (ΔPANSS), the prediction process was performed with a support vector regression model based on a leave-one-out cross-validation framework in primary cohort and was tested in test cohort. The multi-parametric MRI-based radiomic model, including four structural MRI feature from left inferior frontal gyrus, right insula, left middle temporal gyrus and right superior temporal gyrus respectively and six diffusion MRI features from tracts connecting frontal or temporal gyrus possessed a low root mean square error of 15.183 in primary cohort and 14.980 in test cohort. The Pearson's correlation coefficients between predicted and actual values were 0.671 and 0.777 respectively. These results demonstrate that multi-parametric MRI-based radiomic features may predict response to ECT for individual patients. Such features could serve as prognostic neuroimaging biomarkers that provide a critical step toward individualized treatment response prediction in schizophrenia.

Development and assessment of machine learning algorithms for predicting remission after transsphenoidal surgery among patients with acromegaly

PURPOSE

Preoperative prediction of transsphenoidal surgical (TSS) response is important for determining individual treatment strategies for acromegaly. There is currently no accurate predictive model for TSS response for acromegaly. The current study sought to develop and validate machine learning (ML)-based models for preoperative prediction of TSS response for acromegaly.

METHODS

Six hundred sixty-eight patients with acromegaly were enrolled and divided into training (n = 534) and text datasets (n = 134) in this retrospective, data mining and ML study. The forward search algorithm was used to select features, and six ML algorithms were applied to construct TSS response prediction models. The performance of these ML models was validated using receiver operating characteristics analysis. Model calibration, discrimination ability, and clinical usefulness were also assessed.

RESULTS

Three hundred forty-nine (52.2%) patients achieved postoperative remission criteria and exhibited good TSS response. A univariate analysis was conducted and eight features, including age, hypertension, ophthalmic disorders, GH, IGF-1, nadir GH, maximal tumor diameter, and Knosp grade, were significantly associated with the TSS response in patients with acromegaly. After feature selection, the gradient boosting decision tree (GBDT), which was constructed with the eight significant features showed the best favorable discriminatory ability both the training (AUC = 0.8555) and validation (AUC = 0.8178) cohorts. The GBDT model showed good discrimination ability and calibration, with the highest levels of accuracy and specificity, and provided better estimates of TTS responses of patients with acromegaly compared with using only the Knosp grade. Decision curve analysis confirmed that the model was clinically useful.

CONCLUSIONS

ML-based models could aid neurosurgeons in the preoperative prediction of TTS response for patients with acromegaly, and could contribute to determining individual treatment strategies.

Multiple Level CT Radiomics Features Preoperatively Predict Lymph Node Metastasis in Esophageal Cancer: A Multicentre Retrospective Study

Background: Lymph node (LN) metastasis is the most important prognostic factor in esophageal squamous cell carcinoma (ESCC). Traditional clinical factor and existing methods based on CT images are insufficiently effective in diagnosing LN metastasis. A more efficient method to predict LN status based on CT image is needed.

Methods: In this multicenter retrospective study, 411 patients with pathologically confirmed ESCC were registered from two hospitals. Quantitative image features including handcrafted-, computer vision-(CV-), and deep-features were extracted from preoperative arterial phase CT images for each patient. A handcrafted-, CV-, and deep-radiomics signature were built, respectively. Then, multiple radiomics models were constructed by merging independent clinical risk factor into radiomics signatures. The performance of models were evaluated with respect to the discrimination, calibration, and clinical usefulness. Finally, an independent external validation cohort was used to validate the model's predictive performance.

Results: Five, seven, and nine features were selected for building handcrafted-, CV-, and deep-radiomics signatures from extracted features, respectively. Those signatures were statistically significant different between LN-positive and LN-negative patients in all cohorts (p < 0.001). The developed multiple level CT radiomics model that integrates multiple radiomics signatures with clinical risk factor, was superior to traditional clinical factors and the results reported by existing methods, and achieved satisfactory discrimination performance with C-statistic of 0.875 in development cohort, 0.874 in internal validation cohort and 0.840 in independent external validation cohort. Nomogram and decision curve analysis (DCA) further confirmed our method may serve as an effective tool for clinicians to evaluate the risk of LN metastasis in patients with ESCC and further choose treatment strategy.

Conclusions: The proposed multiple level CT radiomics model which integrate multiple level radiomics features into clinical risk factor can be used for preoperative predicting LN metastasis of patients with ESCC.

Development and Interpretation of Multiple Machine Learning Models for Predicting Postoperative Delayed Remission of Acromegaly Patients During Long-Term Follow-Up

Background: Some patients with acromegaly do not reach the remission standard in the short term after surgery but achieve remission without additional postoperative treatment during long-term follow-up; this phenomenon is defined as postoperative delayed remission (DR). DR may complicate the interpretation of surgical outcomes in patients with acromegaly and interfere with decision-making regarding postoperative adjuvant therapy. Objective: We aimed to develop and validate machine learning (ML) models for predicting DR in acromegaly patients who have not achieved remission within 6 months of surgery. Methods: We enrolled 306 acromegaly patients and randomly divided them into training and test datasets. We used the recursive feature elimination (RFE) algorithm to select features and applied six ML algorithms to construct DR prediction models. The performance of these ML models was validated using receiver operating characteristics analysis. We used permutation importance, SHapley Additive exPlanations (SHAP), and local interpretable model-agnostic explanation (LIME) algorithms to determine the importance of the selected features and interpret the ML models. Results: Fifty-five (17.97%) acromegaly patients met the criteria for DR, and five features (post-1w rGH, post-1w nGH, post-6m rGH, post-6m IGF-1, and post-6m nGH) were significantly associated with DR in both the training and the test datasets. After the RFE feature selection, the XGboost model, which comprised the 15 important features, had the greatest discriminatory ability (area under the curve = 0.8349, sensitivity = 0.8889, Youden's index = 0.6842). The XGboost model showed good discrimination ability and provided significantly better estimates of DR of patients with acromegaly compared with using only the Knosp grade. The results obtained from permutation importance, SHAP, and LIME algorithms showed that post-6m IGF-1 is the most important feature in XGboost algorithm prediction and showed the reliability and the clinical practicability of the XGboost model in DR prediction. Conclusions: ML-based models can serve as an effective non-invasive approach to predicting DR and could aid in determining individual treatment and follow-up strategies for acromegaly patients who have not achieved remission within 6 months of surgery.

Immune cell infiltration as a biomarker for the diagnosis and prognosis of digestive system cancer

The digestive system cancers are aggressive cancers with the highest mortality worldwide. In this study, we undertook a systematic investigation of the tumor immune microenvironment to identify diagnostic and prognostic biomarkers. The fraction of 22 immune cell types of patients were estimated using CIBERSORT. The least absolute shrinkage and selection operator (LASSO) analysis was carried out to identify important immune predictors. By comparing immune cell compositions in 801 tumor samples and 46 normal samples, we constructed the diagnostic immune score (DIS), showing high specificity and sensitivity in the training (area under the receiver operating characteristic curve [AUC] = 0.929), validation (AUC = 0.935), and different cancer type cohorts (AUC > 0.70 for all). We also established the prognostic immune score (PIS), which was an effective prognostic factor for relapse‐free survival in training, validation, and entire cohorts (P < .05). In addition, PIS provided a higher net benefit than TNM stage. A composite nomogram was built based on PIS and patients' clinical information with well‐fitted calibration curves (c‐index = 0.84). We further used other cohorts from Gene Expression Omnibus databases and obtained similar results, confirming the reliability and validity of the DIS and PIS. In addition, the unsupervised clustering analysis using immune cell proportions revealed 6 immune subtypes, suggesting that the immune types defined as having relatively high levels of M0 or/and M1 macrophages were the high‐risk subtypes of relapse. In conclusion, this study comprehensively analyzed the tumor immune microenvironment and identified DIS and PIS for digestive system cancers.

A Pre-operative Nomogram for Prediction of Lymph Node Metastasis in Bladder Urothelial Carcinoma

The status of lymph node (LN) metastases plays a decisive role in the selection of surgical procedures and post-operative treatment. Several histopathologic features, known as predictors of LN metastasis, are commonly available post-operatively. Medical imaging improved pre-operative diagnosis, but the results are not fully satisfactory due to substantial false positives. Thus, a reliable and robust method for pre-operative assessment of LN status is urgently required. We developed a prediction model in a training set from the TCGA-BLCA cohort including 196 bladder urothelial carcinoma samples with confirmed LN metastasis status. Least absolute shrinkage and selection operator (LASSO) regression was harnessed for dimension reduction, feature selection, and LNM signature building. Multivariable logistic regression was used to develop the prognostic model, incorporating the LNM signature, and a genomic mutation of MLL2, and was presented with a LNM nomogram. The performance of the nomogram was assessed with respect to its calibration, discrimination, and clinical usefulness. Internal validation was evaluated by the testing set from the TCGA cohort and independent validation was assessed by two independent cohorts. The LNM signature, which consisted of 48 selected features, was significantly associated with LN status (p < 0.005 for both the training and testing sets of the TCGA cohort). Predictors contained in the individualized prediction nomogram included the LNM signature and MLL2 mutation status. The model demonstrated good discrimination, with an area under the curve (AUC) of 98.7% (85.3% for testing set) and good calibration with p = 0.973 (0.485 for testing set) in the Hosmer-Lemeshow goodness of fit test. Decision curve analysis demonstrated that the LNM nomogram was clinically useful. This study presents a pre-operative nomogram incorporating a LNM signature and a genomic mutation, which can be conveniently utilized to facilitate pre-operative individualized prediction of LN metastasis in patients with bladder urothelial carcinoma.

Radiomics signature on 3T dynamic contrast-enhanced magnetic resonance imaging for estrogen receptor-positive invasive breast cancers: Preliminary results for correlation with Oncotype DX recurrence scores

To evaluate the ability of a radiomics signature based on 3T dynamic contrast-enhanced (DCE) magnetic resonance imaging (MRI) to distinguish between low and non-low Oncotype DX (OD) risk groups in estrogen receptor (ER)-positive invasive breast cancers.Between May 2011 and March 2016, 67 women with ER-positive invasive breast cancer who performed preoperative 3T MRI and OD assay were included. We divided the patients into low (OD recurrence score [RS] <18) and non-low risk (RS ≥18) groups. Extracted radiomics features included 8 morphological, 76 histogram-based, and 72 higher-order texture features. A radiomics signature (Rad-score) was generated using the least absolute shrinkage and selection operator (LASSO). Univariate and multivariate logistic regression analyses were performed to investigate the association between clinicopathologic factors, MRI findings, and the Rad-score with OD risk groups, and the areas under the receiver operating characteristic curves (AUC) were used to assess classification performance of the Rad-score.The Rad-score was constructed for each tumor by extracting 10 (6.3%) from 158 radiomics features. A higher Rad-score (odds ratio [OR], 65.209; P <.001), Ki-67 expression (OR, 17.462; P = .007), and high p53 (OR = 8.449; P = .077) were associated with non-low OD risk. The Rad-score classified low and non-low OD risk with an AUC of 0.759.The Rad-score showed the potential for discrimination between low and non-low OD risk groups in patients with ER-positive invasive breast cancers.

MRI-based Radiomics nomogram to detect primary rectal cancer with synchronous liver metastases

Synchronous liver metastasis (SLM) remains a major challenge for rectal cancer. Early detection of SLM is a key factor to improve the survival rate of rectal cancer. In this radiomics study, we predicted the SLM based on the radiomics of primary rectal cancer. A total of 328 radiomics features were extracted from the T2WI images of 194 patients. The least absolute shrinkage and selection operator (LASSO) regression was used to reduce the feature dimension and to construct the radiomics signature. after LASSO, principal component analysis (PCA) was used to sort the features of the surplus characteristics, and selected the features of the total contribution of 85%. Then the prediction model was built by linear regression, and the decision curve analysis was used to judge the net benefit of LASSO and PCA. In addition, we used two independent cohorts for training (n = 135) and validation (n = 159). We found that the model based on LASSO dimensionality construction had the maximum net benefit (in the training set (AUC [95% confidence interval], 0.857 [0.787–0.912]) and in the validation set (0.834 [0.714–0.918]). The radiomics nomogram combined with clinical risk factors and LASSO features showed a good predictive performance in the training set (0.921 [0.862–0.961]) and validation set (0.912 [0.809–0.97]). Our study indicated that radiomics based on primary rectal cancer could provide a non-invasive way to predict the risk of SLM in clinical practice.

Machine Learning-Based Radiomics Predicts Radiotherapeutic Response in Patients With Acromegaly

Background: Prediction of radiotherapeutic response before radiotherapy could help determine individual treatment strategies for patients with acromegaly. Objective: To develop and validate a machine-learning-based multiparametric MRI radiomics model to non-invasively predict radiotherapeutic response in patients with acromegaly. Methods: This retrospective study included 57 acromegaly patients who underwent postoperative radiotherapy between January 2008 and January 2016. Manual lesion segmentation and radiomics analysis were performed on each pituitary adenoma, and 1561 radiomics features were extracted from each sequence. A radiomics signature was built with a support vector machine using leave-one-out cross-validation for feature selection. Multivariable logistic regression analysis was used to select appropriate clinicopathological features to construct a clinical model, which was then combined with the radiomics signature to construct a radiomics model. The performance of this radiomic model was assessed using receiver operating characteristics (ROC) analysis and its calibration, discriminating ability, clinical usefulness. Results: At 3-years after radiotherapy, 25 patients had achieved remission and 32 patients had not. The clinical model incorporating seven clinical features had an area under the ROC (AUC) of 0.86 for predicting radiotherapeutic response, and performed better than any single clinical feature. The radiomics signature constructed with six radiomics features had a significantly higher AUC of 0.92. The radiomics model showed good discrimination abilities and calibration, with an AUC of 0.96. Decision curve analysis confirmed the clinical utility of the radiomics model. Conclusion: Using pre-radiotherapy clinical and MRI data, we developed a radiomics model with favorable performance for individualized non-invasive prediction of radiotherapeutic response, which may help in identifying acromegaly patients who are likely to benefit from radiotherapy.

A radiomics-based nomogram for the preoperative prediction of posthepatectomy liver failure in patients with hepatocellular carcinoma

OBJECTIVES

To develop and validate a radiomics-based nomogram for the preoperative prediction of posthepatectomy liver failure (PHLF) in patients with hepatocellular carcinoma (HCC).

METHODS

One hundred twelve consecutive HCC patients who underwent hepatectomy were included in the study pool (training cohort: n = 80, validation cohort: n = 32), and another 13 patients were included in a pilot prospective analysis. A total of 713 radiomics features were extracted from portal-phase computed tomography (CT) images. A logistic regression was used to construct a radiomics score (Rad-score). Then a nomogram, including Rad-score and other risk factors, was built with a multivariate logistic regression model. The discrimination, calibration and clinical utility of nomogram were evaluated.

RESULTS

The Rad-score could predict PHLF with an AUC of 0.822 (95% CI, 0.726-0.917) in the training cohort and of 0.762 (95% CI, 0.576-0.948) in the validation cohort; however, the approach could not completely outmatch the existing methods (CP [Child-Pugh], MELD [Model of End Stage Liver Disease], ALBI [albumin-bilirubin]). The individual predictive nomogram that included the Rad-score, MELD and performance status (PS) showed better discrimination with an AUC of 0.864 (95% CI, 0.786-0.942), which was higher than the AUCs of the conventional methods (nomogram vs CP, MELD, and ALBI at P < 0.001, P < 0.005, and P < 0.005, respectively). In the validation cohort, the nomogram discrimination was also superior to those of the other three methods (AUC: 0.896; 95% CI, 0.774-1.000). The calibration curves showed good agreement in both cohorts, and the decision curve analysis of the entire cohort revealed that the nomogram was clinically useful. A pilot prospective analysis showed that the radiomics nomogram could predict PHLF with an AUC of 0.833 (95% CI, 0.591-1.000).

CONCLUSIONS

A nomogram based on the Rad-score, MELD, and PS can predict PHLF.

Radiomics improves efficiency for differentiating subclinical pheochromocytoma from lipid-poor adenoma: a predictive, preventive and personalized medical approach in adrenal incidentalomas

Objectives

This study aims to define a radiomic signature for pre-operative differentiation between subclinical pheochromocytoma (sPHEO) and lipid-poor adrenal adenoma (LPA) in adrenal incidentaloma. The goal was to apply a predictive, preventive, and personalized medical approach to the management of adrenal tumors.

Patients and methods

This retrospective study consisted of 265 consecutive patients (training cohort, 212 (LPA, 145; sPHEO, 67); validation cohort, 53 (LPA, 36; sPHEO, 17)). Computed tomography (CT) imaging features were evaluated, including long diameter (LD), short diameter (SD), pre-enhanced CT value (CT_pre), enhanced CT value (CT_post), shape, homogeneity, necrosis or cystic degeneration (N/C). Radiomic features were extracted and then were used to construct a radiomic signature (Rad-score) and radiomic nomogram. The area under the receiver operating characteristic curve (AUC) was used to evaluate their performance.

Results

Sixteen of three hundred forty candidate features were used to build a radiomic signature. The signature was significantly different between the sPHEO and LPA groups (AUC: training, 0.907; validation, 0.902). The radiomic nomogram based on enhanced CT features (M1) consisted of Rad-score, LD, SD, CT_pre, shape, homogeneity and N/C (AUC: training, 0.957; validation, 0.967). The pre-enhanced CT features based radiomic nomogram (M2) included Rad-score, LD, SD, CT_pre, shape, and homogeneity (AUC: training, 0.955; validation, 0.958).

Conclusions

Our radiomic nomograms based on pre-enhanced and enhanced CT images distinguished sPHEO from LPA. In addition, the promising result using pre-enhanced CT images for predictive diagnostics is important because patients could avoid the additional radiation and risk associated with enhanced CT.

Radiomics Signature on Magnetic Resonance Imaging: Association with Disease-Free Survival in Patients with Invasive Breast Cancer

Purpose: To develop a radiomics signature based on preoperative MRI to estimate disease-free survival (DFS) in patients with invasive breast cancer and to establish a radiomics nomogram that incorporates the radiomics signature and MRI and clinicopathological findings.Experimental Design: We identified 294 patients with invasive breast cancer who underwent preoperative MRI. Patients were randomly divided into training (n = 194) and validation (n = 100) sets. A radiomics signature (Rad-score) was generated using an elastic net in the training set, and the cutoff point of the radiomics signature to divide the patients into high- and low-risk groups was determined using receiver-operating characteristic curve analysis. Univariate and multivariate Cox proportional hazards model and Kaplan-Meier analysis were used to determine the association of the radiomics signature, MRI findings, and clinicopathological variables with DFS. A radiomics nomogram combining the Rad-score and MRI and clinicopathological findings was constructed to validate the radiomic signatures for individualized DFS estimation.Results: Higher Rad-scores were significantly associated with worse DFS in both the training and validation sets (P = 0.002 and 0.036, respectively). The radiomics nomogram estimated DFS [C-index, 0.76; 95% confidence interval (CI); 0.74-0.77] better than the clinicopathological (C-index, 0.72; 95% CI, 0.70-0.74) or Rad-score-only nomograms (C-index, 0.67; 95% CI, 0.65-0.69).Conclusions: The radiomics signature is an independent biomarker for the estimation of DFS in patients with invasive breast cancer. Combining the radiomics nomogram improved individualized DFS estimation. Clin Cancer Res; 24(19); 4705-14. ©2018 AACR.

CT-based radiomic model predicts high grade of clear cell renal cell carcinoma

PURPOSE

To compare the predictive models that can incorporate a set of CT image features for preoperatively differentiating the high grade (Fuhrman III-IV) from low grade (Fuhrman I-II) clear cell renal cell carcinoma (ccRCC).

MATERIAL AND METHODS

One hundred and fourteen patients with ccRCC treated with a partial or radical nephrectomy were enrolled in the training cohort. The six non-texture features, including Pseudocapsule, Round mass, maximal tumor diameter (Diameter_max), intratumoral artery (Artery_tumor), enhancement value of the tumor (TEV) and relative TEV (rTEV), were assessed for each tumor. The texture features were extracted from the CT images of the section with the largest area of renal mass at both corticomedullary and nephrographic phases. The least absolute shrinkage and selection operator (LASSO) was used to screen the most valuable texture features to calculate a texture score (Texture-score) for each patient. A logistic regression model was used in the training cohort to discriminate the high from low grade ccRCC at nephrectomy. The predictors would include all non-texture features in Model 1, all non-texture features and Texture-score in Model 2, and Texture-score in Model 3. The performance of the predictive models were tested and compared in an independent validation cohort composed of 92 cases with ccRCC.

RESULTS

Inter-rater agreement was good for each non-texture feature and Texture-score (the concordance correlation coefficient or Kappa coefficient > 0.70). The Texture-score was calculated via a linear combination of the 4 selected texture features. The three models shown good discrimination of the high from low grade ccRCC in the training cohort and the area under receiver operating characteristic curve (AUC) was 0.826 in Mode 1, 0.878 in Model 2 and 0.843 in Model 3, and a significant different AUC was found between Model 1 and Model 2. Application of the predictive models in the validation cohort still gave a discrimination (AUC > 0.670), and the Texture-score based models with or without the non-texture features (Model 2 and 3) showed a better discrimination of the high from low grade ccRCC (P < 0.05).

CONCLUSION

This study presented the Texture-score based models can facilitate the preoperative discrimination of the high from low grade ccRCC.

Prognostic value of an immunohistochemical signature in patients with esophageal squamous cell carcinoma undergoing radical esophagectomy

Here, we aimed to identify an immunohistochemical (IHC)‐based classifier as a prognostic factor in patients with esophageal squamous cell carcinoma (ESCC). A cohort of 235 patients with ESCC undergoing radical esophagectomy (with complete clinical and pathological information) were enrolled in the study. Using the least absolute shrinkage and selection operator (LASSO) regression model, we extracted six IHC features associated with progression‐free survival (PFS) and then built a classifier in the discovery cohort (n = 141). The prognostic value of this classifier was further confirmed in the validation cohort (n = 94). Additionally, we developed a nomogram integrating the IHC‐based classifier to predict the PFS. We used the IHC‐based classifier to stratify patients into high‐ and low‐risk groups. In the discovery cohort, 5‐year PFS was 22.4% (95% CI: 0.14–0.36) for the high‐risk group and 43.3% (95% CI: 0.32–0.58) for the low‐risk group (P = 0.00064), and in the validation cohort, 5‐year PFS was 20.58% (95% CI: 0.12–0.36) for the high‐risk group and 36.43% (95% CI: 0.22–0.60) for the low‐risk group (P = 0.0082). Multivariable analysis demonstrated that the IHC‐based classifier was an independent prognostic factor for predicting PFS of patients with ESCC. We further developed a nomogram integrating the IHC‐based classifier and clinicopathological risk factors (gender, American Joint Committee on Cancer staging, and vascular invasion status) to predict the 3‐ and 5‐year PFS. The performance of the nomogram was evaluated and proved to be clinically useful. Our 6‐IHC marker‐based classifier is a reliable prognostic tool to facilitate the individual management of patients with ESCC after radical esophagectomy.

Molecular Prognostication in Bladder Cancer

Clinical outcomes for patients with bladder cancer have largely remained unchanged over the last three decades despite improvements in surgical techniques, perioperative therapies, and postoperative management. Current management still heavily relies on pathologic staging that does not always reflect an individual patient's risk. The genesis and progression of bladder cancer is now increasingly recognized as being a result of alterations in several pathways that affect the cell cycle, apoptosis, cellular signaling, gene regulation, immune modulation, angiogenesis, and tumor cell invasion. Multiplexed assessment of biomarkers associated with alterations in these pathways offers novel insights into tumor behavior while identifying panels that are capable of reproducibly predicting patient outcomes. Future management of bladder cancer will likely incorporate such prognostic molecular models for risk stratification and treatment personalization.

Individualized prediction of perineural invasion in colorectal cancer: development and validation of a radiomics prediction model

Objective

To develop and validate a radiomics prediction model for individualized prediction of perineural invasion (PNI) in colorectal cancer (CRC).

Methods

After computed tomography (CT) radiomics features extraction, a radiomics signature was constructed in derivation cohort (346 CRC patients). A prediction model was developed to integrate the radiomics signature and clinical candidate predictors [age, sex, tumor location, and carcinoembryonic antigen (CEA) level]. Apparent prediction performance was assessed. After internal validation, independent temporal validation (separate from the cohort used to build the model) was then conducted in 217 CRC patients. The final model was converted to an easy-to-use nomogram.

Results

The developed radiomics nomogram that integrated the radiomics signature and CEA level showed good calibration and discrimination performance [Harrell's concordance index (c-index): 0.817; 95% confidence interval (95% CI): 0.811-0.823]. Application of the nomogram in validation cohort gave a comparable calibration and discrimination (c-index: 0.803; 95% CI: 0.794-0.812).

Conclusions

Integrating the radiomics signature and CEA level into a radiomics prediction model enables easy and effective risk assessment of PNI in CRC. This stratification of patients according to their PNI status may provide a basis for individualized auxiliary treatment.

The development and validation of a CT-based radiomics signature for the preoperative discrimination of stage I-II and stage III-IV colorectal cancer

Objectives

To investigative the predictive ability of radiomics signature for preoperative staging (I-IIvs.III-IV) of primary colorectal cancer (CRC).

Methods

This study consisted of 494 consecutive patients (training dataset: n=286; validation cohort, n=208) with stage I–IV CRC. A radiomics signature was generated using LASSO logistic regression model. Association between radiomics signature and CRC staging was explored. The classification performance of the radiomics signature was explored with respect to the receiver operating characteristics(ROC) curve.

Results

The 16-feature-based radiomics signature was an independent predictor for staging of CRC, which could successfully categorize CRC into stage I-II and III-IV (p <0.0001) in training and validation dataset. The median of radiomics signature of stage III-IV was higher than stage I-II in the training and validation dataset. As for the classification performance of the radiomics signature in CRC staging, the AUC was 0.792(95%CI:0.741-0.853) with sensitivity of 0.629 and specificity of 0.874. The signature in the validation dataset obtained an AUC of 0.708(95%CI:0.698-0.718) with sensitivity of 0.611 and specificity of 0.680.

Conclusions

A radiomics signature was developed and validated to be a significant predictor for discrimination of stage I-II from III-IV CRC, which may serve as a complementary tool for the preoperative tumor staging in CRC.

A Radiomics Nomogram for the Preoperative Prediction of Lymph Node Metastasis in Bladder Cancer

Purpose: To develop and validate a radiomics nomogram for the preoperative prediction of lymph node (LN) metastasis in bladder cancer.Experimental Design: A total of 118 eligible bladder cancer patients were divided into a training set (n = 80) and a validation set (n = 38). Radiomics features were extracted from arterial-phase CT images of each patient. A radiomics signature was then constructed with the least absolute shrinkage and selection operator algorithm in the training set. Combined with independent risk factors, a radiomics nomogram was built with a multivariate logistic regression model. Nomogram performance was assessed in the training set and validated in the validation set. Finally, decision curve analysis was performed with the combined training and validation set to estimate the clinical usefulness of the nomogram.Results: The radiomics signature, consisting of nine LN status-related features, achieved favorable prediction efficacy. The radiomics nomogram, which incorporated the radiomics signature and CT-reported LN status, also showed good calibration and discrimination in the training set [AUC, 0.9262; 95% confidence interval (CI), 0.8657-0.9868] and the validation set (AUC, 0.8986; 95% CI, 0.7613-0.9901). The decision curve indicated the clinical usefulness of our nomogram. Encouragingly, the nomogram also showed favorable discriminatory ability in the CT-reported LN-negative (cN0) subgroup (AUC, 0.8810; 95% CI, 0.8021-0.9598).Conclusions: The presented radiomics nomogram, a noninvasive preoperative prediction tool that incorporates the radiomics signature and CT-reported LN status, shows favorable predictive accuracy for LN metastasis in patients with bladder cancer. Multicenter validation is needed to acquire high-level evidence for its clinical application. Clin Cancer Res; 23(22); 6904-11. ©2017 AACR.

Radiomics Signature: A Potential Biomarker for the Prediction of Disease-Free Survival in Early-Stage (I or II) Non-Small Cell Lung Cancer

Purpose To develop a radiomics signature to estimate disease-free survival (DFS) in patients with early-stage (stage I-II) non-small cell lung cancer (NSCLC) and assess its incremental value to the traditional staging system and clinical-pathologic risk factors for individual DFS estimation. Materials and Methods Ethical approval by the institutional review board was obtained for this retrospective analysis, and the need to obtain informed consent was waived. This study consisted of 282 consecutive patients with stage IA-IIB NSCLC. A radiomics signature was generated by using the least absolute shrinkage and selection operator, or LASSO, Cox regression model. Association between the radiomics signature and DFS was explored. Further validation of the radiomics signature as an independent biomarker was performed by using multivariate Cox regression. A radiomics nomogram with the radiomics signature incorporated was constructed to demonstrate the incremental value of the radiomics signature to the traditional staging system and other clinical-pathologic risk factors for individualized DFS estimation, which was then assessed with respect to calibration, discrimination, reclassification, and clinical usefulness. Results The radiomics signature was significantly associated with DFS, independent of clinical-pathologic risk factors. Incorporating the radiomics signature into the radiomics-based nomogram resulted in better performance (P < .0001) for the estimation of DFS (C-index: 0.72; 95% confidence interval [CI]: 0.71, 0.73) than with the clinical-pathologic nomogram (C-index: 0.691; 95% CI: 0.68, 0.70), as well as a better calibration and improved accuracy of the classification of survival outcomes (net reclassification improvement: 0.182; 95% CI: 0.02, 0.31; P = .02). Decision curve analysis demonstrated that in terms of clinical usefulness, the radiomics nomogram outperformed the traditional staging system and the clinical-pathologic nomogram. Conclusion The radiomics signature is an independent biomarker for the estimation of DFS in patients with early-stage NSCLC. Combination of the radiomics signature, traditional staging system, and other clinical-pathologic risk factors performed better for individualized DFS estimation in patients with early-stage NSCLC, which might enable a step forward precise medicine. ^© RSNA, 2016 Online supplemental material is available for this article.