Breaking boundaries in radiology: redefining AI diagnostics via raw data ahead of reconstruction.

Objective.In the realm of utilizing artificial intelligence (AI) for medical image analysis, the paradigm of 'signal-image-knowledge' has remained unchanged. However, the process of 'signal to image' inevitably introduces information distortion, ultimately leading to irrecoverable biases in the 'image to knowledge' process. Our goal is to skip reconstruction and build a diagnostic model directly from the raw data (signal).Approach. This study focuses on computed tomography (CT) and its raw data (sinogram) as the research subjects. We simulate the real-world process of 'human-signal-image' using the workflow 'CT-simulated data- reconstructed CT,' and we develop a novel AI predictive model directly targeting raw data (RCTM). This model comprises orientation, spatial, and global analysis modules, embodying the fusion of local to global information extraction from raw data. We selected 1994 patients with retrospective cases of solid lung nodules and modeled different types of data.Main results. We employed predefined radiomic features to assess the diagnostic feature differences caused by reconstruction. The results indicated that approximately 14% of the features had Spearman correlation coefficients below 0.8. These findings suggest that despite the increasing maturity of CT reconstruction algorithms, they still introduce perturbations to diagnostic features. Moreover, our proposed RCTM achieved an area under the curve (AUC) of 0.863 in the diagnosis task, showcasing a comprehensive superiority over models constructed from secondary reconstructed CTs (0.840, 0.822, and 0.825). Additionally, the performance of RCTM closely resembled that of models constructed from original CT scans (0.868, 0.878, and 0.866).Significance. The diagnostic and therapeutic approach directly based on CT raw data can enhance the precision of AI models and the concept of 'signal-to-image' can be extended to other types of imaging. AI diagnostic models tailored to raw data offer the potential to disrupt the traditional paradigm of 'signal-image-knowledge', opening up new avenues for more accurate medical diagnostics.

A novel tumor purity and immune infiltration-related model for predicting distant metastasis-free survival in prostate cancer.

BACKGROUND: umor cells, immune cells and stromal cells jointly modify tumor development and progression. We aim to explore the potential effects of tumor purity on the immune microenvironment, genetic landscape and prognosis in prostate cancer (PCa). METHODS: Tumor purity of prostate cancer patients was extracted from The cancer genome atlas (TCGA). Immune cellular proportions were calculated by the CIBERSORT. To identify critical modules related to tumor purity, we used weighted gene co-expression network analysis (WGCNA). Using STRING and Cytoscape, protein-protein interaction (PPI) networks were constructed and analyzed. A Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway, Disease Ontology (DO), and Gene Set Enrichment Analysis (GSEA) enrichment analysis of identified modules was conducted. To identify the expression of key genes at protein levels, we used the Human Protein Atlas (HPA) platform. RESULTS: A model of tumor purity score (TPS) was constructed in the gene expression omnibus series (GSE) 116,918 cohort. TCGA cohort served as a validation set and was employed to validate the TPS. TPS model, as an independent prognostic factor of distant metastasis-free survival (DMFS) in PCa. Patients had higher tumor purity and better prognosis in the low-TPS group. Tumor purity was related to the infiltration of mast cells and macrophage cells positively, whereas related to the infiltration of dendritic cells, T cells and B cells negatively in PCa. The nomogram based on TPS, Age, Gleason score and T stage had a good predictive value and could evaluate the prognosis of PCa metastasis. GO and KEGG enrichment analyses showed that hub genes mainly participate in T cell activation and T-helper lymphocytes (TH) differentiation. Hub genes were mainly enriched in primary immunodeficiency disease, according to DO analysis. SLAMF8 was identified as the most critical gene by Cytoscape and HPA analysis. CONCLUSIONS: Dynamic changes in the immune microenvironment associated with tumor purity could correlate with a poor DMFS of low-purity PCa. The TPS can predict the DMFS of PCa. In addition, prostate cancer metastases may be related to immunosuppression caused by a disorder of the immune microenvironment.

Metabolic-related gene signature model forecasts biochemical relapse in primary prostate cancer.

Metabolism plays an important role in the pathogenesis of prostate cancer (PCa). Hence, we explored candidate metabolic-related genes attributed to biochemical relapse (BCR) of PCa. Gene expression profile and clinical parameters were downloaded from GSE70769 as a "training set". Using univariate Cox and LASSO-COX regression models, risk scores (RSs) were constructed. Kaplan-Meier (K-M) survival and time-dependent receiver operating characteristic (t-ROC) curves were employed. Univariate and multivariate Cox models were utilized to validate prognostic factors for biochemical relapse-free survival (BCRFS). Nomogram was plotted to facilitate clinical application. The dataset obtained from GSE70768 served as "validation set". RSs were constructed by using 7 metabolic-related genes. RSs could significantly predict 1, 3, 5-year BCRFS (AUCs for training set: 0.810-0.836; AUC for validation set: 0.673-0.827). Nomograms could effectively predicted BCRFS (training set: C-index=0.831; validation set: C-index=0.737). RSs model is an independent prognostic factor for BCR, holding greater predictive value than traditional clinicopathological parameters. Clinical Relevance- We built the prognostic nomogram based on metabolic-related gene signatures and clinicopathological features. The nomogram might further optimize biochemical relapse risk stratification for prostate cancer patients with crucial accuracy.

Deep learning for predicting immunotherapeutic efficacy in advanced non-small cell lung cancer patients: a retrospective study combining progression-free survival risk and overall survival risk.

Background: Radiomics based on computed tomography (CT) images is potential in promoting individualized treatment of non-small cell lung cancer (NSCLC), however, its role in immunotherapy needs further exploration. The aim of this study was to develop a CT-based radiomics score to predict the efficacy of immune checkpoint inhibitor (ICI) monotherapy in patients with advanced NSCLC. Methods: Two hundred and thirty-six ICI-treated patients were retrospectively included and divided into a training cohort (n=188) and testing cohort (n=48) at a ratio of 8 to 2. The efficacy outcomes of ICI were evaluated based on overall survival (OS) and progression-free survival (PFS). We designed a survival network and combined it with a Cox regression model to obtain patients' OS risk score (OSRS) and PFS risk score (PFSRS). Results: Based on OSRS and PFSRS, patients were divided into high- and low-risk groups in the training cohort and the test cohort with distinctly different [training cohort, log-rank P<0.001, hazard ratio (HR): 4.14; test cohort, log-rank P=0.014, HR: 4.54] and PFS (training cohort, log-rank P<0.001, HR: 4.52; test cohort, log-rank P<0.001, HR: 6.64). Further joint evaluation of OSRS and PFSRS showed that both were significant in the Cox regression model (P<0.001), and multi-overall survival risk score (MOSRS) displayed more outstanding stratification capabilities than OSRS in both the training (P<0.001) and test cohorts (P=0.002). None of the clinical characteristics were significant in the Cox regression model, and the score that predicted the best immune response was not as good as the risk score from follow-up information in the performance of prognostic stratification. Conclusions: We developed a CT imaging-based score with the potential to become an independent prognostic factor to screen patients who would benefit from ICI treatment, which suggested that CT radiomics could be applied for individualized immunotherapy of NSCLC. Our findings should be further validated by future larger multicenter study.

Mining whole-lung information by artificial intelligence for predicting EGFR genotype and targeted therapy response in lung cancer: a multicohort study.

BACKGROUND: Epidermal growth factor receptor (EGFR) genotype is crucial for treatment decision making in lung cancer, but it can be affected by tumour heterogeneity and invasive biopsy during gene sequencing. Importantly, not all patients with an EGFR mutation have good prognosis with EGFR-tyrosine kinase inhibitors (TKIs), indicating the necessity of stratifying for EGFR-mutant genotype. In this study, we proposed a fully automated artificial intelligence system (FAIS) that mines whole-lung information from CT images to predict EGFR genotype and prognosis with EGFR-TKI treatment. METHODS: We included 18 232 patients with lung cancer with CT imaging and EGFR gene sequencing from nine cohorts in China and the USA, including a prospective cohort in an Asian population (n=891) and The Cancer Imaging Archive cohort in a White population. These cohorts were divided into thick CT group and thin CT group. The FAIS was built for predicting EGFR genotype and progression-free survival of patients receiving EGFR-TKIs, and it was evaluated by area under the curve (AUC) and Kaplan-Meier analysis. We further built two tumour-based deep learning models as comparison with the FAIS, and we explored the value of combining FAIS and clinical factors (the FAIS-C model). Additionally, we included 891 patients with 56-panel next-generation sequencing and 87 patients with RNA sequencing data to explore the biological mechanisms of FAIS. FINDINGS: FAIS achieved AUCs ranging from 0·748 to 0·813 in the six retrospective and prospective testing cohorts, outperforming the commonly used tumour-based deep learning model. Genotype predicted by the FAIS-C model was significantly associated with prognosis to EGFR-TKIs treatment (log-rank p<0·05), an important complement to gene sequencing. Moreover, we found 29 prognostic deep learning features in FAIS that were able to identify patients with an EGFR mutation at high risk of TKI resistance. These features showed strong associations with multiple genotypes (p<0·05, t test or Wilcoxon test) and gene pathways linked to drug resistance and cancer progression mechanisms. INTERPRETATION: FAIS provides a non-invasive method to detect EGFR genotype and identify patients with an EGFR mutation at high risk of TKI resistance. The superior performance of FAIS over tumour-based deep learning methods suggests that genotype and prognostic information could be obtained from the whole lung instead of only tumour tissues. FUNDING: National Natural Science Foundation of China.

A generative adversarial network-based CT image standardization model for predicting progression-free survival of lung cancer.

Progression-free survival (PFS) prediction using computed tomography (CT) images is important for treatment planning in lung cancer. However, the generalization ability of current analysis methods is usually affected by the scanning parameters of CT images, such as slice thickness and reconstruction kernel. In this paper, we proposed a generative adversarial network (GAN)-based model to convert heterogenous CT images into standardized CT images with uniform slice thickness and reconstruction kernel to increase the generalization of the predictive model. This model was trained in 173 patients with multiple CT sequences including both thin/thick voxel-spacing and sharp/soft reconstruction kernel. Afterward, we built a 3D-CNN model to predict the individualized 1year PFS of lung cancer using the standardized CT images in 281 patients. Finally, we evaluated the predictive model by 5-fold cross-validation and the mean area under the receiver operating characteristic curve (AUC). After transforming to the heterogenous CT images into the uniform thin-spacing and sharp kernel CT images, the AUC value of the 3D-CNN model improved from 0.614 to 0.686. Furthermore, this model can stratify the patients into high-risk and low-risk groups, where patients in these two groups showed significant difference in PFS (P < 0.001).

A Quarter-split Domain-adaptive Network for EGFR Gene Mutation Prediction in Lung Cancer by Standardizing Heterogeneous CT image.

Epidermal growth factor receptor (EGFR) gene mutation status is crucial for the treatment planning of lung cancer. The gold standard for detecting EGFR mutation status relies on invasive tumor biopsy and expensive gene sequencing. Recently, computed tomography (CT) images and deep learning have shown promising results in non-invasively predicting EGFR mutation in lung cancer. However, CT scanning parameters such as slice thickness vary largely between different scanners and centers, making the deep learning models very sensitive to noise and therefore not robust in clinical practice. In this study, we propose a novel QuarterNetadaptive model to predict EGFR mutation in lung cancer, which is robust to CT images of different thicknesses. We propose two components: 1) a quarter-split network to sequentially learn local lung features from different lung lobes and global lung features; 2) a domain adaptive strategy to learn CT thickness-invariant features. Furthermore, we collected a large dataset including 1413 patients with both EGFR gene sequencing and CT images of various thicknesses to evaluate the performance of the proposed model. Finally, the QuarterNetadaptive model achieved AUC over 0.88 regarding CT images of different thicknesses, which improves largely than state-of-the-art methods.Clinical relevance-We proposed a non-invasive model to detect EGFR gene mutation in lung cancer, which is robust to CT images of different thicknesses and can assist lung cancer treatment planning.

Gene signatures predict biochemical recurrence-free survival in primary prostate cancer patients after radical therapy.

BACKGROUND: This study evaluated the predictive value of gene signatures for biochemical recurrence (BCR) in primary prostate cancer (PCa) patients. METHODS: Clinical features and gene expression profiles of PCa patients were attained from Gene Expression Omnibus (GEO) and The Cancer Genome Atlas (TCGA) datasets, which were further classified into a training set (n = 419), a validation set (n = 403). The least absolute shrinkage and selection operator Cox (LASSO-Cox) method was used to select discriminative gene signatures in training set for biochemical recurrence-free survival (BCRFS). Selected gene signatures established a risk score system. Univariate and multivariate analyses of prognostic factors about BCRFS were performed using the Cox proportional hazards regression models. A nomogram based on multivariate analysis was plotted to facilitate clinical application. Kyoto Encyclopedia of Gene and Genomes (KEGG) and Gene Ontology (GO) analyses were then executed for differentially expressed genes (DEGs). RESULTS: Notably, the risk score could significantly identify BCRFS by time-dependent receiver operating characteristic (t-ROC) curves in the training set (3-year area under the curve (AUC) = 0.820, 5-year AUC = 0.809) and the validation set (3-year AUC = 0.723, 5-year AUC = 0.733). CONCLUSIONS: Clinically, the nomogram model, which incorporates Gleason score and the risk score, could effectively predict BCRFS and potentially be utilized as a useful tool for the screening of BCRFS in PCa.

3D Deep Learning Model for the Pretreatment Evaluation of Treatment Response in Esophageal Carcinoma: A Prospective Study (ChiCTR2000039279).

PURPOSE: To develop and validate a pretreatment computed tomography (CT)-based deep-learning (DL) model for predicting the treatment response to concurrent chemoradiation therapy (CCRT) among patients with locally advanced thoracic esophageal squamous cell carcinoma (TESCC). METHODS AND MATERIALS: We conducted a prospective, multicenter study on the therapeutic efficacy of CCRT among TESCC patients across 9 hospitals in China (ChiCTR2000039279). A total of 306 patients with locally advanced TESCC diagnosed by histopathology from August 2015 to May 2020 were included in this study. A 3-dimensional DL radiomics model (3D-DLRM) was developed and validated based on pretreatment CT images to predict the response to CCRT. Furthermore, the prediction performance of the newly developed 3D-DLRM was analyzed according to 3 categories: radiation therapy plan, radiation field, and prescription dose used. RESULTS: The 3D-DLRM achieved good prediction performance, with areas under the receiver operating characteristic curve of 0.897 (95% confidence interval, 0.840-0.959) for the training cohort and 0.833 (95% confidence interval, 0.654-1.000) for the validation cohort. Specifically, the 3D-DLRM accurately predicted patients who would not respond to CCRT, with a positive predictive value (PPV) of 100% for the validation cohort. Moreover, the 3D-DLRM performed well in all 3 categories, each with areas under the receiver operating characteristic curve of >0.8 and positive predictive values of approximately 100%. CONCLUSION: The proposed pretreatment CT-based 3D-DLRM provides a potential tool for predicting the response to CCRT among patients with locally advanced TESCC. With the help of precise pretreatment prediction, we may guide the individualized treatment of patients and improve survival.

Specific Borrmann classification in advanced gastric cancer by an ensemble multilayer perceptron network: a multicenter research.

PURPOSE: Borrmann classification in advanced gastric cancer (AGC) is necessarily associated with personalized surgical strategy and prognosis. But few radiomics research studies have focused on specific Borrmann classification, and there is yet no consensus regarding what machine learning methods should be the most effective. METHODS: A combined size of 889 AGC patients was retrospectively enrolled from two centers. Radiomic features were extracted from tumors manually delineated on preoperative computed tomography images. Two classification experiments (Borrmann I/II/III vs. IV and Borrmann II vs. III) were conducted. In each task, we combined three common feature selection methods and five typical machine learning classifiers to construct 15 basic classification models, and then fed the 15 predictions to a designed multilayer perceptron (MLP) network. RESULTS: In internal and external validation cohorts, the proposed ensemble MLP yielded good performance with area under curves of 0.767 and 0.702 for Borrmann I/II/III vs. IV, as well as 0.768 and 0.731 for Borrmann II vs. III. Considering the imbalanced distribution of four Borrmann types (I, 2.9%; II, 12.8%; III, 69.5%; IV, 14.7%), the ensemble MLP surpassed the overfitting barrier and attained fine specificity (0.667 and 0.750 for Borrmann I/II/III vs. IV; 0.714 and 0.620 for Borrmann II vs. III) and sensitivity (0.795 and 0.610 for Borrmann I/II/III vs. IV; 0.652 and 0.703 for Borrmann II vs. III). Also, survival analysis showed that patients could be significantly risk stratified by MLP predicted types in both experiments (p < 0.0001, log-rank test). CONCLUSIONS: This study proposed an MLP-based ensemble learning architecture, which could identify Borrmann type IV automatically and improve the differentiation of Borrmann type II from III. The study provided a new view for specific Borrmann classification in clinical practice.

Exploring the predictive value of additional peritumoral regions based on deep learning and radiomics: A multicenter study.

PURPOSE: The present study assessed the predictive value of peritumoral regions on three tumor tasks, and further explored the influence of peritumors with different sizes. METHODS: We retrospectively collected 333 samples of gastrointestinal stromal tumors from the Second Affiliated Hospital of Zhejiang University School of Medicine, and 183 samples of gastrointestinal stromal tumors from Tianjin Medical University Cancer Hospital. We also collected 211 samples of laryngeal carcinoma and 233 samples of nasopharyngeal carcinoma from the First Affiliated Hospital of Jinan University. The tasks of three tumor datasets were risk assessment (gastrointestinal stromal tumor), T3/T4 staging prediction (laryngeal carcinoma), and distant metastasis prediction (nasopharyngeal carcinoma), respectively. First, deep learning and radiomics were respectively used to construct peritumoral models, to study whether the peritumor had predictive value on three tumor datasets. Furthermore, we defined different sizes peritumors including fixed size (not considering tumor size) and adaptive size (according to average tumor radius) to explore the influence of peritumor of different sizes and types of tumors. Finally, we visualized the deep learning and radiomic models to observe the influence of the peritumor in three datasets. RESULTS: The performance of intra-peritumors are better than intratumors alone in three datasets. Specifically, the comparisons of area under receiver operating characteristic curve in the testing set between intra-peritumoral and intratumoral models are: 0.908 vs 0.873 (P value: 0.037) in gastrointestinal stromal tumor datasets, 0.796 vs 0.756 (P value: 0.188) in laryngeal carcinoma datasets and 0.660 vs 0.579 (P value: 0.431) in nasopharyngeal carcinoma datasets. Furthermore, for gastrointestinal stromal tumor datasets, deep learning is more stable to learn peritumors with both fixed and adaptive size than radiomics. For laryngeal carcinoma datasets, the intra-peritumoral radiomic model could make model performance more balanced. For nasopharyngeal carcinoma datasets, radiomics is also more suitable for modeling peritumors than deep learning. The size of the peritumor is critical in this task, and only the performance of 1.5 mm-4.5 mm peritumors is stable. CONCLUSIONS: Our results indicate that peritumors have additional predictive value in three tumor datasets through deep learning or radiomics. The definitions of the peritumoral region and artificial intelligence method also have great influence on the performance of the peritumor.

Identifying early gastric cancer under magnifying narrow-band images with deep learning: a multicenter study.

BACKGROUND AND AIMS: Narrow-band imaging with magnifying endoscopy (ME-NBI) has shown advantages in the diagnosis of early gastric cancer (EGC). However, proficiency in diagnostic algorithms requires substantial expertise and experience. In this study, we aimed to develop a computer-aided diagnostic model for EGM (EGCM) to analyze and assist in the diagnosis of EGC under ME-NBI. METHODS: A total of 1777 ME-NBI images from 295 cases were collected from 3 centers. These cases were randomly divided into a training cohort (n = 170), an internal test cohort (ITC, n = 73), and an external test cohort (ETC, n = 52). EGCM based on VGG-19 architecture (Visual Geometry Group [VGG], Oxford University, Oxford, UK) with a single fully connected 2-classification layer was developed through fine-tuning and validated on all cohorts. Furthermore, we compared the model with 8 endoscopists with varying experience. Primary comparison measures included accuracy, area under the receiver operating characteristic curve (AUC), sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). RESULTS: EGCM acquired AUCs of .808 in the ITC and .813 in the ETC. Moreover, EGCM achieved similar predictive performance as the senior endoscopists (accuracy: .770 vs .755, P = .355; sensitivity: .792 vs .767, P = .183; specificity: .745 vs .742, P = .931) but better than the junior endoscopists (accuracy: .770 vs .728, P < .05). After referring to the results of EGCM, the average diagnostic ability of the endoscopists was significantly improved in terms of accuracy, sensitivity, PPV, and NPV (P < .05). CONCLUSIONS: EGCM exhibited comparable performance with senior endoscopists in the diagnosis of EGC and showed the potential value in aiding and improving the diagnosis of EGC by endoscopists.

Predicting response to immunotherapy in advanced non-small-cell lung cancer using tumor mutational burden radiomic biomarker.

BACKGROUND: Tumor mutational burden (TMB) is a significant predictor of immune checkpoint inhibitors (ICIs) efficacy. This study investigated the correlation between deep learning radiomic biomarker and TMB, including its predictive value for ICIs treatment response in patients with advanced non-small-cell lung cancer (NSCLC). METHODS: CT images from 327 patients with TMB data (TMB median=6.067 mutations per megabase (range: 0 to 42.151)) were retrospectively collected and randomly divided into a training (n=236), validation (n=26), and test cohort (n=65). We used 3D-densenet to estimate the target tumor area, which used 1020 deep learning features to distinguish High-TMB from Low-TMB patients and establish the TMB radiomic biomarker (TMBRB). The TMBRB was developed in the training cohort combined with validation cohort and evaluated in the test cohort. The predictive value of TMBRB was assessed in a cohort of 123 NSCLC patients who had received ICIs (survival median=462 days (range: 16 to 1128)). RESULTS: TMBRB discriminated between High-TMB and Low-TMB patients in the training cohort (area under the curve (AUC): 0.85, 95% CI: 0.84 to 0.87))and test cohort (AUC: 0.81, 95% CI: 0.77 to 0.85). In this study, the predictive value of TMBRB was better than that of a histological subtype (AUC of training cohort: 0.75, 95% CI: 0.72 to 0.77; AUC of test cohort: 0.71, 95% CI: 0.66 to 0.76) or Radiomic model (AUC of training cohort: 0.75, 95% CI: 0.72 to 0.77; AUC of test cohort: 0.74, 95% CI: 0.69 to 0.79). When predicting immunotherapy efficacy, TMBRB divided patients into a high- and low-risk group with distinctly different overall survival (OS; HR: 0.54, 95% CI: 0.31 to 0.95; p=0.030) and progression-free survival (PFS; HR: 1.78, 95% CI: 1.07 to 2.95; p=0.023). Moreover, TMBRB had a better predictive ability when combined with the Eastern Cooperative Oncology Group performance status (OS: p=0.007; PFS: p=0.003). Visual analysis revealed that tumor microenvironment was important for predicting TMB. CONCLUSION: By combining deep learning technology and CT images, we developed an individual non-invasive biomarker that could distinguish High-TMB from Low-TMB, which might inform decisions on the use of ICIs in patients with advanced NSCLC.

CT-Based Radiomic Signature as a Prognostic Factor in Stage IV ALK-Positive Non-small-cell Lung Cancer Treated With TKI Crizotinib: A Proof-of-Concept Study.

Objectives: To identify a computed tomography (CT)-based radiomic signature for predicting progression-free survival (PFS) in stage IV anaplastic lymphoma kinase (ALK)-positive non-small-cell lung cancer (NSCLC) patients treated with tyrosine kinase inhibitor (TKI) crizotinib. Materials and Methods: This retrospective proof-of-concept study included a cohort of 63 stage IV ALK-positive NSCLC patients who had received TKI crizotinib therapy for model construction and validation. Another independent cohort including 105 stage IV EGFR-positive NSCLC patients was also used for external validation in EGFR-TKI treatment. We initially extracted 481 quantitative three-dimensional features derived from manually segmented tumor volumes of interest. Pearson's correlation analysis along with the least absolute shrinkage and selection operator (LASSO) penalized Cox proportional hazards regression was successively performed to select critical radiomic features. A CT-based radiomic signature for PFS prediction was obtained using multivariate Cox regression. The performance evaluation of the radiomic signature was conducted using the concordance index (C-index), time-dependent receiver operating characteristic (ROC) analysis, and Kaplan-Meier survival analysis. Results: A radiomic signature containing three features showed significant prognostic performance for ALK-positive NSCLC patients in both the training cohort (C-index, 0.744; time-dependent AUC, 0.895) and the validation cohort (C-index, 0.717; time-dependent AUC, 0.824). The radiomic signature could significantly risk-stratify ALK-positive NSCLC patients (hazard ratio, 2.181; P < 0.001) and outperformed other prognostic factors. However, no significant association with PFS was captured for the radiomic signature in the EGFR-positive NSCLC cohort (log-rank tests, P = 0.41). Conclusions: The CT-based radiomic features can capture valuable information regarding the tumor phenotype. The proposed radiomic signature was found to be an effective prognostic factor in stage IV ALK mutated nonsynchronous nodules in NSCLC patients treated with a TKI.

A deep learning radiomics model for preoperative grading in meningioma.

OBJECTIVES: To noninvasively differentiate meningioma grades by deep learning radiomics (DLR) model based on routine post-contrast MRI. METHODS: We enrolled 181 patients with histopathologic diagnosis of meningioma who received post-contrast MRI preoperative examinations from 2 hospitals (99 in the primary cohort and 82 in the validation cohort). All the tumors were segmented based on post-contrast axial T1 weighted images (T1WI), from which 2048 deep learning features were extracted by the convolutional neural network. The random forest algorithm was used to select features with importance values over 0.001, upon which a deep learning signature was built by a linear discriminant analysis classifier. The performance of our DLR model was assessed by discrimination and calibration in the independent validation cohort. For comparison, a radiomic model based on hand-crafted features and a fusion model were built. RESULTS: The DLR signature comprised 39 deep learning features and showed good discrimination performance in both the primary and validation cohorts. The area under curve (AUC), sensitivity, and specificity for predicting meningioma grades were 0.811(95% CI, 0.635-0.986), 0.769, and 0.898 respectively in the validation cohort. DLR performance was superior over the hand-crafted features. Calibration curves of DLR model showed good agreements between the prediction probability and the observed outcome of high-grade meningioma. CONCLUSIONS: Using routine MRI data, we developed a DLR model with good performance for noninvasively individualized prediction of meningioma grades, which achieved a quantization capability superior over the hand-crafted features. This model has potential to guide and facilitate the clinical decision-making of whether to observe or to treat patients by providing prognostic information.

Radiomic nomogram for prediction of axillary lymph node metastasis in breast cancer.

OBJECTIVE: To develop a radiomic nomogram for preoperative prediction of axillary lymph node (LN) metastasis in breast cancer patients. METHODS: Preoperative magnetic resonance imaging data from 411 breast cancer patients was studied. Patients were assigned to either a training cohort (n = 279) or a validation cohort (n = 132). Eight hundred eight radiomic features were extracted from the first phase of T1-DCE images. A support vector machine was used to develop a radiomic signature, and logistic regression was used to develop a nomogram. RESULTS: The radiomic signature based on 12 LN status-related features was constructed to predict LN metastasis, its prediction ability was moderate, with an area under the curve (AUC) of 0.76 and 0.78 in training and validation cohorts, respectively. Based on a radiomic signature and clinical features, a nomogram was developed and showed excellent predictive ability for LN metastasis (AUC 0.84 and 0.87 in training and validation sets, respectively). Another radiomic signature was constructed to distinguish the number of metastatic LNs (less than 2 positive nodes/more than 2 positive nodes), which also showed moderate performance (AUC 0.79). CONCLUSIONS: We developed a nomogram and a radiomic signature that can be used to identify LN metastasis and distinguish the number of metastatic LNs (less than 2 positive nodes/more than 2 positive nodes). Both nomogram and radiomic signature can be used as tools to assist clinicians in assessing LN metastasis in breast cancer patients. KEY POINTS: • ALNM is an important factor affecting breast cancer patients' treatment and prognosis. • Traditional imaging examinations have limited value for evaluating axillary LNs status. • We developed a radiomic nomogram based on MR imagings to predict LN metastasis.

Predicting EGFR mutation status in lung adenocarcinoma on computed tomography image using deep learning.

Epidermal growth factor receptor (EGFR) genotyping is critical for treatment guidelines such as the use of tyrosine kinase inhibitors in lung adenocarcinoma. Conventional identification of EGFR genotype requires biopsy and sequence testing which is invasive and may suffer from the difficulty of accessing tissue samples. Here, we propose a deep learning model to predict EGFR mutation status in lung adenocarcinoma using non-invasive computed tomography (CT).We retrospectively collected data from 844 lung adenocarcinoma patients with pre-operative CT images, EGFR mutation and clinical information from two hospitals. An end-to-end deep learning model was proposed to predict the EGFR mutation status by CT scanning.By training in 14 926 CT images, the deep learning model achieved encouraging predictive performance in both the primary cohort (n=603; AUC 0.85, 95% CI 0.83-0.88) and the independent validation cohort (n=241; AUC 0.81, 95% CI 0.79-0.83), which showed significant improvement over previous studies using hand-crafted CT features or clinical characteristics (p<0.001). The deep learning score demonstrated significant differences in EGFR-mutant and EGFR-wild type tumours (p<0.001).Since CT is routinely used in lung cancer diagnosis, the deep learning model provides a non-invasive and easy-to-use method for EGFR mutation status prediction.

Unsupervised Deep Learning Features for Lung Cancer Overall Survival Analysis.

Lung cancer overall survival analysis using computed tomography (CT) images plays an important role in treatment planning. Most current analysis methods involve hand-crafted image features for survival time prediction. However, hand-crafted features require domain knowledge and may lack specificity to lung cancer. Advanced self-learning models such as deep learning have showed superior performance in many medical image tasks, but they require large amount of data which is difficult to collect for survival analysis because of the long follow-up time. Although data with survival time is difficult to acquire, it is relatively easy to collect lung cancer patients without survival time. In this paper, we proposed an unsupervised deep learning method to take advantage of the unlabeled data for survival analysis, and demonstrated better performance than using hand-crafted features. We proposed a residual convolutional auto encoder and trained the model using images from 274 patients without survival time. Afterwards, we extracted deep learning features through the encoder model, and constructed a Cox proportional hazards model on 129 patients with survival time. The experiment results showed that our unsupervised deep learning feature gained better performance (C-Index = 0.70) than using hand-crafted features (C-Index = 0.62). Furthermore, we divided the patients into two groups according to their Cox hazard value. Kaplan-Meier analysis indicated that our model can divide patients into high and low risk groups and the survival time of these two groups had significant difference (p < 0.01).

Can CT-based radiomics signature predict KRAS/NRAS/BRAF mutations in colorectal cancer?

OBJECTIVES: To investigate whether CT-based radiomics signature can predict KRAS/NRAS/BRAF mutations in colorectal cancer (CRC). METHODS: This retrospective study consisted of a primary cohort (n = 61) and a validation cohort (n = 56) with pathologically confirmed CRC. Patients underwent KRAS/NRAS/BRAF mutation tests and contrast-enhanced CT before treatment. A total of 346 radiomics features were extracted from portal venous-phase CT images of the entire primary tumour. Associations between the genetic mutations and clinical background, tumour staging, and histological differentiation were assessed using univariate analysis. RELIEFF and support vector machine methods were performed to select key features and build a radiomics signature. RESULTS: The radiomics signature was significantly associated with KRAS/NRAS/BRAF mutations (P < 0.001). The area under the curve, sensitivity, and specificity for predicting KRAS/NRAS/BRAF mutations were 0.869, 0.757, and 0.833 in the primary cohort, respectively, while they were 0.829, 0.686, and 0.857 in the validation cohort, respectively. Clinical background, tumour staging, and histological differentiation were not associated with KRAS/NRAS/BRAF mutations in both cohorts (P>0.05). CONCLUSIONS: The proposed CT-based radiomics signature is associated with KRAS/NRAS/BRAF mutations. CT may be useful for analysis of tumour genotype in CRC and thus helpful to determine therapeutic strategies. KEY POINTS: • Key features were extracted from CT images of the primary colorectal tumour. • The proposed radiomics signature was significantly associated with KRAS/NRAS/BRAF mutations. • In the primary cohort, the proposed radiomics signature predicted mutations. • Clinical background, tumour staging, and histological differentiation were unable to predict mutations.