Radiomics and radiogenomics in lung cancer: A review for the clinician

[Research Advances and Obstacles of CT-based Radiomics in Diagnosis and Treatment of Lung Cancer]

影像组学是一种基于多模态医学影像处理分析的技术，该技术能够基于高性能计算机及算法从目前普遍使用的计算机断层扫描（computed tomography, CT）、磁共振图像（magnetic resonance imaging, MRI）和正电子发射/断层图像（positron emission tomography/computed tomography, PET/CT）中自动提取海量数据进行分析，对疾病的早期诊断、良恶性肿瘤鉴别、疾病治疗全程管理，个体化精准治疗等需求提供更多有价值信息。近年来，许多研究表明基于CT的影像组学技术在肺癌的早期诊断、基因表型预测、疗效预测及预后评估均有良好的应用价值，且影像学检查具有无创、经济、可重复等优势。其对临床的指导价值已有所展露，在肺癌的个体化、精准化治疗和研究方面具有较大价值，但是，影像组学特征的重复性和一致性问题以及在肺部肿瘤图像提取中的特征筛选还需进一步研究。

The Growth Trend Predictions in Pulmonary Ground Glass Nodules Based on Radiomic CT Features

Background: The management of ground glass nodules (GGNs) remains a distinctive challenge. This study is aimed at comparing the predictive growth trends of radiomic features against current clinical features for the evaluation of GGNs.

Methods: A total of 110 GGNs in 85 patients were included in this retrospective study, in which follow up occurred over a span ≥2 years. A total of 396 radiomic features were manually segmented by radiologists and quantitatively analyzed using an Analysis Kit software. After feature selection, three models were developed to predict the growth of GGNs. The performance of all three models was evaluated by a receiver operating characteristic (ROC) curve. The best performing model was also assessed by calibration and clinical utility.

Results: After using a stepwise multivariate logistic regression analysis and dimensionality reduction, the diameter and five specific radiomic features were included in the clinical model and the radiomic model. The rad-score [odds ratio (OR) = 5.130; P < 0.01] and diameter (OR = 1.087; P < 0.05) were both considered as predictive indicators for the growth of GGNs. Meanwhile, the area under the ROC curve of the combined model reached 0.801. The high degree of fitting and favorable clinical utility was detected using the calibration curve with the Hosmer-Lemeshow test and the decision curve analysis was utilized for the nomogram.

Conclusions: A combined model using the current clinical features alongside the radiomic features can serve as a powerful tool to assist clinicians in guiding the management of GGNs.

A Novel Model Based on CXCL8-Derived Radiomics for Prognosis Prediction in Colorectal Cancer

Introduction: Prognosis prediction is essential to improve therapeutic strategies and to achieve better clinical outcomes in colorectal cancer (CRC) patients. Radiomics based on high-throughput mining of quantitative medical imaging is an emerging field in recent years. However, the relationship among prognosis, radiomics features, and gene expression remains unknown. Methods: We retrospectively analyzed 141 patients (from study 1) diagnosed with CRC from February 2018 to October 2019 and randomly divided them into training (N = 99) and testing (N = 42) cohorts. Radiomics features in venous phase image were extracted from preoperative computed tomography (CT) images. Gene expression was detected by RNA-sequencing on tumor tissues. The least absolute shrinkage and selection operator (LASSO) regression model was used for selecting imaging features and building the radiomics model. A total of 45 CRC patients (study 2) with immunohistochemical (IHC) staining of CXCL8 diagnosed with CRC from January 2014 to October 2018 were included in the independent testing cohort. A clinical model was validated for prognosis prediction in prognostic testing cohort (163 CRC patients from 2014 to 2018, study 3). We performed a combined radiomics model that was composed of radiomics score, tumor stage, and CXCL8-derived radiomics model to make comparison with the clinical model. Results: In our study, we identified the CXCL8 as a hub gene in affecting prognosis, which is mainly through regulating cytokine-cytokine receptor interaction and neutrophil migration pathway. The radiomics model incorporated 12 radiomics features screened by LASSO according to CXCL8 expression in the training cohort and showed good performance in testing and IHC testing cohorts. Finally, the CXCL8-derived radiomics model combined with tumor stage performed high ability in predicting the prognosis of CRC patients in the prognostic testing cohort, with an area under the curve (AUC) of 0.774 [95% confidence interval (CI): 0.674-0.874]. Kaplan-Meier analysis of the overall survival probability in CRC patients stratified by combined model revealed that high-risk patients have a poor prognosis compared with low-risk patients (Log-rank P < 0.0001). Conclusion: We demonstrated that the radiomics model reflected by CXCL8 combined with tumor stage information is a reliable approach to predict the prognosis in CRC patients and has a potential ability in assisting clinical decision-making.

Radiomics of Liver Metastases: A Systematic Review

Simple Summary

Patients with liver metastases can be scheduled for different therapies (e.g., chemotherapy, surgery, radiotherapy, and ablation). The choice of the most appropriate treatment should rely on adequate understanding of tumor biology and prediction of survival, but reliable biomarkers are lacking. Radiomics is an innovative approach to medical imaging: it identifies invisible-to-the-human-eye radiological patterns that can predict tumor aggressiveness and patients outcome. We reviewed the available literature to elucidate the role of radiomics in patients with liver metastases. Thirty-two papers were analyzed, mostly (56%) concerning metastases from colorectal cancer. Even if available studies are still preliminary, radiomics provided effective prediction of response to chemotherapy and of survival, allowing more accurate and earlier prediction than standard predictors. Entropy and homogeneity were the radiomic features with the strongest clinical impact. In the next few years, radiomics is expected to give a consistent contribution to the precision medicine approach to patients with liver metastases.

Abstract

Multidisciplinary management of patients with liver metastases (LM) requires a precision medicine approach, based on adequate profiling of tumor biology and robust biomarkers. Radiomics, defined as the high-throughput identification, analysis, and translational applications of radiological textural features, could fulfill this need. The present review aims to elucidate the contribution of radiomic analyses to the management of patients with LM. We performed a systematic review of the literature through the most relevant databases and web sources. English language original articles published before June 2020 and concerning radiomics of LM extracted from CT, MRI, or PET-CT were considered. Thirty-two papers were identified. Baseline higher entropy and lower homogeneity of LM were associated with better survival and higher chemotherapy response rates. A decrease in entropy and an increase in homogeneity after chemotherapy correlated with radiological tumor response. Entropy and homogeneity were also highly predictive of tumor regression grade. In comparison with RECIST criteria, radiomic features provided an earlier prediction of response to chemotherapy. Lastly, texture analyses could differentiate LM from other liver tumors. The commonest limitations of studies were small sample size, retrospective design, lack of validation datasets, and unavailability of univocal cut-off values of radiomic features. In conclusion, radiomics can potentially contribute to the precision medicine approach to patients with LM, but interdisciplinarity, standardization, and adequate software tools are needed to translate the anticipated potentialities into clinical practice.

Sensitivity analysis of FDG PET tumor voxel cluster radiomics and dosimetry for predicting mid-chemoradiation regional response of locally advanced lung cancer

We investigated the sensitivity of regional tumor response prediction to variability in voxel clustering techniques, imaging features, and machine learning algorithms in 25 patients with locally advanced non-small cell lung cancer (LA-NSCLC) enrolled on the FLARE-RT clinical trial. Metabolic tumor volumes (MTV) from pre-chemoradiation (PETpre) and mid-chemoradiation fluorodeoxyglucose-positron emission tomography (FDG PET) images (PETmid) were subdivided into K-means or hierarchical voxel clusters by standardized uptake values (SUV) and 3D-positions. MTV cluster separability was evaluated by CH index, and morphologic changes were captured by Dice similarity and centroid Euclidean distance. PETpre conventional features included SUVmean, MTV/MTV cluster size, and mean radiation dose. PETpre radiomics consisted of 41 intensity histogram and 3D texture features (PET Oncology Radiomics Test Suite) extracted from MTV or MTV clusters. Machine learning models (multiple linear regression, support vector regression, logistic regression, support vector machines) of conventional features or radiomic features were constructed to predict PETmid response. Leave-one-out-cross-validated root-mean-squared-error (RMSE) for continuous response regression (ΔSUVmean) and area-under-receiver-operating-characteristic-curve (AUC) for binary response classification were calculated. K-means MTV 2-clusters (MTVhi, MTVlo) achieved maximum CH index separability (Friedman p < 0.001). Between PETpre and PETmid, MTV cluster pairs overlapped (Dice 0.70-0.87) and migrated 0.6-1.1 cm. PETmid ΔSUVmean response prediction was superior in MTV and MTVlo (RMSE = 0.17-0.21) compared to MTVhi (RMSE = 0.42-0.52, Friedman p < 0.001). PETmid ΔSUVmean response class prediction performance trended higher in MTVlo (AUC = 0.83-0.88) compared to MTVhi (AUC = 0.44-0.58, Friedman p = 0.052). Models were more sensitive to MTV/MTV cluster regions (Friedman p = 0.026) than feature sets/algorithms (Wilcoxon signed-rank p = 0.36). Top-ranked radiomic features included GLZSM-LZHGE (large-zone-high-SUV), GTSDM-CP (cluster-prominence), GTSDM-CS (cluster-shade) and NGTDM-CNT (contrast). Top-ranked features were consistent between MTVhi and MTVlo cluster pairs but varied between MTVhi-MTVlo clusters, reflecting distinct regional radiomic phenotypes. Variability in tumor voxel cluster response prediction can inform robust radiomic target definition for risk-adaptive chemoradiation in patients with LA-NSCLC. FLARE-RT trial: NCT02773238.

Standardization of CT radiomics features for multi-center analysis: impact of software settings and parameters

The aim of this multicentric study is an inter-center benchmarking, to assess how different set tools applied to the same radiomics workflow affected the radiomics features (RFs) values. This topic is of key importance to start collaboration between different centers and to bring radiomic studies from benchmark to bedside. A per-lesion analysis was performed on 56 metastases (mts) selected from 14 patients. A single radiologist performed the segmentation of all mts, and RFs were extracted from the same segmentation of each mts, using two different software and file formats. Potential sources of discrepancies were evaluated. The intraclass correlation coefficient was used to describe how strongly the same radiomic measurements calculated in the two different centers resemble each other. Moreover, means of the relative changes of each RF were calculated, compared and gradually reduced. We showed that, after matching all formulas, discrepancies in RFs calculation between two centers ranged from 1% to 277%. Therefore, we evaluated other sources of variability using a stepwise approach, which led us to reduce the inter-center discrepancies to 0% for 22/25 RFs and below 2% for 3 RFs out of 25. In this study we demonstrated that different radiomic applications and masks formats might strongly impact the computation of some RFs. Therefore, when dealing with multi-center studies it is mandatory to adopt all strategies that can help in limiting the differences, thus keeping in mind the feasibility of these strategies in large cohort studies.

The natural language explanation algorithms for the lung cancer computer-aided diagnosis system

Two algorithms for explaining decisions of a lung cancer computer-aided diagnosis system are proposed. Their main peculiarity is that they produce explanations of diseases in the form of special sentences via natural language. The algorithms consist of two parts. The first part is a standard local post-hoc explanation model, for example, the well-known LIME, which is used for selecting important features from a special feature representation of the segmented lung suspicious objects. This part is identical for both algorithms. The second part is a model which aims to connect selected important features and to transform them to explanation sentences in natural language. This part is implemented differently for both algorithms. The training phase of the first algorithm uses a special vocabulary of simple phrases which produce sentences and their embeddings. The second algorithm significantly simplifies some parts of the first algorithm and reduces the explanation problem to a set of simple classifiers. The basic idea behind the improvement is to represent every simple phrase from vocabulary as a class of the "sparse" histograms. An implementation of the second algorithm is shown in detail.

Radiomics-Based Prediction of Overall Survival in Lung Cancer Using Different Volumes-Of-Interest

Lung cancer accounts for the largest amount of deaths worldwide with respect to the other oncological pathologies. To guarantee the most effective cure to patients for such aggressive tumours, radiomics is increasing as a novel and promising research field that aims at extracting knowledge from data in terms of quantitative measures that are computed from diagnostic images, with prognostic and predictive ends. This knowledge could be used to optimize current treatments and to maximize their efficacy. To this end, we hereby study the use of such quantitative biomarkers computed from CT images of patients affected by Non-Small Cell Lung Cancer to predict Overall Survival. The main contributions of this work are two: first, we consider different volumes of interest for the same patient to find out whether the volume surrounding the visible lesions can provide useful information; second, we introduce 3D Local Binary Patterns, which are texture measures scarcely explored in radiomics. As further validation, we show that the proposed signature outperforms not only the features automatically computed by a deep learning-based approach, but also another signature at the state-of-the-art using other handcrafted features.

Inflammatory lesions and brain tumors: is it possible to differentiate them based on texture features in magnetic resonance imaging?

Background:

Neuroimaging strategies are essential to locate, to elucidate the etiology, and to the follow up of brain disease patients. Magnetic resonance imaging (MRI) provides good cerebral soft-tissue contrast detection and diagnostic sensitivity. Inflammatory lesions and tumors are common brain diseases that may present a similar pattern of a cerebral ring enhancing lesion on MRI, and non-enhancing core (which may reflect cystic components or necrosis) leading to misdiagnosis. Texture analysis (TA) and machine learning approaches are computer-aided diagnostic tools that can be used to assist radiologists in such decisions.

Methods:

In this study, we combined texture features with machine learning (ML) methods aiming to differentiate brain tumors from inflammatory lesions in magnetic resonance imaging. Retrospective examination of 67 patients, with a pattern of a cerebral ring enhancing lesion, 30 with inflammatory, and 37 with tumoral lesions were selected. Three different MRI sequences and textural features were extracted using gray level co-occurrence matrix and gray level run length. All diagnoses were confirmed by histopathology, laboratorial analysis or MRI.

Results:

The features extracted were processed for the application of ML methods that performed the classification. T1-weighted images proved to be the best sequence for classification, in which the differentiation between inflammatory and tumoral lesions presented high accuracy (0.827), area under ROC curve (0.906), precision (0.837), and recall (0.912).

Conclusion:

The algorithm obtained textures capable of differentiating brain tumors from inflammatory lesions, on T1-weghted images without contrast medium using the Random Forest machine learning classifier.

Radiomics at a Glance: A Few Lessons Learned from Learning Approaches

Simple Summary

Radiomics has become a prominent component of medical imaging research and many studies show its specific value as a support tool for clinical decision-making processes. Radiomic data are typically analyzed with statistical and machine learning methods, which change depending on the disease context and the imaging modality. We found a certain bias in the literature towards the use of such methods and believe that this limitation may influence the capacity of producing accurate and reliable decisions. Therefore, in view of the relevance of various types of learning methods, we report their significance and discuss their unrevealed potential.

Abstract

Processing and modeling medical images have traditionally represented complex tasks requiring multidisciplinary collaboration. The advent of radiomics has assigned a central role to quantitative data analytics targeting medical image features algorithmically extracted from large volumes of images. Apart from the ultimate goal of supporting diagnostic, prognostic, and therapeutic decisions, radiomics is computationally attractive due to specific strengths: scalability, efficiency, and precision. Optimization is achieved by highly sophisticated statistical and machine learning algorithms, but it is especially deep learning that stands out as the leading inference approach. Various types of hybrid learning can be considered when building complex integrative approaches aimed to deliver gains in accuracy for both classification and prediction tasks. This perspective reviews some selected learning methods by focusing on both their significance for radiomics and their unveiled potential.

Pilot Study of CT-Based Radiomics Model for Early Evaluation of Response to Immunotherapy in Patients With Metastatic Melanoma

Objective: Determine the performance of a computed tomography (CT) -based radiomics model in predicting early response to immunotherapy in patients with metastatic melanoma. Methods: This retrospective study examined 50 patients with metastatic melanoma who received immunotherapy treatment in our hospital with an anti-programmed cell death-1 (PD-1) agent or an inhibitor of cytotoxic T lymphocyte antigen-4 (CTLA-4). Thirty-four patients who received an anti-PD-1 agent were in the training sample and 16 patients who received a CTLA-4 inhibitor were in the validation sample. Patients with true progressive disease (PD) were in the poor response group, and those with pseudoprogression, complete response (CR), partial response (PR), or stable disease (SD) were in the good response group. CT images were examined at baseline and after the first and second cycles of treatment, and the imaging data were extracted for radiomics modeling. Results: The radiomics model based on pre-treatment, post-treatment, and delta features provided the best results for predicting response to immunotherapy. Receiver operating characteristic (ROC) analysis for good response indicated an area under the curve (AUC) of 0.882 for the training group and an AUC of 0.857 for the validation group. The sensitivity, specificity, and accuracy of model were 85.70% (6/7), 66.70% (6/9), and 75% (12/16) for predicting a good response. Conclusion: A CT-based radiomics model for metastatic melanoma has the potential to predict early response to immunotherapy and to identify pseudoprogression.

Recent and Current Advances in FDG-PET Imaging within the Field of Clinical Oncology in NSCLC: A Review of the Literature

Lung cancer is the leading cause of cancer-related deaths around the world, the most common type of which is non-small-cell lung cancer (NSCLC). Computed tomography (CT) is required for patients with NSCLC, but often involves diagnostic issues and large intra- and interobserver variability. The anatomic data obtained using CT can be supplemented by the metabolic data obtained using fluorodeoxyglucose F 18 (FDG) positron emission tomography (PET); therefore, the use of FDG-PET/CT for staging NSCLC is recommended, as it provides more accuracy than either modality alone. Furthermore, FDG-PET/magnetic resonance imaging (MRI) provides useful information on metabolic activity and tumor cellularity, and has become increasingly popular. A number of studies have described FDG-PET/MRI as having a high diagnostic performance in NSCLC staging. Therefore, multidimensional functional imaging using FDG-PET/MRI is promising for evaluating the activity of the intratumoral environment. Radiomics is the quantitative extraction of imaging features from medical scans. The chief advantages of FDG-PET/CT radiomics are the ability to capture information beyond the capabilities of the human eye, non-invasiveness, the (virtually) real-time response, and full-field analysis of the lesion. This review summarizes the recent advances in FDG-PET imaging within the field of clinical oncology in NSCLC, with a focus on surgery and prognostication, and investigates the site-specific strengths and limitations of FDG-PET/CT. Overall, the goal of treatment for NSCLC is to provide the best opportunity for long-term survival; therefore, FDG-PET/CT is expected to play an increasingly important role in deciding the appropriate treatment for such patients.

Application and limitation of radiomics approach to prognostic prediction for lung stereotactic body radiotherapy using breath-hold CT images with random survival forest: A multi-institutional study

PURPOSE

To predict local recurrence (LR) and distant metastasis (DM) in early stage non-small cell lung cancer (NSCLC) patients after stereotactic body radiotherapy (SBRT) in multiple institutions using breath-hold computed tomography (CT)-based radiomic features with random survival forest.

METHODS

A total of 573 primary early stage NSCLC patients who underwent SBRT between January 2006 and March 2016 and met the eligibility criteria were included in this study. Patients were divided into two datasets: training (464 patients in 10 institutions) and test (109 patients in one institution) datasets. A total of 944 radiomic features were extracted from manually segmented gross tumor volumes (GTVs). Feature selection was performed by analyzing inter-segmentation reproducibility, GTV correlation, and inter-feature redundancy. Nine clinical factors, including histology and GTV size, were also used. Three prognostic models (clinical, radiomic, and combined) for LR and DM were constructed using random survival forest (RSF) to deal with total death as a competing risk in the training dataset. Robust models with optimal hyper-parameters were determined using fivefold cross-validation. The patients were dichotomized into two groups based on the median value of the patient-specific risk scores (high- and low-risk score groups). Gray's test was used to evaluate the statistical significance between the two risk score groups. The prognostic power was evaluated by the concordance index with the 95% confidence intervals (CI) via bootstrapping (2000 iterations).

RESULTS

The concordance indices at 3 yr of clinical, radiomic, and combined models for LR were 0.57 [CI: 0.39-0.75], 0.55 [CI: 0.38-0.73], and 0.61 [CI: 0.43-0.78], respectively, whereas those for DM were 0.59 [CI: 0.54-0.79], 0.67 [CI: 0.54-0.79], and 0.68 [CI: 0.55-0.81], respectively, in the test dataset. The combined DM model significantly discriminated its cumulative incidence between high- and low-risk score groups (P < 0.05). The variable importance of RSF in the combined model for DM indicated that two radiomic features were more important than other clinical factors. The feature maps generated on the basis of the most important radiomic feature had visual difference between high- and low-risk score groups.

CONCLUSIONS

The radiomics approach with RSF for competing risks using breath-hold CT-based radiomic features might predict DM in early stage NSCLC patients who underwent SBRT although that may not have potential to predict LR.

Intratumor Heterogeneity Assessed by 18F-FDG PET/CT Predicts Treatment Response and Survival Outcomes in Patients with Hodgkin Lymphoma

RATIONALE AND OBJECTIVES

Radiomic analysis of ¹⁸F-fluorodeoxyglucose positron emission tomography/computed tomography (¹⁸F-FDG PET/CT) images enables the extraction of quantitative information of intratumour heterogeneity. This study investigated whether the baseline ¹⁸F-FDG PET/CT radiomics can predict treatment response and survival outcomes in patients with Hodgkin lymphoma (HL).

MATERIALS AND METHODS

Thirty-five patients diagnosed with HL who underwent ¹⁸F-FDG PET/CT scans before and during chemotherapy were retrospectively enrolled in this investigation. For each patient, we extracted 709 radiomic features from pretreatment PET/CT images. Clinical variables (age, gender, B symptoms, bulky tumor, and disease stage) and radiomic signatures (intensity, texture, and wavelet) were analyzed according to response to therapy, progression-free survival (PFS), and overall survival (OS). Receiver operating characteristic curve, logistic regression, and Cox proportional hazards model were used to examine potential predictive and prognostic factors.

RESULTS

High-intensity run emphasis (HIR) of PET and run-length nonuniformity (RLNU) of CT extracted from gray-level run-length matrix (GLRM) in high-frequency wavelets were independent predictive factors for the treatment response (odds ratio [OR] = 36.4, p = 0.014; OR = 30.4, p = 0.020). Intensity nonuniformity (INU) of PET and wavelet short run emphasis (SRE) of CT from GLRM and Ann Arbor stage were independently related to PFS (hazard ratio [HR] = 9.29, p = 0.023; HR = 18.40, p = 0.012; HR = 7.46, p = 0.049). Zone-size nonuniformity (ZSNU) of PET from gray-level size zone matrix (GLSZM) was independently associated with OS (HR = 41.02, p = 0.001). Based on these factors, a prognostic stratification model was devised for the risk stratification of patients. The proposed model allowed the identification of four risk groups for PFS and OS (p < 0.001 and p < 0.001).

CONCLUSION

HIR_{_}GLRM_PET and RLNU_{_}GLRM_CT in high-frequency wavelets serve as independent predictive factors for treatment response. ZSNU_{_}GLSZM_PET, INU_{_}GLRM_PET, and wavelet SRE_{_}GLRM_CT serve as independent prognostic factors for survival outcomes. The present study proposes a prognostic stratification model that may be clinically beneficial in guiding risk-adapted treatment strategies for patients with HL.

Multi-energy computed tomography and material quantification: Current barriers and opportunities for advancement

Computed tomography (CT) technology has rapidly evolved since its introduction in the 1970s. It is a highly important diagnostic tool for clinicians as demonstrated by the significant increase in utilization over several decades. However, much of the effort to develop and advance CT applications has been focused on improving visual sensitivity and reducing radiation dose. In comparison to these areas, improvements in quantitative CT have lagged behind. While this could be a consequence of the technological limitations of conventional CT, advanced dual-energy CT (DECT) and photon-counting detector CT (PCD-CT) offer new opportunities for quantitation. Routine use of DECT is becoming more widely available and PCD-CT is rapidly developing. This review covers efforts to address an unmet need for improved quantitative imaging to better characterize disease, identify biomarkers, and evaluate therapeutic response, with an emphasis on multi-energy CT applications. The review will primarily discuss applications that have utilized quantitative metrics using both conventional and DECT, such as bone mineral density measurement, evaluation of renal lesions, and diagnosis of fatty liver disease. Other topics that will be discussed include efforts to improve quantitative CT volumetry and radiomics. Finally, we will address the use of quantitative CT to enhance image-guided techniques for surgery, radiotherapy and interventions and provide unique opportunities for development of new contrast agents.

Radiomic features of primary tumor by lung cancer stage: analysis in BRAF mutated non-small cell lung cancer

Background

The clinical features and traditional semantic imaging characteristics of BRAF-mutated non-small cell lung cancer (NSCLC) have been previously reported. The radiomic features of BRAF-mutated NSCLC and their role in predicting cancer stage, however, have yet to be investigated. This study’s goal is to assess the differences in CT radiomic features of primary NSCLC driven by BRAF mutation and stratified by tumor-node-metastasis (TNM) staging.

Methods

Our IRB approved study included 62 patients with BRAF mutations (V600 in 27 and non-V600 in 35 patients), who underwent contrast-enhanced chest CT. Tumor stage was determined based on the 8^th edition of TNM staging. Two thoracic radiologists assessed the primary tumor imaging features such, including tumor size (maximum and minimum dimensions) and density (Hounsfield units, HU). De-identified transverse CT images (DICOM) were processed with 3D slicer (Version 4.7) for manual lesion segmentation and estimation of radiomic features. Descriptive statistics, multivariate logistic regression, and receiver operating characteristics (ROC) were performed.

Results

There were significant differences in the radiomic features based on cancer stages I-IV with the most significant differences between stage IV and stage I lesions [AUC 0.94 (95% CI: 0.86–0.99), P<0.04]. There were also significant differences in radiomic features between stage IV and combined stages I-III [40/113 radiomic features; AUC 0.71 (95% CI: 0.59–0.85); P<0.04–0.0001]. None of the clinical (0/6) or imaging (0/3) features were significantly different between stage IV and combined stages I–III.

Conclusions

The radiomic features of primary tumor in BRAF driven NSCLC significantly vary with cancer stage, independent of standard imaging and clinical features.

[18F]FDG PET immunotherapy radiomics signature (iRADIOMICS) predicts response of non-small-cell lung cancer patients treated with pembrolizumab

Background

Immune checkpoint inhibitors have changed the paradigm of cancer treatment; however, non-invasive biomarkers of response are still needed to identify candidates for non-responders. We aimed to investigate whether immunotherapy [¹⁸F]FDG PET radiomics signature (iRADIOMICS) predicts response of metastatic non-small-cell lung cancer (NSCLC) patients to pembrolizumab better than the current clinical standards.

Patients and methods

Thirty patients receiving pembrolizumab were scanned with [¹⁸F]FDG PET/CT at baseline, month 1 and 4. Associations of six robust primary tumour radiomics features with overall survival were analysed with Mann-Whitney U-test (MWU), Cox proportional hazards regression analysis, and ROC curve analysis. iRADIOMICS was constructed using univariate and multivariate logistic models of the most promising feature(s). Its predictive power was compared to PD-L1 tumour proportion score (TPS) and iRECIST using ROC curve analysis. Prediction accuracies were assessed with 5-fold cross validation.

Results

The most predictive were baseline radiomics features, e.g. Small Run Emphasis (MWU, p = 0.001; hazard ratio = 0.46, p = 0.007; AUC = 0.85 (95% CI 0.69–1.00)). Multivariate iRADIOMICS was found superior to the current standards in terms of predictive power and timewise with the following AUC (95% CI) and accuracy (standard deviation): iRADIOMICS (baseline), 0.90 (0.78–1.00), 78% (18%); PD-L1 TPS (baseline), 0.60 (0.37–0.83), 53% (18%); iRECIST (month 1), 0.79 (0.62–0.95), 76% (16%); iRECIST (month 4), 0.86 (0.72–1.00), 76% (17%).

Conclusions

Multivariate iRADIOMICS was identified as a promising imaging biomarker, which could improve management of metastatic NSCLC patients treated with pembrolizumab. The predicted non-responders could be offered other treatment options to improve their overall survival.

[Formula: see text]: deep learning-based radiomics for the time-to-event outcome prediction in lung cancer

Hand-crafted radiomics has been used for developing models in order to predict time-to-event clinical outcomes in patients with lung cancer. Hand-crafted features, however, are pre-defined and extracted without taking the desired target into account. Furthermore, accurate segmentation of the tumor is required for development of a reliable predictive model, which may be objective and a time-consuming task. To address these drawbacks, we propose a deep learning-based radiomics model for the time-to-event outcome prediction, referred to as DRTOP that takes raw images as inputs, and calculates the image-based risk of death or recurrence, for each patient. Our experiments on an in-house dataset of 132 lung cancer patients show that the obtained image-based risks are significant predictors of the time-to-event outcomes. Computed Tomography (CT)-based features are predictors of the overall survival (OS), with the hazard ratio (HR) of 1.35, distant control (DC), with HR of 1.06, and local control (LC), with HR of 2.66. The Positron Emission Tomography (PET)-based features are predictors of OS and recurrence free survival (RFS), with hazard ratios of 1.67 and 1.18, respectively. The concordance indices of [Formula: see text], [Formula: see text], and [Formula: see text] for predicting the OS, DC, and RFS show that the deep learning-based radiomics model is as accurate or better in predicting predefined clinical outcomes compared to hand-crafted radiomics, with concordance indices of [Formula: see text], [Formula: see text], and [Formula: see text], for predicting the OS, DC, and RFS, respectively. Deep learning-based radiomics has the potential to offer complimentary predictive information in the personalized management of lung cancer patients.

Pre-treatment CT imaging in stage IIIA lung cancer: Can we predict local recurrence after definitive chemoradiotherapy?

OBJECTIVES

The aim of this study was to delineate computed tomography (CT) features of stage IIIA non-small cell lung cancers on pre-treatment staging studies and identify features that could predict local recurrence after definitive concurrent chemoradiotherapy.

MATERIALS AND METHODS

We retrospectively reviewed pre- and post-treatment CT scans for 91 patients with Stage IIIA non-small cell lung cancer undergoing definitive concurrent chemoradiotherapy. Pre-treatment CT qualitative features were evaluated by consensus. The primary endpoint was local recurrence as determined on post-treatment CT scans along with the radiotherapy fields. Local recurrence was defined as intrathoracic in-field and marginal as opposed to out-of-field failures. Competing risk regressions were used to examine associations between CT features and recurrence.

RESULTS

The median follow-up was 51.5 months (range 2.4-111.2). Median overall survival was 25.6 months (95% CI: 20.4-30). At last follow-up, 72 (79.1%) patients had died, 48 (52.7%) had in-field recurrence, and 30 (32.9%) presented with out-of-field recurrence. On pre-treatment CT scans, tumors presenting as pulmonary consolidations (hazard ratio = 2.34, 95% CI: 1.05-5.22; p 0.038) were more likely to have in-field failure. Tumors with 50-100% necrosis (hazard ratio = 0.15, 95% CI: 0.02-1.06) were associated with decreased out-of-field failure (overall p = 0.038). However, these were rare features in our sample which limit the ability of these features to be associated with such outcomes.

CONCLUSIONS

Pre-treatment CT features alone are limited in predicting locoregional recurrence. Larger studies using quantitative tools are needed to predict such outcomes.

Semi-supervised labelling of the femur in a whole-body post-mortem CT database using deep learning

A deep learning pipeline was developed and used to localize and classify a variety of implants in the femur contained in whole-body post-mortem computed tomography (PMCT) scans. The results provide a proof-of-principle approach for labelling content not described in medical/autopsy reports. The pipeline, which incorporated residual networks and an autoencoder, was trained and tested using n = 450 full-body PMCT scans. For the localization component, Dice scores of 0.99, 0.96, and 0.98 and mean absolute errors of 3.2, 7.1, and 4.2 mm were obtained in the axial, coronal, and sagittal views, respectively. A regression analysis found the orientation of the implant to the scanner axis and also the relative positioning of extremities to be statistically significant factors. For the classification component, test cases were properly labelled as nail (N⁺), hip replacement (H⁺), knee replacement (K⁺) or without-implant (I^-) with an accuracy >97%. The recall for I^- and H⁺ cases was 1.00, but fell to 0.82 and 0.65 for cases with K⁺ and N⁺. This semi-automatic approach provides a generalized structure for image-based labelling of features, without requiring time-consuming segmentation.

Quantitative imaging feature pipeline: a web-based tool for utilizing, sharing, and building image-processing pipelines

Quantitative image features that can be computed from medical images are proving to be valuable biomarkers of underlying cancer biology that can be used for assessing treatment response and predicting clinical outcomes. However, validation and eventual clinical implementation of these tools is challenging due to the absence of shared software algorithms, architectures, and the tools required for computing, comparing, evaluating, and disseminating predictive models. Similarly, researchers need to have programming expertise in order to complete these tasks. The quantitative image feature pipeline (QIFP) is an open-source, web-based, graphical user interface (GUI) of configurable quantitative image-processing pipelines for both planar (two-dimensional) and volumetric (three-dimensional) medical images. This allows researchers and clinicians a GUI-driven approach to process and analyze images, without having to write any software code. The QIFP allows users to upload a repository of linked imaging, segmentation, and clinical data or access publicly available datasets (e.g., The Cancer Imaging Archive) through direct links. Researchers have access to a library of file conversion, segmentation, quantitative image feature extraction, and machine learning algorithms. An interface is also provided to allow users to upload their own algorithms in Docker containers. The QIFP gives researchers the tools and infrastructure for the assessment and development of new imaging biomarkers and the ability to use them for single and multicenter clinical and virtual clinical trials.

Radiological-pathological correlation of subsolid pulmonary nodules: A single centre retrospective evaluation of the 2011 IASLC adenocarcinoma classification system

INTRODUCTION

The 2011 IASLC classification system proposes guidelines for radiologists and pathologists to classify adenocarcinomas spectrum lesions as preinvasive, minimally invasive adenocarcinoma (MIA), or invasive adenocarcinoma (IA). IA portends the worst clinical prognosis, and the imaging distinction between MIA and IA is controversial.

MATERIALS AND METHODS

Subsolid pulmonary nodules resected by microcoil localization over a three-year period were retrospectively reviewed by three chest radiologists and a pulmonary pathologist. Nodules were classified radiologically based on preoperative computed tomography (CT), with the solid nodule component measured on mediastinal windows applied to high-frequency lung kernel reconstructions, and pathologically according to 2011 IASLC criteria. Radiology interobserver and radiological-pathological variability of nodule classification, and potential reasons for nodule classification discordance were assessed.

RESULTS

Seventy-one subsolid nodules in 67 patients were included. The average size of invasive disease focus at histopathology was 5 mm (standard deviation 5 mm). Radiology interobserver agreement of nodule classification was good (Cohen's Kappa = 0.604, 95 % CI: 0.447 to 0.761). Agreement between consensus radiological interpretation and pathological category was fair (Cohen's Kappa = 0.236, 95 % CI: 0.054-0.421). Radiological and pathological nodule classification were concordant in 52 % (37 of 71) of nodules. The IASLC proposed CT solid component cut-off of 5 mm to distinguish MIA and IA yielded a sensitivity of 59 % and specificity of 80 %. Common reasons for nodule classification discordance included multiple solid components within a nodule on CT, scar and stromal collapse at pathology, and measurement variability.

CONCLUSION

Solid component(s) within persistent part-solid pulmonary nodules raise suspicion for invasive adenocarcinoma. Preoperative imaging classification is frequently discordant from final pathology, reflecting interpretive and technical challenges in radiological and pathological analysis.

Silencing of BRF2 inhibits the growth and metastasis of lung cancer cells

Transcription factor II B (TFIIB)‑related factor 2 (BRF2) is involved in the development of cancer, but its role in lung cancer is underreported. The present study aimed to explore the role of BRF2 in the regulation of lung cancer cells. Immunofluorescence staining and immunohistochemistry were performed to detect BRF2 protein expression in human lung cancer cells and tissues. Following cell transfection with small interfering RNA for silencing BRF2, the cell proliferation was examined by Cell Counting Kit‑8 and MTT assays. Cell apoptosis, migration and invasion were determined by flow cytometry, wound‑healing and Transwell assay. The expression levels of Akt, phosphorylated (p)‑Akt, Bax, E‑cadherin, Bcl‑2, N‑cadherin, Snail and epidermal growth factor receptor (EGFR) in human lung cancer A549 cells were detected by western blotting. The results demonstrated that BRF2 expression was increased in human lung cancer cells and tissues, and that silencing of BRF2 promoted cell apoptosis but inhibited cell proliferation and migration. The protein expression levels of Akt, E‑cadherin, p‑Akt, Bcl‑2, N‑cadherin, Snail and EGFR in A549 cells were inhibited by silencing of BRF2, while expression levels of Bax and E‑cadherin were increased by silencing BRF2. In conclusion, BRF2 demonstrates high expression in lung cancer and silencing of BRF2 inhibits the growth and metastasis of lung cancer cells. The current findings provide a novel approach for the treatment of lung cancer.

Deep learning combined with radiomics may optimize the prediction in differentiating high-grade lung adenocarcinomas in ground glass opacity lesions on CT scans

PURPOSE

Adenocarcinoma (ADC) is the most common histological subtype of lung cancers in non-small cell lung cancer (NSCLC) in which ground glass opacifications (GGOs) found on computed tomography (CT) scans are the most common lesions. However, the presence of a micropapillary or a solid component is identified as an independent predictor of prognosis, suggesting a more extensive resection. The purpose of our study is to explore imaging phenotyping using a method combining radiomics with deep learning (RDL) to predict high-grade patterns within lung ADC.

METHODS

Included in this study were 111 patients differentiated as having GGOs and pathologically confirmed ADC. Four different groups of methods were compared to classify the GGOs for the prediction of the pathological subtypes of high-grade lung ADCs in definitive hematoxylin and eosin stain, including radiomics with gray-level features, radiomics with textural features, deep learning method, and the RDL.

RESULTS

We evaluated the performance of different models on 111 NSCLC patients using 4-fold cross-validation. The proposed RDL has achieved an overall accuracy of 0.913, which significantly outperforms the other methods (p <  0.01, analysis of variation, ANOVA). In addition, we also verified the generality and practical effectiveness of these models on an independent validation dataset of 28 patients. The results showed that our RDL framework with an accuracy of 0.966 significantly surpassed other methods.

CONCLUSION

High-grade lung ADC based on histologic pattern spectrum in GGO lesions might be predicted by the framework combining radiomics with deep learning, which reveals advantage over radiomics alone.

Contrast-enhanced T1-weighted image radiomics of brain metastases may predict EGFR mutation status in primary lung cancer

Identification of EGFR mutations is critical to the treatment of primary lung cancer and brain metastases (BMs). Here, we explored whether radiomic features of contrast-enhanced T1-weighted images (T1WIs) of BMs predict EGFR mutation status in primary lung cancer cases. In total, 1209 features were extracted from the contrast-enhanced T1WIs of 61 patients with 210 measurable BMs. Feature selection and classification were optimized using several machine learning algorithms. Ten-fold cross-validation was applied to the T1WI BM dataset (189 BMs for training and 21 BMs for the test set). Area under receiver operating characteristic curves (AUC), accuracy, sensitivity, and specificity were calculated. Subgroup analyses were also performed according to metastasis size. For all measurable BMs, random forest (RF) classification with RF selection demonstrated the highest diagnostic performance for identifying EGFR mutation (AUC: 86.81). Support vector machine and AdaBoost were comparable to RF classification. Subgroup analyses revealed that small BMs had the highest AUC (89.09). The diagnostic performance for large BMs was lower than that for small BMs (the highest AUC: 78.22). Contrast-enhanced T1-weighted image radiomics of brain metastases predicted the EGFR mutation status of lung cancer BMs with good diagnostic performance. However, further study is necessary to apply this algorithm more widely and to larger BMs.

A Computed Tomography-Based Radiomics Nomogram to Preoperatively Predict Tumor Necrosis in Patients With Clear Cell Renal Cell Carcinoma

Objective: To develop and validate a radiomics nomogram for preoperative prediction of tumor necrosis in patients with clear cell renal cell carcinoma (ccRCC). Methods: In total, 132 patients with pathologically confirmed ccRCC in one hospital were enrolled as a training cohort, while 123 ccRCC patients from second hospital served as the independent validation cohort. Radiomic features were extracted from corticomedullary and nephrographic phase contrast-enhanced computed tomography (CT) images. A radiomics signature based on optimal features selected by consistency analysis and the least absolute shrinkage and selection operator was developed. An image features model was constructed based on independent image features according to visual assessment. By integrating the radiomics signature and independent image features, a radiomics nomograph was constructed. The predictive performance of the above models was evaluated using receiver operating characteristic (ROC) curve analysis. Furthermore, the nomogram was assessed using calibration curve and decision curve analysis. Results: Thirty-seven features were used to establish a radiomics signature, which demonstrated better predictive performance than did the image features model constructed using tumor size and intratumoral vessels in the training and validation cohorts (p <0.05). The radiomics nomogram demonstrated satisfactory discrimination in the training (area under the ROC curve [AUC] 0.93 [95% CI 0.87-0.96]) and validation (AUC 0.87 [95% CI 0.79-0.93]) cohorts and good calibration (Hosmer-Lemeshow p>0.05). Decision curve analysis verified that the radiomics nomogram had the best clinical utility compared with the other models. Conclusion: The radiomics nomogram developed in the present study is a promising tool to predict tumor necrosis and facilitate preoperative clinical decision-making for patients with ccRCC.

Radiotranscriptomics signature-based predictive nomograms for radiotherapy response in patients with nonsmall cell lung cancer: Combination and association of CT features and serum miRNAs levels

Purpose

We aimed to establish radiotranscriptomics signatures based on serum miRNA levels and computed tomography (CT) texture features and develop nomogram models for predicting radiotherapy response in patients with nonsmall cell lung cancer (NSCLC).

Methods

We first used established radioresistant NSCLC cell lines for miRNA selection. At the same time, patients (103 for training set and 71 for validation set) with NSCLC were enrolled. Their pretreatment contrast‐enhanced CT texture features were extracted and their serum miRNA levels were obtained. Then, radiotranscriptomics feature selection was implemented with the least absolute shrinkage and selection operator (LASSO), and signatures were generated by logistic or Cox regression for objective response rate (ORR), overall survival (OS), and progression‐free survival (PFS). Afterward, radiotranscriptomics signature‐based nomograms were constructed and assessed for clinical use.

Results

Four miRNAs and 22 reproducible contrast‐enhanced CT features were used for radiotranscriptomics feature selection and we generated ORR‐, OS‐, and PFS‐ related radiotranscriptomics signatures. In patients with NSCLC who received radiotherapy, the radiotranscriptomics signatures were independently associated with ORR, OS, and PFS in both the training (OR: 2.94, P < .001; HR: 2.90, P < .001; HR: 3.58, P = .001) and validation set (OR: 2.94, P = .026; HR: 2.14, P = .004; HR: 2.64, P = .016). We also obtained a satisfactory nomogram for ORR. The C‐index values for the ORR nomogram were 0.86 [95% confidence interval (CI), 0.75 to 0.92] in the training set and 0.81 (95% CI, 0.69 to 0.89) in the validation set. The calibration‐in‐the‐large and calibration slope performed well. Decision curve analysis indicated a satisfactory net benefit.

Conclusions

The radiotranscriptomics signature could be an independent biomarker for evaluating radiotherapeutic responses in patients with NSCLC. The radiotranscriptomics signature‐based nomogram could be used to predict patients’ ORR, which would represent progress in individualized medicine.

Additional value of metabolic parameters to PET/CT-based radiomics nomogram in predicting lymphovascular invasion and outcome in lung adenocarcinoma

PURPOSE

Lymphovascular invasion (LVI) impairs surgical outcomes in lung adenocarcinoma (LAC) patients. Preoperative prediction of LVI is challenging by using traditional clinical and imaging parameters. The purpose of this study was to investigate the value of the radiomics nomogram integrating clinical factors, CT features, and maximum standardized uptake value (SUVmax) to predict LVI and outcome in LAC and to evaluate the additional value of the SUVmax to the PET/CT-based radiomics nomogram.

METHODS

A total of 272 LAC patients (87 LVI-present LACs and 185 LVI-absent LACs) with PET/CT scans were retrospectively enrolled, and 160 patients with SUVmax ≥ 2.5 of them were used for PET radiomics analysis. Clinical data and CT features were analyzed to select independent LVI predictors. The performance of the independent LVI predictors and SUVmax was evaluated. Two-dimensional (2D) and three-dimensional (3D) CT radiomics signatures (RSs) and PET-RS were constructed with the least absolute shrinkage and selection operator algorithm and radiomics scores (Rad-scores) were calculated. The radiomics nomograms, incorporating Rad-score and independent clinical and CT factors, with SUVmax (RNWS) or without SUVmax (RNWOS) were built. The performance of the models was assessed with respect to calibration, discrimination, and clinical usefulness. All the clinical, PET/CT, pathologic, therapeutic, and radiomics parameters were assessed to identify independent predictors of progression-free survival (PFS).

RESULTS

CT morphology was the independent LVI predictor. SUVmax provided better discrimination capability compared with CT morphology in the training set (P < 0.001) and test set (P = 0.042). A total of 1409 CT and PET radiomics features were extracted and reduced to 8, 8, and 10 features to build the 2D CT-RS, 3D CT-RS, and the PET-RS, respectively. There was no significant difference in AUC between the 2D-RS and 3D-RS (P > 0.05), and 2D CT-RS showed a relatively higher AUC than 3D CT-RS. The CT-RS, the CT-RNWOS, and the CT-RNWS showed good discrimination in the training set (AUC [area under the curve], 0.799, 0.796, and 0.851, respectively) and the test set (AUC, 0.818, 0.822, and 0.838, respectively). There was significant difference in AUC between the CT-RNWS and CT-RNWOS (P = 0.044) in the training set. Decision curve analysis (DCA) demonstrated the CT-RNWS outperformed the CT-RS and the CT-RNWOS in terms of clinical usefulness. Furthermore, DCA showed the PETCT-RNWS provided the highest net benefit compared with the PET-RNWS and CT-RNWS. PFS was significantly different between the pathologic and RNWS-predicted LVI-present and LVI-absent patients (P < 0.001). Carbohydrate antigen 125 (CA125), carcinoembryonic antigen (CEA), neuron-specific enolase (NSE), pathologic LVI, histologic subtype, and SUVmax were independent predictors of PFS in the 244 CT-RNWS-predicted cohort; and CA125, NSE, pathologic LVI, and SUVmax were the independent predictors of PFS in the 141 PETCT-RNWS-predicted cohort.

CONCLUSIONS

The radiomics nomogram, incorporating Rad-score, clinical and PET/CT parameters, shows favorable predictive efficacy for LVI status in LAC. Pathologic LVI and SUVmax are associated with LAC prognosis.

The texture analysis as a predictive method in the assessment of the cytological specimen of CT-guided FNAC of the lung cancer

The lung cancer is the principle cause of the worldwide deaths and its prognosis is poor with a 5-year overall survival rate. Computed tomography (CT) gives many information about the prognosis, but the problem is the subject interpretation of the findings. Thanks to the computer-aided diagnosis/detection (CAD), it is possible to reduce the second opinion. "Radiomics" is an extension of CAD and overlaps the quantitative imaging data of the CT texture analysis (CTTA) with the clinical information, increasing the power and precision of the decision going through the personalized medicine. The aim of this study is to describe the role of the radiomics in the characterization of the pulmonary nodule. For this study, we retrospectively analyzed the images of the 87 NSCLC patients with a waiver of informed consent from the Institutional Review Board (IRB) at the Campania University "Luigi Vanvitelli" of Naples. All tumors were semiautomatically segmented by a radiologist with 10 years of experience using three diameters (AW Server 3.2). The examinations were acquired using 128 MDCT (GSI CT, GE) with a peak tube voltage of 120 kVp, tube current of 100 or 200 mA, and rotation times of 0.5 or 0.8 s. To confirm the imaging results, the FNAC was performed and for every nodule the following parameters were extracted: the presence of the solid component (named = 1), papillary component (named = 2), and mixed component (named = 3). Feature calculation was performed using the HealthMyne software and Integrated Platform That Enables Better Patient Management Decisions For Oncology. The radiologist uses the Rapid Precise Metrics (RPM)™ functionality to identify a lesion with the algorithm and these methods are put to work. The correlation between each feature and the tumor volume was calculated using a two-step cluster statistical analysis. In this retrospective study, in one year from 2018 to 2019 20 patients with lung adenocarcinoma confirmed with FNAC were enrolled. The pathologic results were subdivided into three categories: the solid architecture (group 1), papillary architecture (group 2), and mixed architecture (group 3). Nine lesions resulted with component 1, seven patients with component 2, and 3 patients with component 3. Eight females and 12 males with a median age 61 and 15 years (mean ± SD = 67.4 ± 9.7 years, range 39-73 years) were enrolled. The two results suggest, with p < 0.05, that the GGO variable is a good discriminating estimator of the kurtosis variable: GGO = "no" implies a high kurtosis value, while GGO = "yes" implies a low value. The numerous data obtained from the automatic analysis allow to have a fertile ground on which to develop a new concept of medicine which is precision medicine. The limit of this study is the poor sample. In the future, in order to have a more mature and consolidated discipline, it is necessary to increase the large scale of observations with further studies to establish the rigorous evaluation criteria. In order for radiomics to mature as a discipline in the future, it will be necessary to develop studies that consolidate its role to standardize the collected data.

A Shallow Convolutional Neural Network Predicts Prognosis of Lung Cancer Patients in Multi-Institutional CT-Image Data

Lung cancer is the most common fatal malignancy in adults worldwide, and non-small cell lung cancer (NSCLC) accounts for 85% of lung cancer diagnoses. Computed tomography (CT) is routinely used in clinical practice to determine lung cancer treatment and assess prognosis. Here, we developed LungNet, a shallow convolutional neural network for predicting outcomes of NSCLC patients. We trained and evaluated LungNet on four independent cohorts of NSCLC patients from four medical centers: Stanford Hospital (n = 129), H. Lee Moffitt Cancer Center and Research Institute (n = 185), MAASTRO Clinic (n = 311) and Charité - Universitätsmedizin (n=84). We show that outcomes from LungNet are predictive of overall survival in all four independent survival cohorts as measured by concordance indices of 0.62, 0.62, 0.62 and 0.58 on cohorts 1, 2, 3, and 4, respectively. Further, the survival model can be used, via transfer learning, for classifying benign vs malignant nodules on the Lung Image Database Consortium (n = 1010), with improved performance (AUC=0.85) versus training from scratch (AUC=0.82). LungNet can be used as a noninvasive predictor for prognosis in NSCLC patients and can facilitate interpretation of CT images for lung cancer stratification and prognostication.

Development and evaluation of a venous computed tomography radiomics model to predict lymph node metastasis from non-small cell lung cancer

The objective of this study was to develop a venous computed tomography (CT)-based radiomics model to predict the lymph node metastasis (LNM) in patients with non-small cell lung cancer (NSCLC). A total of 411 consecutive patients with NSCLC underwent tumor resection and lymph node (LN) dissection from January 2018 to September 2018 in our hospital. A radiologist with 20 years of diagnostic experience retrospectively reviewed all CT scans and classified all visible LNs into LNM and non-LNM groups without the knowledge of pathological diagnosis. A logistic regression model (radiomics model) in classification of pathology-confirmed NSCLC patients with and without LNM was developed on radiomics features for NSCLC patients. A morphology model was also developed on qualitative morphology features in venous CT scans. A training group included 288 patients (99 with and 189 without LNM) and a validation group included 123 patients (42 and 81, respectively). The receiver operating characteristic curve was performed to discriminate LNM (+) from LNM (-) for CT-reported status, the morphology model and the radiomics model. The area under the curve value in LNM classification on the training group was significantly greater at 0.79 (95% confidence interval [CI]: 0.77-0.81) by use of the radiomics model (build by best 10 features in predicting LNM) compared with 0.51 by CT-reported LN status (P < .001) or 0.66 (95% CI: 0.64-0.68) by morphology model (build by tumor size and spiculation) (P < .001). Similarly, the area under the curve value on the validation group was 0.73 (95% CI: 0.70-0.76) by the radiomics model, compared with 0.52 or 0.63 (95% CI: 0.60-0.66) by the other 2 (both P < .001). A radiomics model shows excellent performance for predicting LNM in NSCLC patients. This predictive radiomics model may benefit patients to get better treatments such as an appropriate surgery.

CT Texture Analysis Challenges: Influence of Acquisition and Reconstruction Parameters: A Comprehensive Review

Texture analysis in medical imaging is a promising tool that is designed to improve the characterization of abnormal images from patients, to ultimately serve as a predictive or prognostic biomarker. However, the nature of image acquisition itself implies variability in each pixel/voxel value that could jeopardize the usefulness of texture analysis in the medical field. In this review, a search was performed to identify current published data for computed tomography (CT) texture reproducibility and variability. On the basis of this analysis, the critical steps were identified with a view of using texture analysis as a reliable tool in medical imaging. The need to specify the CT scanners used and the associated parameters in published studies is highlighted. Harmonizing acquisition parameters between studies is a crucial step for future texture analysis.

Integrated CT imaging and tissue immune features disclose a radio-immune signature with high prognostic impact on surgically resected NSCLC

OBJECTIVES

Qualitative and quantitative CT imaging features might intercept the multifaceted tumor immune microenvironment (TIME), providing a non-invasive approach to design new prognostic models in NSCLC patients.

MATERIALS AND METHODS

Our study population consisted of 100 surgically resected NSCLC patients among which 31 served as a validation cohort for quantitative image analysis. TIME was classified according to PD-L1 expression and the magnitude of Tumor Infiltrating Lymphocytes (TILs) and further defined as hot or cold by the tissue analysis of effector (CD8-to-CD3^high/PD-1-to-CD8^low) or inert (CD8-to-CD3^low/PD-1-to-CD8^high) phenotypes. CT datasets acted as source for qualitative (semantic, CT-SFs) and quantitative (radiomic, CT-RFs) features which were correlated with clinico-pathological and TIME profiles to determine their impact on survival outcome.

RESULTS

Specific CT-SFs (texture [TXT], effect [EFC] and margins [MRG]) strongly correlated to PD-L1 and TILs status and showed significant impact on survival outcome (TXT, HR:3.39, 95 % CI 1.12-10-27, P < 0.05; EFC, HR:0.41, 95 % CI 0.18-0.93, P < 0.05; MRG, HR:1.93, 95 % CI 0.88-4.25, P = 0.09). Seven CT derived radiomic features were able to sharply discriminate cases with hot (inflamed) vs cold (desert) TIME, which also exhibited opposite OS (long vs short, HR:0.09, 95 % CI 0.04-0.23, P < 0.001) and DFS (long vs short, HR:0.31, 95 % CI 0.16-0.58, P < 0.001). Moreover, we identified 6 prognostic radiomic features among which ClusterProminence displayed the highest statistical significance (HR:0.13, 95 % CI 0.06-0.31, P < 0.001). These findings were independently validated in an additional cohort of NSCLC (HR:0.11, 95 % CI 0.03-0.40, P = 0.001). Finally, in our training cohort we developed a multiparametric prognostic model, interlacing TIME and clinico-pathological characteristics with CT-SFs (ROC curve AUC:0.83, 95 % CI 0.71-0.92, P < 0.001) or CT-RFs (AUC: 0.91, 95 % CI 0.83-0.99, P < 0.001), which appeared to outperform pTNM staging (AUC: 0.66, 95 % CI 0.51-0.80, P < 0.05) in the risk assessment of NSCLC.

CONCLUSION

Higher order CT extracted features associated with specific TIME profiles may reveal a radio-immune signature with prognostic impact on resected NSCLC.

An individualised radiomics composite model predicting prognosis of stage 1 solid lung adenocarcinoma

AIM

To develop and evaluate a radiomics composite model for predicting disease-free survival (DFS) in stage I solid lung adenocarcinoma, and compare it to a simple radiomics model.

MATERIALS AND METHODS

Patients of pathological stage I solid lung adenocarcinoma treated with lobectomy (n = 119) were enrolled retrospectively. Three hundred and ninety-seven radiomics features per lesion were extracted from enhanced chest computed tomography (CT) imaging. Spearman's correlation coefficient and the LASSO (least absolute shrinkage and selection operator) regression model were used to reduce the dimension and select radiomics features. Univariate or multivariate logistic regression was used to build prediction models. A survival curve based on the radiomics composite model was plotted with Kaplan-Meier survival analysis to stratify the risk of recurrence. The confusion matrix, receiver operating characteristic (ROC) curve, and decision curve analysis were used to evaluate the performance of the prediction models.

RESULTS

Recurrence occurred in 22.6% of patients. The survival curve of the radiomics composite model could accurately differentiate high-risk from low-risk patients. In the validation sets, the areas under the ROC curves (AUCs) of the pathological TNM stage (8th IASLC), clinicopathological model, radiomics model, and radiomics composite model were 0.587 (95% confidence interval [CI] 0.502-0.650), 0.629 (95% CI 0.558-0.682), 0.726 (95% CI 0.681-0.770), and 0.849 (95% CI 0.783-0.898), respectively.

CONCLUSION

The prognosis of stage I solid lung adenocarcinoma predicted by an individualised radiomics composite model was more accurate than that of the simple radiomics model.

Radiogenomic-Based Survival Risk Stratification of Tumor Habitat on Gd-T1w MRI Is Associated with Biological Processes in Glioblastoma

PURPOSE

To (i) create a survival risk score using radiomic features from the tumor habitat on routine MRI to predict progression-free survival (PFS) in glioblastoma and (ii) obtain a biological basis for these prognostic radiomic features, by studying their radiogenomic associations with molecular signaling pathways.

EXPERIMENTAL DESIGN

Two hundred three patients with pretreatment Gd-T1w, T2w, T2w-FLAIR MRI were obtained from 3 cohorts: The Cancer Imaging Archive (TCIA; n = 130), Ivy GAP (n = 32), and Cleveland Clinic (n = 41). Gene-expression profiles of corresponding patients were obtained for TCIA cohort. For every study, following expert segmentation of tumor subcompartments (necrotic core, enhancing tumor, peritumoral edema), 936 3D radiomic features were extracted from each subcompartment across all MRI protocols. Using Cox regression model, radiomic risk score (RRS) was developed for every protocol to predict PFS on the training cohort (n = 130) and evaluated on the holdout cohort (n = 73). Further, Gene Ontology and single-sample gene set enrichment analysis were used to identify specific molecular signaling pathway networks associated with RRS features.

RESULTS

Twenty-five radiomic features from the tumor habitat yielded the RRS. A combination of RRS with clinical (age and gender) and molecular features (MGMT and IDH status) resulted in a concordance index of 0.81 (P < 0.0001) on training and 0.84 (P = 0.03) on the test set. Radiogenomic analysis revealed associations of RRS features with signaling pathways for cell differentiation, cell adhesion, and angiogenesis, which contribute to chemoresistance in GBM.

CONCLUSIONS

Our findings suggest that prognostic radiomic features from routine Gd-T1w MRI may also be significantly associated with key biological processes that affect response to chemotherapy in GBM.

The Natural Growth of Subsolid Nodules Predicted by Quantitative Initial CT Features: A Systematic Review

Background: The detection rate for pulmonary nodules, particularly subsolid nodules (SSNs), has been significantly improved. The purpose of this review is to summarize the relationship between quantitative features of initial CT imaging and the subsequent natural growth of SSNs to explore potential reasons for these findings.

Methods: Relevant studies were collected from a literature search of PubMed, Embase, Web of Science, and Cochrane. Data extraction was performed on the patients' basic information, CT methods, and acquisition methods, including quantitative CT features, and statistical methods.

Results: A total of 10 relevant articles were included in our review, which included 850 patients with 1,026 SSNs. Overall, the results were variable, and the key findings were as follows. Seven studies looked at the relationship between the diameter and growth of SSNs, showing that SSNs with larger diameters were associated with increased growth. An additional three studies which focused on the relationship between CT attenuation and the growth of SSNs showed that SSNs with a high CT attenuation were associated with increased growth.

Conclusion: CT attenuation may be useful in predicting the natural growth of SSNs, and mean CT attenuation may be more useful in predicting the natural growth of pure ground glass nodules (GGNs) than part-solid GGNs. While evaluation by diameter did have some limitations, it demonstrates value in predicting the growth of SSNs.

PET/CT Radiomics in Lung Cancer: An Overview

Quantitative extraction of imaging features from medical scans (‘radiomics’) has attracted a lot of research attention in the last few years. The literature has consistently emphasized the potential use of radiomics for computer-assisted diagnosis, as well as for predicting survival and response to treatment. Radiomics is appealing in that it enables full-field analysis of the lesion, provides nearly real-time results, and is non-invasive. Still, a lot of studies suffer from a series of drawbacks such as lack of standardization and repeatability. Such limitations, along with the unmet demand for large enough image datasets for training the algorithms, are major hurdles that still limit the application of radiomics on a large scale. In this paper, we review the current developments, potential applications, limitations, and perspectives of PET/CT radiomics with specific focus on the management of patients with lung cancer.

Stable and discriminating radiomic predictor of recurrence in early stage non-small cell lung cancer: Multi-site study

OBJECTIVES

To evaluate whether combining stability and discriminability criteria in building radiomic classifiers will improve the prognosis of cancer recurrence in early stage non-small cell lung cancer on non-contrast computer tomography (CT).

MATERIALS AND METHODS

CT scans of 610 patients with early stage (IA, IB, IIA) NSCLC from four independent cohorts were evaluated. A total of 350 patients from Cleveland Clinic Foundation and University of Pennsylvania were divided into two equal sets for training (D₁) and validation set (D₂). 80 patients from The Cancer Genome Atlas Lung Adenocarcinoma and Squamous Cell Carcinoma and 195 patients from The Cancer Imaging Archive, were used as independent second (D₃) and third (D₄) validation sets. A linear discriminant analysis (LDA) classifier was built based on the most stable and discriminate features. In addition, a radiomic risk score (RRS) was generated by using least absolute shrinkage and selection operator, Cox regression model to predict time to progression (TTP) following surgery.

RESULTS

A feature selection strategy focusing on both feature discriminability and stability resulted in the classifier having a higher discriminability on validation datasets compared to the discriminability alone criteria in discriminating cancer recurrence (D₂, AUC of 0.75 vs. 0.65; D₃, 0.74 vs. 0.62; D₄, 0.76 vs. 0.63). The RRS generated by most stable-discriminating features was significantly associated with TTP compared to discriminating alone criteria (HR = 1.66, C-index of 0.72 vs. HR = 1.04, C-index of 0.62).

CONCLUSION

Accounting for both stability and discriminability yielded a more generalizable classifier for predicting cancer recurrence and TTP in early stage NSCLC.

[Artificial intelligence in lung imaging]

CLINICAL/METHODICAL ISSUE

Artificial intelligence (AI) has the potential to improve diagnostic accuracy and management in patients with lung disease through automated detection, quantification, classification, and prediction of disease progression.

STANDARD RADIOLOGICAL METHODS

Owing to unspecific symptoms, few well-defined CT disease patterns, and varying prognosis, interstitial lungs disease represents a focus of AI-based research.

METHODICAL INNOVATIONS

Supervised and unsupervised machine learning can identify CT disease patterns using features which may allow the analysis of associations with specific diseases and outcomes.

PERFORMANCE

Machine learning on the one hand improves computer-aided detection of pulmonary nodules. On the other hand it enables further characterization of pulmonary nodules, which may improve resource effectiveness regarding lung cancer screening programs.

ACHIEVEMENTS

There are several challenges regarding AI-based CT data analysis. Besides the need for powerful algorithms, expert annotations and extensive training data sets that reflect physiologic and pathologic variability are required for effective machine learning. Comparability and reproducibility of AI research deserve consideration due to a lack of standardization in this emerging field.

PRACTICAL RECOMMENDATIONS

This review article presents the state of the art and the challenges concerning AI in lung imaging with special consideration of interstitial lung disease, and detection and consideration of pulmonary nodules.

State-of-the-art in radiomics of hepatocellular carcinoma: a review of basic principles, applications, and limitations

Radiomics is a new field in medical imaging with the potential of changing medical practice. Radiomics is characterized by the extraction of several quantitative imaging features which are not visible to the naked eye from conventional imaging modalities, and its correlation with specific relevant clinical endpoints, such as pathology, therapeutic response, and survival. Several studies have evaluated the use of radiomics in patients with hepatocellular carcinoma (HCC) with encouraging results, particularly in the pretreatment prediction of tumor biological characteristics, risk of recurrence, and survival. In spite of this, there are limitations and challenges to be overcome before the implementation of radiomics into clinical routine. In this article, we will review the concepts of radiomics and their current potential applications in patients with HCC. It is important that the multidisciplinary team involved in the treatment of patients with HCC be aware of the basic principles, benefits, and limitations of radiomics in order to achieve a balanced interpretation of the results toward a personalized medicine.

Correction for Magnetic Field Inhomogeneities and Normalization of Voxel Values Are Needed to Better Reveal the Potential of MR Radiomic Features in Lung Cancer

Purpose: To design and validate a preprocessing procedure dedicated to T2-weighted MR images of lung cancers so as to improve the ability of radiomic features to distinguish between adenocarcinoma and other histological types.

Materials and Methods: A discovery set of 52 patients with advanced lung cancer who underwent T2-weighted MR imaging at 3 Tesla in a single center study from August 2017 to May 2019 was used. Findings were then validated using a validation set of 19 additional patients included from May to October 2019. Tumor type was obtained from the pathology report after trans-thoracic needle biopsy, metastatic lymph node or metastasis samples, or surgical excisions. MR images were preprocessed using N4ITK bias field correction and by normalizing voxel intensities with fat as a reference region. Segmentation and extraction of radiomic features were performed with LIFEx software on the raw images, on the N4ITK-corrected images and on the fully preprocessed images. Two analyses were conducted where radiomic features were extracted: (1) from the whole tumor volume (3D analysis); (2) from all slices encompassing the tumor (2D analysis). Receiver operating characteristic (ROC) analysis was used to identify features that could distinguish between adenocarcinoma and other histological types. Sham experiments were also designed to control the number of false positive findings.

Results: There were 31 (12) adenocarcinomas and 21 (7) other histological types in the discovery (validation) set. In 2D, preprocessing increased the number of discriminant radiomic features from 8 without preprocessing to 22 with preprocessing. 2D analysis yielded more features able to identify adenocarcinoma than 3D analysis (12 discriminant radiomic features after preprocessing in 3D). Preprocessing did not increase false positive findings as no discriminant features were identified in any of the sham experiments. The greatest sensitivity of the 2D analysis applied to preprocessed data was confirmed in the validation set.

Conclusion: Correction for magnetic field inhomogeneities and normalization of voxel values are essential to reveal the full potential of radiomic features to identify the tumor histological type from MR T2-weighted images, with classification performance similar to those reported in PET/CT and in multiphase CT in lung cancers.

What's New on Quantitative CT Analysis as a Tool to Predict Growth in Persistent Pulmonary Subsolid Nodules? A Literature Review

Pulmonary subsolid nodules (SSNs) are observed not infrequently on thin-section chest computed tomography (CT) images. SSNs persisting after a follow-up period of three to six months have a high likelihood of being pre-malignant or malignant lesions. Malignant SSNs usually represent the histologic spectrum of pulmonary adenocarcinomas, and pulmonary adenocarcinomas presenting as SSNs exhibit quite heterogeneous behavior. In fact, while most lesions show an indolent course and may grow very slowly or remain stable for many years, others may exhibit significant growth in a relatively short time. Therefore, it is not yet clear which persistent SSNs should be surgically removed and for how many years stable SSNs should be monitored. In order to solve these two open issues, the use of quantitative analysis has been proposed to define the "tailored" management of persistent SSNs. The main purpose of this review was to summarize recent results about quantitative CT analysis as a diagnostic tool for predicting the behavior of persistent SSNs. Thus, a literature search was conducted in PubMed/MEDLINE, Scopus, and Web of Science databases to find original articles published from January 2014 to October 2019. The results of the selected studies are presented and compared in a narrative way.

CT Densitometry and Morphology of Radiofrequency-Ablated Stage IA Non-Small Cell Lung Cancer: Results from the American College of Surgeons Oncology Group Z4033 (Alliance) Trial

PURPOSE

To evaluate tumor and ablation zone morphology and densitometry related to tumor recurrence in participants with Stage IA non-small cell lung cancer undergoing radiofrequency ablation in a prospective, multicenter trial.

MATERIALS AND METHODS

Forty-five participants (median 76 years old; 25 women; 20 men) from 16 sites were followed for 2 years (December 2006 to November 2010) with computed tomography (CT) densitometry. Imaging findings before and after ablation were recorded, including maximum CT attenuation (in Hounsfield units) at precontrast and 45-, 90-, 180-, and 300-s postcontrast.

RESULTS

Every 1-cm increase in the largest axial diameter of the ablation zone at 3-months' follow-up compared to the index tumor reduced the odds of 2-year recurrence by 52% (P = .02). A 1-cm difference performed the best (sensitivity, 0.56; specificity, 0.93; positive likelihood ratio of 8). CT densitometry precontrast and at 45 seconds showed significantly different enhancement patterns in a comparison among pretreated lung cancer (delta = +61.2 HU), tumor recurrence (delta = +57 HU), and treated tumor/ablation zone (delta [change in attenuation] = +16.9 HU), (P < .0001). Densitometry from 45 to 300 s was also different among pretreated tumor (delta = -6.8 HU), recurrence (delta = -11.2 HU), and treated tumor (delta = +12.1 HU; P = .01). Untreated and residual tumor demonstrated washout, whereas treated tumor demonstrated increased attenuation.

CONCLUSIONS

An ablation zone ≥1 cm larger than the initial tumor, based on 3-month follow-up imaging, is recommended to decrease odds of recurrence. CT densitometry can delineate tumor versus treatment zones.

Effects of simulated dose variation on contrast-enhanced CT-based radiomic analysis for Non-Small Cell Lung Cancer

PURPOSE

To examine the potential effect of CT dose variation on radiomic features in vivo using simulated contrast-enhanced CT dose reduction in patients with non-small lung cell cancer (NSCLC).

METHODS

In this retrospective study, we included 69 patients (25 females, 44 males, median age 66 years) with histologically proven NSCLC who underwent a whole contrast-enhanced body FDG-PET/CT for primary staging. To simulate different CT dose levels, we used an algorithm to simulate low-dose CT images based on a noise model derived from phantom experiments. The tumor lesions and reference regions in the paraspinal muscle were manually segmented to obtain three-dimensional regions of interest. Radiomic feature extraction was performed using the PyRadiomics toolbox. The median relative feature value deviation was assessed for each feature and each dose level.

RESULTS

The mean segmented tumor volume was 340 ml. T-stages of the primary tumors were primarily T3/4. For NSCLCs, the median relative feature value deviation in the lowest dose images varied for the calculated features from 52.2% to -49.5%. In general, dose-dependent deviations of feature values showed a monotonous increase or decrease with decreasing dose levels. Statistical analyses revealed significant differences between the dose levels in 91% of features.

CONCLUSIONS

We examined the effects of simulated CT dose reduction on the values of radiomic features in primary NSCLC and showed significant deviations of varying degrees in numerous feature classes. This is a theoretical indicator of potential influence under real conditions, which should be taken into account in clinical use.