Machine learning based on clinico-biological features integrated 18F-FDG PET/CT radiomics for distinguishing squamous cell carcinoma from adenocarcinoma of lung

Automatic lung cancer subtyping using rapid on-site evaluation slides and serum biological markers

BACKGROUND

Rapid on-site evaluation (ROSE) plays an important role during transbronchial sampling, providing an intraoperative cytopathologic evaluation. However, the shortage of cytopathologists limits its wide application. This study aims to develop a deep learning model to automatically analyze ROSE cytological images.

METHODS

The hierarchical multi-label lung cancer subtyping (HMLCS) model that combines whole slide images of ROSE slides and serum biological markers was proposed to discriminate between benign and malignant lesions and recognize different subtypes of lung cancer. A dataset of 811 ROSE slides and paired serum biological markers was retrospectively collected between July 2019 and November 2020, and randomly divided to train, validate, and test the HMLCS model. The area under the curve (AUC) and accuracy were calculated to assess the performance of the model, and Cohen's kappa (κ) was calculated to measure the agreement between the model and the annotation. The HMLCS model was also compared with professional staff.

RESULTS

The HMLCS model achieved AUC values of 0.9540 (95% confidence interval [CI]: 0.9257-0.9823) in malignant/benign classification, 0.9126 (95% CI: 0.8756-0.9365) in malignancy subtyping (non-small cell lung cancer [NSCLC], small cell lung cancer [SCLC], or other malignancies), and 0.9297 (95% CI: 0.9026-0.9603) in NSCLC subtyping (lung adenocarcinoma [LUAD], lung squamous cell carcinoma [LUSC], or NSCLC not otherwise specified [NSCLC-NOS]), respectively. In total, the model achieved an AUC of 0.8721 (95% CI: 0.7714-0.9258) and an accuracy of 0.7184 in the six-class classification task (benign, LUAD, LUSC, NSCLC-NOS, SCLC, or other malignancies). In addition, the model demonstrated a κ value of 0.6183 with the annotation, which was comparable to cytopathologists and superior to trained bronchoscopists and technicians.

CONCLUSION

The HMLCS model showed promising performance in the multiclassification of lung lesions or intrathoracic lymphadenopathy, with potential application to provide real-time feedback regarding preliminary diagnoses of specimens during transbronchial sampling procedures.

CLINICAL TRIAL NUMBER

Not applicable.

Prognostic Significance of 18 F-FDG PET/CT Radiomics in Patients With Resectable Pancreatic Ductal Adenocarcinoma Undergoing Curative Surgery

PURPOSE

This study aimed to investigate the prognostic significance of PET/CT radiomics to predict overall survival (OS) in patients with resectable pancreatic ductal adenocarcinoma (PDAC).

METHODS

We enrolled 627 patients with resectable PDAC who underwent preoperative 18 F-FDG PET/CT and subsequent curative surgery. Radiomics analysis of the PET/CT images for the primary tumor was performed using the Chang-Gung Image Texture Analysis toolbox. Radiomics features were subjected to least absolute shrinkage and selection operator (LASSO) regression to select the most valuable imaging features of OS. The prognostic significance was evaluated by Cox proportional hazards regression analysis. Conventional PET parameters and LASSO score were assessed as predictive factors for OS by time-dependent receiver operating characteristic curve analysis.

RESULTS

During a mean follow-up of 28.8 months, 378 patients (60.3%) died. In the multivariable Cox regression analysis, tumor differentiation, resection margin status, tumor stage, and LASSO score were independent prognostic factors for OS (HR, 1.753, 1.669, 2.655, and 2.946; all P < 0.001, respectively). The time-dependent receiver operating characteristic curve analysis showed that the LASSO score had better predictive performance for OS than conventional PET parameters.

CONCLUSIONS

The LASSO score using the 18 F-FDG PET/CT radiomics of the primary tumor was the independent prognostic factor for predicting OS in patients with resectable PDAC and may be helpful in determining therapeutic and follow-up plans for these patients.

Radiomics in distinguishing between lung adenocarcinoma and lung squamous cell carcinoma: a systematic review and meta-analysis

Objectives

The aim of this study was to systematically review the studies on radiomics models in distinguishing between lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC) and evaluate the classification performance of radiomics models using images from various imaging techniques.

Materials and methods

PubMed, Embase and Web of Science Core Collection were utilized to search for radiomics studies that differentiate between LUAD and LUSC. The assessment of the quality of studies included utilized the improved Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) and Radiomics Quality Score (RQS). Meta-analysis was conducted to assess the classification performance of radiomics models using various imaging techniques.

Results

The qualitative analysis included 40 studies, while the quantitative synthesis included 21 studies. Median RQS for 40 studies was 12 (range -5~19). Sixteen studies were deemed to have a low risk of bias and low concerns regarding applicability. The radiomics model based on CT images had a pooled sensitivity of 0.78 (95%CI: 0.71~0.83), specificity of 0.85 (95%CI:0.73~0.92), and the area under summary receiver operating characteristic curve (SROC-AUC) of 0.86 (95%CI:0.82~0.89). As for PET images, the pooled sensitivity was 0.80 (95%CI: 0.61~0.91), specificity was 0.77 (95%CI: 0.60~0.88), and the SROC-AUC was 0.85 (95%CI: 0.82~0.88). PET/CT images had a pooled sensitivity of 0.87 (95%CI: 0.72~0.94), specificity of 0.88 (95%CI: 0.80~0.93), and an SROC-AUC of 0.93 (95%CI: 0.91~0.95). MRI images had a pooled sensitivity of 0.73 (95%CI: 0.61~0.82), specificity of 0.80 (95%CI: 0.65~0.90), and an SROC-AUC of 0.79 (95%CI: 0.75~0.82).

Conclusion

Radiomics models demonstrate potential in distinguishing between LUAD and LUSC. Nevertheless, it is crucial to conduct a well-designed and powered prospective radiomics studies to establish their credibility in clinical application.

Systematic review registration

https://www.crd.york.ac.uk/PROSPERO/display_record.php?RecordID=412851, identifier CRD42023412851.

Value of 18F-FDG PET/CT-based radiomics features for differentiating primary lung cancer and solitary lung metastasis in patients with colorectal adenocarcinoma

OBJECTIVE

To investigate the value and applicability of 18F-fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) radiomics in differentiating primary lung cancer (PLC) from solitary lung metastasis (SLM) in patients with colorectal cancer (CRC).

MATERIALS AND METHODS

This retrospective study included 103 patients with CRC and solitary pulmonary nodules (SPNs). The least absolute shrinkage and selection operator (LASSO) was used to screen for optimal radiomics features and establish a PET/CT radiomics model. PET/CT Visual and complex models (combining radiomics with PET/CT visual features) were developed. The area under the receiver operating characteristic (ROC) curve (AUC) was used to determine the predictive value and diagnostic efficiency of the models.

RESULTS

The AUC of the PET/CT radiomics model for differentiating PLC from SLM was 0.872 (95% CI: 0.806-0.939), which was not different from that of the visual (0.829 [95% CI: 0.749-0.908; p = .352]). However, the AUC of the complex model (0.936 [95% CI:0.892-0.981]) was significantly higher than that of the PET/CT radiomics (p = .005) and visual model (p = .001). The sensitivity (SEN), specificity (SPE), accuracy (ACC), positive predictive value (PPV), and negative predictive value (NPV) of PET/CT radiomics for differentiating PLC from SLM were 0.720, 0.887, 0.806, 0.857, and 0.770, respectively.

CONCLUSION

PET/CT radiomics can effectively distinguish PLC and SLM in patients with CRC and SPNs and guide the implementation of personalized treatment.

Value of multi-center 18F-FDG PET/CT radiomics in predicting EGFR mutation status in lung adenocarcinoma

BACKGROUND

Accurate, noninvasive, and reliable assessment of epidermal growth factor receptor (EGFR) mutation status and EGFR molecular subtypes is essential for treatment plan selection and individualized therapy in lung adenocarcinoma (LUAD). Radiomics models based on 18F-FDG PET/CT have great potential in identifying EGFR mutation status and EGFR subtypes in patients with LUAD. The validation of multi-center data, model visualization, and interpretation are significantly important for the management, application and trust of machine learning predictive models. However, few EGFR-related research involved model visualization and interpretation, and multi-center trial.

PURPOSE

To develop explainable optimal predictive models based on handcrafted radiomics features (HRFs) extracted from multi-center 18F-FDG PET/CT to predict EGFR mutation status and molecular subtypes in LUAD.

METHODS

Baseline 18F-FDG PET/CT images of 383 LUAD patients from three hospitals and one public data set were collected. Further, 1808 HRFs were extracted from the primary tumor regions using Pyradiomics. Predictive models were built based on cross-combination of seven feature selection methods and seven machine learning algorithms. Yellowbrick and explainable artificial intelligence technology were used for model visualization and interpretation. Receiver operating characteristic curve, classification report and confusion matrix were used for model performance evaluation. Clinical applicability of the optimal models was assessed by decision curve analysis.

RESULTS

STACK feature selection method combined with light gradient boosting machine (LGBM) reached optimal performance in identifying EGFR mutation status ([area under the curve] AUC = 0.81 in the internal test cohort; AUC = 0.62 in the external test cohort). Random forest feature selection method combined with LGBM reached optimal performance in predicting EGFR mutation molecular subtypes (AUC = 0.89 in the internal test cohort; AUC = 0.61 in the external test cohort).

CONCLUSIONS

Explainable machine learning models combined with radiomics features extracted from multi-center/scanner 18F-FDG PET/CT have certain potential to identify EGFR mutation status and subtypes in LUAD, which might be helpful to the treatment of LUAD.

Predicting event-free survival after induction of remission in high-risk pediatric neuroblastoma: combining 123I-MIBG SPECT-CT radiomics and clinical factors

BACKGROUND

Accurately quantifying event-free survival after induction of remission in high-risk neuroblastoma can lead to better subsequent treatment decisions, including whether more aggressive therapy or milder treatment is needed to reduce unnecessary treatment side effects, thereby improving patient survival.

OBJECTIVE

To develop and validate a 123I-metaiodobenzylguanidine (MIBG) single-photon emission computed tomography-computed tomography (SPECT-CT)-based radiomics nomogram and evaluate its value in predicting event-free survival after induction of remission in high-risk neuroblastoma.

MATERIALS AND METHODS

One hundred and seventy-two patients with high-risk neuroblastoma who underwent an 123I-MIBG SPECT-CT examination were retrospectively reviewed. Eighty-seven patients with high-risk neuroblastoma met the final inclusion and exclusion criteria and were randomized into training and validation cohorts in a 7:3 ratio. The SPECT-CT images of patients were visually analyzed to assess the Curie score. The 3D Slicer software tool was used to outline the region of interest of the lumbar 3-5 vertebral bodies on the SPECT-CT images. Radiomics features were extracted and screened, and a radiomics model was constructed with the selected radiomics features. Univariate and multivariate Cox regression analyses were used to determine clinical risk factors and construct the clinical model. The radiomics nomogram was constructed using multivariate Cox regression analysis by incorporating radiomics features and clinical risk factors. C-index and time-dependent receiver operating characteristic curves were used to evaluate the performance of the different models.

RESULTS

The Curie score had the lowest efficacy for the assessment of event-free survival, with a C-index of 0.576 and 0.553 in the training and validation cohorts, respectively. The radiomics model, constructed from 11 radiomics features, outperformed the clinical model in predicting event-free survival in both the training cohort (C-index, 0.780 vs. 0.653) and validation cohort (C-index, 0.687 vs. 0.667). The nomogram predicted the best prognosis for event-free survival in both the training and validation cohorts, with C-indices of 0.819 and 0.712, and 1-year areas under the curve of 0.899 and 0.748, respectively.

CONCLUSION

123I-MIBG SPECT-CT-based radiomics can accurately predict the event-free survival of high-risk neuroblastoma after induction of remission The constructed nomogram may enable an individualized assessment of high-risk neuroblastoma prognosis and assist clinicians in optimizing patient treatment and follow-up plans, thereby potentially improving patient survival.

Clinical application of 18F-fluorodeoxyglucose positron emission tomography/computed tomography radiomics-based machine learning analyses in the field of oncology

Machine learning (ML) analyses using 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography (PET)/computed tomography (CT) radiomics features have been applied in the field of oncology. The current review aimed to summarize the current clinical articles about 18F-FDG PET/CT radiomics-based ML analyses to solve issues in classifying or constructing prediction models for several types of tumors. In these studies, lung and mediastinal tumors were the most commonly evaluated lesions, followed by lymphatic, abdominal, head and neck, breast, gynecological, and other types of tumors. Previous studies have commonly shown that 18F-FDG PET radiomics-based ML analysis has good performance in differentiating benign from malignant tumors, predicting tumor characteristics and stage, therapeutic response, and prognosis by examining significant differences in the area under the receiver operating characteristic curves, accuracies, or concordance indices (> 0.70). However, these studies have reported several ML algorithms. Moreover, different ML models have been applied for the same purpose. Thus, various procedures were used in 18F-FDG PET/CT radiomics-based ML analysis in oncology, and 18F-FDG PET/CT radiomics-based ML models, which are easy and universally applied in clinical practice, would be expected to be established.

AI/ML advances in non-small cell lung cancer biomarker discovery

Lung cancer is the leading cause of cancer deaths among both men and women, representing approximately 25% of cancer fatalities each year. The treatment landscape for non-small cell lung cancer (NSCLC) is rapidly evolving due to the progress made in biomarker-driven targeted therapies. While advancements in targeted treatments have improved survival rates for NSCLC patients with actionable biomarkers, long-term survival remains low, with an overall 5-year relative survival rate below 20%. Artificial intelligence/machine learning (AI/ML) algorithms have shown promise in biomarker discovery, yet NSCLC-specific studies capturing the clinical challenges targeted and emerging patterns identified using AI/ML approaches are lacking. Here, we employed a text-mining approach and identified 215 studies that reported potential biomarkers of NSCLC using AI/ML algorithms. We catalogued these studies with respect to BEST (Biomarkers, EndpointS, and other Tools) biomarker sub-types and summarized emerging patterns and trends in AI/ML-driven NSCLC biomarker discovery. We anticipate that our comprehensive review will contribute to the current understanding of AI/ML advances in NSCLC biomarker research and provide an important catalogue that may facilitate clinical adoption of AI/ML-derived biomarkers.

Clinico-biological-radiomics (CBR) based machine learning for improving the diagnostic accuracy of FDG-PET false-positive lymph nodes in lung cancer

BACKGROUND

The main problem of positron emission tomography/computed tomography (PET/CT) for lymph node (LN) staging is the high false positive rate (FPR). Thus, we aimed to explore a clinico-biological-radiomics (CBR) model via machine learning (ML) to reduce FPR and improve the accuracy for predicting the hypermetabolic mediastinal-hilar LNs status in lung cancer than conventional PET/CT.

METHODS

A total of 260 lung cancer patients with hypermetabolic mediastinal-hilar LNs (SUVmax ≥ 2.5) were retrospectively reviewed. Patients were treated with surgery with systematic LN resection and pathologically divided into the LN negative (LN-) and positive (LN +) groups, and randomly assigned into the training (n = 182) and test (n = 78) sets. Preoperative CBR dataset containing 1738 multi-scale features was constructed for all patients. Prediction models for hypermetabolic LNs status were developed using the features selected by the supervised ML algorithms, and evaluated using the classical diagnostic indicators. Then, a nomogram was developed based on the model with the highest area under the curve (AUC) and the lowest FPR, and validated by the calibration plots.

RESULTS

In total, 109 LN- and 151 LN + patients were enrolled in this study. 6 independent prediction models were developed to differentiate LN- from LN + patients using the selected features from clinico-biological-image dataset, radiomics dataset, and their combined CBR dataset, respectively. The DeLong test showed that the CBR Model containing all-scale features held the highest predictive efficiency and the lowest FPR among all of established models (p < 0.05) in both the training and test sets (AUCs of 0.90 and 0.89, FPRs of 12.82% and 6.45%, respectively) (p < 0.05). The quantitative nomogram based on CBR Model was validated to have a good consistency with actual observations.

CONCLUSION

This study presents an integrated CBR nomogram that can further reduce the FPR and improve the accuracy of hypermetabolic mediastinal-hilar LNs evaluation than conventional PET/CT in lung cancer, thereby greatly reducing the risk of overestimation and assisting for precision treatment.

Radiomic analysis will add differential diagnostic value of benign and malignant pulmonary nodules: a hybrid imaging study based on [18F]FDG and [18F]FLT PET/CT

PURPOSE

To investigate the clinical value of radiomic analysis on [18F]FDG and [18F]FLT PET on the differentiation of [18F]FDG-avid benign and malignant pulmonary nodules (PNs).

METHODS

Data of 113 patients with inconclusive PNs based on preoperative [18F]FDG PET/CT who underwent additional [18F]FLT PET/CT scans within a week were retrospectively analyzed in the present study. Three methods of analysis including visual analysis, radiomic analysis based on [18F]FDG PET/CT images alone, and radiomic analysis based on dual-tracer PET/CT images were evaluated for differential diagnostic value of benign and malignant PNs.

RESULTS

A total of 678 radiomic features were extracted from volumes of interest (VOIs) of 123 PNs. Fourteen valuable features were thereafter selected. Based on a visual analysis of [18F]FDG PET/CT images, the diagnostic accuracy, sensitivity, and specificity were 61.6%, 90%, and 28.8%, respectively. For the test set, the area under the curve (AUC), sensitivity, and specificity of the radiomic models based on [18F]FDG PET/CT plus [18F]FLT signature were equal or better than radiomics based on [18F]FDG PET/CT only (0.838 vs 0.810, 0.778 vs 0.778, 0.750 vs 0.688, respectively).

CONCLUSION

Radiomic analysis based on dual-tracer PET/CT images is clinically promising and feasible for the differentiation between benign and malignant PNs.

CLINICAL RELEVANCE STATEMENT

Radiomic analysis will add differential diagnostic value of benign and malignant pulmonary nodules: a hybrid imaging study based on [18F]FDG and [18F]FLT PET/CT.

KEY POINTS

• Radiomics brings new insights into the differentiation of benign and malignant pulmonary nodules beyond the naked eyes. • Dual-tracer imaging shows the biological behaviors of cancerous cells from different aspects. • Radiomics helps us get to the histological view in a non-invasive approach.

CT Radiomics Predict EGFR-T790M Resistance Mutation in Advanced Non-Small Cell Lung Cancer Patients After Progression on First-line EGFR-TKI

RATIONALE AND OBJECTIVES

We aim to explore the value of chest CT radiomics in predicting the epidermal growth factor receptor (EGFR)-T790M resistance mutation of advanced non-small cell lung cancer (NSCLC) patients after the failure of first-line EGFR-tyrosine kinase inhibitor (EGFR-TKI).

MATERIALS AND METHODS

A total of 211 and 135 advanced NSCLC patients with tumor tissue-based (Cohort-1) or circulating tumor DNA (ctDNA)-based (Cohort-2) EGFR-T790M testing were included, respectively. Cohort-1 was used for modeling and Cohort-2 was for models' validation. Radiomic features were extracted from tumor lesions on chest nonenhanced CT (NECT) and/or contrast-enhanced CT (CECT). We used eight feature selectors and eight classifier algorithms to establish radiomic models. Models were evaluated by area under the receiver operating characteristic curve (AUC), calibration curve, and decision curve analysis (DCA).

RESULTS

CT morphological manifestations of peripheral location and pleural indentation sign were associated with EGFR-T790M. For NECT, CECT, and NECT+CECT radiomic features, the feature selector and classifier algorithms of LASSO and Stepwise logistic regression, Boruta and SVM, and LASSO and SVM were chosen to develop the optimal model, respectively (AUC: 0.844, 0.811, and 0.897). All models performed well in calibration curves and DCA. Independent validation of models in Cohort-2 revealed that both NECT and CECT models individually had limited power for predicting EGFR-T790M mutation detected by ctDNA (AUC: 0.649, 0.675), while the NECT+CECT radiomic model had a satisfactory AUC (0.760).

CONCLUSION

This study proved the feasibility of using CT radiomic features to predict the EGFR-T790M resistance mutation, which could be helpful in guiding personalized therapeutic strategies.

Machine Learning Techniques to Predict Timeliness of Care among Lung Cancer Patients

Delays in the assessment, management, and treatment of lung cancer patients may adversely impact prognosis and survival. This study is the first to use machine learning techniques to predict the quality and timeliness of care among lung cancer patients, utilising data from the Victorian Lung Cancer Registry (VLCR) between 2011 and 2022, in Victoria, Australia. Predictor variables included demographic, clinical, hospital, and geographical socio-economic indices. Machine learning methods such as random forests, k-nearest neighbour, neural networks, and support vector machines were implemented and evaluated using 20% out-of-sample cross validations via the area under the curve (AUC). Optimal model parameters were selected based on 10-fold cross validation. There were 11,602 patients included in the analysis. Evaluated quality indicators included, primarily, overall proportion achieving "time from referral date to diagnosis date ≤ 28 days" and proportion achieving "time from diagnosis date to first treatment date (any intent) ≤ 14 days". Results showed that the support vector machine learning methods performed well, followed by nearest neighbour, based on out-of-sample AUCs of 0.89 (in-sample = 0.99) and 0.85 (in-sample = 0.99) for the first indicator, respectively. These models can be implemented in the registry databases to help healthcare workers identify patients who may not meet these indicators prospectively and enable timely interventions.

Novel tools for early diagnosis and precision treatment based on artificial intelligence

Lung cancer has the highest mortality rate among all cancers in the world. Hence, early diagnosis and personalized treatment plans are crucial to improving its 5-year survival rate. Chest computed tomography (CT) serves as an essential tool for lung cancer screening, and pathology images are the gold standard for lung cancer diagnosis. However, medical image evaluation relies on manual labor and suffers from missed diagnosis or misdiagnosis, and physician heterogeneity. The rapid development of artificial intelligence (AI) has brought a whole novel opportunity for medical task processing, demonstrating the potential for clinical application in lung cancer diagnosis and treatment. AI technologies, including machine learning and deep learning, have been deployed extensively for lung nodule detection, benign and malignant classification, and subtype identification based on CT images. Furthermore, AI plays a role in the non-invasive prediction of genetic mutations and molecular status to provide the optimal treatment regimen, and applies to the assessment of therapeutic efficacy and prognosis of lung cancer patients, enabling precision medicine to become a reality. Meanwhile, histology-based AI models assist pathologists in typing, molecular characterization, and prognosis prediction to enhance the efficiency of diagnosis and treatment. However, the leap to extensive clinical application still faces various challenges, such as data sharing, standardized label acquisition, clinical application regulation, and multimodal integration. Nevertheless, AI holds promising potential in the field of lung cancer to improve cancer care.

Preoperative prediction of microvascular invasion and perineural invasion in pancreatic ductal adenocarcinoma with 18F-FDG PET/CT radiomics analysis

AIM

To develop and validate a predictive model based on 2-[¹⁸F]-fluoro-2-deoxy-d-glucose (¹⁸F-FDG) positron-emission tomography (PET)/computed tomography (CT) radiomics features and clinicopathological parameters to preoperatively identify microvascular invasion (MVI) and perineural invasion (PNI), which are important predictors of poor prognosis in patients with pancreatic ductal adenocarcinoma (PDAC).

MATERIALS AND METHODS

Preoperative ¹⁸F-FDG PET/CT images and clinicopathological parameters of 170 patients in PDAC were collected retrospectively. The whole tumour and its peritumoural variants (tumour dilated with 3, 5, and 10 mm pixels) were applied to add tumour periphery information. A feature-selection algorithm was employed to mine mono-modality and fused feature subsets, then conducted binary classification using gradient boosted decision trees.

RESULTS

For MVI prediction, the model performed best on a fused subset of ¹⁸F-FDG PET/CT radiomics features and two clinicopathological parameters, with an area under the receiver operating characteristic curve (AUC) of 83.08%, accuracy of 78.82%, recall of 75.08%, precision of 75.5%, and F1-score of 74.59%. For PNI prediction, the model achieved best prediction results only on the subset of PET/CT radiomics features, with AUC of 94%, accuracy of 89.33%, recall of 90%, precision of 87.81%, and F1 score of 88.35%. In both models, 3 mm dilation on the tumour volume produced the best results.

CONCLUSIONS

The radiomics predictors from preoperative ¹⁸F-FDG PET/CT imaging exhibited instructive predictive efficacy in the identification of MVI and PNI status preoperatively in PDAC. Peritumoural information was shown to assist in MVI and PNI predictions.

Classification of Histological Types and Stages in Non-small Cell Lung Cancer Using Radiomic Features Based on CT Images

Non-invasive diagnostic method based on radiomic features in patients with non-small cell lung cancer (NSCLC) has attracted attention. This study aimed to develop a CT image-based model for both histological typing and clinical staging of patients with NSCLC. A total of 309 NSCLC patients with 537 CT series from The Cancer Imaging Archive (TCIA) database were included in this study. All patients were randomly divided into the training set (247 patients, 425 CT series) and testing set (62 patients, 112 CT series). A total of 107 radiomic features were extracted. Four classifiers including random forest, XGBoost, support vector machine, and logistic regression were used to construct the classification model. The classification model had two output layers: histological type (adenocarcinoma, squamous cell carcinoma, and large cell) and clinical stage (I, II, and III) of NSCLC patients. The area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) with 95% confidence interval (CI) were utilized to evaluate the performance of the model. Seven features were selected for inclusion in the classification model. The random forest model had the best classification ability compared with other classifiers. The AUC of the RF model for histological typing and clinical staging of NSCLC patients in the testing set was 0.700 (95% CI, 0.641-0.759) and 0.881 (95% CI, 0.842-0.920), respectively. The CT image-based radiomic feature model had good classification ability for both histological typing and clinical staging of patients with NSCLC.

Value of contrast-enhanced magnetic resonance imaging-T2WI-based radiomic features in distinguishing lung adenocarcinoma from lung squamous cell carcinoma with solid components >8 mm

Background

Radiomics is one of the research frontiers in the field of imaging and has excellent diagnostic performance. However, there is a lack of magnetic resonance imaging (MRI)-based omics studies on identifying pathological subtypes of lung cancer. Here we explored the value of the contrast-enhanced MRI-T2-weighted imaging (T2WI)-based radiomic analysis in distinguishing adenocarcinoma (Ade) from squamous cell carcinoma (Squ) with solid components >8 mm.

Methods

A retrospective analysis was performed of a total of 71 lung cancer patients who undergoing contrast-enhanced MRI and computed tomography (CT) before treatment, and the nodules had solid components ≥8 mm in our center from January 2020 to September 2021. All enrolled patients were divided into Squ and Ade groups according to the pathological results. In addition, the two groups were randomly divided into training set and validation set in a ratio of about 7:3. Radiomics software was used to extract the relevant radiomic features. The least absolute shrinkage and selection operator (Lasso) was used to screen radiomic features that were most relevant to lung cancer subtypes, thus calculating the radiomic scores (Rad-score) and constructing the radiomic models. Multivariate logistic regression was used to combine relevant clinical features with Rad-score to form combined model nomograms. The receiver operating characteristic (ROC) curves. the area under the ROC curve (AUC), the decision curve analysis (DCA) and the DeLong's test were used to evaluate the clinical application potentials.

Results

The sensitivity and specificity of the clinical model based on smoking was 75.0% and 93.8%. The AUC of the constructed magnetic resonance (MR)-Rad model for differentiating the pathological subtypes of lung cancer was 0.8651 in the validation sets. The AUC of the CT-Rad model in the validation set were 0.9286. The combined model constructed by combining clinical features and Rad-score had AUC of 0.8016, for identifying the 2 pathological subtypes of lung cancer in the validation set. There was no significant difference in diagnostic performance between MR-Rad model and CT-Rad model (P>0.05).

Conclusions

The MR-Rad model has a diagnostic performance similar to that of CT-Rad model, while the diagnostic performance of the combined mode was better than the single MR model.

Machine learning-based radiomic computed tomography phenotyping of thymic epithelial tumors: Predicting pathological and survival outcomes

OBJECTIVE

For patients with thymic epithelial tumors, accurately predicting clinicopathological outcomes remains challenging. We aimed to investigate the performance of machine learning-based radiomic computed tomography phenotyping for predicting pathological (World Health Organization [WHO] type and TNM stage) and survival outcomes (overall and progression-free survival) in patients with thymic epithelial tumors.

METHODS

This retrospective study included patients with thymic epithelial tumors between January 2001 and January 2022. The radiomic features were extracted from preoperative unenhanced computed tomography images. After strict feature selection, random forest and random survival forest models were fitted to predict pathological and survival outcomes, respectively. The model performance was assessed by the area under the curve (AUC) and validated internally by the bootstrap method.

RESULTS

In total, 124 patients with a median age of 61 years were included. The radiomics random forest models of WHO type and TNM stage showed satisfactory performance with an AUC_WHO of 0.898 (95% CI, 0.753-1.000) and an AUC_TNM of 0.766 (95% CI, 0.642-0.886). For overall survival and progression-free survival prediction, the radiomics random survival forest models showed good performance (integrated AUCs, 0.923; 95% CI, 0.691-1.000 and 0.702; 95% CI, 0.513-0.875, respectively), and the integrated AUCs increased to 0.935 (95% CI, 0.705-1.000) and 0.811 (95% CI, 0.647-0.942), respectively, when combined with clinicopathological features.

CONCLUSIONS

Machine learning-based radiomic computed tomography phenotyping might allow for the satisfactory prediction of pathological and survival outcomes and further improve prognostic performance when integrated with clinicopathological features in patients with thymic epithelial tumors.

The role of artificial intelligence based on PET/CT radiomics in NSCLC: Disease management, opportunities, and challenges

Objectives

Lung cancer has been widely characterized through radiomics and artificial intelligence (AI). This review aims to summarize the published studies of AI based on positron emission tomography/computed tomography (PET/CT) radiomics in non-small-cell lung cancer (NSCLC).

Materials and methods

A comprehensive search of literature published between 2012 and 2022 was conducted on the PubMed database. There were no language or publication status restrictions on the search. About 127 articles in the search results were screened and gradually excluded according to the exclusion criteria. Finally, this review included 39 articles for analysis.

Results

Classification is conducted according to purposes and several studies were identified at each stage of disease:1) Cancer detection (n=8), 2) histology and stage of cancer (n=11), 3) metastases (n=6), 4) genotype (n=6), 5) treatment outcome and survival (n=8). There is a wide range of heterogeneity among studies due to differences in patient sources, evaluation criteria and workflow of radiomics. On the whole, most models show diagnostic performance comparable to or even better than experts, and the common problems are repeatability and clinical transformability.

Conclusion

AI-based PET/CT Radiomics play potential roles in NSCLC clinical management. However, there is still a long way to go before being translated into clinical application. Large-scale, multi-center, prospective research is the direction of future efforts, while we need to face the risk of repeatability of radiomics features and the limitation of access to large databases.

Distinction of Lymphoma from Sarcoidosis on 18F-FDG PET/CT: Evaluation of Radiomics-Feature-Guided Machine Learning Versus Human Reader Performance

Sarcoidosis and lymphoma often share common features on ¹⁸F-FDG PET/CT, such as intense hypermetabolic lesions in lymph nodes and multiple organs. We aimed at developing and validating radiomics signatures to differentiate sarcoidosis from Hodgkin lymphoma (HL) and diffuse large B-cell lymphoma (DLBCL). Methods: We retrospectively collected 420 patients (169 sarcoidosis, 140 HL, and 111 DLBCL) who underwent pretreatment ¹⁸F-FDG PET/CT at the University Hospital of Liege. The studies were randomly distributed to 4 physicians, who gave their diagnostic suggestion among the 3 diseases. The individual and pooled performance of the physicians was then calculated. Interobserver variability was evaluated using a sample of 34 studies interpreted by all physicians. Volumes of interest were delineated over the lesions and the liver using MIM software, and 215 radiomics features were extracted using the RadiomiX Toolbox. Models were developed combining clinical data (age, sex, and weight) and radiomics (original and tumor-to-liver TLR radiomics), with 7 different feature selection approaches and 4 different machine-learning (ML) classifiers, to differentiate sarcoidosis and lymphomas on both lesion-based and patient-based approaches. Results: For identifying lymphoma versus sarcoidosis, physicians' pooled sensitivity, specificity, area under the receiver-operating-characteristic curve (AUC), and accuracy were 0.99 (95% CI, 0.97-1.00), 0.75 (95% CI, 0.68-0.81), 0.87 (95% CI, 0.84-0.90), and 89.3%, respectively, whereas for identifying HL in the tumor population, it was 0.58 (95% CI, 0.49-0.66), 0.82 (95% CI, 0.74-0.89), 0.70 (95% CI, 0.64-0.75) and 68.5%, respectively. Moderate agreement was found among observers for the diagnosis of lymphoma versus sarcoidosis and HL versus DLBCL, with Fleiss κ-values of 0.66 (95% CI, 0.45-0.87) and 0.69 (95% CI, 0.45-0.93), respectively. The best ML models for identifying lymphoma versus sarcoidosis showed an AUC of 0.94 (95% CI, 0.93-0.95) and 0.85 (95% CI, 0.82-0.88) in lesion- and patient-based approaches, respectively, using TLR radiomics (plus age for the second). To differentiate HL from DLBCL, we obtained an AUC of 0.95 (95% CI, 0.93-0.96) in the lesion-based approach using TLR radiomics and 0.86 (95% CI, 0.80-0.91) in the patient-based approach using original radiomics and age. Conclusion: Characterization of sarcoidosis and lymphoma lesions is feasible using ML and radiomics, with very good to excellent performance, equivalent to or better than that of physicians, who showed significant interobserver variability in their assessment.

Development and validation of a radiomic nomogram based on pretherapy dual-energy CT for distinguishing adenocarcinoma from squamous cell carcinoma of the lung

Objective

Based on pretherapy dual-energy computed tomography (DECT) images, we developed and validated a nomogram combined with clinical parameters and radiomic features to predict the pathologic subtypes of non-small cell lung cancer (NSCLC) - adenocarcinoma (ADC) and squamous cell carcinoma (SCC).

Methods

A total of 129 pathologically confirmed NSCLC patients treated at the Second Affiliated Hospital of Nanchang University from October 2017 to October 2021 were retrospectively analyzed. Patients were randomly divided in a ratio of 7:3 (n=90) into training and validation cohorts (n=39). Patients' pretherapy clinical parameters were recorded. Radiomics features of the primary lesion were extracted from two sets of monoenergetic images (40 keV and 100 keV) in arterial phases (AP) and venous phases (VP). Features were selected successively through the intra-class correlation coefficient (ICC) and the least absolute shrinkage and selection operator (LASSO). Multivariate logistic regression analysis was then performed to establish predictive models. The prediction performance between models was evaluated and compared using the receiver operating characteristic (ROC) curve, DeLong test, and Akaike information criterion (AIC). A nomogram was developed based on the model with the best predictive performance to evaluate its calibration and clinical utility.

Results

A total of 87 ADC and 42 SCC patients were enrolled in this study. Among the five constructed models, the integrative model (AUC: Model 4 = 0.92, Model 5 = 0.93) combining clinical parameters and radiomic features had a higher AUC than the individual clinical models or radiomic models (AUC: Model 1 = 0.84, Model 2 = 0.79, Model 3 = 0.84). The combined clinical-venous phase radiomics model had the best predictive performance, goodness of fit, and parsimony; the area under the ROC curve (AUC) of the training and validation cohorts was 0.93 and 0.90, respectively, and the AIC value was 60.16. Then, this model was visualized as a nomogram. The calibration curves demonstrated it's good calibration, and decision curve analysis (DCA) proved its clinical utility.

Conclusion

The combined clinical-radiomics model based on pretherapy DECT showed good performance in distinguishing ADC and SCC of the lung. The nomogram constructed based on the best-performing combined clinical-venous phase radiomics model provides a relatively accurate, convenient and noninvasive method for predicting the pathological subtypes of ADC and SCC in NSCLC.

Automated data preparation for in vivo tumor characterization with machine learning

Background

This study proposes machine learning-driven data preparation (MLDP) for optimal data preparation (DP) prior to building prediction models for cancer cohorts.

Methods

A collection of well-established DP methods were incorporated for building the DP pipelines for various clinical cohorts prior to machine learning. Evolutionary algorithm principles combined with hyperparameter optimization were employed to iteratively select the best fitting subset of data preparation algorithms for the given dataset. The proposed method was validated for glioma and prostate single center cohorts by 100-fold Monte Carlo (MC) cross-validation scheme with 80-20% training-validation split ratio. In addition, a dual-center diffuse large B-cell lymphoma (DLBCL) cohort was utilized with Center 1 as training and Center 2 as independent validation datasets to predict cohort-specific clinical endpoints. Five machine learning (ML) classifiers were employed for building prediction models across all analyzed cohorts. Predictive performance was estimated by confusion matrix analytics over the validation sets of each cohort. The performance of each model with and without MLDP, as well as with manually-defined DP were compared in each of the four cohorts.

Results

Sixteen of twenty established predictive models demonstrated area under the receiver operator characteristics curve (AUC) performance increase utilizing the MLDP. The MLDP resulted in the highest performance increase for random forest (RF) (+0.16 AUC) and support vector machine (SVM) (+0.13 AUC) model schemes for predicting 36-months survival in the glioma cohort. Single center cohorts resulted in complex (6-7 DP steps) DP pipelines, with a high occurrence of outlier detection, feature selection and synthetic majority oversampling technique (SMOTE). In contrast, the optimal DP pipeline for the dual-center DLBCL cohort only included outlier detection and SMOTE DP steps.

Conclusions

This study demonstrates that data preparation prior to ML prediction model building in cancer cohorts shall be ML-driven itself, yielding optimal prediction models in both single and multi-centric settings.

CT-based radiomics in predicting pathological response in non-small cell lung cancer patients receiving neoadjuvant immunotherapy

Objectives

In radiomics, high-throughput algorithms extract objective quantitative features from medical images. In this study, we evaluated CT-based radiomics features, clinical features, in-depth learning features, and a combination of features for predicting a good pathological response (GPR) in non-small cell lung cancer (NSCLC) patients receiving immunotherapy-based neoadjuvant therapy (NAT).

Materials and methods

We reviewed 62 patients with NSCLC who received surgery after immunotherapy-based NAT and collected clinicopathological data and CT images before and after immunotherapy-based NAT. A series of image preprocessing was carried out on CT scanning images: tumor segmentation, conventional radiomics feature extraction, deep learning feature extraction, and normalization. Spearman correlation coefficient, principal component analysis (PCA), and least absolute shrinkage and selection operator (LASSO) were used to screen features. The pretreatment traditional radiomics combined with clinical characteristics (before_rad_cil) model and pretreatment deep learning characteristics (before_dl) model were constructed according to the data collected before treatment. The data collected after NAT created the after_rad_cil model and after_dl model. The entire model was jointly constructed by all clinical features, conventional radiomics features, and deep learning features before and after neoadjuvant treatment. Finally, according to the data obtained before and after treatment, the before_nomogram and after_nomogram were constructed.

Results

In the before_rad_cil model, four traditional radiomics features ("original_shape_flatness," "wavelet hhl_firer_skewness," "wavelet hlh_firer_skewness," and "wavelet lll_glcm_correlation") and two clinical features ("gender" and "N stage") were screened out to predict a GPR. The average prediction accuracy (ACC) after modeling with k-nearest neighbor (KNN) was 0.707. In the after_rad_cil model, nine features predictive of GPR were obtained after feature screening, among which seven were traditional radiomics features: "exponential_firer_skewness," "exponential_glrlm_runentropy," "log- sigma-5-0-mm-3d_firer_kurtosis," "logarithm_skewness," "original_shape_elongation," "original_shape_brilliance," and "wavelet llh_glcm_clustershade"; two were clinical features: "after_CRP" and "after lymphocyte percentage." The ACC after modeling with support vector machine (SVM) was 0.682. The before_dl model and after_dl model were modeled by SVM, and the ACC was 0.629 and 0.603, respectively. After feature screening, the entire model was constructed by multilayer perceptron (MLP), and the ACC of the GPR was the highest, 0.805. The calibration curve showed that the predictions of the GPR by the before_nomogram and after_nomogram were in consensus with the actual GPR.

Conclusion

CT-based radiomics has a good predictive ability for a GPR in NSCLC patients receiving immunotherapy-based NAT. Among the radiomics features combined with the clinicopathological information model, deep learning feature model, and the entire model, the entire model had the highest prediction accuracy.

Identification of pulmonary adenocarcinoma and benign lesions in isolated solid lung nodules based on a nomogram of intranodal and perinodal CT radiomic features

To develop and validate a predictive model based on clinical radiology and radiomics to enhance the ability to distinguish between benign and malignant solitary solid pulmonary nodules. In this study, we retrospectively collected computed tomography (CT) images and clinical data of 286 patients with isolated solid pulmonary nodules diagnosed by surgical pathology, including 155 peripheral adenocarcinomas and 131 benign nodules. They were randomly divided into a training set and verification set at a 7:3 ratio, and 851 radiomic features were extracted from thin-layer enhanced venous phase CT images by outlining intranodal and perinodal regions of interest. We conducted preprocessing measures of image resampling and eigenvalue normalization. The minimum redundancy maximum relevance (mRMR) and least absolute shrinkage and selection operator (lasso) methods were used to downscale and select features. At the same time, univariate and multifactorial analyses were performed to screen clinical radiology features. Finally, we constructed a nomogram based on clinical radiology, intranodular, and perinodular radiomics features. Model performance was assessed by calculating the area under the receiver operating characteristic curve (AUC), and the clinical decision curve (DCA) was used to evaluate the clinical practicability of the models. Univariate and multivariate analyses showed that the two clinical factors of sex and age were statistically significant. Lasso screened four intranodal and four perinodal radiomic features. The nomogram based on clinical radiology, intranodular, and perinodular radiomics features showed the best predictive performance (AUC=0.95, accuracy=0.89, sensitivity=0.83, specificity=0.96), which was superior to other independent models. A nomogram based on clinical radiology, intranodular, and perinodular radiomics features is helpful to improve the ability to predict benign and malignant solitary pulmonary nodules.

[18F]FDG-PET/CT radiomics and artificial intelligence in lung cancer: Technical aspects and potential clinical applications

Lung cancer is the second most common cancer and the leading cause of cancer-related death worldwide. Molecular imaging using [¹⁸F]fluorodeoxyglucose Positron Emission Tomography and/or Computed Tomography ([¹⁸F]FDG-PET/CT) plays an essential role in the diagnosis, evaluation of response to treatment, and prediction of outcomes. The images are evaluated using qualitative and conventional quantitative indices. However, there is far more information embedded in the images, which can be extracted by sophisticated algorithms. Recently, the concept of uncovering and analyzing the invisible data extracted from medical images, called radiomics, is gaining more attention. Currently, [¹⁸F]FDG-PET/CT radiomics is growingly evaluated in lung cancer to discover if it enhances the diagnostic performance or implication of [¹⁸F]FDG-PET/CT in the management of lung cancer. In this review, we provide a short overview of the technical aspects, as they are discussed in different articles of this special issue. We mainly focus on the diagnostic performance of the [¹⁸F]FDG-PET/CT-based radiomics and the role of artificial intelligence in non-small cell lung cancer, impacting the early detection, staging, prediction of tumor subtypes, biomarkers, and patient's outcomes.

Radiomics in Oncological PET Imaging: A Systematic Review-Part 1, Supradiaphragmatic Cancers

Radiomics is an upcoming field in nuclear oncology, both promising and technically challenging. To summarize the already undertaken work on supradiaphragmatic neoplasia and assess its quality, we performed a literature search in the PubMed database up to 18 February 2022. Inclusion criteria were: studies based on human data; at least one specified tumor type; supradiaphragmatic malignancy; performing radiomics on PET imaging. Exclusion criteria were: studies only based on phantom or animal data; technical articles without a clinically oriented question; fewer than 30 patients in the training cohort. A review database containing PMID, year of publication, cancer type, and quality criteria (number of patients, retrospective or prospective nature, independent validation cohort) was constructed. A total of 220 studies met the inclusion criteria. Among them, 119 (54.1%) studies included more than 100 patients, 21 studies (9.5%) were based on prospectively acquired data, and 91 (41.4%) used an independent validation set. Most studies focused on prognostic and treatment response objectives. Because the textural parameters and methods employed are very different from one article to another, it is complicated to aggregate and compare articles. New contributions and radiomics guidelines tend to help improving quality of the reported studies over the years.

The Machine Learning Model for Distinguishing Pathological Subtypes of Non-Small Cell Lung Cancer

Purpose

Machine learning models were developed and validated to identify lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC) using clinical factors, laboratory metrics, and 2-deoxy-2[¹⁸F]fluoro-D-glucose ([¹⁸F]F-FDG) positron emission tomography (PET)/computed tomography (CT) radiomic features.

Methods

One hundred and twenty non-small cell lung cancer (NSCLC) patients (62 LUAD and 58 LUSC) were analyzed retrospectively and randomized into a training group (n = 85) and validation group (n = 35). A total of 99 feature parameters—four clinical factors, four laboratory indicators, and 91 [¹⁸F]F-FDG PET/CT radiomic features—were used for data analysis and model construction. The Boruta algorithm was used to screen the features. The retained minimum optimal feature subset was input into ten machine learning to construct a classifier for distinguishing between LUAD and LUSC. Univariate and multivariate analyses were used to identify the independent risk factors of the NSCLC subtype and constructed the Clinical model. Finally, the area under the receiver operating characteristic curve (AUC) values, sensitivity, specificity, and accuracy (ACC) was used to validate the machine learning model with the best performance effect and Clinical model in the validation group, and the DeLong test was used to compare the model performance.

Results

Boruta algorithm selected the optimal subset consisting of 13 features, including two clinical features, two laboratory indicators, and nine PEF/CT radiomic features. The Random Forest (RF) model and Support Vector Machine (SVM) model in the training group showed the best performance. Gender (P=0.018) and smoking status (P=0.011) construct the Clinical model. In the validation group, the SVM model (AUC: 0.876, ACC: 0.800) and RF model (AUC: 0.863, ACC: 0.800) performed well, while Clinical model (AUC:0.712, ACC: 0.686) performed moderately. There was no significant difference between the RF and Clinical models, but the SVM model was significantly better than the Clinical model. 

Conclusions

The proposed SVM and RF models successfully identified LUAD and LUSC. The results indicate that the proposed model is an accurate and noninvasive predictive tool that can assist clinical decision-making, especially for patients who cannot have biopsies or where a biopsy fails.

Combined model of radiomics, clinical, and imaging features for differentiating focal pneumonia-like lung cancer from pulmonary inflammatory lesions: an exploratory study

Background

Only few studies have focused on differentiating focal pneumonia-like lung cancer (F-PLC) from focal pulmonary inflammatory lesion (F-PIL). This exploratory study aimed to evaluate the clinical value of a combined model incorporating computed tomography (CT)-based radiomics signatures, clinical factors, and CT morphological features for distinguishing F-PLC and F-PIL.

Methods

In total, 396 patients pathologically diagnosed with F-PLC and F-PIL from two medical institutions between January 2015 and May 2021 were retrospectively analyzed. Patients from center 1 were included in the training (n = 242) and internal validation (n = 104) cohorts. Moreover, patients from center 2 were classified under the external validation cohort (n = 50). The clinical and CT morphological characteristics of both groups were compared first. And then, a clinical model incorporating clinical and CT morphological features, a radiomics model reflecting the radiomics signature of lung lesions, and a combined model were developed and validated, respectively.

Results

Age, gender, smoking history, respiratory symptoms, air bronchogram, necrosis, and pleural attachment differed significantly between the F-PLC and F-PIL groups (all P < 0.05). For the clinical model, age, necrosis, and pleural attachment were the most effective factors to differentiate F-PIL from F-PLC, with the area under the curves (AUCs) of 0.838, 0.819, and 0.717 in the training and internal and external validation cohorts, respectively. For the radiomics model, five radiomics features were found to be significantly related to the identification of F-PLC and F-PIL (all P < 0.001), with the AUCs of 0.804, 0.877, and 0.734 in the training and internal and external validation cohorts, respectively. For the combined model, five radiomics features, age, necrosis, and pleural attachment were independent predictors for distinguishing between F-PLC and F-PIL, with the AUCs of 0.915, 0.899, and 0.805 in the training and internal and external validation cohorts, respectively. The combined model exhibited a better performance than had the clinical and radiomics models.

Conclusions

The combined model, which incorporates CT-based radiomics signatures, clinical factors, and CT morphological characteristics, is effective in differentiating F-PLC from F-PIL.

Development of a PET/CT molecular radiomics-clinical model to predict thoracic lymph node metastasis of invasive lung adenocarcinoma ≤ 3 cm in diameter

Background

To investigate the value of ¹⁸F-FDG PET/CT molecular radiomics combined with a clinical model in predicting thoracic lymph node metastasis (LNM) in invasive lung adenocarcinoma (≤ 3 cm).

Methods

A total of 528 lung adenocarcinoma patients were enrolled in this retrospective study. Five models were developed for the prediction of thoracic LNM, including PET radiomics, CT radiomics, PET/CT radiomics, clinical and integrated PET/CT radiomics-clinical models. Ten PET/CT radiomics features and two clinical characteristics were selected for the construction of the integrated PET/CT radiomics-clinical model. The predictive performance of all models was examined by receiver operating characteristic (ROC) curve analysis, and clinical utility was validated by nomogram analysis and decision curve analysis (DCA).

Results

According to ROC curve analysis, the integrated PET/CT molecular radiomics-clinical model outperformed the clinical model and the three other radiomics models, and the area under the curve (AUC) values of the integrated model were 0.95 (95% CI: 0.93–0.97) in the training group and 0.94 (95% CI: 0.89–0.97) in the test group. The nomogram analysis and DCA confirmed the clinical application value of this integrated model in predicting thoracic LNM.

Conclusions

The integrated PET/CT molecular radiomics-clinical model proposed in this study can ensure a higher level of accuracy in predicting the thoracic LNM of clinical invasive lung adenocarcinoma (≤ 3 cm) compared with the radiomics model or clinical model alone.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13550-022-00895-x.

Evaluation and optimization of radiomics features stability to respiratory motion in 18 F-FDG 3D PET imaging

PURPOSE

To evaluate the impact of respiratory motion on radiomics features in ¹⁸ F-fluoro-2-deoxy-D-glucose three dimensional positron emission tomography (¹⁸ F-FDG 3D PET) imaging and optimize feature stability by combining preprocessing configurations and aggregation strategies.

METHODS

An in-house developed respiratory motion phantom was imaged in 3D PET scanner under nine respiratory patterns including one reference pattern. In total, 487 radiomics features were extracted for each respiratory pattern. Feature stability to respiratory motion was first evaluated by metrics of coefficient of variation (COV) and relative difference (RD) in a fixed preprocessing configuration. Further, one-way ANOVA and trend analysis were performed to evaluate the impact of preprocessing configuration (voxel size, discretization scheme) and aggregation strategy on feature stability. Finally, an optimization framework was proposed by selected feature-specific configurations with minimum COV value, and the diagnostic performance was validated in stable versus unstable features and fixed versus optimal features by a set of 46 patients with lung disease.

RESULTS

PET radiomics features were sensitive to respiratory motion, only 79/487 (16%) features were identified to be very stable in the fixed configuration. Preprocessing configuration and aggregation strategy had an impact on feature stability. For different voxel size, bin number, bin size and aggregation strategy, 188/487 (39%), 70/487 (15%), 148/487 (30%), and 38/95 (29%) features appeared significant changes in feature stability. The optimized configuration had the potential to improve feature stability compared to fixed configuration, with the COV decreased from 22% ±24% to 16% ±20%. Regarding the diagnostic performance, the stable and optimal configuration features outperformed the unstable and fixed configuration features, respectively (AUC 0.88, 0.87 vs. 0.83, 0.85).

CONCLUSIONS

Respiratory motion shows considerable impact on feature stability in 3D PET imaging, while optimizing preprocessing configuration may improve feature stability and diagnostic performance.