Identification of epidermal growth factor receptor mutations in pulmonary adenocarcinoma using dual-energy spectral computed tomography

From Images to Genes: Radiogenomics Based on Artificial Intelligence to Achieve Non-Invasive Precision Medicine in Cancer Patients

With the increasing demand for precision medicine in cancer patients, radiogenomics emerges as a promising frontier. Radiogenomics is originally defined as a methodology for associating gene expression information from high-throughput technologies with imaging phenotypes. However, with advancements in medical imaging, high-throughput omics technologies, and artificial intelligence, both the concept and application of radiogenomics have significantly broadened. In this review, the history of radiogenomics is enumerated, related omics technologies, the five basic workflows and their applications across tumors, the role of AI in radiogenomics, the opportunities and challenges from tumor heterogeneity, and the applications of radiogenomics in tumor immune microenvironment. The application of radiogenomics in positron emission tomography and the role of radiogenomics in multi-omics studies is also discussed. Finally, the challenges faced by clinical transformation, along with future trends in this field is discussed.

Differentiation of pathological subtypes and Ki-67 and TTF-1 expression by dual-energy CT (DECT) volumetric quantitative analysis in non-small cell lung cancer

BACKGROUND

To explore the value of dual-energy computed tomography (DECT) in differentiating pathological subtypes and the expression of immunohistochemical markers Ki-67 and thyroid transcription factor 1 (TTF-1) in patients with non-small cell lung cancer (NSCLC).

METHODS

Between July 2022 and May 2024, patients suspected of lung cancer who underwent two-phase contrast-enhanced DECT were prospectively recruited. Whole-tumor volumetric and conventional spectral analysis were utilized to measure DECT parameters in the arterial and venous phase. The DECT parameters model, clinical-CT radiological features model, and combined prediction model were developed to discriminate pathological subtypes and predict Ki-67 or TTF-1 expression. Multivariate logistic regression analysis was used to identify independent predictors. The diagnostic efficacy was assessed by the area under the receiver operating characteristic curve (AUC) and compared using DeLong's test.

RESULTS

This study included 119 patients (92 males and 27 females; mean age, 63.0 ± 9.4 years) who was diagnosed with NSCLC. When applying the DECT parameters model to differentiate between adenocarcinoma and squamous cell carcinoma, ROC curve analysis indicated superior diagnostic performance for conventional spectral analysis over volumetric spectral analysis (AUC, 0.801 vs. 0.709). Volumetric spectral analysis exhibited higher diagnostic efficacy in predicting immunohistochemical markers compared to conventional spectral analysis (both P < 0.05). For Ki-67 and TTF-1 expression, the combined prediction model demonstrated optimal diagnostic performance with AUC of 0.943 and 0.967, respectively.

CONCLUSIONS

The combined predictive model based on volumetric quantitative analysis in DECT offers valuable information to discriminate immunohistochemical expression status, facilitating clinical decision-making for patients with NSCLC.

Assessment of solitary pulmonary nodules using dual-layer spectral detector computed tomography

We aim to quantitatively investigate the difference between benign and malignant solid pulmonary nodules that appeared on dual-energy spectral computed tomography, and assess the diagnostic accuracy of several parameters derived from computed tomography in differentiating malignant from benign pulmonary nodules. Between September 2021 and December 2022, spectral images of 71 patients (male:female = 44:27, mean age = 71.0 years) confirmed by pathology were retrospectively analyzed in the venous phase. Patients were classified into the malignant group and the benign group. The iodine concentration values of the nodules, normalized iodine concentration of the nodules to the neighboring vessels, virtual monochromatic images of 40 and 80 keV, and slope of the spectral curve were calculated and compared between the benign and malignant groups. Receiver operating characteristic curves and the area under the curve were performed to assess the diagnostic performance of the above parameters. Both virtual monochromatic images and iodine concentration maps prove to be highly useful in differentiating benign and malignant pulmonary nodules. The malignant pulmonary nodules have higher iodine density and slope of the spectral curve than the benign lesions. The combined model of iodine density and curve slope with an optimal cutoff of 0.39 (area under the curve = 0.82) yielded a sensitivity of 95% and a specificity of 63%. Contrast-enhanced dual-energy spectral computed tomography allows promising capability of distinguishing malignant from benign lesions, potential for avoiding unnecessary invasive procedure or surgery.

Differentiation of osteoblastic metastases and bone islands on dual-energy computed tomography in patients with untreated lung cancer

OBJECTIVE

To evaluate the diagnostic efficacy of quantitative dual-energy computed tomography (CT) parameters for distinguishing osteoblastic metastases (OBMs) from bone islands (BIs) in untreated lung cancer.

MATERIAL AND METHODS

Dual-energy CT images of 24 patients with OBMs and 56 patients with BIs obtained between January 2019 and December 2021 were retrospectively analyzed. The CT70keV value, calcium(water) density [Dcalcium(water)], and water(calcium) density [Dwater(calcium)] were analyzed. Diagnostic performance was assessed by measuring the area under the curve (AUC), and specificity, sensitivity, and accuracy were determined.

RESULTS

A total of 70 OBMs and 67 BIs were included. The AUC values of CT70keV, Dcalcium(water), and Dwater(calcium) showed no significant differences (0.950 vs. 0.947 vs. 0.929, respectively; P > 0.05). The optimal CT70keV cutoff value was 885.1 HU, with specificity, sensitivity, and accuracy of 81.4 %, 92.5 %, and 86.9 %, respectively. When using Dcalcium(water) < 254.9 mg/cm3 and Dwater(calcium) < 1250.6 mg/cm3, respectively, 119 of 137 lesions showed consistent diagnostic results (true or false). Sub-analysis of these 119 lesions showed specificity of 92.1 %, which was higher than that of CT70keV (P = 0.021). The AUC, sensitivity, and accuracy were 0.974, 92.9 %, and 92.4 %, respectively, which were not significantly different from those of CT70keV (P = 0.230, 0.906, and 0.220, respectively). Among the 18 lesions showing inconsistent diagnoses, Dcalcium(water) diagnosed 11 lesions correctly, and Dwater(calcium) diagnosed the remaining seven lesions correctly.

CONCLUSION

The combination of Dcalcium(water) and Dwater(calcium) demonstrated a promising role in the differentiation of OBMs from BIs in lung cancer patients.

Diagnosis of newly developed multiple myeloma without bone disease detectable on conventional computed tomography (CT) scan by using dual-energy CT

Objective

To evaluate the diagnostic utility of fat (hydroxyapatite) density [DFat (HAP)] on dual-energy computed tomography (DECT) for identifying clinical diagnosed multiple myeloma without bone disease (MNBD) that is not visible on conventional CT scans.

Material and Methods

In this age-gender-examination sites matched case control prospective study, Chest and/or abdominal images on Revolution CT of MNBDs and control subjects were consecutive enrolled in a 1:2 ratio from October 2022 to November 2023. Multiple myeloma was clinical diagnosed according to criteria of the International Myeloma Working Group. Regions of interest (ROIs) were drawn separately for all thoracolumbar vertebrae in the scanning range by two radiologists. Additionally, a radiologist specializing in musculoskeletal imaging supervised the process. DFat (HAP) was extracted from each ROI. The spine was divided into upper thoracic (UPT), middle and lower thoracic (MLT), thoracolumbar (TL), and middle and lower lumbar (MLL) vertebrae. The area under the receiver operating characteristic curve (AUC) was calculated to evaluate the diagnostic performance of DFat (HAP) in diagnosing multiple myeloma, and the sensitivity, specificity, and accuracy under the optimal cut-off were determined by Youden index (sensitivity + specificity -1).

Results

A total of 32 and MNBD patients and 64 control patients were included. The total number of ROIs outlined included MNBD group (n = 493) and control group (n = 986). For all vertebrae, DFat(HAP) got average performance in the diagnosis of MNBD (AUC = 0.733, p < 0.001) with a cut-off value of 958 (mg/cm3); the sensitivity, specificity, and accuracy were 58.8 %, 77.8 %, and 71.7 %, respectively. Regarding segment analysis, the diagnostic performance was good for all (AUC, 0.803-0.837; p < 0.001) but the UPT segment (AUC = 0.692, p = 0.002). The optimal diagnostic cut-off values for the MLT, TL, and MLL vertebrae were 955 mg/cm3, 947 mg/cm3, and 947 mg/cm3, respectively; the sensitivity, specificity, and accuracy were 80.0 %-87.5 %, 71.9 %-82.6 %, and 77.1 %-81.6 %, respectively.

Conclusion

DECT was effective for detecting MNBD, and better diagnostic results can be obtained by grouping different spine segments.

Three-dimensional iodine mapping quantified by dual-energy CT for predicting programmed death-ligand 1 expression in invasive pulmonary adenocarcinoma

We examined the association between texture features using three-dimensional (3D) io-dine density histogram on delayed phase of dual-energy CT (DECT) and expression of programmed death-ligand 1 (PD-L1) using immunostaining methods in non-small cell lung cancer. Consecutive 37 patients were scanned by DECT. Unenhanced and enhanced (3 min delay) images were obtained. 3D texture analysis was performed for each nodule to obtain 7 features (max, min, median, mean, standard deviation, skewness, and kurtosis) from iodine density mapping and extracellular volume (ECV). A pathologist evaluated a tumor proportion score (TPS, %) using PD-L1 immunostaining: PD-L1 high (TPS ≥ 50%) and low or negative expression (TPS < 50%). Associations between PD-L1 expression and each 8 parameter were evaluated using logistic regression analysis. The multivariate logistic regression analysis revealed that skewness and ECV were independent indicators associated with high PD-L1 expression (skewness: odds ratio [OR]  7.1 [95% CI 1.1, 45.6], p = 0.039; ECV: OR 6.6 [95% CI 1.1, 38.4], p = 0.037). In the receiver-operating characteristic analysis, the area under the curve of the combination of skewness and ECV was 0.83 (95% CI 0.67, 0.93) with sensitivity of 64% and specificity of 96%. Skewness from 3D iodine density histogram and ECV on dual energy CT were significant factors for predicting PD-L1 expression.

Does dual-layer spectral detector CT provide added value in predicting spread through air spaces in lung adenocarcinoma? A preliminary study

OBJECTIVES

To examine the predictive value of dual-layer spectral detector CT (DLCT) for spread through air spaces (STAS) in clinical lung adenocarcinoma.

METHODS

A total of 225 lung adenocarcinoma cases were retrospectively reviewed for demographic, clinical, pathological, traditional CT, and spectral parameters. Multivariable logistic regression analysis was carried out based on three logistic models, including a model using traditional CT features (traditional model), a model using spectral parameters (spectral model), and an integrated model combining traditional CT and spectral parameters (integrated model). Receiver operating characteristic (ROC) curve analysis and decision curve analysis (DCA) were performed to assess these models.

RESULTS

Univariable analysis showed significant differences between the STAS and non-STAS groups in traditional CT features, including nodule density (p < 0.001), pleural indentation types (p = 0.006), air-bronchogram sign (p = 0.031), the presence of spiculation (p < 0.001), long-axis diameter of the entire nodule (LD) (p < 0.001), and consolidation/tumor ratio (CTR) (p < 0.001). Multivariable analysis revealed that LD > 20 mm (odds ratio [OR] = 2.271, p = 0.025) and CTR (OR = 24.208, p < 0.001) were independent predictors in the traditional model, while electronic density (ED) in the venous phase was an independent predictor in the spectral (OR = 1.062, p < 0.001) and integrated (OR = 1.055, p < 0.001) models. The area under the curve (AUC) for the integrated model (0.84) was the highest (spectral model, 0.83; traditional model, 0.80), and the difference between the integrated and traditional models was statistically significant (p = 0.015). DCA showed that the integrated model had superior clinical value versus the traditional model.

CONCLUSIONS

DLCT has added value for STAS prediction in lung adenocarcinoma.

CLINICAL RELEVANCE STATEMENT

Spectral CT has added value for spread through air spaces prediction in lung adenocarcinoma so may impact treatment planning in the future.

KEY POINTS

• Electronic density may be a potential spectral index for predicting spread through air spaces in lung adenocarcinoma. • A combination of spectral and traditional CT features enhances the performance of traditional CT for predicting spread through air spaces.

Predicting epidermal growth factor receptor mutations in non-small cell lung cancer through dual-layer spectral CT: a prospective study

OBJECTIVE

To determine whether quantitative parameters of detector-derived dual-layer spectral computed tomography (DLCT) can reliably identify epidermal growth factor receptor (EGFR) mutation status in patients with non-small cell lung cancer (NSCLC).

METHODS

Patients with NSCLC who underwent arterial phase (AP) and venous phase (VP) DLCT between December 2021 and November 2022 were subdivided into the mutated and wild-type EGFR groups following EGFR mutation testing. Their baseline clinical data, conventional CT images, and spectral images were obtained. Iodine concentration (IC), iodine no water (INW), effective atomic number (Zeff), virtual monoenergetic images, the slope of the spectral attenuation curve (λHU), enhancement degree (ED), arterial enhancement fraction (AEF), and normalized AEF (NAEF) were measured for each lesion.

RESULTS

Ninety-two patients (median age, 61 years, interquartile range [51, 67]; 33 men) were evaluated. The univariate analysis indicated that IC, normalized IC (NIC), INW and ED for the AP and VP, as well as Zeff and λHU for the VP were significantly associated with EGFR mutation status (all p < 0.05). INW(VP) showed the best diagnostic performance (AUC, 0.892 [95% confidence interval {CI}: 0.823, 0.960]). However, neither AEF (p = 0.156) nor NAEF (p = 0.567) showed significant differences between the two groups. The multivariate analysis showed that INW(AP) and NIC(VP) were significant predictors of EGFR mutation status, with the latter showing better performance (p = 0.029; AUC, 0.897 [95% CI: 0.816, 0.951] vs. 0.774 [95% CI: 0.675, 0.855]).

CONCLUSION

Quantitative parameters of DLCT can help predict EGFR mutation status in patients with NSCLC.

CRITICAL RELEVANCE STATEMENT

Quantitative parameters of DLCT, especially NIC(VP), can help predict EGFR mutation status in patients with NSCLC, facilitating appropriate and individualized treatment for them.

KEY POINTS

Determining EGFR mutation status in patients with NSCLC before starting therapy is essential. Quantitative parameters of DLCT can predict EGFR mutation status in NSCLC patients. NIC in venous phase is an important parameter to guide individualized treatment selection for NSCLC patients.

Efficient 18F-fluorodeoxyglucose positron emission tomography/computed tomography-based machine learning model for predicting epidermal growth factor receptor mutations in non-small cell lung cancer

BACKGROUND

Beyond the human eye's limitations, radiomics provides more information that can be used for diagnosis. We develop a personalized and efficient model based on 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography/computed tomography (PET/CT) to predict epidermal growth factor receptor (EGFR) mutations to help identify which non-small cell cancer (NSCLC) patients are candidates for EGFR-tyrosine kinase inhibitors (TKIs) therapy.

METHODS

We retrospectively included 100 patients with NSCLC and randomized them according to 70 patients in the training group and 30 patients in the validation group. The least absolute shrinkage and selection operator logistic regression (LLR) algorithm and support vector machine (SVM) classifier were used to build the models and predict whether EGFR is mutated or not. The predictive efficacy of the LLR algorithm-based model and the SVM classifier-based model was evaluated by plotting the receiver operating characteristic (ROC) curves and calculating the area under the curve (AUC).

RESULTS

The AUC, sensitivity and specificity of our radiomics model by LLR algorithm were 0.792, 0.967, and 0.600 for the training group and 0.643, 1.00, and 0.378 for the validation group, respectively, in predicting EGFR mutations. The AUC was 0.838 for the training group and 0.696 for the validation group after combining radiomics features with clinical features. The prediction results based on the SVM classifier showed that the validation group had the best performance when based on radial kernel function with AUC, sensitivity, and specificity of 0.741, 0.667, and 0.825, respectively.

CONCLUSIONS

Radiomics models based on 18F-FDG PET/CT modeled with different machine learning algorithms can improve the predictive efficacy of the models. Models that combine clinical features are more clinically valuable.

Dual-energy CT-based radiomics in predicting EGFR mutation status non-invasively in lung adenocarcinoma

Objectives

Patients with epidermal growth factor receptor (EGFR) mutations in lung adenocarcinoma (LUAD) can benefit from individualized targeted therapy. This study aims to develop, compare, analyse prediction models based on dual-energy spectral computed tomography (DESCT) and CT-based radiomic features to non-invasively predict EGFR mutation status in LUAD.

Materials and methods

Patients with LUAD (n = 175), including 111 patients with and 64 patients without EGFR mutations, were enrolled in the current study. All patients were randomly divided into a training dataset (122 cases) and validation dataset (53 cases) at a ratio of 7:3. After extracting CT-based radiomic, DESCT and clinical features, we built seven prediction models and a nomogram of the best prediction. Receiver operating characteristic (ROC) curves and the mean area under the curve (AUC) values were used for comparisons among the models to obtain the best prediction model for predicting EGFR mutations.

Results

The best distinguishing ability is the combined model incorporating radiomic, DESCT and clinical features for predicting the EGFR mutation status with an AUC of 0.86 (95 % CI: 0.79-0.92) in the training group and an AUC value of 0.83 (95 % CI: 0.73, 0.96) in the validation group.

Conclusions

Our study provides a predictive nomogram non-invasively with a combination of CT-based radiomic, DESCT and clinical features, which can provide image-based biological information for targeted therapy of LUAD with EGFR mutations.

Dual-energy CT-based radiomics for predicting invasiveness of lung adenocarcinoma appearing as ground-glass nodules

Objectives

To explore the value of radiomics based on Dual-energy CT (DECT) for discriminating preinvasive or MIA from IA appearing as GGNs before surgery.

Methods

The retrospective study included 92 patients with lung adenocarcinoma comprising 30 IA and 62 preinvasive-MIA, which were further divided into a training (n=64) and a test set (n=28). Clinical and radiographic features along with quantitative parameters were recorded. Radiomics features were derived from virtual monoenergetic images (VMI), including 50kev and 150kev images. Intraclass correlation coefficients (ICCs), Pearson's correlation analysis and least absolute shrinkage and selection operator (LASSO) penalized logistic regression were conducted to eliminate unstable and redundant features. The performance of the models was evaluated by area under the curve (AUC) and the clinical utility was assessed using decision curve analysis (DCA).

Results

The DECT-based radiomics model performed well with an AUC of 0.957 and 0.865 in the training and test set. The clinical-DECT model, comprising sex, age, tumor size, density, smoking, alcohol, effective atomic number, and normalized iodine concentration, had an AUC of 0.929 in the training and 0.719 in the test set. In addition, the radiomics model revealed a higher AUC value and a greater net benefit to patients than the clinical-DECT model.

Conclusion

DECT-based radiomics features were valuable in predicting the invasiveness of GGNs, yielding a better predictive performance than the clinical-DECT model.

Diagnosis of benign and malignant thyroid nodules by a dual-layer spectral detector CT-based nomogram

Introduction

Preoperative diagnosis of benign and malignant thyroid nodules is crucial for appropriate clinical treatment and individual patient management. In this study, a double-layer spectral detector computed tomography (DLCT)-based nomogram for the preoperative classification of benign and malignant thyroid nodules was developed and tested.

Methods

A total of 405 patients with pathological findings of thyroid nodules who underwent DLCT preoperatively were retrospectively recruited. They were randomized into a training cohort (n=283) and a test cohort (n=122). Information on clinical features, qualitative imaging features and quantitative DLCT parameters was collected. Univariate and multifactorial logistic regression analyses were used to screen independent predictors of benign and malignant nodules. A nomogram model based on the independent predictors was developed to make individualized predictions of benign and malignant thyroid nodules. Model performance was evaluated by calculating the area under the receiver operating characteristic curve (AUC), calibration curve and decision curve analysis(DCA).

Results

Standardized iodine concentration in the arterial phase, the slope of the spectral hounsfield unit(HU) curves in the arterial phase, and cystic degeneration were identified as independent predictors of benign and malignant thyroid nodules. After combining these three metrics, the proposed nomogram was diagnostically effective, with AUC values of 0.880 for the training cohort and 0.884 for the test cohort. The nomogram showed a better fit (all p &gt; 0.05 by Hosmer−Lemeshow test) and provided a greater net benefit than the simple standard strategy within a large range of threshold probabilities in both cohorts.

Discussion

The DLCT-based nomogram has great potential for the preoperative prediction of benign and malignant thyroid nodules. This nomogram can be used as a simple, noninvasive, and effective tool for the individualized risk assessment of benign and malignant thyroid nodules, helping clinicians make appropriate treatment decisions.

Prediction of EGFR Mutation Status in Non–Small Cell Lung Cancer Based on Ensemble Learning

Objectives: We aimed to identify whether ensemble learning can improve the performance of the epidermal growth factor receptor (EGFR) mutation status predicting model.

Methods: We retrospectively collected 168 patients with non–small cell lung cancer (NSCLC), who underwent both computed tomography (CT) examination and EGFR test. Using the radiomics features extracted from the CT images, an ensemble model was established with four individual classifiers: logistic regression (LR), support vector machine (SVM), random forest (RF), and extreme gradient boosting (XGBoost). The synthetic minority oversampling technique (SMOTE) was also used to decrease the influence of data imbalance. The performances of the predicting model were evaluated using the area under the curve (AUC).

Results: Based on the 26 radiomics features after feature selection, the SVM performed best (AUCs of 0.8634 and 0.7885 on the training and test sets, respectively) among four individual classifiers. The ensemble model of RF, XGBoost, and LR achieved the best performance (AUCs of 0.8465 and 0.8654 on the training and test sets, respectively).

Conclusion: Ensemble learning can improve the model performance in predicting the EGFR mutation status of patients with NSCLC, showing potential value in clinical practice.

Differentiation Between Solitary Pulmonary Inflammatory Lesions and Solitary Cancer Using Gemstone Spectral Imaging

BACKGROUND

The distinction between solitary inflammatory lesion and solitary lung cancer remains a challenge because of their considerable overlapping computed tomography (CT) imaging features.

PURPOSE

This study aimed to verify whether spectral CT parameters can differentiate solitary lung cancer from solitary inflammatory lesions and to find their correlations with lesion size.

METHODS

A total of 78 patients with solitary lung lesions were included in our study. All of them underwent enhanced CT scans with Gemstone Spectral Imaging (GSI) mode, which was one of the dual-energy imaging technologies. According to maximum diameter (Dmax) of the lesion, regions of interest were collected and divided into inflammatory (group I: <3 cm [IA], n = 17; ≥3 cm [IB], n = 14) and cancer groups (group II: <3 cm [IIA], n = 20; ≥3 cm [IIB], n = 27). Computed tomography values (HU40keV, HU70keV), effective atomic number (Zeff), iodine concentration (IC), normalized IC (NIC), and spectral curve slopes (λ30, λ40) of each region of interest were calculated. The NIC was defined as the IC ratio of the lesion to the descending aorta. Mann-Whitney U test was used for intergroup (I vs II, IA vs IIA, IB vs IIB) and intragroup (IA vs IB, IIA vs IIB) comparisons, and receiver operating characteristic curve analysis was performed. Correlation analysis was applied to find the relationship between Dmax and GSI parameters.

RESULTS

No significant correlation was found between GSI parameters and Dmax in the inflammatory group, whereas inverse correlations were found in the cancer group. Gemstone spectral imaging parameters (except HU70keV) of group IIA were significantly higher than those of group IIB. There were significant differences in HU40keV, IC, NIC, λ30, and λ40 between groups IB and IIB under both arterial and venous phase (P values < 0.05), whereas the area under the curve for λ30 under venous phase was largest, and sensitivity and specificity were 96.32% and 85.71%, respectively. However, only HU40keV and HU70keV values under the arterial phase of IIA were significantly higher than those of IA.

CONCLUSIONS

Quantitative parameters of GSI demonstrated an inverse correlation with the lesion size of solitary lung cancer, and GSI parameters can be new ways to differentiate solitary lung cancer from solitary inflammatory lesions.

Comparison of Radiation Dose and Image Quality Between Split-Filter Twin Beam Dual-Energy Images and Single-Energy Images in Single-Source Contrast-Enhanced Chest Computed Tomography

OBJECTIVE

To compare image quality and radiation dose of split-filter TwinBeam dual-energy (SF-TBDE) with those of single-energy images (SECT) in the contrast-enhanced chest computed tomography (CT).

METHODS

Two hundred patients who underwent SF-TBDE (n = 100) and SECT (n = 100) contrast-enhanced chest scanning were retrospectively analyzed. The contrast-to-noise ratio (CNR) and figure of merit (FOM)-CNR of 5 structures (lung, aorta, pulmonary artery, thyroid, and erector spinae) were calculated and subjectively evaluated by 2 independent radiologists. Radiation dose was compared using volume CT dose index and size-specific dose estimate.

RESULTS

The CNR and FOM-CNR of lung and erector spinae in SF-TBDE were higher than those of SECT (P < 0.001). The differences in the subjective image quality between the 2 groups were not significant (P = 0.244). Volume CT dose index and size-specific dose estimate of SF-TBDE were lower than those of SECT (6.60 ± 1.56 vs 7.81 ± 3.02 mGy, P = 0.001; 9.25 ± 1.60 vs. 10.55 ± 3.54; P = 0.001).

CONCLUSIONS

The SF-TBDE CT can provide similar image quality at a lower radiation dose compared with SECT.

Quantitative dual-energy CT for evaluating hepatocellular carcinoma after transarterial chemoembolization

We aimed to investigate the role of the quantitative parameters of dual-energy computed tomography (DECT) in evaluating patients with hepatocellular carcinoma (HCC) treated by transarterial chemoembolization (TACE). We retrospectively identified 80 HCC patients (mean age, 56 years; 61 men) treated by TACE who received contrast-enhanced DECT and were retreated by TACE within 7 days between November 2018 and December 2019. Taking digital subtraction angiography (DSA) and CT images as reference standard, two readers measured and calculated the values of normalized iodine concentration at arterial phase (NICAP), normalized iodine concentration at portal venous phase (NICPP), iodine concentration difference (ICD), arterial iodine fraction (AIF) and slope of the spectral Hounsfield unit curve (λHu) by placing matched regions of interests (ROIs) within the tumor active area (TAA), adjacent normal hepatic parenchyma (ANHP) and tumor necrotic area (TNA). Differences between the parameters were analyzed by the Kruskal-Wallis H test. Receiver operating characteristic analysis of the parameters performance in differentiating the three tissues types was performed. AIF exhibited a good performance in distinguishing TAA (0.93 ± 0.31) and ANHP (0.18 ± 0.14), the areas under the receiver operating characteristic curve (AUC) was 0.989, while the λHu exhibited an excellent performance in distinguishing TAA (3.32 ± 1.24) and TNA (0.29 ± 0.27), with an AUC of 1.000. In conclusion, quantitative DECT can be effectively used to evaluate the tumor viability in HCC patients treated by TACE.

Differentiating malignant and benign necrotic lung lesions using kVp-switching dual-energy spectral computed tomography

BACKGROUND

Necrotic pulmonary lesions manifest as relatively low-density internally on contrast-enhanced computed tomography (CT). However, using CT to differentiate malignant and benign necrotic pulmonary lesions is challenging, as these lesions have similar peripheral enhancement. With the introduction of dual-energy spectral CT (DESCT), more quantitative parameters can be obtained and the ability to differentiate material compositions has been highly promoted. This study investigated the use of kVp-switching DESCT in differentiating malignant from benign necrotic lung lesions.

METHODS

From October 2016 to February 2019, 40 patients with necrotic lung cancer (NLC) and 31 with necrotic pulmonary mass-like inflammatory lesion (NPMIL) were enrolled and underwent DESCT. The clinical characteristics of patients, CT morphological features, and DESCT quantitative parameters of lesions were compared between the two groups. Binary logistic regression analysis was performed to identify the independent prognostic factors differentiating NPMIL from NLC. Receiver operating characteristic (ROC) curves were used to assess the diagnostic performance of single-parameter and multiparametric analyses.

RESULTS

Significant differences in age, C-reactive protein concentration, the slope of the spectral curve from 40 to 65 keV (K_40-65 keV) of necrosis in non-contrast-enhanced scanning (NCS), arterial phase (AP) and venous phase (VP), effective atomic number of necrosis in NCS, and iodine concentration (IC) of the solid component in VP were observed between groups (all p < 0.05). The aforementioned parameters had area under the ROC curve (AUC) values of 0.747, 0.691, 0.841, 0.641, 0.660, 0.828, and 0.754, respectively, for distinguishing between NLC and NPMIL. Multiparametric analysis showed that age, K_40-65 keV of necrosis in NCS, and IC of the solid component in VP were the most effective factors for differentiating NLC from NPMIL, with an AUC of 0.966 and percentage of correct class of 88.7%.

CONCLUSIONS

DESCT can differentiate malignant from benign necrotic lung lesions with a relatively high accuracy.

Dual-Energy CT-Based Nomogram for Decoding HER2 Status in Patients With Gastric Cancer

OBJECTIVE. The purpose of this study was to develop and evaluate a dual-energy CT (DECT)-based nomogram for noninvasive identification of the status of human epidermal growth factor receptor 2 (HER2; also known as ERBB2) expression in gastric cancer (GC). MATERIALS AND METHODS. A total of 206 patients with histologically proven GC who underwent pretreatment DECT were retrospectively recruited and randomly allocated to a training cohort (n = 144) or a test cohort (n = 62). Information on clinical characteristics, qualitative imaging features, and quantitative DECT parameters was collected. Univariate analysis and multivariate logistic regression were implemented to screen independent predictors of HER2 status. An individualized nomogram was built, and its discrimination, calibration, and clinical usefulness were assessed. RESULTS. Tumor location, the iodine concentration of the tumor in the venous phase, and the normalized iodine concentration of the tumor in the venous phase were significant factors predictive of HER2 status (all p < .05). After these three indicators were integrated, the proposed nomogram showed a favorable diagnostic performance, with AUCs of 0.807 (95% CI, 0.718-0.897) in the training cohort and 0.815 (95% CI, 0.661-0.968) in the test cohort. The nomogram showed a preferable fitting (all p > .05 by the Hosmer-Lemeshow test) and would offer more net benefits than simple default strategies within a wide range of threshold probabilities in both cohorts. CONCLUSION. The DECT-based nomogram has great application potential in terms of detecting HER2 status in GC, and can serve as a novel substitute for invasive testing.

Development of a dual-energy spectral computed tomography-based nomogram for the preoperative discrimination of histological grade in colorectal adenocarcinoma patients

Background

The usefulness of a dual-energy spectral computed tomography (DESCT)-based nomogram in discriminating between histological grades of colorectal adenocarcinoma (CRAC) is unclear. This study aimed to develop such a nomogram and assess its ability to preoperatively discriminate between histological grades in CRAC patients.

Methods

Primary tumors monochromatic CT value, iodine concentration (IC) value, and effective atomic number (Eff-Z) in the arterial (AP) and venous phases (VP) were retrospectively compared between patients with high-grade (n=65) and low-grade (n=108) CRAC who underwent preoperative abdominal DESCT. Univariate analysis was used to compare the DESCT parameters and clinical factors between these two patient groups. Statistically significant features in the univariate analysis were included in the multivariate logistic regression model to identify the indicators for building a nomogram that could discriminate between histological grades in CRAC patients. The clinical usefulness of the nomogram and its value for predicting overall survival were statistically evaluated.

Results

The logistic regression analysis showed that age, clinical T stage, clinical N stage, and IC values in AP and VP were significant independent predictors for high-grade CRAC. A quantitative nomogram developed based on these predictors showed excellent performance for discriminating between the histological grades, with an area under the curve (AUC) of 0.886 and excellent agreement in the calibration curve. The Kaplan-Meier curve for overall survival showed that our nomogram identified a significant difference between the high- and low-risk groups [hazard ratio (HR), 2.188; 95% CI, 1.072-4.465; P=0.027).

Conclusions

This study presents a nomogram that incorporates DESCT parameters and clinical factors and can potentially be used as a clinical tool for individual preoperative prediction of CRAC histological grade.

Imaging genomics for accurate diagnosis and treatment of tumors: A cutting edge overview

Imaging genomics refers to the establishment of the connection between invasive gene expression features and non-invasive imaging features. Tumor imaging genomics can not only understand the macroscopic phenotype of tumor, but also can deeply analyze the cellular and molecular characteristics of tumor tissue. In recent years, tumor imaging genomics has been a key in the field of medicine. The incidence of cancer in China has increased significantly, which is the main reason of disease death of urban residents. With the rapid development of imaging medicine, depending on imaging genomics, many experts have made remarkable achievements in tumor screening and diagnosis, prognosis evaluation, new treatment targets and understanding of tumor biological mechanism. This review analyzes the relationship between tumor radiology and gene expression, which provides a favorable direction for clinical staging, prognosis evaluation and accurate treatment of tumors.

Application of Dual-Energy Spectral Computed Tomography to Thoracic Oncology Imaging

Computed tomography (CT) is an important imaging modality in evaluating thoracic malignancies. The clinical utility of dual-energy spectral computed tomography (DESCT) has recently been realized. DESCT allows for virtual monoenergetic or monochromatic imaging, virtual non-contrast or unenhanced imaging, iodine concentration measurement, and effective atomic number (Z_eff map). The application of information gained using this technique in the field of thoracic oncology is important, and therefore many studies have been conducted to explore the use of DESCT in the evaluation and management of thoracic malignancies. Here we summarize and review recent DESCT studies on clinical applications related to thoracic oncology.

Epidermal growth factor receptor mutations in lung adenocarcinoma: associations between dual-energy spectral CT measurements and histologic results

OBJECTIVE

To analyze the relationship between dual-energy spectral CT and epidermal growth factor receptor (EGFR) mutation status in patients with lung adenocarcinoma.

METHODS

The quantitative parameters of spectral CT were analyzed in 208 patients with lung adenocarcinoma. The quantitative parameters including CT_40keV and CT_70keV values, effective atomic number (Zeff), iodine concentration (IC), water concentration (WC), and the slope of the spectral curve (λ _HU) were calculated. Statistical analysis was used to determine the clinical characteristics and quantitative parameters for the diagnosis of EGFR-mutation status. The ROC curves were used to calculate diagnostic efficiency.

RESULTS

Sex (p = 0.027) and smoking history (p = 0.019) differed significantly according to the EGFR-mutation status. Spectral CT quantitative parameters (CT_40keV and CT_70keV values, λ _HU, Zeff and IC) differed significantly between the EGFR mutant and the EGFR wild-type groups (p < 0.05) during the arterial phase (AP) and venous phase (VP). However, WC was not statistically different between the two groups (p > 0.05). ROC curve analysis revealed the combination of the significantly different quantitative parameters provided the best diagnostic performance for determining the EGFR-mutation status (AUC: 76.0%) in the AP, while the AUC during the VP was 75.6%.

CONCLUSION

The quantitative parameters of dual-energy spectral CT have potential value for identifying the EGFR-mutation status.

Development of a dual-energy spectral CT based nomogram for the preoperative discrimination of mutated and wild-type KRAS in patients with colorectal cancer

PURPOSE

To develop a dual-energy spectral CT (DESCT) nomogram for the preoperative identification of KRAS mutation in patients with colorectal cancer (CRC).

METHOD

One hundred and twenty-four patients who underwent energy spectrum CT pre-operatively were recruited and split into mutated KRAS group (n = 50) and wild-type KRAS group (n = 74). DESCT parameters, including monochromatic CT value, iodine concentration, water concentration, and effective atomic number were measured independently by two reviewers in the arterial, venous, and delayed phases. Normalized iodine concentration (NIC) and slope k of the spectral HU curve were calculated. Evaluate other imaging features such as ATL/LTL ratio, tumor gross pattern, pericolorectal fat invasion (PFI) was also performed by these reviewers. Independent predictors for KRAS mutation were screened out using logistic regression, and these predictors were presented as a nomogram. The receiver operating characteristic (ROC) curve, calibration curve and decision curve analysis (DCA) were used to evaluate the clinical usefulness of the nomogram.

RESULTS

The slope k in the arterial phase, effective atomic number in the arterial phase, NIC in the venous phase, ATL/LTL ratio and PFI were significant independent predictors for KRAS mutation. Based on these independent predictors, a quantitative nomogram was developed to predict individual KRAS mutation probability. The nomogram had excellent performance with an AUC of 0.848 and excellent calibration. DCA showed that our nomogram has outstanding clinical utility.

CONCLUSIONS

This study demonstrates that a DESCT based nomogram has potential value for individual preoperative identification of KRAS mutation in CRC patients.

Assessment of Solitary Pulmonary Nodules Based on Virtual Monochrome Images and Iodine-Dependent Images Using a Single-Source Dual-Energy CT with Fast kVp Switching

With lung cancer being the most common malignancy diagnosed worldwide, lung nodule assessment has proved to be one of big challenges of modern medicine. The aim of this study was to examine the usefulness of Dual Energy Computed Tomography (DECT) in solitary pulmonary nodule (SPN) assessment. Between January 2017 and June 2018; 65 patients (42 males and 23 females) underwent DECT scans in the late arterial phase (AP) and venous phase (VP). We concluded that imaging at an energy level of 65 keV was the most accurate in detecting malignancy in solitary pulmonary nodules (SPNs) measuring ≤30 mm in diameter on virtual monochromatic maps. Both virtual monochromatic images and iodine concentration maps prove to be highly useful in differentiating benign and malignant pulmonary nodules. As for iodine concentration maps, the analysis of venous phase images resulted in the highest clinical usefulness. To summarize, DECT may be a useful tool in the differentiation of benign and malignant SPNs. A single-phase DECT examination with scans acquired 90 s after contrast media injection is recommended.

Predicting EGFR mutation subtypes in lung adenocarcinoma using 18F-FDG PET/CT radiomic features

Background

Identification of epidermal growth factor receptor (EGFR) mutation types is crucial before tyrosine kinase inhibitors (TKIs) treatment. Radiomics is a new strategy to noninvasively predict the genetic status of cancer. In this study, we aimed to develop a predictive model based on ¹⁸F-fluorodeoxyglucose positron emission tomography-computed tomography (¹⁸F-FDG PET/CT) radiomic features to identify the specific EGFR mutation subtypes.

Methods

We retrospectively studied ¹⁸F-FDG PET/CT images of 148 patients with isolated lung lesions, which were scanned in two hospitals with different CT scan setting (slice thickness: 3 and 5 mm, respectively). The tumor regions were manually segmented on PET/CT images, and 1,570 radiomic features (1,470 from CT and 100 from PET) were extracted from the tumor regions. Seven hundred and ninety-four radiomic features insensitive to different CT settings were first selected using the Mann white U test, and collinear features were further removed from them by recursively calculating the variation inflation factor. Then, multiple supervised machine learning models were applied to identify prognostic radiomic features through: (I) a multi-variate random forest to select features of high importance in discriminating different EGFR mutation status; (II) a logistic regression model to select features of the highest predictive value of the EGFR subtypes. The EGFR mutation predicting model was constructed from prognostic radiomic features using the popular Xgboost machine-learning algorithm and validated using 3-fold cross-validation. The performance of predicting model was analyzed using the receiver operating characteristic curve (ROC) and measured with the area under the curve (AUC).

Results

Two sets of prognostic radiomic features were found for specific EGFR mutation subtypes: 5 radiomic features for EGFR exon 19 deletions, and 5 radiomic features for EGFR exon 21 L858R missense. The corresponding radiomic predictors achieved the prediction accuracies of 0.77 and 0.92 in terms of AUC, respectively. Combing these two predictors, the overall model for predicting EGFR mutation positivity was also constructed, and the AUC was 0.87.

Conclusions

In our study, we established predictive models based on radiomic analysis of 18F-FDG PET/CT images. And it achieved a satisfying prediction power in the identification of EGFR mutation status as well as the certain EGFR mutation subtypes in lung cancer.

The associations among quantitative spectral CT parameters, Ki-67 expression levels and EGFR mutation status in NSCLC

Dual-energy spectral computed tomography (DESCT) is based on fast switching between high and low voltages from view to view to obtain dual-energy imaging data, and it can generate monochromatic image sets, iodine-based material decomposition images and spectral CT curves. Quantitative spectral CT parameters may be valuable for reflecting Ki-67 expression and EGFR mutation status in non-small-cell lung cancer (NSCLC). We investigated the associations among the quantitative parameters generated in DESCT and Ki-67 expression and EGFR mutation in NSCLC. We studied sixty-five NSCLC patients with preoperative DESCT scans, and their specimens underwent Ki-67 and EGFR evaluations. Statistical analyses were performed to identify the spectral CT parameters for the diagnosis of Ki-67 expression and EGFR mutation status. We found that tumour grade and the slope of the spectral CT curve in the venous phase were the independent factors influencing the Ki-67 expression level, and the area under the curve (AUC) of the slope of the spectral CT curve in the venous phase in the receiver operating characteristic analysis for distinguishing different Ki-67 expression levels was 0.901. Smoking status and the normalized iodine concentration in the venous phase were independent factors influencing EGFR mutation, and the AUC of the two-factor combination for predicting the presence of EGFR mutation was 0.807. These results show that spectral CT parameters may be useful for predicting Ki-67 expression and the presence of EGFR mutation in NSCLC.

Radiogenomic Models Using Machine Learning Techniques to Predict EGFR Mutations in Non-Small Cell Lung Cancer

BACKGROUND

The purpose of this study was to build radiogenomics models from texture signatures derived from computed tomography (CT) and ¹⁸F-FDG PET-CT (FDG PET-CT) images of non-small cell lung cancer (NSCLC) with and without epidermal growth factor receptor (EGFR) mutations.

METHODS

Fifty patients diagnosed with NSCLC between 2011 and 2015 and with known EGFR mutation status were retrospectively identified. Texture features extracted from pretreatment CT and FDG PET-CT images by manual contouring of the primary tumor were used to develop multivariate logistic regression (LR) models to predict EGFR mutations in exon 19 and exon 20.

RESULTS

An LR model evaluating FDG PET-texture features was able to differentiate EGFR mutant from wild type with an area under the curve (AUC), sensitivity, specificity, and accuracy of 0.87, 0.76, 0.66, and 0.71, respectively. The model derived from CT texture features had an AUC, sensitivity, specificity, and accuracy of 0.83, 0.84, 0.73, and 0.78, respectively. FDG PET-texture features that could discriminate between mutations in EGFR exon 19 and 21 demonstrated AUC, sensitivity, specificity, and accuracy of 0.86, 0.84, 0.73, and 0.78, respectively. Based on CT texture features, the AUC, sensitivity, specificity, and accuracy were 0.75, 0.81, 0.69, and 0.75, respectively.

CONCLUSION

Non-small cell lung cancer texture analysis using FGD-PET and CT images can identify tumors with mutations in EGFR. Imaging signatures could be valuable for pretreatment assessment and prognosis in precision therapy.

Synchronous multiple lung cancers presenting as multifocal pure ground glass nodules: are whole-body positron emission tomography/computed tomography and brain enhanced magnetic resonance imaging necessary?

Background

Multifocal ground glass nodules (GGNs) represent a special radiological pattern indicative of synchronous multiple lung cancers (SMLCs), especially adenocarcinoma. However, the necessity of performing whole-body positron emission tomography/computed tomography (PET-CT) scanning and brain enhanced magnetic resonance imaging (MRI) as a staging workup for multifocal pure GGN (pGGN) patients remains unclear. The purpose of this study was to determine the utility of these two imaging scans for patients with multifocal pGGNs.

Methods

This retrospective study was reviewed and approved by the ethics committee of the Cancer Hospital of the Chinese Academy of Medical Sciences. The study cohort was retrospectively selected from patients with multifocal pGGNs who underwent whole-body PET-CT examinations and/or brain enhanced MRIs between January 2010 and February 2019 at our institution. The additional value of the two exams for detecting nodal and distant metastases was evaluated.

Results

In total, 73 patients (male-to-female ratio, 20:53; median age, 57 years) with multifocal pGGNs who underwent whole-body PET-CT (55 patients) and/or brain enhanced MRI (25 patients) were enrolled. No clearly metastatic lesions were detected. Among the enrolled patients, 53 (128 pGGNs) underwent complete surgical resection. All pGGNs were adenocarcinomas and/or preneoplasias, and no lymph node metastases were found on final pathology. Whole-body PET-CT and brain enhanced MRI added no definite benefit compared with chest CT alone before surgery.

Conclusions

Whole-body PET-CT scans and brain enhanced MRIs are not necessary for patients with multifocal pGGNs.

The value of single-source dual-energy CT imaging for discriminating microsatellite instability from microsatellite stability human colorectal cancer

OBJECTIVES

To demonstrate the value of single-source dual-energy computed tomography (ssDECT) imaging for discriminating microsatellite instability (MSI) from microsatellite stability (MSS) colorectal cancer (CRC).

METHODS

Thirty-eight and seventy-six patients with pathologically proven MSI and MSS CRC, respectively, were retrospectively selected and compared. These patients underwent contrast-enhanced abdominal ssDECT scans before any anti-cancer treatment. Effective atomic number (Eff-Z) in precontrast phase, slope k of spectral HU curve in precontrast (k-P), arterial (k-A), venous (k-V), and delayed phase (k-D), normalized iodine concentration in arterial (NIC-A), venous (NIC-V), and delayed phase (NIC-D), of tumors in two groups were measured by two reviewers. Consistency of measurements was tested by intra-class correlation coefficients (ICC). Mann-Whitney U test or Student's t test was used to compare above values between MSI and MSS. Multivariate logistic regression was used to analyze multiple parameters. Receiver operating characteristic curves were calculated to assess diagnostic efficacies.

RESULTS

Interobserver agreement was excellent (ICC > 0.80). MSI CRC had significantly lower values in all measurements (NIC-A, V, D; k-P, A, V, D; Eff-Z) than MSS CRC. For discriminating MSI from MSS CRC, the area under curve (AUC) using k-A was the highest (AUC, 0.803; sensitivity, 72.4%; specificity, 76.3%). The multivariate logistic regression (selection method, Enter) with combined ssDECT parameters (NIC-A, NIC-V, NIC-D, Eff-Z, k-P, k-A, k-V, k-D) significantly improved diagnostic capability with AUC of 0.886 (sensitivity, 81.6%; specificity, 81.6%).

CONCLUSIONS

The combination of multiple parameters in ssDECT imaging by multivariate logistic regression provides relatively high diagnostic accuracy for discriminating MSI from MSS CRC.

KEY POINTS

• ssDECT generates multiple parameters for discriminating CRC with MSI from MSS. • ssDECT measurements for MSI CRC were significantly lower than MSS CRC. • Combination of ssDECT parameters further improves diagnostic capability for differentiation.