Assessment of high-grade pattern in lung invasive nonmucinous adenocarcinoma based on multiparametric features from dual-layer detector spectral computed tomography combined with extracellular volume fraction

Introduction

Lung cancer remains the leading cause of cancer-related mortality worldwide, with approximately 2.5 million new cases reported in 2022 (1). Among its various subtypes, non-small cell lung cancer (NSCLC) accounts for the majority of cases, with invasive nonmucinous adenocarcinoma (INMA) being the most prevalent histologic type. INMA encompasses a spectrum of growth patterns—including lepidic, acinar, papillary, solid, and micropapillary—which frequently coexist within the same tumor (2,3).

In 2020, the International Association for the Study of Lung Cancer (IASLC) introduced a novel grading system for INMA that incorporates the presence of high-grade patterns (HGPs), which are defined as solid, micropapillary, or complex glandular components (4). This grading system has demonstrated superior prognostic value for predicting recurrence-free and overall survival compared to traditional histologic subtyping alone (4,5). Notably, emerging evidence suggests that even a minor proportion of HGP within the tumor is associated with poorer prognosis (6). Consequently, accurate preoperative assessment of HGP status has become essential for risk stratification and for guiding decisions regarding surgical extent and adjuvant chemotherapy, particularly in patients with early-stage disease.

Tumor behavior is profoundly influenced by its microenvironment, specifically the extracellular matrix (ECM). The ECM provides structural support and regulates critical cellular processes through biochemical and biomechanical signaling (7). ECM remodeling facilitates tumor progression by promoting angiogenesis, enabling cancer cell migration through altered matrix stiffness, and creating pathways for metastatic dissemination. Additionally, interactions between tumor cells and ECM components can induce epithelial-mesenchymal transition (EMT), further enhancing invasive potential. Thus, quantifying ECM characteristics offers valuable insights into tumor aggressiveness and differentiation status.

The extracellular volume (ECV) fraction, derived from quantitative imaging, has emerged as a robust, noninvasive surrogate marker for these microstructural changes and specifically reflects microvascular density and matrix fibrosis (7,8). The principle underlying this function relates to ECV being the quantitative reflection for the proportion of tissue volume not occupied by cells—encompassing the interstitial matrix, fluid spaces, and intravascular compartments. Consequently, the expansion of the ECM due to desmoplastic stromal reactions, inflammatory edema, or aberrant angiogenesis results in a significant increase in ECV values. Recent studies have demonstrated that ECV correlates with pathological differentiation and prognosis across various malignancies. Specifically, in a study on thymic epithelial tumors, ECV values derived from equilibrium-phase dual-energy CT were significantly higher in thymic carcinomas (38.2%) compared to thymomas (25.9%), serving as a reliable metric for differentiating these histological subtypes (8). Similarly, in a study on rectal adenocarcinoma, ECV demonstrated high diagnostic efficiency [area under the curve (AUC) =0.892] in distinguishing high-grade from low-grade tumors, with elevated ECV values correlating positively with poorer pathological differentiation (7). This relationship supports the value of ECV as an imaging biomarker for characterizing tumor grade, stiffness, and the underlying biological state of the lung cancer microenvironment.

Dual-layer detector spectral computed tomography (DLCT) is a relatively recent technological advancement that has enabled the simultaneous acquisition of conventional anatomical images and multi-energy spectral data. Unlike conventional CT, which provides only density information, DLCT exploits the energy-dependent attenuation properties of different materials. By separating low- and high-energy photons, DLCT can generate several quantitative parameters: (I) the effective atomic number (Zeff), which reflects the mean atomic composition of tissues and aids in material characterization; (II) iodine concentration (IC), which quantifies tissue perfusion and vascularity; and (III) virtual monoenergetic images, which optimize contrast at specific energy levels (9,10). These spectral parameters have shown promise in characterizing lung diseases, differentiating between benign and malignant nodules, assessing tumor vascularity, and evaluating treatment response (9,10).

Although conventional CT features—such as spiculation, lobulation, and pleural indentation—have been extensively investigated in their ability to differentiate invasive from noninvasive lung adenocarcinomas, studies specifically focusing on the preoperative prediction of HGP status in INMA are relatively scarce (11). Furthermore, no study thus far has comprehensively evaluated the combined value of ECV, multiparametric spectral imaging, and conventional CT features in predicting HGP. Therefore, this study aimed to evaluate the ability of ECV combined with spectral and conventional CT features to reflect the HGP proportion of INMA and to thus generate insights critical to clinical decision-making. We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-508/rc).

Methods

Study patients

This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments and was approved by the Ethics Committee of Weihai Central Hospital Affiliated to Qingdao University (approval No. LL-2024-044). The requirement for individual consent was waived due to the retrospective nature of the analysis. The clinical and imaging data from patients diagnosed with pulmonary INMA from January 2021 to December 2023 at Weihai Central Hospital Affiliated to Qingdao University were retrospectively collected. All patients had pathologically confirmed INMA. The inclusion criteria were as follows: (I) confirmed diagnosis of INMA with a clearly defined histological subtype; (II) spectral CT scan conducted within 1 month before the pathological examination; (III) the absence of any antitumor or other drug treatments before the CT scan; (IV) and detailed laboratory test results within 1 week of the CT scan. Meanwhile, the exclusion criteria were (I) a history of malignant tumors and (II) poor image quality affecting the analysis. Ultimately, 317 patients comprising 107 males and 210 females were included in the study (Figure 1).

Figure 1 Flowchart of patient selection. CT, computed tomography; HGP, high-grade pattern.

CT image acquisition

All the patient image acquisitions were performed with a dual-layer detector spectral CT system (IQon Spectral CT, Philips Healthcare, Best, the Netherlands). The scanning parameters were as follows: a tube voltage of 120 kVp, a tube current modulated by automated radiation dose control, a collimator width of 64×0.625 mm, a pitch factor of 0.798 mm, a rotation time of 0.5 s, and a slice thickness and spacing of 5.0 mm.

During the enhanced scan, a nonionic contrast agent (300 mg/mL iodine; GE HealthCare, Chicago, IL, USA) was injected through the antecubital vein. Subsequently, an 80-mL dose was administered at a 3 mL/s flow rate. The arterial phase (AP) and venous phase (VP) scans were performed with delays of 30 s and 60 s, respectively. The images were reconstructed according to the lung window [window width: 1,500 Hounsfield units (HU); window level: −500 HU] and mediastinal window (window width: 350 HU; window level: 40 HU) with a 512×512 matrix, a layer thickness of 1.0 mm, and a layer spacing of 1.0 mm.

Imaging analysis

The spectrum-based images were processed on the postprocessing workstation (IntelliSpace Portal version 10.1; Philips Healthcare). The following elements were analyzed via conventional CT images: lesion diameter (average of maximum long and vertical short diameters), lesion shape (round or irregular), vascular anomaly syndrome, tumor edge (clear or indistinct), lobulation, spiculation, pleural indentation, and air bronchogram sign. The most significant lesion cross-section was selected, creating a circular region of interest (ROI) that covered one-half to two-thirds of the lesion, with the blood vessels, bronchi, and vacuoles being avoided. The ROI was then transferred to the iodine density (ID) map. During the AP and VP of the lesion, the Zeff, IC, normalized IC (NIC), the slope of the energy spectrum curve, and CT values in the 40 to 200-keV energy range were measured. The NIC was calculated as follows: NIC = IC_lesion/IC_{artery, venous}, where IC_{artery, venous} represents the IC of the aorta or subclavian artery at the same level during both the AP and VP. The slope formula of the energy spectrum curve was as follows: γHU = (|CT_{40 keV} − CT_{120 keV}|)/60. Finally, the ECV was calculated as follows: ECV = (1 − hematocrit) × IC_lesion/IC_aorta, where IC_lesion and IC_aorta are the VP lesions and aortic iodine densities, respectively. Two radiologists, with 5 and 10 years of experience in diagnosing thoracic tumors, respectively, independently examined the images in a blinded manner. Any discrepancies were resolved through discussion, and the mean of their measurements was calculated to determine the final results.

Histologic analysis

Pathological specimens were evaluated by two experienced pathologists with expertise in diagnosing thoracic tumors. The histological subtypes were classified according to the 2021 World Health Organization standards for lung adenocarcinoma. Each lung INMA tissue subtype was then reported in 5% increments, and subtypes exceeding this threshold were recorded. Specimens were categorized into two groups based on the proportion of HGP component: specimens with HGP component ≥20% and specimens with HGP component <20%.

Statistical analysis

All the data were analyzed with R software version 4.4.1 (R Foundation for Statistical Computing, Vienna, Austria; http://www.Rproject.org) and SPSS version 27.0 (IBM Corp., Armonk, NY, USA). The Kolmogorov-Smirnov test was employed to assess the normality of quantitative data. Normally distributed quantitative data are presented as the mean ± standard deviation, and nonnormally distributed quantitative data are presented as the median and interquartile range (IQR). Group comparisons were made via the independent-sample t-test or Mann-Whitney U test. The χ² test was used to compare categorical variables. The variables that showed statistically significant differences in the univariate analysis were included in a multifactor logistic regression analysis. The stepwise forward selection method was employed to identify independent predictive factors for HGP ≥20% in INMA. P<0.05 was considered statistically significant.

Three models were established as follows: a conventional CT feature model, a spectral CT quantitative parameter model, and a combined parameter model. Receiver operating characteristic (ROC) curves were generated to evaluate the effectiveness of each model in diagnosing HGP ≥20% in INMA. The DeLong test was applied to compare differences in the AUC among the models, and the cutoff values for each model were determined via the Youden index.

The Hosmer-Lemeshow test and a calibration curve were applied to evaluate the goodness of fit and the calibration of the joint diagnostic model. Additionally, decision curve analysis (DCA) was used to assess the clinical applicability of the model. The intraclass correlation coefficient (ICC) and kappa coefficients were used to determine the consistency of parameters evaluated by two physicians, with ICC >0.75 indicating good consistency and kappa values >0.60 indicating high consistency.

Results

Baseline characteristics of patients

This study included 317 patients, consisting of 107 males (33.75%) and 210 females (66.25%), with a mean age of 61.24±8.95 years. Among them, 226 (71.29%) patients had a history of smoking. Two radiologists independently evaluated all imaging materials and subsequently reached a consensus on all qualitative and quantitative characteristics. Interobserver agreement for measurements is shown in Table 1. Among the patients, 46 were placed into the HGP ≥20% group and 271 into the HGP <20% group. There were no significant differences in age (P>0.05), gender (P>0.05), or smoking history (P>0.05) between the groups (Table 2). The ECV in the HGP ≥20% group was smaller than that in the HGP <20% group (P<0.05, Table 2).

Relationships between pathology and dual-layer spectral CT findings

The HGP ≥20% group, compared to the HGP <20% group, exhibited a significantly high incidence of lobulation (P<0.05) and spiculation signs (P<0.05), as well as larger lesion diameters (P<0.05). However, there were no significant differences in lesion shape (P>0.05), edge (P>0.05), pleural traction signs (P>0.05), air bronchus signs (P>0.05), or vascular abnormalities (P>0.05) between the groups (Table 3).

Among the eight spectral CT parameters analyzed, the values of the NIC_VP (P<0.05), γHU_AP (P<0.05), γHU_VP (P<0.05), and IC_AP (P<0.05) were significantly higher in the HGP <20% group, while the values of Zeff_AP (P<0.05) and IC_VP (P<0.05) were lower. NIC_AP (P>0.05) and Zeff_VP (P>0.05) were not significantly different (Figures 2,3, and Table 4). These findings highlight the significance of HGP in clinical evaluations.

Figure 2 Spectral CT of an invasive nonmucinous adenocarcinoma with a high-grade pattern <20% in a 60-year-old female patient. The arrows indicate the tumor lesions. (A) Conventional CT plain lung window image; (B) conventional CT plain mediastinal window image; (C) Zeff_AP pseudocolor images; (D) Zeff_VP pseudocolor images; (E) IC_AP image; (F) IC_VP image. Zeff_AP =8.58; Zeff_VP =8.26; IC_AP =1.58 mg/mL; IC_VP =1.67 mg/mL. AP, arterial phase; CT, computed tomography; IC, iodine concentration; VP, venous phase; Zeff, effective atomic number.

Figure 3 Spectral CT of an invasive nonmucinous adenocarcinoma with high-grade pattern ≥20% in a 71-year-old male patient. The arrows indicate the tumor lesions. (A) Conventional CT plain lung window image. (B) Conventional CT plain mediastinal window image. (C) Zeff_AP pseudocolor images. (D) Zeff_VP pseudocolor images. (E) IC_AP image. (F) IC_VP image. Zeff_AP =8.91; Zeff_VP =8.38; IC_AP =1.21 mg/mL; IC_VP =2.17 mg/mL. AP, arterial phase; CT, computed tomography; IC, iodine concentration; VP, venous phase; Zeff, effective atomic number.

Establishment of predictive models

The multifactor logistic analysis included conventional CT features and spectral CT parameters that exhibited significant differences compared to the single-factor analysis. The findings revealed that the independent factors predicting HGP ≥20% in INMA were ECV [odds ratio (OR) =0.004; P=0.036], lesion diameter (OR =3.345; P<0.001), NIC_VP (OR =0.001; P=0.001), Zeff_AP (OR =4.936; P<0.001), IC_AP (OR =0.190; P<0.001), IC_VP (OR =2.420; P=0.004), lobulation sign (OR =3.509; P=0.007), and spiculation sign (OR =3.409; P=0.006, Table 5).

The statistically significant factors from the single-factor analysis were used to establish conventional CT features and spectral CT quantitative parameter models. The independent predictors identified in the multivariate logistic regression analysis were combined to establish a joint model. Furthermore, the Hosmer-Lemeshow test confirmed the absence of overfitting (χ²=10.366, P=0.240). The calibration curve (Figure 4) indicated a strong alignment between the predicted and actual probabilities of HGP ≥20% in INMA, represented by a mean absolute error of 0.023.

Figure 4 Calibration curves for the combined model. There was good agreement between the predicted and actual probability of the proportion of high-grade pattern in invasive nonmucinous adenocarcinoma.

Comparison of the models’ prediction accuracy

The AUC of the conventional CT feature model, the spectral CT quantitative parameter model, and the combined model in predicting HGP ≥20% in INMA were 0.825, 0.829, and 0.909, respectively. Moreover, the AUC of the combined model was significantly higher than that of the conventional model (Z=3.350; P=0.001) and spectral model (Z=2.8167, P=0.005). There were no significant differences between the spectral and conventional models (Z=−0.093, P=0.926, Table 6 and Figure 5). In addition, the DCA demonstrated that the combined model offers a higher clinical benefit rate than do the other two models (Figure 6).

Figure 5 Receiver operating characteristic curves of three models for predicting the proportion of high-grade pattern in invasive nonmucinous adenocarcinoma of the lung. The combined model had the largest AUC. AUC, area under the curve.

Figure 6 Clinical decision curves for the three models. The curves showed a higher net benefit for the combined model than for the other two models. “All” indicates that all patients had invasive nonmucinous adenocarcinoma with a proportion of high-grade pattern ≥20%. “None” indicates that all patients had invasive nonmucinous adenocarcinoma with a proportion of high-grade pattern <20%.

Discussion

Previous studies have confirmed that the morphological characteristics of lung lesions, such as density, size, lobulation sign, spiculation sign, air bronchogram sign, and pleural traction sign, are crucial for predicting HGP in INMA (12-16). For instance, Chen et al. (12) found that tumors with larger diameters and solid components had a higher incidence and proportion of HGP. Moreover, Li et al. (14) reported a specificity of 97.8%, a sensitivity of 54.5%, and AUC of 0.820 when using tumor size to differentiate HGP.

Additionally, Wang et al. (16) discovered that lobulated or spiculated lesions frequently contain higher proportions of HGP, while Dong et al. (15) identified the shape of the lesion as an independent risk factor for HGP. These features suggest the aggressive nature of HGP in INMA, characterized by rapid tumor growth and significant invasion of adjacent lung tissue. In line with previous findings, our study showed that the lesions in the HGP ≥20% group exhibited more lobulation and spiculation signs, along with larger diameters (13,16). Relying solely on conventional features to determine HGP proportions in INMA poses significant limitations. The interaction between tumor cells and their microenvironment is crucial for activating the tumor-associated stromal cells that function in the facilitation of angiogenesis (17,18), which leads to vital changes in ECV. In addition, Fukui et al. (19) found that the ECV at an optimal cutoff of 32.8% is effective in diagnosing pancreatic cancer, with a sensitivity of 61.0% and a specificity of 85.1%. Moreover, DLCT facilitates the reconstruction of parameters such as ID images, enhancing the quantitative analysis of lung tumors (20). This advancement is essential for distinguishing between benign and malignant cases (21,22) and improves the accuracy of the diagnosis, thereby establishing a trajectory for future innovations in the measurement of ECV.

In this study, the magnitude of ECV varied significantly across the groups. In addition, the IC_AP and NIC_VP were lower in the HGP ≥20% group. This potentially resulted from the rapid tumor growth surpassing the formation of new blood vessels, thus promoting the occurrence of ischemia and hypoxia in the local microenvironment, which ultimately reduced the ECV values. These results align with the findings from Li et al. (14). Furthermore, the IC_VP in the HGP ≥20% group was notably higher than that in the HGP <20% group. This may be attributed to the fact that the lung tumors exhibiting a high proportion of HGP also have high cell density and a more compact structure. Previous research indicated that Zeff analysis is a valuable tool for elucidating the characteristics of the tissues (23). Our study found that the Zeff_AP of lesions in the HGP ≥20% group was higher than that of those in the HGP <20% group, suggesting significant differences in cell number, invasion extent, and fiber composition. Recognizing these distinctions is crucial for improving diagnostic and therapeutic approaches.

Certain limitations to this study should be acknowledged. To begin, the data were derived from a single center and included an imbalanced sample size for each group. In addition, the ROI was defined based only on the largest cross-section of the lesion, potentially missing tumor heterogeneity and important CT features, such as the proportion of solid components. Moreover, the CT values were not included. Future research should prioritize multicenter studies with large sample sizes to identify the key parameters and implement external validation to enhance the accuracy of the models.

Conclusions

The combined model incorporating ECV, lesion diameter, lobulation, spiculation, NIC_VP, Zeff_AP, IC_AP, and IC_VP demonstrated significant predictive value for estimating the proportion of HGP in INMA. Compared with models based on conventional CT features or spectral CT quantitative parameters, the multiparameter combined model achieved superior diagnostic performance and may serve as an effective noninvasive tool for predicting the HGP proportion in INMA.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-508/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-508/dss

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-508/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of Weihai Central Hospital Affiliated to Qingdao University (No. LL-2024-044). The requirement for individual consent was waived due to the retrospective nature of the analysis.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Bray F, Laversanne M, Sung H, Ferlay J, Siegel RL, Soerjomataram I, Jemal A. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2024;74:229-63. [Crossref] [PubMed]
2. Hendriks LEL, Remon J, Faivre-Finn C, Garassino MC, Heymach JV, Kerr KM, Tan DSW, Veronesi G, Reck M. Non-small-cell lung cancer. Nat Rev Dis Primers 2024;10:71. [Crossref] [PubMed]
3. Nicholson AG, Tsao MS, Beasley MB, Borczuk AC, Brambilla E, Cooper WA, Dacic S, Jain D, Kerr KM, Lantuejoul S, Noguchi M, Papotti M, Rekhtman N, Scagliotti G, van Schil P, Sholl L, Yatabe Y, Yoshida A, Travis WD. The 2021 WHO Classification of Lung Tumors: Impact of Advances Since 2015. J Thorac Oncol 2022;17:362-87. [Crossref] [PubMed]
4. Hou L, Wang T, Chen D, She Y, Deng J, Yang M, Zhang Y, Zhao M, Zhong Y, Ma M, Zhao G, Chen Y, Xie D, Zhu Y, Chen Q, Wu C, Chen C. Prognostic and predictive value of the newly proposed grading system of invasive pulmonary adenocarcinoma in Chinese patients: a retrospective multicohort study. Mod Pathol 2022;35:749-56. [Crossref] [PubMed]
5. Moreira AL, Ocampo PSS, Xia Y, Zhong H, Russell PA, Minami Y, et al. A Grading System for Invasive Pulmonary Adenocarcinoma: A Proposal From the International Association for the Study of Lung Cancer Pathology Committee. J Thorac Oncol 2020;15:1599-610. [Crossref] [PubMed]
6. Mikubo M, Tamagawa S, Kondo Y, Hayashi S, Sonoda D, Naito M, Shiomi K, Ichinoe M, Satoh Y. Micropapillary and solid components as high-grade patterns in IASLC grading system of lung adenocarcinoma: Clinical implications and management. Lung Cancer 2024;187:107445. [Crossref] [PubMed]
7. Li Q, Bao J, Zhang Y, Dou Y, Liu A, Liu M, Wu H, Wu J, Zhao L, Yang Z, Zhu L, Gao W, Zhao W, Wei M, Hao F. Predictive value of CT-based extracellular volume fraction in the preoperative pathologic grading of rectal adenocarcinoma: A preliminary study. Eur J Radiol 2023;163:110811. [Crossref] [PubMed]
8. Takumi K, Nagano H, Myogasako T, Nakano T, Fukukura Y, Ueda K, Tabata K, Tanimoto A, Yoshiura T. Feasibility of iodine concentration and extracellular volume fraction measurement derived from the equilibrium phase dual-energy CT for differentiating thymic epithelial tumors. Jpn J Radiol 2023;41:45-53. [Crossref] [PubMed]
9. Zhang G, Li S, Yang K, Shang L, Zhang F, Huang Z, Ren J, Zhang Z, Zhou J, Pu H, Man Q, Kong W. The value of dual-energy spectral CT in differentiating solitary pulmonary tuberculosis and solitary lung adenocarcinoma. Front Oncol 2022;12:1000028. [Crossref] [PubMed]
10. Wang T, Fan Z, Zou L, Hou Y. Can quantitative parameters of spectral computed tomography predict lymphatic metastasis in lung cancer? a systematic review and meta-analysis. Radiother Oncol 2023;183:109643. [Crossref] [PubMed]
11. Yu Y, Fu Y, Chen X, Zhang Y, Zhang F, Li X, Zhao X, Cheng J, Wu H. Dual-layer spectral detector CT: predicting the invasiveness of pure ground-glass adenocarcinoma. Clin Radiol 2022;77:e458-65. [Crossref] [PubMed]
12. Chen LW, Yang SM, Wang HJ, Chen YC, Lin MW, Hsieh MS, Song HL, Ko HJ, Chen CM, Chang YC. Prediction of micropapillary and solid pattern in lung adenocarcinoma using radiomic values extracted from near-pure histopathological subtypes. Eur Radiol 2021;31:5127-38. [Crossref] [PubMed]
13. Dong H, Yin LK, Qiu YG, Wang XB, Yang JJ, Lou CC, Ye XD. Prediction of high-grade patterns of stage IA lung invasive adenocarcinoma based on high-resolution CT features: a bicentric study. Eur Radiol 2023;33:3931-40. [Crossref] [PubMed]
14. Li Q, Li X, Li XY, He XQ, Chu ZG, Luo TY. Histological subtypes of solid-dominant invasive lung adenocarcinoma: differentiation using dual-energy spectral CT. Clin Radiol 2021;76:77.e1-7. [Crossref] [PubMed]
15. Dong H, Wang X, Qiu Y, Lou C, Ye Y, Feng H, Ye X, Chen D. Establishment and visualization of a model based on high-resolution CT qualitative and quantitative features for prediction of micropapillary or solid components in invasive lung adenocarcinoma. J Cancer Res Clin Oncol 2023;149:10519-30. [Crossref] [PubMed]
16. Wang F, Wang CL, Yi YQ, Zhang T, Zhong Y, Zhu JJ, Li H, Yang G, Yu TF, Xu H, Yuan M. Comparison and fusion prediction model for lung adenocarcinoma with micropapillary and solid pattern using clinicoradiographic, radiomics and deep learning features. Sci Rep 2023;13:9302. [Crossref] [PubMed]
17. de Visser KE, Joyce JA. The evolving tumor microenvironment: From cancer initiation to metastatic outgrowth. Cancer Cell 2023;41:374-403. [Crossref] [PubMed]
18. Panettieri E, Maki H, Kim BJ, Kang HC, Cox V, Vega EA, Mizuno T, Pant S, Javle M, Vauthey JN, Kawaguchi Y. Arterial enhancement pattern predicts survival in patients with resectable and unresectable intrahepatic cholangiocarcinoma. Surg Oncol 2022;40:101696. [Crossref] [PubMed]
19. Fukui H, Onishi H, Nakamoto A, Tsuboyama T, Ota T, Yano K, Enchi Y, Yamada D, Takeda Y, Kobayashi S, Fukuda Y, Eguchi H, Matsui T, Tatsumi M, Tomiyama N. Pancreatic fibrosis by extracellular volume fraction using Contrast-enhanced computed tomography and relationship with pancreatic cancer. Eur J Radiol 2022;156:110522. [Crossref] [PubMed]
20. Kim C, Kim W, Park SJ, Lee YH, Hwang SH, Yong HS, Oh YW, Kang EY, Lee KY. Application of Dual-Energy Spectral Computed Tomography to Thoracic Oncology Imaging. Korean J Radiol 2020;21:838-50. [Crossref] [PubMed]
21. Zegadło A, Żabicka M, Kania-Pudło M, Maliborski A, Różyk A, Sośnicki W. 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. J Clin Med 2020;9:2514. [Crossref] [PubMed]
22. Liu G, Li M, Li G, Li Z, Liu A, Pu R, Cao H, Liu Y. Assessing the Blood Supply Status of the Focal Ground-Glass Opacity in Lungs Using Spectral Computed Tomography. Korean J Radiol 2018;19:130-8. [Crossref] [PubMed]
23. Deniffel D, Sauter A, Fingerle A, Rummeny EJ, Makowski MR, Pfeiffer D. Improved differentiation between primary lung cancer and pulmonary metastasis by combining dual-energy CT-derived biomarkers with conventional CT attenuation. Eur Radiol 2021;31:1002-10. [Crossref] [PubMed]