Value of spectral parameters in the differential diagnosis of benign and malignant breast nodules in non-enhanced chest CT

Introduction

Breast cancer is the most common cancer in women worldwide, and its early detection is critical to improving patient outcomes (1). In recent years, there has been a significant global increase in the detection rate of breast lesions due to the ever-increasing use of computed tomography (CT) scans as a diagnostic tool (2-4). However, many breast nodules are also discovered incidentally, and previous studies have reported that such nodules have a malignancy rate of 28–32% (5-8). The accurate identification of benign and malignant breast nodules is crucial in minimizing the unnecessary biopsy of benign nodules and reducing the risk of metastasis caused by biopsies. Patients with high-risk breast nodules should undergo surgical resection followed by postoperative chemotherapy, while those with benign nodules can typically be managed with follow-up examinations or minimally invasive surgery. Therefore, appropriate patient selection can prevent overtreatment and unnecessary surgical intervention.

In clinical practice, magnetic resonance imaging (MRI), ultrasound (US), and mammography are commonly used in the diagnosis of breast nodules (9,10), and each has its own advantages and disadvantages. For example, due to its exceptional soft-tissue resolution, MRI has superior sensitivity; however, it also has a prolonged examination time, high cost, and limited suitability for individuals with claustrophobia. Conversely, US may result in the misdiagnosis of certain early malignant nodules due to subjective judgement discrepancies among operators. In addition, high breast density increases the risk of breast cancer, and can mask tumors, and the sensitivity of mammography decreases when applied to dense mammary glands (11,12). Conventional chest CT is extensively employed in routine physical examinations and clinical staging assessments; however, due to its limited soft-tissue resolution and absence of quantitative parameters, it relies solely on morphological criteria for subjective evaluation.

The emerging dual-layer spectral computed tomography (DSCT) technology employs two layers made from different materials of detectors to simultaneously obtain low- and high-energy separation spectral images in all patients using standard CT protocols, including the effective atomic number (Zeff), virtual monoenergetic images (VMIs), and iodine concentration (IC) (13,14). The technique also has the advantages of enhanced visualization of tissue contrast in images, improved objective measurements of image quality, and equivalent subjective image quality at reduced radiation dosage (15,16). Further, it can be used to determine material composition and reduce metal artifacts, and thus has promising prospects for clinical applications, such as enhancing lesion detection sensitivity and qualitative accuracy (13,17-19).

Previous studies have shown that some DSCT quantitative parameters, particularly the IC and the slope of the spectral Hounsfield unit curve (λHU), play a pivotal role in assessing breast nodules, including in distinguishing between benign and malignant breast nodules, as well as predicting the histopathological classifications and its associations with immunohistochemical biomarkers in breast carcinomas (20-23). Further, additional research has shown that the performance of machine-learning models is superior to that of a univariate analysis (24). However, the majority of studies have primarily relied on enhanced CT scans, and limited research has examined the quantitative parameters derived from non-enhanced DSCT scans.

In this study, we hypothesized that the quantitative parameters obtained from DSCT non-enhanced images could serve as valuable discriminators between benign and malignant breast nodules. To validate our hypothesis, we aimed to develop models of DSCT quantitative parameters and typical radiological features for the early identification of malignant breast nodules in patients, and compare the diagnostic efficacy of these models. Such modules could facilitate personalized 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-24-575/rc).

Methods

Patient selection

The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). This retrospective study was approved by the Ethics Review Board of The First Affiliated Hospital of Soochow University (No. 2024423), and the requirement of individual consent for this retrospective analysis was waived. This retrospective cross-sectional study recruited 150 patients with breast nodules at The First Affiliated Hospital of Soochow University from March 2023 to December 2023. Ultimately, a total of 121 breast nodules from 114 patients who underwent DSCT scans and pathological diagnosis were enrolled in the study, including 68 cases in the malignant group, and 53 cases in the benign group. All the patients were women with a mean age ± standard deviation of 51.18±15.45 years and an age range of 13–85 years. The exclusion criteria were as follows: (I) incomplete spectral parameters; (II) patients who had received neoadjuvant chemotherapy before DSCT examination; and/or (III) insufficient image quality for measurement.

DSCT technique

The non-enhanced chest scans of all patients using a DSCT (IQon Spectral CT, Philips Healthcare, Best, The Netherlands) followed the same routine protocol. The scan coverage was taken from the lung apex to the lowest hemi-diaphragm. The following routine scan parameters were used without specific optimization: tube voltage: 120 kV; tube current (mA): adjusted by automatic exposure control technology and the adaptive iterative reconstruction algorithm; detector collimation: 64×0.625 mm; tube speed: 0.5 s/r; screw pitch: 0.925; field of view: 500 mm; and image reconstruction matrix: 500×500.

Image analysis and measurement of DSCT parameters

The radiological features were independently analyzed by two experienced radiologists (with 7 and over 10 years of chest CT diagnosis experience, receptively) who did not have access to any patient information or pathological results. If any discrepancies arose, a consensus was reached through consultation. The radiological features assessed included the morphology (quasi-circular/irregular), lobular sign, spicule sign, and calcification. The non-enhanced CT attenuation values were obtained by placing the regions of interest (ROIs) on the breast nodules, covering no less than 2/3 of the lesion area, avoiding necrosis and calcification. Three ROIs were selected for each section to calculate the average value. Additionally, the maximum diameter of each nodule was measured at the largest cross-section of the nodule.

The spectral-based images were transferred to the Philips Intellispace Portal 9.0 workstation to automatically reconstruct the VMIs (40–200 keV) and Zeff map. The section with the largest mass area was selected to manually delineate the largest possible ROIs to cover no less than 2/3 of the area of the mass, avoiding necrosis and calcification. The DSCT parameters were obtained independently by a radiologist with 2 years of experience. To reduce the measurement variation, three ROIs were selected for each section to calculate the average value. The spectral parameters used in this study were as follows: (I) the CT attenuation values (HU) from 40 to 70 keV; (II) the λHU; and (III) the Zeff. The λHU was calculated as follows: λHU = (CT attenuation value at 70 keV − CT attenuation value at 40 keV)/30.

Statistical analysis

All the statistical analyses were performed using SPSS software (version 26.0, SPSS, IBM). The Chi-square test or Fisher’s exact test was used to compare the qualitative features. While the independent-sample t-test or Mann-Whitney U test was used to compare the quantitative parameters. The variables with statistically significant results in the two groups were included in the logistic regression analysis, and the diagnostic models for DSCT and conventional CT were then established. The area under the curves (AUCs) of the receiver operating characteristic (ROC) curves, with the expressing 95% confidence intervals (95% CIs), were obtained to evaluate the diagnostic ability of the models. The Delong test was used to compare the diagnostic performance of the two models. The Kappa coefficient was used to assess the inter-observer agreement for the typical radiological features extracted by the two readers (a value >0.90 indicated excellent agreement; 0.81–0.90 indicated satisfactory agreement; 0.71–0.80 indicated moderate agreement; 0.60–0.70 indicated inadequate agreement; and <0.60 indicated poor agreement). A two-sided P value <0.05 was considered statistically significant.

Results

Participant characteristics

In total, 150 participants were initially enrolled in our study, of whom 36 were excluded because they had incomplete spectral parameters (n=22), poor-quality images (n=2), had received neoadjuvant chemotherapy before the DSCT scan (n=1), and had invisible breast nodules (n=11). Ultimately, 114 patients with a total of 121 breast nodules were included in the study. The patients were all women with a mean age ± standard deviation of 51.18±15.45 years, and an age range of 13–85 years. The patient selection process is illustrated in Figure 1. According to the pathological results, the lesions were divided into the malignant group (n=68, which included 60 invasive ductal carcinomas, 3 invasive lobular carcinomas, 3 ductal carcinomas in situ, 1 mucinous breast carcinoma, and 1 solid papillary carcinoma) and the benign group (n=53, which included 48 fibroadenomas, 3 cases of mammary duct ectasias, and 2 cases of chronic mastitis).

Figure 1 Patient selection process. DSCT, dual-layer spectral computed tomography.

Comparison of clinical data and radiological features

The kappa values of the tumor morphology, lobular sign, spicule sign, and calcification features were all >0.85, indicating satisfactory agreement between the two readers. The clinical data and typical radiological features of the benign and malignant nodules are summarized in Table 1. Based on our analysis, significant differences were observed between the two groups in terms of age (P<0.001), tumor size (P=0.003), and typical radiological features, including the lobular sign (P<0.001) and spicule sign (P<0.001). However, no significant differences were observed between the two groups in terms of the CT attenuation values, tumor morphology, and calcification (P>0.05).

Comparison of DSCT parameters

The DSCT quantitative parameters of the benign and malignant nodules are summarized in Table 2, and representative images are shown in Figure 2. The Zeff (P<0.001), λHU (P<0.001), and CT attenuation values at 40–70 keV (40 keV, P<0.001; 50 keV, P=0.001; 60 keV, P=0.008; 70 keV, P=0.027) exhibited significant differences between the two groups. Specifically, the Zeff and CT attenuation values at 40 and 50 keV were higher in the malignant group than the benign group, while the λHU was lower in the malignant group than the benign group.

Figure 2 Representative images of two breast nodule patients. (A-D) 40 keV, 70 keV, conventional CT, and effective atomic number map of a 54-year-old woman with breast fibroadenoma. The CT attenuation values at 40 and 70 keV, and the Zeff are 32.9 HU, 46.7 HU, and 7.12, respectively. The margin of the mass was smooth and the density was uniform. 40 keV had higher contrast and tumor visibility than 70 keV and conventional images. (E-H) 40 keV, 70 keV, conventional CT, and effective atomic number map of a 54-year-old woman with breast invasive ductal carcinoma. The CT attenuation values at 40 and 70 keV, and the Zeff are 56.4 HU, 51.3HU, and 7.32, respectively. The lobular sign and spicule sign can be seen. (I) λHU of the benign energy curve is 0.46 HU/keV, while that of the malignant energy curve is –0.14. (J) Photomicrograph confirmed the pathological finding of the malignant tumor as invasive ductal carcinoma (hematoxylin and eosin, ×40). 40 keV, CT attenuation value at 40 keV; 70 keV, CT attenuation value at 70 keV; HU, Hounsfield unit; Zeff, effective atomic number; λHU, slope of the spectral HU curve; CT, computed tomography.

Logistic regression and ROC curve analysis

The logistic regression analysis (Table 3) revealed that DSCT quantitative parameters, including the Zeff [odds ratio (OR): 9.22; 95% CI: 2.11–40.35; P=0.003], λHU (OR: 0.64; 95% CI: 0.52–0.79; P<0.001), and CT attenuation values at 40 keV (OR: 8.69; 95% CI: 3.28–23.06; P<0.001) and 50 keV (OR: 0.01; 95% CI: 0.001–0.07; P<0.001), were risk factors for predicting malignant breast nodules. The above four DSCT parameters were combined to establish a DSCT diagnostic model (combined model 1). Table 4 presents the corresponding AUC, sensitivity, and specificity values for each DSCT quantitative parameter, and the combined DSCT diagnostic model. Notably, the ROC curve analysis revealed that the AUC for the Zeff, λHU, and CT attenuation values at 40 and 50 keV were 0.792 (95% CI: 0.71–0.87), 0.733 (95% CI: 0.64–0.83), 0.735 (95% CI: 0.64–0.83), and 0.686 (95% CI: 0.59–0.78), respectively. The Zeff exhibited superior diagnostic capability compared to the other DSCT parameters. Further, when all four parameters were combined, the AUC achieved a value of 0.899 (95% CI: 0.84–0.96) (Figure 3).

Figure 3 The ROC curve of the DSCT diagnostic model (combined model 1) showed the differential diagnosis efficacy of the Zeff (AUC =0.792), λHU (AUC =0.733), 40 keV (AUC =0.735), 50 keV (AUC =0.686), and the combined DSCT parameters (AUC =0.899). Zeff, effective atomic number; λHU, slope of the spectral HU curve; 40 keV, CT attenuation value at 40 keV; 50 keV, CT attenuation value at 50 keV; ROC, receiver operating characteristic; DSCT, dual-layer spectral computed tomography; AUC, area under the curve.

Subsequently, we integrated the lobular sign, spicule sign, and age to develop a conventional CT diagnostic model (combined model 2). Table 5 summarizes the AUC, sensitivity, and specificity values for the typical radiological features and the conventional CT diagnostic model. In contrast to the conventional model for which the lobular sign had an AUC of 0.705 (95% CI: 0.61–0.80), the spicule sign had an AUC of 0.628 (95% CI: 0.53–0.73), and age had an AUC of 0.690 (95% CI: 0.59–0.79), the conventional CT diagnostic model had an AUC of 0.796 (95% CI: 0.72–0.87) (Figure 4). The Delong test showed that the DSCT parameter diagnostic model outperformed the conventional CT model in terms of predictive performance (P=0.04).

Figure 4 The ROC curve of the conventional CT diagnostic model (combined model 2) showed the differential diagnosis efficacy of the lobular sign (AUC =0.705), needle sign (AUC =0.628), age (AUC =0.690), and combined model (AUC =0.796). ROC, receiver operating characteristic; CT, computed tomography; AUC, area under the curve.

Discussion

Our study showed that the diagnostic model combining multiple DSCT quantitative parameters had superior efficacy in the differential diagnosis of malignant and benign breast nodules than the conventional CT model. Thus the quantitative parameters of DSCT can be a useful supplement to conventional typical radiological features, its can provide more information, and could assist in the identification of breast nodules found by chance that require further diagnostic evaluation. Thus the identification of breast nodules founded by chance which require further diagnostic evaluation would be easier for radiologists. Further, the use of DSCT quantitative parameters could prevent unnecessary examinations and biopsies, which in turn could alleviate the economic burden placed on patients and reduce the risk of injury.

Our study revealed that DSCT quantitative parameters, including the Zeff, CT attenuation values at 40–50 keV, and λHU, can be used independently to distinguish between benign and malignant breast nodules. Notably, the Zeff had the highest efficacy with an AUC of 0.792. Further, consistent with previous research studies (21,22), we found that the Zeff values of the malignant nodules were significantly higher than those of the benign nodules. The Zeff represents the average atomic number of a mixture of various composite substances, reflecting the intrinsic material composition of a tissue or lesion.

In conventional CT imaging, breast nodules often have similar densities and radiological features, leading to overlapping characteristics between early stage malignancies and benign masses. The findings of our study indicated that there was no statistically significant difference in the CT attenuation values between the benign and malignant groups. Further, the Zeff of the malignant group was higher than that of the benign group, which might be due to abnormal cell proliferation in malignant breast nodules and changes in internal composition. Therefore, if distinguishing between benign and malignant masses becomes challenging based solely on conventional CT parameters alone, the Zeff could provide additional valuable information (25,26).

Additionally, a statistically significant difference was observed in the λHU values between the benign and malignant groups due to variations in tissue composition, blood supply, cell growth patterns, and other characteristics. The attenuation caused by X-rays differs under different energy levels, leading to potential differences in the attenuation characteristics of CT attenuation values and spectral curves. As the energy level increases, the X-ray penetration ability strengthens, and the tissues absorb less X-ray radiation, resulting in reduced differences between tissues. Consequently, the spectral curve exhibits a steeper slope in the 40–70 keV range, before gradually flattening as energy levels increase. Thus, for the purposes of this study, the λHU was calculated using CT attenuation values at the 40–70 keV level. Consistent with previous findings, we found that the λHU was lower in the malignant group than the benign group (21,22).

Recently, Metin (27) and Okada (28) showed that low keV VMIs obtained by DSCT exhibited superior ability in the discrimination of breast nodules with significantly higher conspicuity. Specifically, subjective tumor visibility was found to be better at 40 keV and 50 keV than 120 keV (29). The present study found statistically significant differences between the benign and malignant groups at 40 and 50 keV, with the most pronounced difference observed at 40 keV. Moreover, a progressively more prominent distinction was evident as the energy level decreased, which suggests that nodules detected at lower energy levels have greater significance. These findings are consistent with those of previous studies (27,28).

Li (30) and Liao (31) reported that the breasts of younger patients were denser, and breast density decreased with age. However, in recent years, the average age of patients with breast cancer has decreased (32,33). Sometimes, it can be relatively difficult to differentiate tumor tissue from denser mammary glands solely through conventional non-enhanced CT images. The Zeff images generated by DSCT are presented in color, enabling distinct substances to be represented by different colors. Notably, we found that on the Zeff images, certain nodules exhibited greater prominence than normal mammary glands and manifested with enhanced clarity. Any enhancement in the tumor detection rate in dense mammary glands would be beneficial.

The presence of the lobular sign and spicule sign are generally thought to indicate malignant nodules. The appearance of lobular sign is mainly due to the uneven growth rate of a malignant tumor, while the growth rate of a benign tumor is slower and the edge is generally smooth (34). The results of our study also showed the two radiological features were important signs for distinguishing between benign and malignant nodules, and the lobular sign was an independent factor for identifying malignant nodules. However, the evaluation of radiological features depends on the diagnostic ability of radiologists.

The spicule sign is a significant indicator for distinguishing between benign and malignant breast lesions. However, in our study, the AUC of the spicule sign was not particularly high. Indeed, it was lower than the AUC of age. This might be due to a number of factors. First, the study was based on non-enhanced CT, which has limited lesion resolution. Second, previous studies have shown that a tumor with a low grade is more likely to have a spicule margin (35,36). It may be that low-grade tumors have low cellularity and a rich collagen matrix, and thus tend to show a greater desmoplastic reaction in the surrounding tissue than high-grade tumors (36,37). It is also possible that some high-grade tumor cases in the malignant group in our study did not show spicule signs. Third, due to the differences in the enrolled samples, the age of the benign group was lower than that of the malignant group. We intend to expand the sample size in future research.

The diagnostic efficacy of a single DSCT parameter or radiological feature is not particularly pronounced. Notably, we found that while the Zeff and CT attenuation value at 40 keV showed statistically significant differences between the benign and malignant groups, the 95% CIs from the logistic regression analysis were wide (Zeff: 2.11–40.35; 40 keV: 3.28–23.06). This suggests that the diagnostic accuracy based on these single parameters may not be particularly reliable. Further, a wide CI implies a limited sample size. Thus, these parameters, require further validation in a larger cohort.

We subsequently integrated multiple DSCT parameters and radiological features to establish a DSCT diagnostic model (combined model 1) and a conventional CT diagnostic model (combined model 2), which had AUCs of 0.899 (95% CI: 0.84–0.96) and 0.796 (95% CI: 0.72–0.87), respectively. The diagnostic model incorporating multiple parameters had a more refined CI. This indicates that a more comprehensive consideration of multiple parameters should be taken into account in the clinical diagnosis of lesions to enhance diagnostic accuracy. The Delong test revealed that the diagnostic efficacy of the DSCT diagnostic model was significantly higher than that of the conventional CT diagnostic model. Thus, the DSCT diagnostic model could help radiologists to identify benign and malignant breast tumors in non-enhanced images. It could also be a useful, objective supplement to radiological features, and offer a simplified and convenient approach for the diagnosis of clinical breast nodules. Our findings also have significant implications for the differential diagnosis of incidental benign and malignant breast nodules detected during routine physical examinations or other chest CT scans, and could potentially reduce the need for unnecessary biopsy procedures or re-examinations.

This study had several limitations. First, this study was a single-center study with a small sample size, and the specific pathological types of the benign and malignant nodules were not further distinguished. The DSCT quantitative parameters of specific pathological types differ. The limited sample size and the heterogeneity in the data might have resulted in diminished diagnostic reliability for specific parameters. Thus, it is imperative to collect a larger number of samples to enhance the depth of our study, and an external validation cohort is needed to further validate our results. Second, this retrospective study might have introduced selection bias, and some patients in the malignant group underwent DSCT scans to evaluate the stage of breast tumors and potential metastasis, which should be taken into consideration in the analysis and interpretation of the results. Third, the parameters measured from the maximum cross-section of the tumor were exclusively selected for analysis in this study. In future studies, it is anticipated that software or radiomics analysis could be employed to comprehensively analyze entire tumors.

Conclusions

Compared with conventional CT, DSCT can provide more diverse quantitative parameters for the diagnosis and differentiation of diseases. These DSCT parameters could also be used in the diagnosis of benign and malignant breast nodules detected incidentally during routine physical examinations or other chest CT scans, and could reduce unnecessary puncture biopsy procedures or re-examinations.

Acknowledgments

The authors would like to express their gratitude to Dr. Chunhong Hu, Dr. Ximing Wang, and Dr. Junpeng Luo from the Department of Radiology at The First Affiliated Hospital of Soochow University for their invaluable support throughout this study. Additionally, the authors would like to thank the dedicated staff in the Radiology Technical Group for their assistance in facilitating data collection. We are also grateful for the technical guidance provided by the Philips CT team.

Funding: This study was supported by Jiangsu Province Capability Improvement Project through Science, Technology and Education (Jiangsu Provincial Medical Key Discipline Cultivation Unit) (No. JSDW202242), and Suzhou Key Laboratory of Medical Imaging (No. SZS2024032).

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-24-575/rc

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-24-575/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. The study was approved by the Ethics Review Board of The First Affiliated Hospital of Soochow University (No. 2024423), and the requirement of individual consent for this retrospective analysis was waived. The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013).

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. Krug KB, Houbois C, Grinstein O, Borggrefe J, Puesken M, Hanstein B, Malter W, Maintz D, Hellmich M. Focal Breast Lesions in Clinical CT Examinations of the Chest: A Retrospective Analysis. Rofo 2017;189:977-89. [Crossref] [PubMed]
3. Agliata MF, Calabrò N, Tricca S, Rampi AM, Gambaro ACL, Ferrante D, Carriero A. Mammary nodules as incidental findings on chest computed tomography: a retrospective analysis on their frequency and predictive value. Radiol Med 2023;128:912-21. [Crossref] [PubMed]
4. Park J, Kim C, Cha YK, Chung MJ. Missed and Detected Incidental Breast Cancers on Contrast Enhanced Chest CT: Detection Rates and CT Features. Diagnostics (Basel) 2023.
5. Monzawa S, Washio T, Yasuoka R, Mitsuo M, Kadotani Y, Hanioka K. Incidental detection of clinically unexpected breast lesions by computed tomography. Acta Radiol 2013;54:374-9. [Crossref] [PubMed]
6. Porter G, Steel J, Paisley K, Watkins R, Holgate C. Incidental breast masses detected by computed tomography: are any imaging features predictive of malignancy? Clin Radiol 2009;64:529-33. [Crossref] [PubMed]
7. Moyle P, Sonoda L, Britton P, Sinnatamby R. Incidental breast lesions detected on CT: what is their significance? Br J Radiol 2010;83:233-40. [Crossref] [PubMed]
8. Hussain A, Gordon-Dixon A, Almusawy H, Sinha P, Desai A. The incidence and outcome of incidental breast lesions detected by computed tomography. Ann R Coll Surg Engl 2010;92:124-6. [Crossref] [PubMed]
9. Wang B, Chen YY, Yang S, Chen ZW, Luo J, Cui XW, Dietrich CF, Yi AJ. Combined Use of Shear Wave Elastography, Microvascular Doppler Ultrasound Technique, and BI-RADS for the Differentiation of Benign and Malignant Breast Masses. Front Oncol 2022;12:906501. [Crossref] [PubMed]
10. Behrens S, Laue H, Althaus M, Boehler T, Kuemmerlen B, Hahn HK, Peitgen HO. Computer assistance for MR based diagnosis of breast cancer: present and future challenges. Comput Med Imaging Graph 2007;31:236-47. [Crossref] [PubMed]
11. Boyd NF, Guo H, Martin LJ, Sun L, Stone J, Fishell E, Jong RA, Hislop G, Chiarelli A, Minkin S, Yaffe MJ. Mammographic density and the risk and detection of breast cancer. N Engl J Med 2007;356:227-36. [Crossref] [PubMed]
12. Kerlikowske K, Zhu W, Tosteson AN, Sprague BL, Tice JA, Lehman CD, Miglioretti DLBreast Cancer Surveillance Consortium. Identifying women with dense breasts at high risk for interval cancer: a cohort study. Ann Intern Med 2015;162:673-81. [Crossref] [PubMed]
13. So A, Nicolaou S. Spectral Computed Tomography: Fundamental Principles and Recent Developments. Korean J Radiol 2021;22:86-96. [Crossref] [PubMed]
14. Hua CH, Shapira N, Merchant TE, Klahr P, Yagil Y. Accuracy of electron density, effective atomic number, and iodine concentration determination with a dual-layer dual-energy computed tomography system. Med Phys 2018;45:2486-97. [Crossref] [PubMed]
15. Sodickson AD, Keraliya A, Czakowski B, Primak A, Wortman J, Uyeda JW. Dual energy CT in clinical routine: how it works and how it adds value. Emerg Radiol 2021;28:103-17. [Crossref] [PubMed]
16. Wortman JR, Shyu JY, Dileo J, Uyeda JW, Sodickson AD. Dual-energy CT for routine imaging of the abdomen and pelvis: radiation dose and image quality. Emerg Radiol 2020;27:45-50. [Crossref] [PubMed]
17. Große Hokamp N, Neuhaus V, Abdullayev N, Laukamp K, Lennartz S, Mpotsaris A, Borggrefe J. Reduction of artifacts caused by orthopedic hardware in the spine in spectral detector CT examinations using virtual monoenergetic image reconstructions and metal-artifact-reduction algorithms. Skeletal Radiol 2018;47:195-201. [Crossref] [PubMed]
18. Chae HD, Hong SH, Shin M, Choi JY, Yoo HJ. Combined use of virtual monochromatic images and projection-based metal artifact reduction methods in evaluation of total knee arthroplasty. Eur Radiol 2020;30:5298-307. [Crossref] [PubMed]
19. Yoo HJ, Hong SH, Chung BM, Moon SJ, Choi JY, Chae HD, Chang MY. Metal Artifact Reduction in Virtual Monoenergetic Spectral Dual-Energy CT of Patients With Metallic Orthopedic Implants in the Distal Radius. AJR Am J Roentgenol 2018;211:1083-91. [Crossref] [PubMed]
20. Barbara Krug K, Schömig-Markiefka B, Campbell GM, Püsken M, Maintz D, Schlamann M, Klein K, Gabriel Schafigh D, Malter W, Hellmich M. Correlation of CT-data derived from multiparametric dual-layer CT-maps with immunohistochemical biomarkers in invasive breast carcinomas. Eur J Radiol 2022;156:110544. [Crossref] [PubMed]
21. Demirler Şimşir B, Krug KB, Burke C, Hellmich M, Maintz D, Coche E. Possibility to discriminate benign from malignant breast lesions detected on dual-layer spectral CT-evaluation. Eur J Radiol 2021;142:109832. [Crossref] [PubMed]
22. Wang X, Liu D, Zeng X, Jiang S, Li L, Yu T, Zhang J. Dual-energy CT quantitative parameters for the differentiation of benign from malignant lesions and the prediction of histopathological and molecular subtypes in breast cancer. Quant Imaging Med Surg 2021;11:1946-57. [Crossref] [PubMed]
23. Volterrani L, Gentili F, Fausto A, Pelini V, Megha T, Sardanelli F, Mazzei MA. Dual-Energy CT for Locoregional Staging of Breast Cancer: Preliminary Results. AJR Am J Roentgenol 2020;214:707-14. [Crossref] [PubMed]
24. Lan X, Wang X, Qi J, Chen H, Zeng X, Shi J, Liu D, Shen H, Zhang J. Application of machine learning with multiparametric dual-energy computed tomography of the breast to differentiate between benign and malignant lesions. Quant Imaging Med Surg 2022;12:810-22. [Crossref] [PubMed]
25. Zheng X, Ma Y, Xiong S, Chen X, Zheng W, Cui Y, Huang G. Clinical value of dual-layer spectral detector CT in distinguishing diagnosis of pulmonary primary malignant tumor, chronic inflammation and tuberculosis. Jie Fang Jun Yi Xue Za Zhi 2022;47:1133-43.
26. Jia Y, Xiao X, Sun Q, Jiang H. CT spectral parameters and serum tumour markers to differentiate histological types of cancer histology. Clin Radiol 2018;73:1033-40. [Crossref] [PubMed]
27. Metin Y, Metin NO, Özdemir O, Taşçı F, Kul S. The role of low keV virtual monochromatic imaging in increasing the conspicuity of primary breast cancer in dual-energy spectral thoracic CT examination for staging purposes. Acta Radiol 2020;61:168-74. [Crossref] [PubMed]
28. Okada K, Matsuda M, Tsuda T, Kido T, Murata A, Nishiyama H, Nishiyama K, Yamasawa H, Kamei Y, Kurata M, Fukushima M, Kitazawa R, Mochizuki T. Dual-energy computed tomography for evaluation of breast cancer: value of virtual monoenergetic images reconstructed with a noise-reduced monoenergetic reconstruction algorithm. Jpn J Radiol 2020;38:154-64. [Crossref] [PubMed]
29. Martin SS, Albrecht MH, Wichmann JL, Hüsers K, Scholtz JE, Booz C, Bodelle B, Bauer RW, Metzger SC, Vogl TJ, Lehnert T. Value of a noise-optimized virtual monoenergetic reconstruction technique in dual-energy CT for planning of transcatheter aortic valve replacement. Eur Radiol 2017;27:705-14. [Crossref] [PubMed]
30. Li JX, Xie FJ, Chen CH, Chen KM, Tsai CJ. Dual-Energy Computed Tomography for Evaluation of Breast Cancer Follow-Ups: Comparison of Virtual Monoenergetic Images and Iodine-Map. Diagnostics (Basel) 2022.
31. Liao YS, Zhang JY, Hsu YC, Hong MX, Lee LW. Age-Specific Breast Density Changes in Taiwanese Women: A Cross-Sectional Study. Int J Environ Res Public Health 2020;
32. Britt KL, Cuzick J, Phillips KA. Key steps for effective breast cancer prevention. Nat Rev Cancer 2020;20:417-36. [Crossref] [PubMed]
33. Torre LA, Bray F, Siegel RL, Ferlay J, Lortet-Tieulent J, Jemal A. Global cancer statistics, 2012. CA Cancer J Clin 2015;65:87-108. [Crossref] [PubMed]
34. Zhang Q, Gao Q, Cui W, Guo L, Wang D, Zhang H. The value of the nomogram model for predicting benign and malignant pulmonary solid nodules based on spectral CT quantitative parameters. Journal of Clinical Radiology 2022;41:2215-8.
35. Han M, Kim TH, Kang DK, Kim KS, Yim H. Prognostic role of MRI enhancement features in patients with breast cancer: value of adjacent vessel sign and increased ipsilateral whole-breast vascularity. AJR Am J Roentgenol 2012;199:921-8. [Crossref] [PubMed]
36. Rong XC, Kang YH, Shi GF, Ren JL, Liu YH, Li ZG, Yang G. The use of mammography-based radiomics nomograms for the preoperative prediction of the histological grade of invasive ductal carcinoma. J Cancer Res Clin Oncol 2023;149:11635-45. [Crossref] [PubMed]
37. Stavros AT, Thickman D, Rapp CL, Dennis MA, Parker SH, Sisney GA. Solid breast nodules: use of sonography to distinguish between benign and malignant lesions. Radiology 1995;196:123-34. [Crossref] [PubMed]