Quantitative CT analysis in lung cancer research: a bibliometric analysis based on Web of Science [2005–2025]

Introduction

Lung cancer is the most prevalent malignant tumour in terms of incidence and mortality worldwide (1). The improvement in 5-year survival rates relies heavily on early diagnosis and precision treatment (2). Computed tomography (CT) serves as the core imaging modality for lung cancer screening, diagnosis, and follow-up. Research has demonstrated that the implementation of low-dose CT in screening high-risk populations can significantly reduce lung cancer-related mortality (3,4). However, visual analysis of lung cancer CT images is subject to subjective variation (5). Consequently, research into CT quantitative analysis techniques has attracted considerable attention.

Quantitative CT analysis possesses objective and reproducible characteristics (6), and its application in lung cancer research has yielded substantial results. Traditional quantitative CT parameters provide fundamental yet critical objective indicators for the diagnosis and assessment of lung cancer. With technological advancements, the emerging field of radiomics has emerged through the high-throughput extraction of vast quantitative features (such as texture features) (7). By transforming visual images into quantitatively mineable data, radiomics enables non-invasive assessment of tumour heterogeneity, histological subtypes, and genetic mutation status (8,9). In recent years, leveraging artificial intelligence (AI) and deep learning, quantitative CT analysis has achieved unprecedented levels of automation, precision, and high throughput. Deep learning models, particularly convolutional neural networks, have demonstrated outstanding performance in the automated detection and segmentation of pulmonary nodules, significantly enhancing diagnostic efficiency while reducing human error (10). More remarkably, CT quantitative analysis AI models have demonstrated the potential to surpass human experts in lung cancer diagnosis and risk stratification (11).

However, quantitative CT analysis has advanced so rapidly in the field of lung cancer that the literature has proliferated and the knowledge system has become highly complex. Conventional narrative or systematic reviews, while valuable for synthesizing content and critically appraising study quality, cannot systematically quantify publication patterns, collaboration structures, citation networks, or temporal shifts in research focus. In contrast, bibliometric analysis is specifically designed to address these questions. It can reveal evolutionary trajectories, detect emerging research fronts through burst analysis, map international and institutional collaboration networks, identify influential works, and pinpoint underrepresented research areas (12).

To date, no study has provided a systematic, quantitative, and reproducible overview of the entire landscape of quantitative CT analysis in lung cancer. Therefore, conducting a bibliometric analysis in this field is timely and necessary. Such an analysis will help researchers and clinicians understand the field’s structure, track its development, and identify future directions. In this study, we performed a systematic bibliometric analysis of 1,618 publications on the application of quantitative CT analysis in lung cancer, retrieved from the Web of Science Core Collection (WoSCC) between 2005 and 2025. This study aims to summarize and identify key research areas and developmental trends in quantitative CT analysis for lung cancer, thereby providing a reference for future research. We present this article in accordance with the BIBLIO reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0845/rc).

Methods

Data acquisition and search strategy

We have chosen to use only the WoSCC for our literature search, primarily for the following reasons: WoSCC is the most widely recognised standard database in the field of bibliometrics; it offers high-quality indexing, comprehensive citation data and standardised formatting, thereby ensuring the reproducibility and international comparability of the analysis; This study focuses on global development trajectories and mainstream international trends; however, non-English databases such as CNKI and Wanfang differ significantly from WoSCC in terms of keyword indexing, abstract coverage and citation records. Including them would increase data heterogeneity; furthermore, due to resource constraints, we are currently unable to conduct cross-database comparisons. We acknowledge that this approach may result in the omission of some high-quality non-English literature, and have outlined this limitation in the discussion section.

Since 2005, the volume of literature in this field has increased dramatically. The year 2005 was selected as the starting point because it marks the launch of large-scale low-dose CT lung cancer screening trials (such as the National Lung Screening Trial) and the emergence of early studies on quantitative CT parameters, which laid the foundation for subsequent developments in radiomics and AI. The search cut-off date of 3 October 2025 reflects the date on which the literature search was actually conducted. Furthermore, the 2005–2025 timeframe fully captures the transition of quantitative CT analysis for lung cancer from traditional volumetric measurements to deep learning-based automation, constituting a meaningful and self-contained research phase. Consequently, we extracted publications released between 1 January 2005 and 3 October 2025. The search strategy employed was: (TS = (“Lung Neoplasms” OR “neoplasm* pulmonary” OR “pulmonary neoplasm*” OR “neoplasm* lung” OR “lung neoplasm” OR “lung cancer*” OR “cancer* lung” OR “cancer of lung” OR “pulmonary cancer*” OR “cancer* pulmonary” OR “cancer of the lung”)) AND TS = (“Quantitative CT” OR “Quantitative Computed Tomography” OR “QCT” OR “CT densitometry” OR “Quantitative analysis” OR “Quantitative measurement” OR “Quantitative assessment” OR “Quantitative imaging” OR “3D Quantitative CT”).

Inclusion criteria: articles relevant to the research focus and published in open-access journals. Studies explicitly involving quantitative CT analysis in lung cancer (including but not limited to texture analysis, radiomics, and the application of machine learning or deep learning to CT images) were considered relevant.

Exclusion criteria: conference papers (n=65), conference abstracts (n=27), book chapters (n=9), early online publications (n=9), editorial materials (n=7), letters (n=4), retractions (n=4), corrections (n=3), retraction notices (n=1), and non-English language articles (n=23).

The screening process was as follows: two researchers screened the literature based on search terms, with each reviewing titles, abstracts, keywords, and other relevant information to identify eligible studies. The initial inter-reviewer agreement rate was 95%. In the event of disagreement, a third reviewer was consulted. Figure 1 illustrates the research steps of this study. This process identified 1,770 articles published between 1 January 2005 and 3 October 2025 that met the inclusion criteria. Following application of the exclusion criteria, 1,618 articles were ultimately selected for analysis. All data were downloaded directly from databases; therefore, no ethical declaration or approval was required (Figure 1).

Figure 1 Flowchart of the literature searching and screening in the study.

Data analysis

This study visually presents the current research status in the field of quantitative CT analysis of lung cancer through graphs and tables. We analysed 1,618 included papers using CiteSpace (version 6.3 R1 Basic). For the author collaboration network, the node selection criterion was the g-index (k=20). For keyword clustering and timeline analysis, a 5-year time slice was used. The keyword citation burst detection used the default parameters of CiteSpace (γ=0.5, minimum duration =2 years) to identify emerging research frontiers. The number of national publications, total citations, H-index, author-related citation indicators, author centrality, journal publication volume, and journal citation indicators were extracted and organized into tables from CiteSpace. The H-index measures the quantity and citation impact of research outputs (13), and centrality reflects the bridge position of a node in the network (14). The annual publication volume data were sorted using Microsoft Office Excel 2017 and generated a bar chart to show the trend of publication volume changes; a bubble chart was generated using the bioinformatics platform webpage (https://www.bioinformatics.com.cn/) to show the publication volume of each country. To help readers unfamiliar with bibliometric methods understand the key indicators in this study, Table 1 defines indicators such as publication volume, citation count, average citation per paper, H-index, centrality, and citation burst intensity. Table 1 provides the definitions and academic significance of each indicator.

Results

Analysis of posting volume

Our analysis indicates that from 2005 to 2017, the number of published papers increased significantly, peaking at 119 in 2017 (Figure 2A). From 2017 to 2025, the number of published papers fluctuated before declining. In terms of national publication volumes, the United States and China are the world’s most prolific contributors (Table 1), though their developmental trajectories exhibit marked temporal differences (Figure 2B). The United States commenced its growth in publication volume earlier, peaking in 2017 before experiencing a fluctuating decline, with a substantial reduction by 2025. In contrast, China’s growth in publication volume commenced later but persisted for a longer duration, peaking only in 2023 and maintaining a relatively high level by 2025. We analysed the academic output and impact of different nations (Table 2). The United States leads in both publications and influence, topping the list with 514 papers and 35,507 citations, an H-index of 88, and an average citation count of 69.08 per paper. China ranks second with 441 papers, though its citation count and average citation per paper are relatively lower at 8,710 and 19.76, respectively. Japan and Germany ranked third and fourth with 132 and 108 papers, respectively, receiving 4,475 and 5,203 citations, averaging 33.9 and 48.18 citations per paper. Italy placed fifth with 100 papers, garnering 5,027 citations, averaging 50.27 citations per paper.

Figure 2 Annual publication volume analysis chart. (A) Global annual publication volume and trend chart. (B) Bubble chart of the top 5 countries by publication volume, where circle size and colour denote publication volume. Circles increasing from small to large and colour shifting from yellow to cyan indicate rising publication volume for that country. NP, number of publications.

Author collaboration analysis

Author collaboration analysis indicates that Robert J. Gillies possesses the highest centrality (0.03) and citation count [14]. Binsheng Zhao and James L. Mulshine share identical centrality (0.02), ranking second. Whilst Philippe Lambin exhibits the lowest centrality, his citation count [10] ranks second (Table 3). The author collaboration network diagram reveals that James L. Mulshine, Andrew J. Buckler, and Binsheng Zhao engaged in mutual collaboration, though this occurred relatively early, between 2005 and 2010. Additionally, Robert J. Gillies, Philippe Lambin, and Hugo J. W. L. Aerts also collaborated, with their interactions occurring around 2017 to 2019 (Figure 3).

Figure 3 Author collaboration network diagram. Each node represents an author, with node size indicating the author’s influence. Connecting lines denote collaborations between authors, with colours ranging from purple to red representing the time period from 2005 to 2025. L/N, links per node; LBY, look-back years; LRF, link retaining factor.

Journal publication and citation analysis

Journal analysis indicates that PLoS One published the highest number of articles (n=39), though its average citations per article (39.74) and impact factor (2.6) were comparatively low. Medical Physics and Scientific Reports demonstrated relatively balanced metrics, with publication volume, average citations per article, and impact factor all ranking mid-table. The European Journal of Nuclear Medicine and Molecular Imaging published slightly fewer articles (n=24), yet achieved the highest citation per article (62.88) and the highest impact factor (7.6). Subsequently, Frontiers in Oncology also published 24 articles, with a citation per article of 45.25 and an impact factor of 3.3 (Table 4).

Keyword cluster analysis

Keyword analysis was conducted using CiteSpace, with years per slice set to 5 and scale factor K set to 10. Results indicate that this research domain comprises six primary clusters: #0 quantitative computed tomography, #1 breath analysis, #2 non-small cell lung cancer, #3 CT perfusion, #4 tumour heterogeneity, and #5 subsolid nodule. Research hotspots predominantly emerged between 2005 and 2015, gradually diminishing from 2015 to 2025. Recent key research themes centre on Machine learning and deep learning within cluster #0 quantitative computed tomography, alongside AI within cluster #4 tumour heterogeneity, with 2020 being the most prominent year (Figure 4). Figure 5 displays the 25 keywords exhibiting the strongest citation bursts. Deep learning, machine learning, and AI rank as the top 3 keywords with the most pronounced bursts, also being the newest keywords emerging since 2020. Metastasis, non-small cell lung cancer, open label, and impact likewise represent the newest keywords since 2020, though their citation bursts are comparatively weaker. Additionally, earlier keywords such as tumour heterogeneity, features, and positron emission tomography exhibit strong bursts in citation frequency (Figure 5).

Figure 4 Keyword clustering timeline chart, depicting the time period from 2005 to 2025 from left to right, with identical colours representing the same cluster. L/N, links per node; LBY, look-back years; LRF, link retaining factor.

Figure 5 The 25 most frequently cited keywords. CT, computed tomography; FDG, fluorodeoxyglucose; PET, positron emission tomography.

Tabulation and summarizing the findings

The publication trend from 2005 to 2025 can be divided into two stages: a continuous growth from 2005 to 2017 (with a peak of 119 articles in 2017), and a fluctuating decline from 2017 to 2025, despite the breakthroughs in AI, machine learning, and deep learning. This plateau may reflect a maturation of the field, with research transitioning from method development to validation and clinical translation studies, though this interpretation requires further investigation. The United States and China are the most productive countries, with the peak in the United States occurring earlier (in 2017), and the peak in China occurring later (in 2023). Robert J. Gillies has the highest author centrality. The European Journal of Nuclear Medicine and Molecular Imaging has the highest impact factor and average citation per article. Keyword clustering identified six groups, with recent hotspots focusing on machine/deep learning and AI (peak in 2020). Citation bursts confirmed that “deep learning”, “machine learning”, and “artificial intelligence” are the strongest recent bursts. Regarding Table 5 (existing reviews), three representative reviews were selectively presented to illustrate the key transitions in the field (traditional parameters → radiomics → deep learning); they were identified through targeted searches, and other reviews were excluded due to narrow themes or lack of representativeness.

Synthesis of findings

A bibliometric analysis of 1,618 studies revealed a clear transition of lung cancer CT quantitative analysis through three technological stages: traditional volume measurement [2005–2010], image-based machine learning [2011–2017], and deep learning automation [2018–2025]. This evolution was supported by multiple analytical dimensions. First, the publication trend showed a peak in 2017, coinciding with the transition from the second stage to the third stage. Second, the author collaboration network highlighted the active collaboration of key researchers (such as Gillies, Lambin, Aerts) between 2017 and 2019, which was precisely the period when deep learning began to dominate. Third, the burst analysis of keywords confirmed that “deep learning”, “machine learning”, and “artificial intelligence” were the strongest and most recent burst words, while earlier burst words such as “tumour heterogeneity” and “features” were consistent with the previous stages. Fourth, the journal impact factor indicators showed that high-impact journals (such as the “European Journal of Nuclear Medicine and Molecular Imaging”) were more inclined to publish articles related to radiomics and AI, reflecting the shift in the field towards advanced quantitative methods.

Discussion

This study systematically reviewed research on quantitative CT analysis in lung cancer from 2005 to 2025 using bibliometric methods. As far as we know, this is the first bibliometric study specifically dedicated to quantitative CT analysis of lung cancer. The trend we observed is in line with the overall development trend of medical imaging research or the field of AI in the medical domain. Over the two-decade period, publication volume in this field underwent two distinct phases: a substantial increase followed by a modest decline. Annual publication numbers rose progressively from 2005 to 2017, peaking in 2017 (119 papers). This trend may be attributed to rising lung cancer detection rates and the widespread adoption of CT screening (15,16). From 2017 to 2025, a slight decline was observed, which may be attributed to the field of research entering a new phase characterised by an exploratory period, necessitating further in-depth investigation. Although breakthroughs have been made in AI, machine learning and deep learning, this plateau phase may reflect the maturity of the field. Research is shifting from method development to validation and clinical translation.

Analysis of national publication volumes indicates that China and the United States produce the highest numbers of papers, yet their development trajectories exhibit marked temporal divergence. The United States experienced an earlier rise in publication volume than China, followed by a significant decline after reaching its peak. In contrast, China’s publication volume has only seen a modest decrease, potentially attributable to its large patient population and substantial regional disparities (17,18). The difference in average citations per article between the United States and China (69.08 for the United States vs. 19.76 for China) might be due to the earlier research maturity and foundational contributions of the research teams in the United States. Language and database indexing biases might have affected the citation visibility of some Chinese studies. These differences do not imply that the quality of Chinese research is lower; rather, they highlight complementary advantages. Analysis of author collaboration networks reveals that the collaborative networks spanning 2005–2010 and 2017–2019 comprise distinct networks. The collaboration network featuring Robert J. Gillies, who exhibits the highest centrality, pertains to the 2017–2019 period, indicating that newer research groups collaborate more closely. Analysis of journal publication volumes indicates that PLoS One published the highest number of articles (n=39), though its average citations per article were relatively low (39.74). Conversely, the European Journal of Nuclear Medicine and Molecular Imaging published slightly fewer articles (n=24) but achieved the highest average citations per article (62.88). This suggests that within the field of quantitative CT analysis for lung cancer, a high volume of publications does not necessarily equate to high individual article impact.

Keyword analysis indicates that during the early stages, terms such as tumour heterogeneity, features, and positron emission tomography exhibited the most pronounced surge in citations, signifying that research primarily focused on enhancing CT imaging techniques and quantifying tumour characteristics (19-21). In recent years, with the advancement of AI, the primary research keywords in CT quantitative analysis for lung cancer have shifted towards machine learning and deep learning within the context of quantitative CT, and AI within the context of tumour heterogeneity. This indicates that quantitative analysis utilising machine learning and deep learning, alongside the application of AI to identify tumour heterogeneity in lung cancer, has become a current research focus (22-24). Moreover, metastasis, non-small cell lung cancer, open-label and impact have also been key terms in this field in recent years. This indicates that current research further employs quantitative CT analysis to investigate tumour metastasis (25), whilst ongoing studies are also dedicated to obtaining validation in prospective clinical trials (such as open-label trials) (26). Since 2020, deep learning and AI have become the dominant research direction with profound clinical significance: AI-based quantitative CT analysis has the potential to achieve automatic detection, segmentation and risk stratification of pulmonary nodules. However, clinical translation faces multiple obstacles, including the lack of standardization in CT acquisition and feature extraction, poor generalization of models due to single-centre retrospective data, “black box” interpretability issues, and regulatory barriers. Overcoming these obstacles requires prospective multi-centre trials, unified protocols, and interpretable AI frameworks.

This study has several limitations. Firstly, only the literature indexed in the WoSCC was included, which might have excluded some studies that were not included in WoSCC. It is worth noting that non-English databases [such as China National Knowledge Infrastructure (CNKI), Wanfang] and non-English literature were not retrieved. This introduces language and selection biases, potentially systematically ignoring a large number of Chinese and other non-English studies. These factors may affect the validity of the conclusions regarding international research output and collaboration levels. Secondly, the screening process mainly relied on the literature information and abstracts extracted from WoSCC, and did not read the full texts of all papers, which may lead to the omission of certain information. Thirdly, bibliometric indicators (such as publication volume, citation count, H-index, centrality) reflect academic influence and collaboration networks, but are not equivalent to research quality or clinical evidence strength. High citations may result from self-promotion, controversial findings, or review articles, rather than methodological rigor. Therefore, our results should be interpreted as a mapping of research trends and networks, rather than an assessment of the validity of individual studies.

Conclusions

CT quantitative analysis has been extensively studied in the field of lung cancer, becoming an indispensable component in both diagnosis and treatment. With technological advancements, AI, machine learning, and deep learning have emerged as primary research focuses within this domain. The application of these novel CT quantitative analysis techniques may provide fresh insights for lung cancer diagnosis and treatment, thereby improving patient prognosis. However, despite the rapid progress in methodology, there are still several key limitations. Most studies are retrospective and single-centre; the CT acquisition and feature extraction protocols lack standardization; the external validation of AI models is scarce, severely hindering clinical translation. To promote the development of this field, the following actionable suggestions are proposed: (I) CT scan parameters (such as slice thickness and reconstruction algorithm) and imaging biomarker extraction guidelines should be unified. (II) AI-related research should focus on developing interpretable models to build the trust of clinicians. (III) More prospective multi-centre trials and external validation of existing AI models in diverse populations are needed. Addressing these gaps will accelerate the transformation of quantitative CT analysis towards precision medicine for lung cancer.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the BIBLIO reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0845/rc

Peer Review File: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0845/prf

Funding: This work was supported by grants from Yunnan Fundamental Research Projects (Nos. 202101AY070001-170 and 202301AY070001-085), The Pioneer Incubator Fund of Tumor Anesthesia and Analgesia Committee of China Anti-Cancer Association (No. GWRX-CR-2024-1), and Kunming Medical University Student Innovation and Entrepreneurship Training Program Project (No. JX-2025CYD025).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0845/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.

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, Ferlay J, Soerjomataram I, et al. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2018;68:394-424. [Crossref] [PubMed]
2. Huang S, Yang J, Shen N, et al. Artificial intelligence in lung cancer diagnosis and prognosis: Current application and future perspective. Semin Cancer Biol 2023;89:30-7. [Crossref] [PubMed]
3. Balekian AA, Tanner NT, Fisher JM, et al. Factors Associated with a Positive Baseline Screening Exam Result in the National Lung Screening Trial. Ann Am Thorac Soc 2016;13:1568-74. [Crossref] [PubMed]
4. Reck M, Dettmer S, Kauczor HU, et al. Lung Cancer Screening With Low-Dose Computed Tomography. Dtsch Arztebl Int 2023;120:387-92. [Crossref] [PubMed]
5. Seki N, Eguchi K, Kaneko M, et al. Stage-size relationship in long-term repeated CT screening for lung cancer: Anti-lung cancer association project. J Clin Oncol 2009;27:1540.
6. Koo HJ, Sung YS, Shim WH, et al. Quantitative Computed Tomography Features for Predicting Tumor Recurrence in Patients with Surgically Resected Adenocarcinoma of the Lung. PLoS One 2017;12:e0167955. [Crossref] [PubMed]
7. Thawani R, McLane M, Beig N, et al. Radiomics and radiogenomics in lung cancer: A review for the clinician. Lung Cancer 2018;115:34-41. [Crossref] [PubMed]
8. Avanzo M, Stancanello J, Pirrone G, et al. Radiomics and deep learning in lung cancer. Strahlenther Onkol 2020;196:879-87. [Crossref] [PubMed]
9. Rinaldi L, Guerini Rocco E, Spitaleri G, et al. Association between Contrast-Enhanced Computed Tomography Radiomic Features, Genomic Alterations and Prognosis in Advanced Lung Adenocarcinoma Patients. Cancers (Basel) 2023;15:4553. [Crossref] [PubMed]
10. Zhao W, Ma J, Zhao L, et al. PUNDIT: Pulmonary nodule detection with image category transformation. Med Phys 2023;50:2914-27. [Crossref] [PubMed]
11. Choi H, Kim H, Hong W, et al. Prediction of visceral pleural invasion in lung cancer on CT: deep learning model achieves a radiologist-level performance with adaptive sensitivity and specificity to clinical needs. Eur Radiol 2021;31:2866-76. [Crossref] [PubMed]
12. Greener S. Evaluating literature with bibliometrics. Interactive Learning Environments 2022;30:1168-9.
13. Hirsch JE. An index to quantify an individual's scientific research output. Proc Natl Acad Sci U S A 2005;102:16569-72. [Crossref] [PubMed]
14. Abbasi A, Hossain L, Leydesdorff L. Betweenness centrality as a driver of preferential attachment in the evolution of research collaboration networks. J Informetr 2012;6:403-12.
15. Blandin Knight S, Crosbie PA, Balata H, et al. Progress and prospects of early detection in lung cancer. Open Biol 2017;7:170070. [Crossref] [PubMed]
16. Shlomi D, Ben-Avi R, Balmor GR, et al. Screening for lung cancer: time for large-scale screening by chest computed tomography. Eur Respir J 2014;44:217-38. [Crossref] [PubMed]
17. Liu X, Yang Q, Pan L, et al. Burden of respiratory tract cancers in China and its provinces, 1990-2021: a systematic analysis of the Global Burden of Disease Study 2021. Lancet Reg Health West Pac 2025;55:101485. [Crossref] [PubMed]
18. Cao W, Qin K, Liu B, et al. Cancer statistics in China: Epidemiology, risk factors, and prevention. Chin J Cancer Res 2025;37:912-28. [Crossref] [PubMed]
19. Chicklore S, Goh V, Siddique M, et al. Quantifying tumour heterogeneity in 18F-FDG PET/CT imaging by texture analysis. Eur J Nucl Med Mol Imaging 2013;40:133-40. [Crossref] [PubMed]
20. Wu J, Aguilera T, Shultz D, et al. Early-Stage Non-Small Cell Lung Cancer: Quantitative Imaging Characteristics of (18)F Fluorodeoxyglucose PET/CT Allow Prediction of Distant Metastasis. Radiology 2016;281:270-8. [Crossref] [PubMed]
21. Carvalho S, Leijenaar RTH, Troost EGC, et al. 18F-fluorodeoxyglucose positron-emission tomography (FDG-PET)-Radiomics of metastatic lymph nodes and primary tumor in non-small cell lung cancer (NSCLC) - A prospective externally validated study. PLoS One 2018;13:e0192859. [Crossref] [PubMed]
22. Astaraki M, Yang G, Zakko Y, et al. A Comparative Study of Radiomics and Deep-Learning Based Methods for Pulmonary Nodule Malignancy Prediction in Low Dose CT Images. Front Oncol 2021;11:737368. [Crossref] [PubMed]
23. Han Y, Ma Y, Wu Z, et al. Histologic subtype classification of non-small cell lung cancer using PET/CT images. Eur J Nucl Med Mol Imaging 2021;48:350-60. [Crossref] [PubMed]
24. Hertel A, Streuer A, Diehl S, et al. Targeting tumoral heterogeneity in lung cancer: a novel, CT-texture-guided targeted biopsy approach with exome sequencing. NPJ Precis Oncol 2025;9:342. [Crossref] [PubMed]
25. Bilous M, Serdjebi C, Boyer A, et al. Quantitative mathematical modeling of clinical brain metastasis dynamics in non-small cell lung cancer. Sci Rep 2019;9:13018. [Crossref] [PubMed]
26. Yang H, Chen H, Ni R, et al. Circulating genetically abnormal cells combined with artificial intelligence for accurate and non-invasive early detection on NSCLC. J Clin Oncol 2021;39:3056.