Artificial intelligence in medical education for pulmonary nodule management: a narrative review

Introduction

Lung cancer remains the most common malignant tumor worldwide, accounting for the highest incidence and mortality rates among all cancers (1). The early detection of pulmonary nodules through systematic screening and subsequent diagnostic evaluation play a pivotal role in improving patient prognosis. Low-dose computed tomography (LDCT) serves as a critical screening tool for early detection of pulmonary nodules due to its high sensitivity (2). The diagnostic specificity of LDCT screening for lung cancer ranges widely from 26.4% to 99.7%, primarily due to differences in nodule size, composition, and imaging protocols. In parallel, false-positive rates—defined as any outcome prompting further evaluation without a final cancer diagnosis—range from 0.6% to 49.3%, depending on the criteria used to define a positive finding, including size thresholds, volume doubling time (VDT), and follow-up protocols (3). For high-risk populations identified by initial LDCT screening, intermittent LDCT follow-up monitoring is typically recommended. For larger nodules (≥15 mm), immediate further examinations such as contrast-enhanced computed tomography (CT) and flurodeoxyglucose positron emission tomography-computed tomography (¹⁸FDG-PET/CT) are required, with transthoracic needle biopsy performed when necessary (4). However, current diagnostic techniques have significant limitations: PET-CT demonstrates no significant advantage in the differential diagnosis of ground-glass opacity (GGO) nodules and pulmonary nodules with solid components ≤8 mm, while its sensitivity and specificity in distinguishing benign from malignant solitary pulmonary nodules >8 mm are also suboptimal (5,6); the diagnosis rate of endobronchial ultrasound-guided transbronchial needle aspiration for malignant diseases was 0.79 [95% confidence interval (CI): 0.72–0.88] (7). Therefore, to meet clinical needs and improve health economic benefits, there is an urgent need to develop novel, non-invasive, and complementary pulmonary nodule management strategies. The ideal management protocol should provide timely and effective intervention at the early stage of lung cancer while minimizing overtreatment of benign nodules.

In recent years, artificial intelligence (AI) has rapidly emerged and been applied to various medical scenarios, demonstrating significant advantages in healthcare data processing. With the improvement of computational power and the advent of powerful graphics processing units (GPUs), traditional machine learning and its extended branch—deep learning—have enabled efficient analysis of large-scale medical imaging datasets, leading to the widespread application of AI technology in the early detection and diagnosis of pulmonary nodules (8). This technological advancement is particularly evident in areas such as early identification, automated nodule detection, characterization of benign and malignant features, and prediction of pathological subtypes. Insights gained from AI technology can provide valuable references for clinical decision-making, facilitating the development of more effective treatment and follow-up strategies. Notably, AI technology can even extend to the prediction of patient outcomes, thereby supporting more comprehensive and precise healthcare decision-making processes.

However, integrating AI into clinical practice requires a paradigm shift in medical education. As AI becomes increasingly embedded in healthcare, medical professionals must acquire essential skills, including data literacy, the ability to interpret AI systems, and collaborative skills to work alongside intelligent systems. Simultaneously, medical education needs to keep pace with the rapid advancement of technology by incorporating AI-related content into new educational frameworks, fostering critical thinking, and emphasizing the ethical implications of AI in medicine. This article systematically summarizes the role of AI in the diagnosis and treatment of pulmonary nodules, explores the challenges and opportunities it brings to medical education, and proposes strategies for training the next generation of physicians. We present this article in accordance with the Narrative Review reporting checklist (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-313/rc).

Methods

The PubMed and Google Scholar databases were mainly searched in March 2025 with the following search terms to identify papers of interest: “Pulmonary nodule detection and identification”, “AI-assisted pulmonary nodule diagnosis”, “AI-aids medical teaching of pulmonary nodule diagnosis”, “AI-assisted medical imaging teaching interpretation”, and “AI-supported clinical medical teaching”. We reviewed meta-analyses, original articles, and review articles. We rigorously selected literature published between 2015 and 2025 to ensure the inclusion of up-to-date evidence supporting our conclusions. Moreover, publications not written in English were excluded. The search strategy is summarized in Table 1.

Applications of AI technology in pulmonary nodule management

AI-assisted identification and detection of pulmonary nodule

AI demonstrates significant advantages in assisting physicians with pulmonary nodule identification, with its high sensitivity and rapid image-reading capabilities making it highly valuable in lung cancer screening. In a study on baseline lung cancer screening using LDCT, AI demonstrated a markedly lower rate of negative misclassifications compared to radiologists (0.8% vs. 11.1%) and substantially reduced the referral risk associated with missed diagnoses from 11.8% to 2.9% (9). Even early commercial AI systems developed for chest X-ray interpretation achieved diagnostic accuracy comparable to that of board-certified radiologists, with some algorithms demonstrating superior performance [area under the receiver operating characteristic curve (AUC): 0.86–0.93; P value range:<0.01 to 0.40] (10). Furthermore, deep learning-based computer-aided diagnosis (CAD) systems—including architectures such as Faster Region-based convolutional neural networks (CNNs), Single Shot MultiBox Detector (SSD), and You Only Look Once version 3 (YOLOv3)—exhibited remarkable sensitivity in pulmonary nodule detection on chest CT scans, with true positive rates consistently exceeding 90% (11). However, the implementation of AI-assisted diagnostic systems in pulmonary nodule detection has been associated with reduced diagnostic specificity, potentially leading to a substantial increase in unnecessary CT follow-up examinations. Epidemiological projections estimate that this could result in approximately 59,700 to 79,600 additional patients per million population undergoing potentially avoidable diagnostic procedures, many of whom may ultimately be diagnosed as non-cancer cases (12).

Moreover, AI exhibits clear advantages over manual measurements in calculating pulmonary nodule volume and estimating VDT, particularly for small nodules with a maximum diameter of <10 mm (13,14). And the recognition model of inert lung nodule (VDT >600 days) established based on deep learning research of neural network has a diagnostic accuracy of 77.50–81.13% (AUC 0.77) (15).

AI-assisted differential diagnosis of benign and malignant pulmonary nodule

AI-assisted pulmonary nodule classification technology has evolved from traditional computer vision algorithms to deep learning (CNN and transfer learning), and further to multi-view knowledge fusion and end-to-end automatic learning, becoming a powerful decision-support tool in clinical auxiliary diagnosis. Early CAD methods primarily relied on handcrafted feature extraction techniques such as morphological features, gray-level co-occurrence matrix (GLCM), and local binary pattern (LBP), combined with support vector machine (SVM) or random forest (RF) for classification. However, these methods were limited by their feature representation capabilities and generalizability (16). With the advancement of deep learning technologies, CNN have been widely adopted. Researchers have utilized models such as ResNet, VGG, or AlexNet for transfer learning, combined with 3D CNN or multi-view CNN for feature extraction and classification, significantly improving diagnostic accuracy (17). A study applied a CAD system to 1,000 LDCT images from physical examinations (LNPE1000 dataset). By employing 3D CNN, 3D ResNet, and Fully Connected Neural Network, the system achieved automated and efficient detection of pulmonary nodules, particularly those measuring 3–6 mm, with an accuracy of 0.879—comparable to that of expert radiologists. Furthermore, the system demonstrated an accuracy of 0.911 in differentiating intrapulmonary from pleural nodules, and a precision of 0.950 in distinguishing GGO from non-GGO nodules (18). The ensemble learning approach, which integrates 2D CNN (ResNet-50) and 3D CNN (Inception-v1), demonstrates superior performance in assessing the malignancy risk of pulmonary nodules under LDCT screening conditions compared to the PanCan statistical model (AUC =0.93 vs. 0.90). In certain cases, this approach achieves diagnostic accuracy comparable to that of radiologists (19). Currently, AI diagnostic technology is advancing toward end-to-end automated learning, integrating advanced techniques such as reinforcement learning and Transformer architectures to enhance multi-modal data fusion and intelligent diagnostic capabilities. In the future, AI technology is expected to further integrate large-scale pre-trained models to improve diagnostic accuracy and clinical interpretability (20).

For nodules diagnosed as malignant, AI can leverage deep learning and radiomics approaches to high-throughput extract multi-modal data from imaging, pathology, and molecular levels, enabling the differentiation of growth patterns among various histological subtypes, precise identification of pathological subtypes and staging, and facilitating the selection of treatment modalities (21,22). Taking common clinical cases such as lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC) as examples, a comprehensive predictive model incorporating tumor markers, ¹⁸FDG/PET-CT and CT radiomic features, established based on deep learning networks, has achieved an AUC of 0.901 (95% CI: 0.840–0.957) (23). Furthermore, deep learning has made it possible for the noninvasive identification of clinicopathologic subtypes (small cell lung cancer, LUAD, and LUSC), with a deep CNN-based model achieving a sensitivity of 0.90 and specificity of 0.44 on CT images from 301 patients (24). A deep learning model for evaluating visceral pleural invasion in early-stage LUAD demonstrated performance comparable to that of radiologists in terms of AUC (AUC =0.75, 95% CI: 0.67–0.84) and outperformed radiologists in partial AUC assessments within the high-specificity range (90–100%) (P<0.05) (25). The continued advancement of AI holds promises for revolutionizing the precision diagnosis and treatment of lung cancer, enhancing the accuracy of patient management.

AI-assisted personalized treatment of pulmonary nodule

AI is capable of extracting critical features from multi-modal data, including tumor micro-morphology, genetic mutations, and pathological information, to predict patient treatment responses for precision therapies such as targeted therapy and immunotherapy, thereby enabling personalized treatment strategies. Based on CT imaging, a radiomics-based machine learning model has been developed to predict the EGFR T790M mutation in non-small cell lung cancer (NSCLC) patients at diagnosis, achieving an AUC of 0.80 (95% CI: 0.79–0.81) with radiomics features alone, and 0.86 (95% CI: 0.87–0.88) when combined with clinical features (26). For predicting anaplastic lymphoma kinase (ALK) gene fusion in NSCLC patients, the model’s AUC further improves to 0.8481, and ALK-positive patients exhibit significantly better progression-free survival (16.8 vs. 7.5 months, P=0.01) (27).

Moreover, AI enables non-invasive prediction of key immunotherapy biomarkers, including programmed death ligand 1 (PD-L1) expression levels and tumor mutation burden, and accurately distinguishes between high and low treatment responders through imaging features, thereby aiding immunotherapy decision-making. AI models assist in NSCLC pathological analysis, improving the consistency of TPS scoring among pathologists from 81.4% to 90.2% (P<0.001), reducing subjective variability in PD-L1 assessment and enhancing the precision of treatment decision-making (28). In the context of targeted therapy and immunotherapy, AI radiomics models not only predict EGFR-TKI efficacy but also utilize whole-lung CT imaging features to assess the risk of treatment resistance (29,30). For predicting the efficacy of EGFR-TKI and immune checkpoint inhibitor, an AI prognostic model based on EfficientNet has achieved performance comparable to that of radiologists, highlighting its potential to enhance diagnostic and therapeutic capabilities in primary healthcare settings (31).

Challenges and opportunities of AI in medical education for pulmonary nodule management

With the groundbreaking advancements of AI technology in medical imaging analysis, disease prediction, and clinical decision support, medical education is gradually transitioning from traditional models to data-driven, personalized, and intelligent approaches. The integration of AI has not only significantly improved the accuracy and efficiency of clinical diagnosis and decision-making but also generated a demand for new medical skills such as data management, machine learning, and human-computer interaction. Furthermore, AI has facilitated the deep integration of medicine with disciplines such as computer science, engineering, and ethics, providing vast opportunities for the development of cutting-edge applications, including intelligent diagnostic devices, virtual assistants, and clinical decision support systems.

However, the rapid application of AI in medicine has not been without challenges. In chest CT screening, AI assistance may improve sensitivity but tends to reclassify more nodules into high-risk categories, thereby increasing false-positive rates and leading to unnecessary follow-up (12). Although AI diagnostic systems have received relatively high usability scores (with an average “good” system usability scale score of 74.3±11.9), improving the efficiency and diagnostic accuracy of radiologists, the learning curve is steep. Senior physicians with over 10 years of experience tend to be more cautious about using AI, indicating a need to optimize the AI learning experience and strengthen related training (32).

These challenges reflect broader systemic issues in medical education. Through rapid processing of large medical datasets, AI has fostered a new collaborative model between clinicians and intelligent systems. To adapt, medical education must integrate AI-focused curricula and develop interdisciplinary frameworks that equip physicians for AI-driven healthcare (33). Key priorities include standardizing curriculum content and defining core competencies, as current teaching approaches remain fragmented and inconsistent (34-36). Additionally, limited access to qualified instructors, annotated datasets, computational tools, and simulation platforms continues to hinder effective AI-based instruction (37,38). As AI becomes more central to clinical decision-making, it also challenges traditional physician roles, raising concerns about autonomy and professional identity (39). Therefore, future clinicians must be trained to collaborate with AI systems, critically evaluate their outputs, and maintain empathy, ethical judgment, and patient-centered care (40).

Talent development strategies for pulmonary nodule management in the AI era

Training objectives and core competencies

The cultivation of diagnostic talents in the AI era should focus on developing interdisciplinary professionals with expertise in both medicine and AI. Training should equip physicians with core competencies in both clinical medicine and data science to understand the principles and limitations of AI in diagnostics. Learners should be able to integrate AI tools with clinical judgment, critically interpret AI outputs, and uphold humanistic care and patient trust in technology-assisted decision-making. Specifically, the training program should emphasize the following core competencies:

• Firstly, AI literacy and data literacy, including understanding the basic principles and performance metrics of commonly used diagnostic AI algorithms, as well as the ability to identify potential biases and errors (41-43).
• Second, comprehensive diagnostic thinking skills, which emphasize integrating AI-generated insights with human medical knowledge and cultivating the ability to extract critical diagnostic information from vast amounts of data (44,45).
• Third, humanistic care and ethical decision-making abilities, which involve upholding professional integrity in a highly technical medical environment, emphasizing doctor-patient communication, empathy, and thoughtful consideration of ethical issues in AI-assisted diagnostics (36).

It is imperative to ensure that trainees fully comprehend that even in the era of AI, physicians remain ultimately responsible for clinical decisions and must uphold their enduring role as compassionate guardians of care.

Curriculum reform integrating AI

With rapid AI advancements, there is growing support for urgent curriculum reforms that integrate AI-related content into medical training, equipping physicians for complex clinical decision-making. To this end, medical curricula should incorporate data science, machine learning, and other relevant topics to help students develop a profound understanding of the foundational theories and application scenarios of AI (46). Specific measures may include adding courses in medical informatics or biostatistics during the basic phase, along with clinical-phase electives on medical AI that introduce advancements such as AI-assisted imaging and intelligent diagnostic support. Assessment methods should also be adjusted accordingly, emphasizing students’ ability to apply real-world data and intelligent systems, and shifting medical training from memorization-based learning to competency-based approaches that prioritize information integration and utilization over textbook recall (47).

Interdisciplinary collaboration and training

In the context of integrating AI into medical practice, interdisciplinary training is crucial for developing “Medicine + AI” composite talents in diagnostics. Internally within medicine, digital education and AI-assisted tools foster in-depth interactions among respiratory medicine, thoracic surgery, oncology, radiology, pathology, immunology, and other disciplines, enhancing collaborative diagnostic capabilities (48). Additionally, incorporating AI into medical education curricula helps students grasp both the strengths and limitations of AI technologies. Specific strategies include introducing courses on data science and machine learning, initiating research projects that combine medical and engineering expertise, providing practical engineering experiences for medical students, and encouraging participation in interdisciplinary exchange programs (44). Furthermore, specialized training programs that integrate medicine with engineering, requiring students to have solid undergraduate backgrounds in mathematics and computer science, provide systematic dual-degree (MD and PhD in engineering) education (49). This approach embeds engineering perspectives and innovation into medical training, ultimately cultivating diagnosticians with broad horizons and diverse competencies.

Strengthening practical training and AI application

Practical education should leverage AI technologies to enhance trainees’ diagnostic reasoning and skill proficiency. AI can significantly contribute to practice scenarios, effectively developing diagnostic reasoning and clinical skills (50). Currently, AI-driven simulation systems generate diverse medical imaging cases, enabling students to practice recognizing various pathological signs, thus refining their diagnostic accuracy and competence (51). The introduction of AI-assisted diagnostic systems in teaching hospitals has been shown to significantly improve the accuracy of pulmonary nodule detection among medical imaging students (by approximately 15–20%), effectively compensating for the lack of experience among beginners and bringing their diagnostic performance closer to that of resident physicians (52). In standardized residency training programs, the integration of intelligent tools such as clinical decision-support AI and medical imaging diagnostic AI into simulated real-world clinical scenarios has become an important trend in modern medical education. However, a previous study has revealed a decline in diagnostic accuracy among students after the removal of AI assistance, indicating that AI is better suited as a supplementary tool for beginners to accumulate clinical experience rather than a replacement for traditional teaching methods (53). Particularly in advanced learning stages, the timely correction of diagnostic errors and the provision of systematic guidance still rely on experienced clinical teaching teams and standardized training feedback mechanisms (54). Therefore, it is recommended to adopt a hybrid “AI-assisted + traditional teaching” model in medical education to leverage the technical advantages of AI while fostering clinical reasoning, thereby optimizing educational outcomes.

To enhance AI-integrated training, educational tools should provide immediate feedback to help trainees address knowledge gaps, while also cultivating their ability to critically interpret and, when necessary, question AI-generated outputs. Such approaches foster trainees who can collaborate effectively with AI while maintaining independent clinical judgment.

Enhancing faculty development and continuing education

In promoting the reforms and innovations mentioned above, faculty development and continuing education for existing healthcare professionals are equally critical (55). Medical educators must acquire a comprehensive understanding of the advancements in AI technology and its applications in the medical field to effectively integrate the latest developments into classroom teaching practices (56,57). Universities and teaching hospitals should also provide faculty with training and professional development opportunities, such as participation in medical AI seminars, workshops, or visits to relevant research institutions, to ensure that educators continuously update their knowledge base.

Moreover, a systematic continuing education mechanism should be established for practicing clinical professionals to enhance their learning efficiency and skill levels through personalized online learning, virtual experimental teaching, and intelligent assessment. It is incumbent upon medical institutions and professional associations to organize continuing education courses, workshops and online training programs on a regular basis in order to assist physicians in continuously improving their competencies in relation to AI (58). To illustrate this point, it should be noted that certain radiology residency programs have introduced an AI fundamentals training camp into their regular curriculum, which is comprised of lectures, reviews of the literature, and programming practices, with a view to deepening trainees’ understanding of AI (59).

Conclusions

The application of AI in the diagnosis and treatment of pulmonary nodules is progressively transitioning from technological exploration to clinical practice. Leveraging deep learning and radiomics, AI has markedly enhanced the efficiency and accuracy of early detection of pulmonary nodules. However, significant challenges persist, including issues related to limited generalizability due to dependence on extensive annotated datasets, difficulties in detecting subtle early-stage lesions, and requirements for substantial data quality (60). Furthermore, AI-driven precise pathological subtype predictions, which hold potential to significantly inform surgical decisions, still require validation through prospective multicenter clinical trials.

The deep integration of AI into medicine has also presented new challenges to medical education. Traditional curricula need to evolve to integrate AI technologies comprehensively, fostering interdisciplinary expertise and human-AI collaboration skills (61). To address these demands, educational programs should incorporate AI-related competencies, including data science and machine learning, and emphasize the ethical implications and limitations inherent in AI-assisted diagnosis. This approach not only enhances clinicians’ efficiency but also preserves their capability for independent decision-making and humanistic patient care.

Future research should prioritize multimodal data integration algorithms, such as combining radiomics with deep learning techniques, to improve diagnostic accuracy. Additionally, overcoming the inherent clinical limitations of AI necessitates greater educational emphasis on the cooperative dynamics between humans and AI, ensuring clinicians retain critical judgment and empathetic care. Ultimately, optimized AI integration promises to advance diagnostic and therapeutic strategies, benefiting a broader patient population.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the Narrative Review reporting checklist. Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-313/rc

Peer Review File: Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-313/prf

Funding: This work was supported by the Beijing Municipal Natural Science Foundation Project (No. L248072) and Peking Union Medical College Hospital Outstanding Young Talent Development Program (No. UBJ10528).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-313/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, Laversanne M, Sung H, et al. 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. National Lung Screening Trial Research Team. Reduced lung-cancer mortality with low-dose computed tomographic screening. N Engl J Med 2011;365:395-409. [Crossref] [PubMed]
3. Jonas DE, Reuland DS, Reddy SM, et al. Screening for Lung Cancer With Low-Dose Computed Tomography: Updated Evidence Report and Systematic Review for the US Preventive Services Task Force. JAMA 2021;325:971-87. [Crossref] [PubMed]
4. Ettinger D, Wood D, Aisner D, et al. editors. NCCN guidelines for non-small cell lung cancer. Plymouth Meeting. PA: National Comprehensive Cancer Network; 2020.
5. Chun EJ, Lee HJ, Kang WJ, et al. Differentiation between malignancy and inflammation in pulmonary ground-glass nodules: The feasibility of integrated (18)F-FDG PET/CT. Lung Cancer 2009;65:180-6. [Crossref] [PubMed]
6. Gilbert FJ, Harris S, Miles KA, et al. Comparative accuracy and cost-effectiveness of dynamic contrast-enhanced CT and positron emission tomography in the characterisation of solitary pulmonary nodules. Thorax 2022;77:988-96. [Crossref] [PubMed]
7. Lou L, Huang X, Tu J, et al. Endobronchial ultrasound-guided transbronchial needle aspiration in peripheral pulmonary lesions: a systematic review and meta-analysis. Clin Exp Metastasis 2023;40:45-52. [Crossref] [PubMed]
8. de Margerie-Mellon C, Chassagnon G. Artificial intelligence: A critical review of applications for lung nodule and lung cancer. Diagn Interv Imaging 2023;104:11-7. [Crossref] [PubMed]
9. Walstra ANH, Lancaster HL, Heuvelmans MA, et al. Feasibility of AI as first reader in the 4-IN-THE-LUNG-RUN lung cancer screening trial: impact on negative-misclassifications and clinical referral rate. Eur J Cancer 2025;216:115214. [Crossref] [PubMed]
10. van Leeuwen KG, Schalekamp S, Rutten MJCM, et al. Comparison of Commercial AI Software Performance for Radiograph Lung Nodule Detection and Bone Age Prediction. Radiology 2024;310:e230981. [Crossref] [PubMed]
11. Ahmed I, Chehri A, Jeon G, et al. Automated Pulmonary Nodule Classification and Detection Using Deep Learning Architectures. IEEE/ACM Trans Comput Biol Bioinform 2023;20:2445-56. [Crossref] [PubMed]
12. Geppert J, Asgharzadeh A, Brown A, et al. Software using artificial intelligence for nodule and cancer detection in CT lung cancer screening: systematic review of test accuracy studies. Thorax 2024;79:1040-9. [Crossref] [PubMed]
13. Revel MP, Lefort C, Bissery A, et al. Pulmonary nodules: preliminary experience with three-dimensional evaluation. Radiology 2004;231:459-66. [Crossref] [PubMed]
14. Qi LL, Wu BT, Tang W, et al. Long-term follow-up of persistent pulmonary pure ground-glass nodules with deep learning-assisted nodule segmentation. Eur Radiol 2020;30:744-55. [Crossref] [PubMed]
15. Wang B, Zhang H, Li W, et al. Neural network-based model for evaluating inert nodules and volume doubling time in T1 lung adenocarcinoma: a nested case-control study. Front Oncol 2023;13:1037052. [Crossref] [PubMed]
16. Nishio M, Sugiyama O, Yakami M, et al. Computer-aided diagnosis of lung nodule classification between benign nodule, primary lung cancer, and metastatic lung cancer at different image size using deep convolutional neural network with transfer learning. PLoS One 2018;13:e0200721. [Crossref] [PubMed]
17. Shin HC, Roth HR, Gao M, et al. Deep Convolutional Neural Networks for Computer-Aided Detection: CNN Architectures, Dataset Characteristics and Transfer Learning. IEEE Trans Med Imaging 2016;35:1285-98. [Crossref] [PubMed]
18. Han Y, Qi H, Wang L, et al. Pulmonary nodules detection assistant platform: An effective computer aided system for early pulmonary nodules detection in physical examination. Comput Methods Programs Biomed 2022;217:106680. [Crossref] [PubMed]
19. Venkadesh KV, Setio AAA, Schreuder A, et al. Deep Learning for Malignancy Risk Estimation of Pulmonary Nodules Detected at Low-Dose Screening CT. Radiology 2021;300:438-47. [Crossref] [PubMed]
20. 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]
21. Wu W, Parmar C, Grossmann P, et al. Exploratory Study to Identify Radiomics Classifiers for Lung Cancer Histology. Front Oncol 2016;6:71. [Crossref] [PubMed]
22. He J, Wang B, Tao J, et al. Accurate classification of pulmonary nodules by a combined model of clinical, imaging, and cell-free DNA methylation biomarkers: a model development and external validation study. Lancet Digit Health 2023;5:e647-56. [Crossref] [PubMed]
23. Ren C, Zhang J, Qi M, et al. Machine learning based on clinico-biological features integrated (18)F-FDG PET/CT radiomics for distinguishing squamous cell carcinoma from adenocarcinoma of lung. Eur J Nucl Med Mol Imaging 2021;48:1538-49. [Crossref] [PubMed]
24. Aydın N, Çelik Ö, Aslan AF, et al. Detection of Lung Cancer on Computed Tomography Using Artificial Intelligence Applications Developed by Deep Learning Methods and the Contribution of Deep Learning to the Classification of Lung Carcinoma. Curr Med Imaging 2021;17:1137-41. [Crossref] [PubMed]
25. 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]
26. Lu J, Ji X, Liu X, et al. Machine learning-based radiomics strategy for prediction of acquired EGFR T790M mutation following treatment with EGFR-TKI in NSCLC. Sci Rep 2024;14:446. [Crossref] [PubMed]
27. Song Z, Liu T, Shi L, et al. The deep learning model combining CT image and clinicopathological information for predicting ALK fusion status and response to ALK-TKI therapy in non-small cell lung cancer patients. Eur J Nucl Med Mol Imaging 2021;48:361-71. [Crossref] [PubMed]
28. Choi S, Cho SI, Ma M, et al. Artificial intelligence-powered programmed death ligand 1 analyser reduces interobserver variation in tumour proportion score for non-small cell lung cancer with better prediction of immunotherapy response. Eur J Cancer 2022;170:17-26. [Crossref] [PubMed]
29. Song J, Shi J, Dong D, et al. A New Approach to Predict Progression-free Survival in Stage IV EGFR-mutant NSCLC Patients with EGFR-TKI Therapy. Clin Cancer Res 2018;24:3583-92. [Crossref] [PubMed]
30. Wang S, Yu H, Gan Y, et al. Mining whole-lung information by artificial intelligence for predicting EGFR genotype and targeted therapy response in lung cancer: a multicohort study. Lancet Digit Health 2022;4:e309-19. [Crossref] [PubMed]
31. Xu N, Wang J, Dai G, et al. EfficientNet-Based System for Detecting EGFR-Mutant Status and Predicting Prognosis of Tyrosine Kinase Inhibitors in Patients with NSCLC. J Imaging Inform Med 2024;37:1086-99. [Crossref] [PubMed]
32. Liu W, Wu Y, Zheng Z, et al. Evaluating artificial intelligence's role in lung nodule diagnostics: A survey of radiologists in two pilot tertiary hospitals in China. J Clin Imaging Sci 2024;14:31. [Crossref] [PubMed]
33. Healy WJ, Musani A, Fallaw DJ, et al. Emerging Role of Artificial Intelligence in Academic Pulmonary Medicine. South Med J 2024;117:369-70. [Crossref] [PubMed]
34. Sapci AH, Sapci HA. Artificial Intelligence Education and Tools for Medical and Health Informatics Students: Systematic Review. JMIR Med Educ 2020;6:e19285. [Crossref] [PubMed]
35. Srivastava TK, Waghmare L. Implications of Artificial Intelligence (AI) on Dynamics of Medical Education and Care: A Perspective. Journal of Clinical and Diagnostic Research 2020;14:JI01-JI02.
36. Pinnock R, McDonald J, Ritchie D, et al. Humans and machines: Moving towards a more symbiotic approach to learning clinical reasoning. Med Teach 2020;42:246-51. [Crossref] [PubMed]
37. Rincón EHH, Jimenez D, Aguilar LAC, et al. Mapping the use of artificial intelligence in medical education: a scoping review. BMC Med Educ 2025;25:526. [Crossref] [PubMed]
38. McCoy LG, Nagaraj S, Morgado F, et al. What do medical students actually need to know about artificial intelligence? NPJ Digit Med 2020;3:86. [Crossref] [PubMed]
39. Ackerhans S, Huynh T, Kaiser C, et al. Exploring the role of professional identity in the implementation of clinical decision support systems-a narrative review. Implement Sci 2024;19:11. [Crossref] [PubMed]
40. van der Niet AG, Bleakley A. Where medical education meets artificial intelligence: 'Does technology care?'. Med Educ 2021;55:30-6. [Crossref] [PubMed]
41. Park SH, Do KH, Kim S, et al. What should medical students know about artificial intelligence in medicine? J Educ Eval Health Prof 2019;16:18. [Crossref] [PubMed]
42. Rampton V, Mittelman M, Goldhahn J. Implications of artificial intelligence for medical education. Lancet Digit Health 2020;2:e111-2. [Crossref] [PubMed]
43. Lareyre F, Adam C, Carrier M, et al. Artificial Intelligence for Education of Vascular Surgeons. Eur J Vasc Endovasc Surg 2020;59:870-1. [Crossref] [PubMed]
44. Paranjape K, Schinkel M, Nannan Panday R, et al. Introducing Artificial Intelligence Training in Medical Education. JMIR Med Educ 2019;5:e16048. [Crossref] [PubMed]
45. Mir MM, Mir GM, Raina NT, et al. Application of Artificial Intelligence in Medical Education: Current Scenario and Future Perspectives. J Adv Med Educ Prof 2023;11:133-40. [Crossref] [PubMed]
46. Seth P, Hueppchen N, Miller SD, et al. Data Science as a Core Competency in Undergraduate Medical Education in the Age of Artificial Intelligence in Health Care. JMIR Med Educ 2023;9:e46344. [Crossref] [PubMed]
47. Friederichs H, Friederichs WJ, März M. ChatGPT in medical school: how successful is AI in progress testing? Med Educ Online 2023;28:2220920. [Crossref] [PubMed]
48. van Kooten MJ, Tan CO, Hofmeijer EIS, et al. A framework to integrate artificial intelligence training into radiology residency programs: preparing the future radiologist. Insights Imaging 2024;15:15. [Crossref] [PubMed]
49. Brouillette M. AI added to the curriculum for doctors-to-be. Nat Med 2019;25:1808-9. [Crossref] [PubMed]
50. van de Venter R, Skelton E, Matthew J, et al. Artificial intelligence education for radiographers, an evaluation of a UK postgraduate educational intervention using participatory action research: a pilot study. Insights Imaging 2023;14:25. [Crossref] [PubMed]
51. Borgbjerg J, Thompson JD, Salte IM, et al. Towards AI-augmented radiology education: a web-based application for perception training in chest X-ray nodule detection. Br J Radiol 2023;96:20230299. [Crossref] [PubMed]
52. Wang D, Huai B, Ma X, et al. Application of artificial intelligence-assisted image diagnosis software based on volume data reconstruction technique in medical imaging practice teaching. BMC Med Educ 2024;24:405. [Crossref] [PubMed]
53. Lyu X, Dong L, Fan Z, et al. Artificial intelligence-based graded training of pulmonary nodules for junior radiology residents and medical imaging students. BMC Med Educ 2024;24:740. [Crossref] [PubMed]
54. Chassagnon G, Billet N, Rutten C, et al. Learning from the machine: AI assistance is not an effective learning tool for resident education in chest x-ray interpretation. Eur Radiol 2023;33:8241-50. [Crossref] [PubMed]
55. Kolachalama VB, Garg PS. Machine learning and medical education. NPJ Digit Med 2018;1:54. [Crossref] [PubMed]
56. Narayanan S, Ramakrishnan R, Durairaj E, et al. Artificial Intelligence Revolutionizing the Field of Medical Education. Cureus 2023;15:e49604. [Crossref] [PubMed]
57. Wei H, Qiu J, Yu H, et al. Medco: Medical education copilots based on a multi-agent framework. In: European Conference on Computer Vision. Springer, 2024: 119–35.
58. Zhang X, Jiang L, Liang M, et al. Current Situation and Prospects of Education in Medical Imaging Artificial Intelligence. In: Liu S. editor. Artificial Intelligence in Medical Imaging in China. Singapore: Springer Nature Singapore; 2024:415-29.
59. Hu R, Rizwan A, Hu Z, et al. An Artificial Intelligence Training Workshop for Diagnostic Radiology Residents. Radiol Artif Intell 2023;5:e220170. [Crossref] [PubMed]
60. Tandon YK, Bartholmai BJ, Koo CW. Putting artificial intelligence (AI) on the spot: machine learning evaluation of pulmonary nodules. J Thorac Dis 2020;12:6954-65. [Crossref] [PubMed]
61. Sun L, Yin C, Xu Q, et al. Artificial intelligence for healthcare and medical education: a systematic review. Am J Transl Res 2023;15:4820-8.