Sybil: A Validated Deep Learning Model to Predict Future Lung Cancer Risk From a Single Low-Dose Chest Computed Tomography

Broadening the Net: Overcoming Challenges and Embracing Novel Technologies in Lung Cancer Screening

Lung cancer is one of the leading causes of cancer-related mortality worldwide, with most cases diagnosed at advanced stages where curative treatment options are limited. Low-dose computed tomography (LDCT) for lung cancer screening (LCS) of individuals selected based on age and smoking history has shown a significant reduction in lung cancer-specific mortality. The number needed to screen to prevent one death from lung cancer is lower than that for breast cancer, cervical cancer, and colorectal cancer. Despite the substantial impact on reducing lung cancer-related mortality and proof that LCS with LDCT is effective, uptake of LCS has been low and LCS eligibility criteria remain imperfect. While LCS programs have historically faced patient recruitment challenges, research suggests that there are novel opportunities to both identify and improve screening for at-risk populations. In this review, we discuss the global obstacles to implementing LCS programs and strategies to overcome barriers in resource-limited settings. We explore successful approaches to promote LCS through robust engagement with community partners. Finally, we examine opportunities to enhance LCS in at-risk populations not captured by current eligibility criteria, including never smokers and individuals with a family history of lung cancer, with a focus on early detection through novel artificial intelligence technologies.

Driving Knowledge to Action: Building a Better Future With Artificial Intelligence-Enabled Multidisciplinary Oncology

Artificial intelligence (AI) is transforming multidisciplinary oncology at an unprecedented pace, redefining how clinicians detect, classify, and treat cancer. From earlier and more accurate diagnoses to personalized treatment planning, AI's impact is evident across radiology, pathology, radiation oncology, and medical oncology. By leveraging vast and diverse data-including imaging, genomic, clinical, and real-world evidence-AI algorithms can uncover complex patterns, accelerate drug discovery, and help identify optimal treatment regimens for each patient. However, realizing the full potential of AI also necessitates addressing concerns regarding data quality, algorithmic bias, explainability, privacy, and regulatory oversight-especially in low- and middle-income countries (LMICs), where disparities in cancer care are particularly pronounced. This study provides a comprehensive overview of how AI is reshaping cancer care, reviews its benefits and challenges, and outlines ethical and policy implications in line with ASCO's 2025 theme, Driving Knowledge to Action. We offer concrete calls to action for clinicians, researchers, industry stakeholders, and policymakers to ensure that AI-driven, patient-centric oncology is accessible, equitable, and sustainable worldwide.

Artificial Intelligence and Machine Learning Innovations to Improve Design and Representativeness in Oncology Clinical Trials

The integration of artificial intelligence (AI) and machine learning (ML) in oncology clinical trials is rapidly evolving alongside the broader field. For example, AI-driven adaptive trial designs may allow for real-time modifications based on emerging safety and efficacy signals, enabling more responsive and efficient trials. AI-powered diagnostic tools, including radiomics, computational pathology, and spatial omics, can improve trial patient selection and response assessments. ML-based patient outcome simulations can similarly enhance patient stratification strategies and statistical power. Application of AI can also improve the accessibility of real-world data, including opportunities to enhance data extraction, standardization, and harmonization of data from routine clinical practice. Data generated from digital health technologies (eg, wearable devices, electronic sensors, computing platforms, software applications) may enable a more comprehensive understanding of patient populations to support clinical trials from enrollment to assessment. Automation of trial operations and data management can also improve data fidelity and decrease investigator burden, which has the potential to streamline trial execution and increase potential use of decentralization. There are ongoing efforts to enhance regulatory clarity, mitigate bias, and uphold ethical use of these novel technologies. In this article, we review use cases of AI and ML in oncology clinical trials, including their role in patient recruitment, trial design and operations, data management, and diagnostics. Although these technologies can have applications across all phases of drug development including early discovery, we focus on phase II and III trials, where AI and ML may have a pronounced ability to enhance trial efficiency, patient stratification, and regulatory decision making. By integrating AI and ML, clinical trials can become more adaptive, data-driven, and inclusive in the pursuit of improving patient outcomes.

China Protocol for early screening, precise diagnosis, and individualized treatment of lung cancer

Early screening, diagnosis, and treatment of lung cancer are pivotal in clinical practice since the tumor stage remains the most dominant factor that affects patient survival. Previous initiatives have tried to develop new tools for decision-making of lung cancer. In this study, we proposed the China Protocol, a complete workflow of lung cancer tailored to the Chinese population, which is implemented by steps including early screening by evaluation of risk factors and three-dimensional thin-layer image reconstruction technique for low-dose computed tomography (Tre-LDCT), accurate diagnosis via artificial intelligence (AI) and novel biomarkers, and individualized treatment through non-invasive molecule visualization strategies. The application of this protocol has improved the early diagnosis and 5-year survival rates of lung cancer in China. The proportion of early-stage (stage I) lung cancer has increased from 46.3% to 65.6%, along with a 5-year survival rate of 90.4%. Moreover, especially for stage IA1 lung cancer, the diagnosis rate has improved from 16% to 27.9%; meanwhile, the 5-year survival rate of this group achieved 97.5%. Thus, here we defined stage IA1 lung cancer, which cohort benefits significantly from early diagnosis and treatment, as the "ultra-early stage lung cancer", aiming to provide an intuitive description for more precise management and survival improvement. In the future, we will promote our findings to multicenter remote areas through medical alliances and mobile health services with the desire to move forward the diagnosis and treatment of lung cancer.

Patient Reactions to Artificial Intelligence-Clinician Discrepancies: Web-Based Randomized Experiment

BACKGROUND

As the US Food and Drug Administration (FDA)-approved use of artificial intelligence (AI) for medical imaging rises, radiologists are increasingly integrating AI into their clinical practices. In lung cancer screening, diagnostic AI offers a second set of eyes with the potential to detect cancer earlier than human radiologists. Despite AI's promise, a potential problem with its integration is the erosion of patient confidence in clinician expertise when there is a discrepancy between the radiologist's and the AI's interpretation of the imaging findings.

OBJECTIVE

We examined how discrepancies between AI-derived recommendations and radiologists' recommendations affect patients' agreement with radiologists' recommendations and satisfaction with their radiologists. We also analyzed how patients' medical maximizing-minimizing preferences moderate these relationships.

METHODS

We conducted a randomized, between-subjects experiment with 1606 US adult participants. Assuming the role of patients, participants imagined undergoing a low-dose computerized tomography scan for lung cancer screening and receiving results and recommendations from (1) a radiologist only, (2) AI and a radiologist in agreement, (3) a radiologist who recommended more testing than AI (ie, radiologist overcalled AI), or (4) a radiologist who recommended less testing than AI (ie, radiologist undercalled AI). Participants rated the radiologist on three criteria: agreement with the radiologist's recommendation, how likely they would be to recommend the radiologist to family and friends, and how good of a provider they perceived the radiologist to be. We measured medical maximizing-minimizing preferences and categorized participants as maximizers (ie, those who seek aggressive intervention), minimizers (ie, those who prefer no or passive intervention), and neutrals (ie, those in the middle).

RESULTS

Participants' agreement with the radiologist's recommendation was significantly lower when the radiologist undercalled AI (mean 4.01, SE 0.07, P<.001) than in the other 3 conditions, with no significant differences among them (radiologist overcalled AI [mean 4.63, SE 0.06], agreed with AI [mean 4.55, SE 0.07], or had no AI [mean 4.57, SE 0.06]). Similarly, participants were least likely to recommend (P<.001) and positively rate (P<.001) the radiologist who undercalled AI, with no significant differences among the other conditions. Maximizers agreed with the radiologist who overcalled AI (β=0.82, SE 0.14; P<.001) and disagreed with the radiologist who undercalled AI (β=-0.47, SE 0.14; P=.001). However, whereas minimizers disagreed with the radiologist who overcalled AI (β=-0.43, SE 0.18, P=.02), they did not significantly agree with the radiologist who undercalled AI (β=0.14, SE 0.17, P=.41).

CONCLUSIONS

Radiologists who recommend less testing than AI may face decreased patient confidence in their expertise, but they may not face this same penalty for giving more aggressive recommendations than AI. Patients' reactions may depend in part on whether their general preferences to maximize or minimize align with the radiologists' recommendations. Future research should test communication strategies for radiologists' disclosure of AI discrepancies to patients.

Histological proven AI performance in the UKLS CT lung cancer screening study: Potential for workload reduction

PURPOSE

Artificial intelligence (AI) could reduce lung cancer screening computer tomography (CT)-reading workload if used as a first-reader, ruling-out negative CT-scans at baseline. Evidence is lacking to support AI performance when compared to gold-standard lung cancer outcomes. This study validated the performance of a commercially available AI software in the UK lung cancer screening (UKLS) trial dataset, with comparison to human reads and histological lung cancer outcomes, and estimated CT-reading workload reduction.

METHODS

1252 UKLS-baseline-CT-scans were evaluated independently by AI and human readers. AI performance was evaluated on two-levels. Firstly, AI classification and individual reads were compared to a EU reference standard (based on NELSON2.0-European Position Statement) determined by a European expert panel blinded from individual results. A positive misclassification was defined as a nodule positive read ≥ 100mm3 and no/<100mm3 nodules in the expert read; A negative misclassification was defined as a nodule negative read, whereas an indeterminate or positive finding in the expert read. Secondly, AI nodule classification was compared to gold-standard histological lung cancer outcomes. CT-reading workload reduction was calculated from AI negative CT-scans when AI was used as first-reader.

RESULTS

Expert panel reference standard reported 815 (65 %) negative and 437 (35 %) indeterminate/positive CT-scans in the dataset of 1252 UKLS-participants. Compared to the reference standard, AI resulted in less misclassification than human reads, NPV 92·0 %(90·2 %-95·3 %). On comparison to gold-standard, AI detected all 31 baseline-round lung cancers, but classified one as negative due to the 100mm3 threshold, NPV 99·8 %(99·0 %-99·9 %). Estimated maximum CT-reading workload reduction was 79 %.

CONCLUSION

Implementing AI as first-reader to rule-out negative CT-scans, shows considerable potential to reduce CT-reading workload and does not lead to missed lung cancers.

Translating premalignant biology to accelerate non-small-cell lung cancer interception

Over the past decade, substantial progress has been made in the development of targeted and immune-based therapies for patients with advanced non-small-cell lung cancer. To further improve outcomes for patients with lung cancer, identifying and intercepting disease at the earliest and most curable stages are crucial next steps. With the recent implementation of low-dose computed tomography scan screening in populations at high risk, there is an emerging unmet need for new diagnostic, prognostic and therapeutic tools to help treat patients suspected of harbouring premalignant lesions and minimally invasive non-small-cell lung cancer. Continued advances in the identification of the earliest drivers of lung carcinogenesis are poised to address these unmet needs. Employing multimodal approaches to chart the temporal and spatial maps of the molecular events driving lung premalignant lesion progression will refine our understanding of early carcinogenesis. Elucidating the molecular drivers of premalignancy is critical to the development of biomarkers to detect those incubating a premalignant lesion, to stratify risk for progression to invasive cancer and to identify novel therapeutic targets to intercept that process. In this Review, we summarize emerging insights into the earliest cellular and molecular events associated with lung squamous and adenocarcinoma carcinogenesis and highlight the growing opportunity for translating these insights into clinical tools for early detection and disease interception to transform the outcomes for those at risk for lung cancer.

Optimizing Strategy for Lung Cancer Screening: From Risk Prediction to Clinical Decision Support

PURPOSE

Low-dose computed tomography (LDCT) screening is effective in reducing lung cancer mortality by detecting the disease at earlier, more treatable stages. However, high false-positive rates and the associated risks of subsequent invasive diagnostic procedures present significant challenges. This study proposes an advanced pipeline that integrates machine learning (ML) and causal inference techniques to optimize lung cancer screening decisions.

MATERIALS AND METHODS

Using real-world data from the OneFlorida+ Clinical Research Consortium, we developed ML models to predict individual lung cancer risk and estimate the benefits of LDCT screening. Explainable artificial intelligence techniques were applied to identify key risk factors, ensuring transparency and trust in the model's predictions. Causal ML methods were used to estimate individualized treatment effects of LDCT screening, answering the critical what-if question regarding risk reduction from LDCT.

RESULTS

We defined a high-risk cohort of 5,947 patients who underwent LDCT, along with matched controls, to evaluate the framework. Our models demonstrated predictive performance with AUCs of 0.777 and 0.793 for 1-year and 3-year risk predictions, respectively. Causal modeling showed a consistent reduction in lung cancer risk across different subgroups due to LDCT. Specifically, the doubly robust model showed an average risk reduction of 9.5% for males and 12% for females. Age-stratified results indicated a 9.5% reduction for individuals age 50-60 years, a 7.5% reduction for those age 60-70 years, and the largest reduction of 15.1% for the 70-80 age group.

CONCLUSION

Integrating ML and causal inference into clinical workflows offers a robust tool for enhancing lung cancer screening. This pipeline provides accurate risk assessments and actionable insights tailored to individuals, empowering clinicians and patients to make informed screening decisions. The differential risk reduction across subgroups highlights the importance of personalized screening in improving outcomes for populations at risk of lung cancer.

Natural history models for lung Cancer: A scoping review

INTRODUCTION

Natural history models (NHMs) of lung cancer (LC) simulate the disease's natural progression providing a baseline for assessing the impact of interventions. NHMs have been increasingly used to inform public health policies, highlighting their utility. The objective of this scoping review was to summarize existing LC NHMs, identify their limitations, and propose a framework for future NHM development.

METHODS

We searched MEDLINE, Embase, Web of Science, and IEEE Xplore from their inception to October 5, 2023, for peer-reviewed, full-length articles with an LC NHM. Model characteristics, their applications, data sources used, and limitations were extracted and narratively synthesized.

RESULTS

From 238 publications, 69 publications were included in our review, corresponding to 22 original LC NHMs and 47 model applications. The majority of the models (n = 15, 68 %) used a microsimulation approach. NHM parameters were predominately informed by cancer registries, trial and institutional data, and literature. Model quality and performance were evaluated in 8 (36 %) models. Twenty (91 %) models included at least one carcinogenesis risk factor-primarily age, sex, and smoking history. Three (14 %) LC NHMs modeled progression in never-smokers; one (5 %) addressed recurrence. Non-tobacco smoking, nodule type, and biomarker expression were not considered in existing NHMs. Based on our findings, we proposed a framework for future LC NHM development which incorporates recurrence, nodule type differentiation, biomarker expression levels, biological factors, and non-smoking-related risk factors.

CONCLUSION

Regular updating and future research are warranted to address limitations in existing NHMs thereby ensuring relevance and accuracy of modeling approaches in the evolving LC landscape.

Significance of Image Reconstruction Parameters for Future Lung Cancer Risk Prediction Using Low-Dose Chest Computed Tomography and the Open-Access Sybil Algorithm

PURPOSE

Sybil is a validated publicly available deep learning-based algorithm that can accurately predict lung cancer risk from a single low-dose computed tomography (LDCT) scan. We aimed to study the effect of image reconstruction parameters and CT scanner manufacturer on Sybil's performance.

MATERIALS AND METHODS

Using LDCTs of a subset of the National Lung Screening Trial participants, which we previously used for internal validation of the Sybil algorithm (test set), we ran the Sybil algorithm on LDCT series pairs matched on kilovoltage peak, milliampere-seconds, reconstruction interval, reconstruction diameter, and either reconstruction filter or axial slice thickness. We also evaluated the cumulative effect of these parameters by combining the best- and the worst-performing parameters. A subanalysis compared Sybil's performance by CT manufacturer. We considered any LDCT positive if future lung cancer was subsequently confirmed by biopsy or surgical resection. The areas under the curve (AUCs) for each series pair were compared using DeLong's test.

RESULTS

There was no difference in Sybil's performance between 1049 pairs of standard versus bone reconstruction filter (AUC at 1 year 0.84 [95% confidence interval (CI): 0.70-0.99] vs 0.86 [95% CI: 0.75-0.98], P = 0.87) and 1961 pairs of standard versus lung reconstruction filter (AUC at 1 year 0.98 [95% CI: 0.97-0.99] vs 0.98 [95% CI: 0.96-0.99], P = 0.81). Similarly, there was no difference in 1288 pairs comparing 2-mm versus 5-mm axial slice thickness (AUC at 1 year 0.98 [95% CI: 0.94-1.00] vs 0.99 [95% CI: 0.97-0.99], P = 0.68). The best-case scenario combining a lung reconstruction filter with 2-mm slice thickness compared with the worst-case scenario combining a bone reconstruction filter with 2.5-mm slice thickness uncovered a significantly different performance at years 2-4 ( P = 0.03). Subanalysis showed no significant difference in performance between Siemens and Toshiba scanners.

CONCLUSIONS

Sybil's predictive performance for future lung cancer risk is robust across different reconstruction filters and axial slice thicknesses, demonstrating its versatility in various imaging settings. Combining favorable reconstruction parameters can significantly enhance predictive ability at years 2-4. The absence of significant differences between Siemens and Toshiba scanners further supports Sybil's versatility.

Lung Cancer Risk Prediction in Patients with Persistent Pulmonary Nodules Using the Brock Model and Sybil Model

BACKGROUND/OBJECTIVES

Persistent pulmonary nodules are at higher risk of developing into lung cancers. Assessing their future cancer risk is essential for successful interception. We evaluated the performance of two risk prediction models for persistent nodules in hospital-based cohorts: the Brock model, based on clinical and radiological characteristics, and the Sybil model, a novel deep learning model for lung cancer risk prediction.

METHODS

Patients with persistent pulmonary nodules-defined as nodules detected on at least two computed tomography (CT) scans, three months apart, without evidence of shrinkage-were included in the retrospective (n = 130) and prospective (n = 301) cohorts. We analyzed the correlations between demographic factors, nodule characteristics, and Brock scores and assessed the performance of both models. We also built machine learning models to refine the risk assessment for our cohort.

RESULTS

In the retrospective cohort, Brock scores ranged from 0% to 85.82%. In the prospective cohort, 62 of 301 patients were diagnosed with lung cancer, displaying higher median Brock scores than those without lung cancer diagnosis (18.65% vs. 4.95%, p < 0.001). Family history, nodule size ≥10 mm, part-solid nodule types, and spiculation were associated with the risks of lung cancer. The Brock model had an AUC of 0.679, and Sybil's AUC was 0.678. We tested five machine learning models, and the logistic regression model achieved the highest AUC at 0.729.

CONCLUSIONS

For patients with persistent pulmonary nodules in real-world cancer hospital-based cohorts, both the Brock and Sybil models had values and limitations for lung cancer risk prediction. Optimizing predictive models in this population is crucial for improving early lung cancer detection and interception.

Artificial intelligence utilization in cancer screening program across ASEAN: a scoping review

BACKGROUND

Cancer remains a significant health challenge in the ASEAN region, highlighting the need for effective screening programs. However, approaches, target demographics, and intervals vary across ASEAN member states, necessitating a comprehensive understanding of these variations to assess program effectiveness. Additionally, while artificial intelligence (AI) holds promise as a tool for cancer screening, its utilization in the ASEAN region is unexplored.

PURPOSE

This study aims to identify and evaluate different cancer screening programs across ASEAN, with a focus on assessing the integration and impact of AI in these programs.

METHODS

A scoping review was conducted using PRISMA-ScR guidelines to provide a comprehensive overview of cancer screening programs and AI usage across ASEAN. Data were collected from government health ministries, official guidelines, literature databases, and relevant documents. The use of AI in cancer screening reviews involved searches through PubMed, Scopus, and Google Scholar with the inclusion criteria of only included studies that utilized data from the ASEAN region from January 2019 to May 2024.

RESULTS

The findings reveal diverse cancer screening approaches in ASEAN. Countries like Myanmar, Laos, Cambodia, Vietnam, Brunei, Philippines, Indonesia and Timor-Leste primarily adopt opportunistic screening, while Singapore, Malaysia, and Thailand focus on organized programs. Cervical cancer screening is widespread, using both opportunistic and organized methods. Fourteen studies were included in the scoping review, covering breast (5 studies), cervical (2 studies), colon (4 studies), hepatic (1 study), lung (1 study), and oral (1 study) cancers. Studies revealed that different stages of AI integration for cancer screening: prospective clinical evaluation (50%), silent trial (36%) and exploratory model development (14%), with promising results in enhancing cancer screening accuracy and efficiency.

CONCLUSION

Cancer screening programs in the ASEAN region require more organized approaches targeting appropriate age groups at regular intervals to meet the WHO's 2030 screening targets. Efforts to integrate AI in Singapore, Malaysia, Vietnam, Thailand, and Indonesia show promise in optimizing screening processes, reducing costs, and improving early detection. AI technology integration enhances cancer identification accuracy during screening, improving early detection and cancer management across the ASEAN region.

A comparison of an integrated and image-only deep learning model for predicting the disappearance of indeterminate pulmonary nodules

BACKGROUND

Indeterminate pulmonary nodules (IPNs) require follow-up CT to assess potential growth; however, benign nodules may disappear. Accurately predicting whether IPNs will resolve is a challenge for radiologists. Therefore, we aim to utilize deep-learning (DL) methods to predict the disappearance of IPNs.

MATERIAL AND METHODS

This retrospective study utilized data from the Dutch-Belgian Randomized Lung Cancer Screening Trial (NELSON) and Imaging in Lifelines (ImaLife) cohort. Participants underwent follow-up CT to determine the evolution of baseline IPNs. The NELSON data was used for model training. External validation was performed in ImaLife. We developed integrated DL-based models that incorporated CT images and demographic data (age, sex, smoking status, and pack years). We compared the performance of integrated methods with those limited to CT images only and calculated sensitivity, specificity, and area under the receiver operating characteristic curve (AUC). From a clinical perspective, ensuring high specificity is critical, as it minimizes false predictions of non-resolving nodules that should be monitored for evolution on follow-up CTs. Feature importance was calculated using SHapley Additive exPlanations (SHAP) values.

RESULTS

The training dataset included 840 IPNs (134 resolving) in 672 participants. The external validation dataset included 111 IPNs (46 resolving) in 65 participants. On the external validation set, the performance of the integrated model (sensitivity, 0.50; 95 % CI, 0.35-0.65; specificity, 0.91; 95 % CI, 0.80-0.96; AUC, 0.82; 95 % CI, 0.74-0.90) was comparable to that solely trained on CT image (sensitivity, 0.41; 95 % CI, 0.27-0.57; specificity, 0.89; 95 % CI, 0.78-0.95; AUC, 0.78; 95 % CI, 0.69-0.86; P = 0.39). The top 10 most important features were all image related.

CONCLUSION

Deep learning-based models can predict the disappearance of IPNs with high specificity. Integrated models using CT scans and clinical data had comparable performance to those using only CT images.

Radiomics for lung cancer diagnosis, management, and future prospects

Lung cancer remains the leading cause of cancer-related mortality worldwide, with its early detection and effective treatment posing significant clinical challenges. Radiomics, the extraction of quantitative features from medical imaging, has emerged as a promising approach for enhancing diagnostic accuracy, predicting treatment responses, and personalising patient care. This review explores the role of radiomics in lung cancer diagnosis and management, with methods ranging from handcrafted radiomics to deep learning techniques that can capture biological intricacies. The key applications are highlighted across various stages of lung cancer care, including nodule detection, histology prediction, and disease staging, where artificial intelligence (AI) models demonstrate superior specificity and sensitivity. The article also examines future directions, emphasising the integration of large language models, explainable AI (XAI), and super-resolution imaging techniques as transformative developments. By merging diverse data sources and incorporating interpretability into AI models, radiomics stands poised to redefine clinical workflows, offering more robust and reliable tools for lung cancer diagnosis, treatment planning, and outcome prediction. These advancements underscore radiomics' potential in supporting precision oncology and improving patient outcomes through data-driven insights.

Demographic bias of expert-level vision-language foundation models in medical imaging

Advances in artificial intelligence (AI) have achieved expert-level performance in medical imaging applications. Notably, self-supervised vision-language foundation models can detect a broad spectrum of pathologies without relying on explicit training annotations. However, it is crucial to ensure that these AI models do not mirror or amplify human biases, disadvantaging historically marginalized groups such as females or Black patients. In this study, we investigate the algorithmic fairness of state-of-the-art vision-language foundation models in chest x-ray diagnosis across five globally sourced datasets. Our findings reveal that compared to board-certified radiologists, these foundation models consistently underdiagnose marginalized groups, with even higher rates seen in intersectional subgroups such as Black female patients. Such biases present over a wide range of pathologies and demographic attributes. Further analysis of the model embedding uncovers its substantial encoding of demographic information. Deploying medical AI systems with biases can intensify preexisting care disparities, posing potential challenges to equitable healthcare access and raising ethical questions about their clinical applications.

Artificial intelligence-assisted point-of-care devices for lung cancer

Lung cancer is the leading cause of cancer-related deaths worldwide, primarily due to late-stage detection, which limits treatment options. Early detection and screening can increase survival rates, but traditional medical imaging methods are costly and inconvenient. Point-of-care biosensors present a promising alternative, being user-friendly, less labor-intensive, and minimally invasive. With high sensitivity and selectivity, these biosensors detect lung cancer-associated biomarkers, including protein and nucleic acid, in biological fluids such as serum, urine, and saliva. Integrating artificial intelligence (AI) with biosensors has further improved their performance. AI algorithms can analyze complex data, differentiate lung cancer patients from healthy individuals, and even predict the risk of cancer metastasis. Despite these advancements, a comprehensive review of AI-coupled biosensors for lung cancer screening and detection has not yet been conducted. The clinical translation of these biosensors is challenged by a lack of standardization in biomarker selection, the number of biomarkers tested, and the determination of clinical cut-off values. This review focuses on recent advances in biosensors for lung cancer screening and detection, the challenges in their clinical application, and the role of AI in improving biosensor performance. Additionally, it explores future perspectives on the evolution of AI-assisted biosensors into comprehensive health monitoring systems, aiming to bridge the gap between technological innovation and practical clinical use.

Performance of Lung Cancer Prediction Models for Screening-detected, Incidental, and Biopsied Pulmonary Nodules

Purpose To evaluate the performance of eight lung cancer prediction models on patient cohorts with screening-detected, incidentally detected, and bronchoscopically biopsied pulmonary nodules. Materials and Methods This study retrospectively evaluated promising predictive models for lung cancer prediction in three clinical settings: lung cancer screening with low-dose CT, incidentally detected pulmonary nodules, and nodules deemed suspicious enough to warrant a biopsy. The area under the receiver operating characteristic curve of eight validated models, including logistic regressions on clinical variables and radiologist nodule characterizations, artificial intelligence (AI) on chest CT scans, longitudinal imaging AI, and multimodal approaches for prediction of lung cancer risk was assessed in nine cohorts (n = 898, 896, 882, 219, 364, 117, 131, 115, 373) from multiple institutions. Each model was implemented from their published literature, and each cohort was curated from primary data sources collected over periods from 2002 to 2021. Results No single predictive model emerged as the highest-performing model across all cohorts, but certain models performed better in specific clinical contexts. Single-time-point chest CT AI performed well for screening-detected nodules but did not generalize well to other clinical settings. Longitudinal imaging and multimodal models demonstrated comparatively good performance on incidentally detected nodules. When applied to biopsied nodules, all models showed low performance. Conclusion Eight lung cancer prediction models failed to generalize well across clinical settings and sites outside of their training distributions. Keywords: Diagnosis, Classification, Application Domain, Lung Supplemental material is available for this article. © RSNA, 2025 See also commentary by Shao and Niu in this issue.

Progression-Free Survival Prediction Model Based on AI-Enhanced Dynamic Radiomics for Personalized EGFR-TKI Treatment Monitoring Patients With Lung Adenocarcinoma

BACKGROUND AND OBJECTIVE

Epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKIs) are the standard first-line treatment for patients with advanced lung adenocarcinoma (LUAD) with EGFR mutations. However, treatment effectiveness varies widely among individuals, and effective models to predict treatment response are lacking. This study aims to establish a progression-free survival (PFS) prediction model based on dynamic changes in pre- and post-treatment CT scans combined with patients' clinical features.

METHODS

A total of 183 patients with advanced LUAD who received first-line treatment at Peking University Third Hospital from January 2013 to December 2022 were enrolled. A 3D-UNet model was fine-tuned using data from 405 patients with non-small cell lung cancer for advanced lesion segmentation. Clinical and radiomic features extracted using 3D models from 80 EGFR-mutant LUAD patients were used to develop PFS prediction models with a deep-learning binary classification model. The accuracy, specificity, sensitivity, AUC, and F1 score of the models were validated in patients with mutant and wild-type EGFR.

RESULTS

In the EGFR-mutant test set (N = 53), the AUC for the 9-month and 12-month progression prediction models were 0.858 (95% CI, 0.707-0.972) and 0.873 (95% CI, 0.747-0.974). Their accuracies were 81.1% (95% CI, 69.8%-90.6%) and 84.9% (95% CI, 73.6%-94.3%), specificities were 87.5% and 72.2%, sensitivities were 80.0% and 91.4%, and F1 scores were 0.878 and 0.889, respectively.

CONCLUSION

This study developed treatment response prediction models for EGFR-mutant LUAD patients. These models demonstrated strong predictive value for PFS in patients treated with EGFR-TKIs, potentially enabling a more efficient personalized CT scan schedule.

Finding Consensus on Trust in AI in Health Care: Recommendations From a Panel of International Experts

BACKGROUND

The integration of artificial intelligence (AI) into health care has become a crucial element in the digital transformation of health systems worldwide. Despite the potential benefits across diverse medical domains, a significant barrier to the successful adoption of AI systems in health care applications remains the prevailing low user trust in these technologies. Crucially, this challenge is exacerbated by the lack of consensus among experts from different disciplines on the definition of trust in AI within the health care sector.

OBJECTIVE

We aimed to provide the first consensus-based analysis of trust in AI in health care based on an interdisciplinary panel of experts from different domains. Our findings can be used to address the problem of defining trust in AI in health care applications, fostering the discussion of concrete real-world health care scenarios in which humans interact with AI systems explicitly.

METHODS

We used a combination of framework analysis and a 3-step consensus process involving 18 international experts from the fields of computer science, medicine, philosophy of technology, ethics, and social sciences. Our process consisted of a synchronous phase during an expert workshop where we discussed the notion of trust in AI in health care applications, defined an initial framework of important elements of trust to guide our analysis, and agreed on 5 case studies. This was followed by a 2-step iterative, asynchronous process in which the authors further developed, discussed, and refined notions of trust with respect to these specific cases.

RESULTS

Our consensus process identified key contextual factors of trust, namely, an AI system's environment, the actors involved, and framing factors, and analyzed causes and effects of trust in AI in health care. Our findings revealed that certain factors were applicable across all discussed cases yet also pointed to the need for a fine-grained, multidisciplinary analysis bridging human-centered and technology-centered approaches. While regulatory boundaries and technological design features are critical to successful AI implementation in health care, ultimately, communication and positive lived experiences with AI systems will be at the forefront of user trust. Our expert consensus allowed us to formulate concrete recommendations for future research on trust in AI in health care applications.

CONCLUSIONS

This paper advocates for a more refined and nuanced conceptual understanding of trust in the context of AI in health care. By synthesizing insights into commonalities and differences among specific case studies, this paper establishes a foundational basis for future debates and discussions on trusting AI in health care.

Deep Learning in Thoracic Oncology: Meta-Analytical Insights into Lung Nodule Early-Detection Technologies

Background/Objectives: Detecting lung nodules on computed tomography (CT) images is critical for diagnosing thoracic cancers. Deep learning models, particularly convolutional neural networks (CNNs), show promise in automating this process. This systematic review and meta-analysis aim to evaluate the diagnostic accuracy of these models, focusing on lesion-wise sensitivity as the primary metric. Methods: A comprehensive literature search was conducted, identifying 48 studies published up to 7 November 2023. The pooled diagnostic performance was assessed using a random-effects model, with lesion-wise sensitivity as the key outcome. Factors influencing model performance, including participant demographics, dataset privacy, and data splitting methods, were analyzed. Methodological rigor was maintained through the Checklist for Artificial Intelligence in Medical Imaging (CLAIM) and Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) tools. Trial Registration: This review is registered with PROSPERO under CRD42023479887. Results: The meta-analysis revealed a pooled sensitivity of 79% (95% CI: 72-86%) for independent datasets and 85% (95% CI: 83-88%) across all datasets. Variability in performance was associated with dataset characteristics and study methodologies. Conclusions: While deep learning models demonstrate significant potential in lung nodule detection, the findings highlight the need for more diverse datasets, standardized evaluation protocols, and interventional studies to enhance generalizability and clinical applicability. Further research is necessary to validate these models across broader patient populations.

Medical multimodal multitask foundation model for lung cancer screening

Lung cancer screening (LCS) reduces mortality and involves vast multimodal data such as text, tables, and images. Fully mining such big data requires multitasking; otherwise, occult but important features may be overlooked, adversely affecting clinical management and healthcare quality. Here we propose a medical multimodal-multitask foundation model (M3FM) for three-dimensional low-dose computed tomography (CT) LCS. After curating a multimodal multitask dataset of 49 clinical data types, 163,725 chest CT series, and 17 tasks involved in LCS, we develop a scalable multimodal question-answering model architecture for synergistic multimodal multitasking. M3FM consistently outperforms the state-of-the-art models, improving lung cancer risk and cardiovascular disease mortality risk prediction by up to 20% and 10% respectively. M3FM processes multiscale high-dimensional images, handles various combinations of multimodal data, identifies informative data elements, and adapts to out-of-distribution tasks with minimal data. In this work, we show that M3FM advances various LCS tasks through large-scale multimodal and multitask learning.

Diagnostic Performance of the Modified Lung CT Screening Reporting and Data System in a TB-Endemic Country: The Korean National Lung Cancer Screening Program

BACKGROUND

In 2019, Korea initiated the world's first national low-dose CT imaging lung cancer screening (LCS) program, adapting the Lung CT Screening Reporting and Data System (Lung-RADS) to counteract the high false-positive rates driven by prevalent TB.

RESEARCH QUESTION

Does the modified Lung-RADS enhance screening specificity while maintaining sensitivity?

STUDY DESIGN AND METHODS

This nationwide, retrospective cohort study included high-risk individuals aged 54 to 74 years with active tobacco use of at least 30 pack-years participating in the national LCS program from 2019 through 2020. The modified Lung-RADS 1.0 introduced category 2b for nodules matching the size of categories 3 or 4, but showing benign features like granulomas and juxtapleural nodules, and enhanced details for category 4X. Lung cancer diagnosis rates within 1 year of screening and the diagnostic performance of the modified and original Lung-RADS were evaluated.

RESULTS

Among 152,918 participants (98.2% male; mean [SD] age, 61.7 [5.3] years), lung cancer was diagnosed in 0.68% of participants (1,047 of 152,918). A linear trend in cancer rates across Lung-RADS categories was noted (P < .001). Category 2b showed a higher cancer rate than category 2 (0.25% [45 of 18,120] vs 0.14% [33 of 23,467]; P = .01), but lower than category 3 (0.53% [37 of 7,009]; P = .001). Category 4X showed a cancer rate of 36.88% (416 of 1,128). The modified Lung-RADS demonstrated improved specificity (91.96% [139,664 of 151,871] vs 80.06% [121,589 of 151,871]; P < .001) compared with the original criteria. Although sensitivity showed a modest decrease (81.9% [858 of 1,047] vs 86.2% [903 of 1,047]; P < .001), the modification substantially reduced the follow-up burden, decreasing the number of positive screening results needed to detect 1 case of cancer from 34.5 to 15.2. The positive predictive value improved significantly (from 2.90% [903 of 31,185] to 6.57% [858 of 13,065]; P < .001), whereas the negative predictive value remained consistently high (modified, 99.86% [139,664 of 139,853] vs original, 99.88% [121,589 of 121,733]; P = .23).

INTERPRETATION

Korea's modified Lung-RADS enhanced screening efficiency through improved specificity, despite a small reduction in sensitivity.

Deep learning-based lung cancer risk assessment using chest computed tomography images without pulmonary nodules ≥8 mm

Background

Low-dose chest computed tomography (LDCT) screening improves early detection of lung cancer but poses challenges such as false positives and overdiagnosis, especially for nodules smaller than 8 mm where follow-up guidelines are unclear. Traditional risk prediction models have limitations, and deep learning (DL) algorithms offer potential improvements but often require large datasets. This study aimed to develop a DL-based, label-free lung cancer risk prediction model using alternative LDCT images and validate it in individuals without non-calcified solid pulmonary nodules larger than 8 mm.

Methods

We utilized LDCT scans from individuals without non-calcified solid nodules larger than 8 mm to develop a DL-based lung cancer risk prediction model. An alternative training dataset included 1,064 LDCT scans: 380 from patients with pathologically confirmed lung cancer and 684 from control individuals without lung cancer development over 5 years. For the lung cancer group, only the contralateral lung (without the tumor) was analyzed to represent high-risk individuals without large nodules. The LDCT scans were randomly divided into training and validation sets in a 3:1 ratio. Four three-dimensional (3D) convolutional neural networks (CNNs; 3D-CNN, MobileNet v2, SEResNet18, EfficientNet-B0) were trained using densely connected U-Net (DenseUNet)-segmented lung parenchyma images. The models were validated on a real-world test dataset comprising 1,306 LDCT scans (1,254 low-risk and 52 high-risk individuals) and evaluated using the area under the receiver operating characteristic (ROC) curve (AUC), Brier scores, and calibration measures.

Results

In the validation dataset, the AUC values were 0.801 for 3D-CNN, 0.802 for MobileNet v2, 0.755 for EfficientNet-B0, and 0.833 for SEResNet18. Corresponding Brier scores were 0.169, 0.175, 0.217, and 0.156, respectively, indicating good calibration, especially for SEResNet18. In the test dataset, the AUC values were 0.769 for 3D-CNN, 0.753 for MobileNet v2, 0.681 for EfficientNet-B0, and 0.820 for SEResNet18, with Brier scores of 0.169, 0.180, 0.202, and 0.138, respectively. The SEResNet18 model demonstrated the best performance, achieving the highest AUC and lowest Brier score in both validation and test datasets.

Conclusions

Our study demonstrated that DL-based, label-free lung cancer risk prediction models using alternative LDCT images can effectively predict lung cancer development in individuals without non-calcified solid pulmonary nodules larger than 8 mm. By analyzing lung parenchyma on LDCT images without relying on nodule detection, these models may enhance the efficiency of LDCT screening programs. Further prospective studies are needed to determine their clinical utility and impact on screening protocols, and validation in larger, diverse populations is necessary to ensure generalizability.

A review of deep learning for brain tumor analysis in MRI

Recent progress in deep learning (DL) is producing a new generation of tools across numerous clinical applications. Within the analysis of brain tumors in magnetic resonance imaging, DL finds applications in tumor segmentation, quantification, and classification. It facilitates objective and reproducible measurements crucial for diagnosis, treatment planning, and disease monitoring. Furthermore, it holds the potential to pave the way for personalized medicine through the prediction of tumor type, grade, genetic mutations, and patient survival outcomes. In this review, we explore the transformative potential of DL for brain tumor care and discuss existing applications, limitations, and future directions and opportunities.

Advances and challenges in precision imaging

Technological innovations in genomics and related fields have facilitated large sequencing efforts, supported new biological discoveries in cancer, and spawned an era of liquid biopsy biomarkers. Despite these advances, precision oncology has practical constraints, partly related to cancer's biological diversity and spatial and temporal complexity. Advanced imaging technologies are being developed to address some of the current limitations in early detection, treatment selection and planning, drug delivery, and therapeutic response, as well as difficulties posed by drug resistance, drug toxicity, disease monitoring, and metastatic evolution. We discuss key areas of advanced imaging for improving cancer outcomes and survival. Finally, we discuss practical challenges to the broader adoption of precision imaging in the clinic and the need for a robust translational infrastructure.

Permutation Test for Image-on-Scalar Regression With an Application to Breast Cancer

Image based screening is now routinely available for early detection of cancer and other diseases. Quantitative analysis for effects of risk factors on digital images is important to extract biological insights for modifiable factors in prevention studies and understand pathways for targets in preventive drugs. However, current approaches are restricted to summary measures within the image with the assumption that all relevant features needed to characterize an image can be identified and appropriately quantified. Motivated by data challenges in breast cancer, we propose a nonparametric statistical framework for risk factor screening that uses the whole mammogram image as outcome. The proposed permutation test allows assessment of whether a set of scalar risk factors is associated with the whole image in the presence of correlated residuals across the spatial domain. We provide extensive simulation studies and illustrate an application to the Joanne Knight Breast Health Cohort using the mammogram imaging data.

The anesthesiologist's guide to critically assessing machine learning research: a narrative review

Artificial Intelligence (AI), especially Machine Learning (ML), has developed systems capable of performing tasks that require human intelligence. In anesthesiology and other medical fields, AI applications can improve the precision and efficiency of daily clinical practice, and can also facilitate a personalized approach to patient care, which can lead to improved outcomes and quality of care. ML has been successfully applied in various settings of daily anesthesiology practice, such as predicting acute kidney injury, optimizing anesthetic doses, and managing postoperative nausea and vomiting. The critical evaluation of ML models in healthcare is crucial to assess their validity, safety, and clinical applicability. Evaluation metrics allow an objective statistical assessment of model performance. Tools such as Shapley Values (SHAP) help interpret how individual variables contribute to model predictions. Transparency in reporting is key in maintaining trust in these technologies and to ensure their use follows ethical principles, aiming to reduce safety concerns while also benefiting patients. Understanding evaluation metrics is essential, as they provide detailed information on model performance and their ability to discriminate between individual class rates. This article offers a comprehensive framework in assessing the validity, applicability, and limitations of models, guiding responsible and effective integration of ML technologies into clinical practice. A balance between innovation, patient safety and ethical considerations must be pursued.

Global burden of lung cancer in 2022 and projections to 2050: Incidence and mortality estimates from GLOBOCAN

BACKGROUND

Lung cancer continues to pose a serious global public health challenge. Timely evidence on the global epidemiological profile of the disease is crucial to facilitate the implementation to lung cancer control efforts. This study provides updated global estimates for lung cancer incidence and mortality in 2022, along with projections for new cases and deaths up to 2050.

METHODS

In the population-based study, we extracted data about lung cancer new cases and deaths from GLOBOCAN 2022 database across 185 countries or territories. We analyzed age-standardized rates by sex, country, region, and human development index (HDI). Projected new cases and deaths for 2050 were estimated using global demographic projections.

RESULTS

In 2022, lung cancer stood as the most frequently diagnosed cancer and the primary cause of cancer-related deaths on a global scale with approximately 2.48 million new cases and 1.8 million deaths, respectively. The incidence and mortality rates of lung cancer exhibited disparities in sex and world regions. Furthermore, incidence and mortality rates increasing as HDI increased. If the incidence and mortality rates remain stable as in 2022, the burden of lung cancer is projected to increase to 4·62 million new cases and 3·55 million deaths by 2050.

CONCLUSIONS

Lung cancer is the predominant form of cancer and the foremost contributor to cancer-related mortality in 2022 with notable geographical, sex, and socioeconomic disparities.

Classification of Caries Based on CBCT: A Deep Learning Network Interpretability Study

This study aimed to create a caries classification scheme based on cone-beam computed tomography (CBCT) and develop two deep learning models to improve caries classification accuracy. A total of 2713 axial slices were obtained from CBCT images of 204 carious teeth. Both classification models were trained and tested using the same pretrained classification networks on the dataset, including ResNet50_vd, MobileNetV3_large_ssld, and ResNet50_vd_ssld. The first model was used directly to classify the original images (direct classification model). The second model incorporated a presegmentation step for interpretation (interpretable classification model). Performance evaluation metrics including accuracy, precision, recall, and F1 score were calculated. The Local Interpretable Model-agnostic Explanations (LIME) method was employed to elucidate the decision-making process of the two models. In addition, a minimum distance between caries and pulp was introduced for determining the treatment strategies for type II carious teeth. The direct model that utilized the ResNet50_vd_ssld network achieved top accuracy, precision, recall, and F1 score of 0.700, 0.786, 0.606, and 0.616, respectively. Conversely, the interpretable model consistently yielded metrics surpassing 0.917, irrespective of the network employed. The LIME algorithm confirmed the interpretability of the classification models by identifying key image features for caries classification. Evaluation of treatment strategies for type II carious teeth revealed a significant negative correlation (p < 0.01) with the minimum distance. These results demonstrated that the CBCT-based caries classification scheme and the two classification models appeared to be acceptable tools for the diagnosis and categorization of dental caries.

Lung cancer screening in India: Preparing for the future using smart tools & biomarkers to identify highest risk individuals

There is a growing burden of lung cancer cases in India, incidence projected to increase from 63,708 cases (2015) to 81,219 cases (2025). The increasing numbers are attributed to smoking (India currently has nearly 100 million adult smokers) and environmental pollution. Most patients present with advanced disease (80-85% are incurable), causing nearly 60,000 annual deaths from lung cancer. Early detection through lung cancer screening (LCS) can result in curative therapies for earlier stages of lung cancer and improved survival. Annual low-dose computerized tomography (LDCT) is the standard method for LCS. Usually, high-risk populations (age>50 yr and >20 pack-years of smoking) are considered for LCS, but even such focused screening may be challenging in resource-limited countries like India. However, developing a smart LCS programme with high yield may be possible by leveraging demographic and genomic data, use of smart tools, and judicious use of blood-based biomarkers. Developing this model over the next several years will facilitate a structured cancer screening programme for populations at the highest risk of lung cancer. In this paper, we discuss the demographics of lung cancer in India and its relation to smoking patterns. Further, we elaborate on the potential applications and challenges of bringing a smart approach to LCS in high-risk populations in India.

Lung cancer research and treatment: global perspectives and strategic calls to action

BACKGROUND

Lung cancer remains a critical public health issue, presenting multifaceted challenges in prevention, diagnosis, and treatment. This article aims to review the current landscape of lung cancer research and management, delineate the persistent challenges, and outline pragmatic solutions.

MATERIALS AND METHODS

Global experts from academia, regulatory agencies such as the Food and Drug Administration (FDA) and the European Medicines Agency (EMA), the National Cancer Institute (NCI), professional societies, the pharmaceutical and biotech industries, and patient advocacy groups were gathered by the New York Lung Cancer Foundation to review the state of the art in lung cancer and to formulate calls to action.

RESULTS

Improving lung cancer management and research involves promoting tobacco cessation, identifying individuals at risk who could benefit from early detection programs, and addressing treatment-related toxicities. Efforts should focus on conducting well-designed trials to determine the optimal treatment sequence. Research into innovative biomarkers and therapies is crucial for more personalized treatment. Ensuring access to appropriate care for all patients, whether enrolled in clinical trials or not, must remain a priority.

CONCLUSIONS

Lung cancer is a major health burden worldwide, and its treatment has become increasingly complex over the past two decades. Improvement in lung cancer management and research requires unified messaging and global collaboration, expanded education, and greater access to screening, biomarker testing, treatment, as well as increased representativeness, participation, and diversity in clinical trials.

Opportunity and Opportunism in Artificial Intelligence-Powered Data Extraction: A Value-Centered Approach

Radiologists' traditional role in the diagnostic process is to respond to specific clinical questions and reduce uncertainty enough to permit treatment decisions to be made. This charge is rapidly evolving due to forces such as artificial intelligence (AI), big data (opportunistic imaging, imaging prognostication), and advanced diagnostic technologies. A new modernistic paradigm is emerging whereby radiologists, in conjunction with computer algorithms, will be tasked with extracting as much information from imaging data as possible, often without a specific clinical question being posed and independent of any stated clinical need. In addition, AI algorithms are increasingly able to predict long-term outcomes using data from seemingly normal examinations, enabling AI-assisted prognostication. As these algorithms become a standard component of radiology practice, the sheer amount of information they demand will increase the need for streamlined workflows, communication, and data management techniques. In addition, the provision of such information raises reimbursement, liability, and access issues. Guidelines will be needed to ensure that all patients have access to the benefits of this new technology and guarantee that mined data do not inadvertently create harm. In this Review, we discuss the challenges and opportunities relevant to radiologists in this changing landscape, with an emphasis on ensuring that radiologists provide high-value care.

The Pivotal Role of Baseline LDCT for Lung Cancer Screening in the Era of Artificial Intelligence

In this narrative review, we address the ongoing challenges of lung cancer (LC) screening using chest low-dose computerized tomography (LDCT) and explore the contributions of artificial intelligence (AI), in overcoming them. We focus on evaluating the initial (baseline) LDCT examination, which provides a wealth of information relevant to the screening participant's health. This includes the detection of large-size prevalent LC and small-size malignant nodules that are typically diagnosed as LCs upon growth in subsequent annual LDCT scans. Additionally, the baseline LDCT examination provides valuable information about smoking-related comorbidities, including cardiovascular disease, chronic obstructive pulmonary disease, and interstitial lung disease (ILD), by identifying relevant markers. Notably, these comorbidities, despite the slow progression of their markers, collectively exceed LC as ultimate causes of death at follow-up in LC screening participants. Computer-assisted diagnosis tools currently improve the reproducibility of radiologic readings and reduce the false negative rate of LDCT. Deep learning (DL) tools that analyze the radiomic features of lung nodules are being developed to distinguish between benign and malignant nodules. Furthermore, AI tools can predict the risk of LC in the years following a baseline LDCT. AI tools that analyze baseline LDCT examinations can also compute the risk of cardiovascular disease or death, paving the way for personalized screening interventions. Additionally, DL tools are available for assessing osteoporosis and ILD, which helps refine the individual's current and future health profile. The primary obstacles to AI integration into the LDCT screening pathway are the generalizability of performance and the explainability.

Predicting radiation pneumonitis in lung cancer using machine learning and multimodal features: a systematic review and meta-analysis of diagnostic accuracy

OBJECTIVES

To evaluate the diagnostic accuracy of machine learning models incorporating multimodal features for predicting radiation pneumonitis in lung cancer through a systematic review and meta-analysis.

METHODS

Relevant studies were identified through a systematic search of PubMed, Web of Science, Embase, and the Cochrane Library from October 2003 to December 2023. Additional studies were located by reviewing bibliographies and relevant websites. Two independent researchers screened titles, abstracts, and full-text articles according to predefined inclusion and exclusion criteria. Data extraction was performed using standardized forms, and study quality was assessed using the Quality Assessment of Diagnostic Accuracy Studies-2 tool. The primary outcomes, including combined sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), diagnostic odds ratio (DOR), and area under the curve (AUC), were calculated using STATA MP-64 software(Stata Corporation LLC, College Station, USA) with a random-effects model. Meta-analysis was conducted to synthesize diagnostic accuracy measures, and analyses of heterogeneity and publication bias were performed.

RESULTS

A total of 1,406 patients with primary lung cancer were included in this systematic review, drawing data from 9 studies. The pooled analysis revealed a sensitivity of 0.74 [0.58-0.85] and a specificity of 0.91 [0.87-0.95] for machine learning models in diagnosing radiation pneumonitis. The positive likelihood ratio (PLR) was 8.69 [5.21-14.50], the negative likelihood ratio (NLR) was 0.28 [0.16-0.49], and the diagnostic odds ratio (DOR) was 30.73 [11.96-78.97]. The area under the curve (AUC) was 0.93 [0.90-0.95], indicating excellent diagnostic performance. Meta-regression analysis identified that the number of machine learning models, year of publication, and study design contributed to heterogeneity among studies. No evidence of publication bias was found. Overall, machine learning models incorporating multimodal characteristics demonstrated 75% accuracy in predicting moderate to severe radiation pneumonitis.

CONCLUSION

In conclusion, by integrating the current machine learning (ML) algorithm's ability in big data mining, a predictive model can be constructed by combining multi-modal features such as genetics, imaging, and cell factors. By selecting multiple machine learning algorithm frameworks and competing for the best combination model based on research goals, the reliability and accuracy of the radiation pneumonitis prediction model can be greatly improved.

TRIAL REGISTRATION

PROSPERO (CRD42024497599).

Enhancing neuro-oncology care through equity-driven applications of artificial intelligence

The disease course and clinical outcome for brain tumor patients depend not only on the molecular and histological features of the tumor but also on the patient's demographics and social determinants of health. While current investigations in neuro-oncology have broadly utilized artificial intelligence (AI) to enrich tumor diagnosis and more accurately predict treatment response, postoperative complications, and survival, equity-driven applications of AI have been limited. However, AI applications to advance health equity in the broader medical field have the potential to serve as practical blueprints to address known disparities in neuro-oncologic care. In this consensus review, we will describe current applications of AI in neuro-oncology, postulate viable AI solutions for the most pressing inequities in neuro-oncology based on broader literature, propose a framework for the effective integration of equity into AI-based neuro-oncology research, and close with the limitations of AI.

Body composition as a biomarker for assessing future lung cancer risk

Purpose

To investigate if body composition is a biomarker for assessing the risk of developing lung cancer.

Materials and Methods

Low-dose computed tomography (LDCT) scans from the Pittsburgh Lung Screening Study (PLuSS) (n=3,635, 22 follow-up years) and NLST-ACRIN (n=16,435, 8 follow-up years) cohorts were used in the study. Artificial intelligence (AI) algorithms were developed to automatically segment and quantify subcutaneous adipose tissue (SAT), visceral adipose tissue (VAT), intramuscular adipose tissue (IMAT), skeletal muscle (SM), and bone. Cause-specific Cox proportional hazards models were used to evaluate the hazard ratios (HRs). Standard time-dependent receiver operating characteristic (ROC) analysis was used to evaluate the prognostic ability of different models over time.

Results

The final composite models were formed by seven variables: age (HR=1.20), current smoking status (HR=1.59), bone volume (HR=1.79), SM density (HR=0.29), IMAT ratio (HR=0.33), IMAT density (HR=0.56), and SAT volume (HR=0.56). The models trained on the PLuSS cohort achieved a mean AUC of 0.76 (95% CI: 0.74-0.79) over 21 follow-up years and 0.70 (95% CI: 0.66-0.74) over the first 7 follow-up years for predicting lung cancer development within the PLuSS cohort. In contrast, models trained on the PLuSS cohort alone, as well as in combination with the NLST cohorts, achieved an AUC ranging from 0.61 to 0.68 in the NLST cohort over a 7-year follow-up period.

Conclusion

Body composition assessed on LDCT is a significant predictor of lung cancer risk and could improve the effectiveness of LDCT lung cancer screening by optimizing the screening eligibility and frequency.

Summary statement

Body composition assessed on LDCT is a significant predictor of lung cancer risk and could improve the effectiveness of LDCT lung cancer screening by optimizing the screening eligibility and frequency.

Key Points

This study unveils the significant associations between body tissues and lung cancer risk.The prediction models based on body composition alone, as well as the combination of demographics and body composition features can effectively identify patients at higher risk of developing lung cancer.

Lung cancer screening - Time for an update?

Lung cancer screening can reduce the mortality of lung cancer, the leading cause of cancer death worldwide. Real world screening experience highlights areas for improvement in a complex and changing world, particularly ethnic disparity, and the potential for new and emerging risk factors, in addition to well known risk of smoking and asbestos exposure. Biomarkers offer the promise of objective risk assessment but are not yet ready for clinical practice. This review discusses some of the major issues faced by lung cancer screening and the potential role for biomarkers.

Using a Polygenic Risk Score to Improve the Risk Prediction of Non-Small Cell Lung Cancer in Taiwan

PURPOSE

Low-dose computed tomography (LDCT) can help reducing lung cancer mortality. In Taiwan, the existing screening criteria revolve around smoking habits and family history of lung cancer. The role of genetic variation in non-small cell lung cancer (NSCLC) development is increasingly recognized. In this study, we aimed to investigate the potential benefits of polygenic risk scores (PRSs) in predicting NSCLC and enhancing the effectiveness of screening programs.

METHODS

We conducted a retrospective cohort study that included participants without prior diagnosis of lung cancer and later received LDCT for lung cancer screening. Genetic data for these participants were obtained from the project of Taiwan Precision Medicine Initiative. We adopted the model of genome-wide association study-derived PRS calculation using 19 susceptibility loci associated with the risk of NSCLC as reported by Dai et al.

RESULTS

We studied a total of 2,287 participants (1,197 male, 1,090 female). More female participants developed NSCLC during the follow-up period (4.4% v 2.5%, P = .015). The only risk factor of NSCLC diagnosis among male participants was age. Among female participants, independent risk factors of NSCLC diagnosis were age (adjusted hazard ratio [aHR], 1.08 [95% CI, 1.04 to 1.11]), a family history of lung cancer (aHR, 3.21 [95% CI, 1.78 to 5.77]), and PRS fourth quartile (aHR, 2.97 [95% CI, 1.25 to 7.07]). We used the receiver operating characteristics to show an AUC value of 0.741 for the conventional model. With the further incorporation of PRS, the AUC rose to 0.778.

CONCLUSION

The evaluation of PRS for NSCLC prediction holds promise for enhancing the effectiveness of lung cancer screening in Taiwan especially in women. By incorporating genetic information, screening criteria can be tailored to identify individuals at higher risks of NSCLC.

Non-small-cell lung cancer

Non-small-cell lung cancer (NSCLC) is one of the most frequent cancer types and is responsible for the majority of cancer-related deaths worldwide. The management of NSCLC has improved considerably, especially in the past 10 years. The systematic screening of populations at risk with low-dose CT, the implementation of novel surgical and radiotherapeutic techniques and a deeper biological understanding of NSCLC that has led to innovative systemic treatment options have improved the prognosis of patients with NSCLC. In non-metastatic NSCLC, the combination of various perioperative strategies and adjuvant immunotherapy in locally advanced disease seem to enhance cure rates. In metastatic NSCLC, the implementation of novel drugs might prolong disease control together with preserving quality of life. The further development of predictive clinical and genetic markers will be essential for the next steps in individualized treatment concepts.

Lung cancer screening in never smokers

PURPOSE OF REVIEW

Low-dose computed tomography (LDCT) lung cancer screening has been established in smokers, but its role in never smokers remains unclear. The differences in lung cancer biology between smokers and nonsmokers highlight the importance of a discriminated approach. This overview focuses on the emerging data and implementation challenges for LDCT screening in nonsmokers.

RECENT FINDINGS

The first LDCT screening study in nonsmokers enriched with risk factors demonstrated a lung cancer detection rate double that of the phase 3 trials in smokers. The relative risk of lung cancer detected by LDCT has also been found to be similar amongst female never smokers and male ever smokers in Asia. Majority of lung cancers detected through LDCT screening are stage 0/1, leading to concerns of overdiagnosis. Risk prediction models to enhance individual selection and nodule management could be useful to enhance the utility of LDCT screening in never smokers.

SUMMARY

With appropriate risk stratification, LDCT screening in never smokers may attain similar efficacy as compared to smokers. A global effort is needed to generate evidence surrounding optimal screening strategies, as well as health and economic benefits to determine the suitability of widespread implementation.

Development and validation of an integrated system for lung cancer screening and post-screening pulmonary nodules management: a proof-of-concept study (ASCEND-LUNG)

Background

In order to address the low compliance and dissatisfied specificity of low-dose computed tomography (LDCT), efficient and non-invasive approaches are needed to complement its limitations for lung cancer screening and management. The ASCEND-LUNG study is a prospective two-stage case-control study designed to evaluate the performance of a liquid biopsy-based comprehensive lung cancer screening and post-screening pulmonary nodules management system.

Methods

We aimed to develop a comprehensive lung cancer system called Peking University Lung Cancer Screening and Management System (PKU-LCSMS) which comprises a lung cancer screening model to identify specific populations requiring LDCT and an artificial intelligence-aided (AI-aided) pulmonary nodules diagnostic model to classify pulmonary nodules following LDCT. A dataset of 465 participants (216 cancer, 47 benign, 202 non-cancer control) were used for the two models' development phase. For the lung cancer screening model development, cancer participants were randomly split at a ratio of 1:1 into the train and validation cohorts, and then non-cancer controls were age-matched to the cancer cases in a 1:1 ratio. Similarly, for the AI-aided pulmonary nodules model, cancer and benign participants were also randomly divided at a ratio of 2:1 into the train and validation cohorts. Subsequently, during the model validation phase, sensitivity and specificity were validated using an independent validation cohort consisting of 291 participants (140 cancer, 25 benign, 126 non-cancer control). Prospectively collected blood samples were analyzed for multi-omics including cell-free DNA (cfDNA) methylation, mutation, and serum protein. Computerized tomography (CT) images data was also obtained. Paired tissue samples were additionally analyzed for DNA methylation, DNA mutation, and messenger RNA (mRNA) expression to further explore the potential biological mechanisms. This study is registered with ClinicalTrials.gov, NCT04817046.

Findings

Baseline blood samples were evaluated for the whole screening and diagnostic process. The cfDNA methylation-based lung cancer screening model exhibited the highest area under the curve (AUC) of 0.910 (95% CI, 0.869-0.950), followed by the protein model (0.891 [95% CI, 0.845-0.938]) and lastly the mutation model (0.577 [95% CI, 0.482-0.672]). Further, the final screening model, which incorporated cfDNA methylation and protein features, achieved an AUC of 0.963 (95% CI, 0.942-0.984). In the independent validation cohort, the multi-omics screening model showed a sensitivity of 99.2% (95% CI, 0.957-1.000) at a specificity of 56.3% (95% CI, 0.472-0.652). For the AI-aided pulmonary nodules diagnostic model, which incorporated cfDNA methylation and CT images features, it yielded a sensitivity of 81.1% (95% CI, 0.732-0.875), a specificity of 76.0% (95% CI, 0.549-0.906) in the independent validation cohort. Furthermore, four differentially methylated regions (DMRs) were shared in the lung cancer screening model and the AI-aided pulmonary nodules diagnostic model.

Interpretation

We developed and validated a liquid biopsy-based comprehensive lung cancer screening and management system called PKU-LCSMS which combined a blood multi-omics based lung cancer screening model incorporating cfDNA methylation and protein features and an AI-aided pulmonary nodules diagnostic model integrating CT images and cfDNA methylation features in sequence to streamline the entire process of lung cancer screening and post-screening pulmonary nodules management. It might provide a promising applicable solution for lung cancer screening and management.

Funding

This work was supported by Science, Science, Technology & Innovation Project of Xiongan New Area, Beijing Natural Science Foundation, CAMS Innovation Fund for Medical Sciences (CIFMS), Clinical Medicine Plus X-Young Scholars Project of Peking University, the Fundamental Research Funds for the Central Universities, Research Unit of Intelligence Diagnosis and Treatment in Early Non-small Cell Lung Cancer, Chinese Academy of Medical Sciences, National Natural Science Foundation of China, Peking University People's Hospital Research and Development Funds, National Key Research and Development Program of China, and the fundamental research funds for the central universities.

Premalignant Progression in the Lung: Knowledge Gaps and Novel Opportunities for Interception of Non-Small Cell Lung Cancer. An Official American Thoracic Society Research Statement

Rationale: Despite significant advances in precision treatments and immunotherapy, lung cancer is the most common cause of cancer death worldwide. To reduce incidence and improve survival rates, a deeper understanding of lung premalignancy and the multistep process of tumorigenesis is essential, allowing timely and effective intervention before cancer development. Objectives: To summarize existing information, identify knowledge gaps, formulate research questions, prioritize potential research topics, and propose strategies for future investigations into the premalignant progression in the lung. Methods: An international multidisciplinary team of basic, translational, and clinical scientists reviewed available data to develop and refine research questions pertaining to the transformation of premalignant lung lesions to advanced lung cancer. Results: This research statement identifies significant gaps in knowledge and proposes potential research questions aimed at expanding our understanding of the mechanisms underlying the progression of premalignant lung lesions to lung cancer in an effort to explore potential innovative modalities to intercept lung cancer at its nascent stages. Conclusions: The identified gaps in knowledge about the biological mechanisms of premalignant progression in the lung, together with ongoing challenges in screening, detection, and early intervention, highlight the critical need to prioritize research in this domain. Such focused investigations are essential to devise effective preventive strategies that may ultimately decrease lung cancer incidence and improve patient outcomes.

The Promise and Challenges of AI Integration in Ovarian Cancer Screenings

PURPOSE

Ovarian cancer is oftendiagnosed late due to vague symptoms, leading to poor survival rate. Improved screening tests could mitigate this issue. This narrative review examines the potential and challenges of integrating artificial intelligence (A.I.) into ovarian cancer screenings, with a focus on improving early detection, diagnosis, and personalized risk assessment.

METHOD

A comprehensive review of existing literature was conducted, analyzing studies and discussions within the scientific community.

RESULTS

A.I. shows promise in significantly improving the ovarian cancer screening processes, increasing accuracy, efficiency, and resource allocation. However, data quality and bias issues pose considerable challenges, potentially leading to healthcare disparities.

CONCLUSIONS

Integrating A.I. into ovarian cancer screenings offers potential benefits but comes with significant challenges. By promoting diverse data collection, engaging with underrepresented groups, and ensuring ethical data use, A.I. can be harnessed for more accurate and equitable ovarian cancer diagnoses.