How Advancing AI Can Change the Course of Lung Cancer Treatment
Authors: Amina Khalpey, PhD, Brynne Rozell, DO, Kirtana Roopan, Jessa Deckwa, BS, Zain Khalpey, MD, PhD, FACS
Lung cancer is a leading cause of cancer-related deaths worldwide. Early detection and accurate assessment of lung cancer are critical for improving patient outcomes. Artificial Intelligence (AI) holds the potential to revolutionize lung cancer detection, diagnosis, and treatment. In this white paper, we explore the role of AI in lung cancer, its current applications, challenges, and future prospects. The motivation for this work is deeply personal, as the author’s father had stage IIb left upper lobe lung adenocarcinoma and underwent chemotherapy. By advancing AI in the field of lung cancer, we hope to contribute to a future where fewer families must endure the emotional and physical toll of this devastating disease.
1.1 Personal motivation
My father was diagnosed with stage IIb left upper lobe lung adenocarcinoma, which required chemotherapy. His journey through diagnosis and treatment has inspired me to explore ways to improve lung cancer care. In this white paper, I delve into the potential of AI in revolutionizing the detection, diagnosis, and treatment of lung cancer.
2. Lung cancer: current challenges
Lung cancer is a significant global health burden, with high morbidity and mortality rates. Early detection is crucial for improving patient outcomes. However, current diagnostic tools often involve invasive procedures and are associated with various challenges, including false positives, unnecessary interventions, and radiation exposure.
2. AI in lung cancer detection and diagnosis
2.1 AI-driven imaging analysis
Screening for lung cancer requires in-depth, time-consuming analysis in which radiologists must discern the location, size, border, shape, and type of each pulmonary nodule found (1). AI-driven imaging analysis is a promising approach for facilitating and improving lung cancer detection. Deep learning algorithms, such as convolutional neural networks (CNNs), have shown potential in accurately identifying lung nodules in computed tomography (CT) scans (2). AI can also be useful for lung cancer staging in positron emission tomography (PET) and CT image analysis. However, such algorithms can have trouble differentiating structures like bronchi, blood vessels, and lymph nodes of lung tissue from pulmonary nodules, which can lead to misdiagnosis (3).
2.2 AI in pathology
Tissue biopsy is another important tool in lung cancer diagnosis, but pathology of samples can be challenging to discern manually because of the various subtypes of lung cancer. AI can analyze abnormalities in biopsies, facilitating pathological analysis, reducing the incidence of false-negatives, and advancing the accuracy and efficiency of lung cancer classification. Automated pathological classification algorithms have been developed using deep convolutional neural networks (DCNN). These models are trained on cytological images generated by Generative Adversarial Networks (GAN), resulting in efficient and effective AI for the lung cancer diagnosis (3).
2.3 Multi-modal AI approaches
Integrating multi-modal data sources, such as demographics, medical history, and genetic factors, can enhance the performance and accuracy of AI-driven lung cancer risk prediction algorithms (4).
2.4 Explainable AI (xAI)
Incorporating xAI techniques can help healthcare providers understand and trust AI-generated predictions, allowing for more informed decision-making and improved patient outcomes (Holzinger et al). For example, computer-aided detection systems, such as those used for lung cancer screening, are based on lung segmentation, pulmonary nodule detection, and nodule classification steps. Lung segmentation focuses the CT image on a specific region for closer analysis. The nodule detection step identifies potentially malignant lung nodules (1). Understanding these fundamental designs behind AI algorithms through xAI techniques will allow a smoother integration of AI into healthcare.
3. AI in lung cancer treatment
3.1. AI-driven precision medicine
AI can support precision medicine by identifying genetic and molecular factors that may influence treatment response, enabling tailored therapies for individual patients (6).
3.2. AI-based treatment planning
Current treatments for lung cancer include radiotherapy, drug treatment, and/or surgical resection. The analysis of the cancer stage, location of the tumor, along with genetic and histological changes needed to determine the best course of treatment is time-consuming and inefficient for physicians to do manually. AI that models this analysis can help physicians identify important patient information and relevant evidence to make treatment plans more efficiently (3). AI-driven treatment planning tools can optimize radiation therapy delivery, minimizing damage to healthy tissue and improving treatment efficacy (7).
4. Challenges and future prospects
4.1 Data privacy and security
Developing AI systems in big data settings for lung cancer care requires addressing data privacy and security concerns, including data anonymization and secure storage and sharing of sensitive information. For instance, there is risk of unintentional disclosure of personally identifying health information. Additionally, new technologies constantly arising, like GPS and drones, may make maintaining ethical standards a challenge and force a change in generally accepted privacy standards (8).
4.2. Regulatory considerations
Regulatory frameworks must evolve to accommodate AI-driven diagnostic and treatment tools, ensuring patient safety and adherence to ethical guidelines.
4.3. Clinical integration and adoption
Integrating AI technologies into clinical practice requires addressing barriers to adoption, including provider resistance, cost, and infrastructure requirements. Often the mechanism behind the output of AI algorithms is obscure, which can cause reluctance among the people whose jobs it is meant to facilitate. It is important that healthcare workers understand how AI technologies work in order to encourage trust and a more synergetic integration into healthcare.
4.4. Collaboration and interdisciplinary research
To maximize the impact of AI in lung cancer care, collaboration between AI experts, medical professionals, and other stakeholders is essential. Such collaboration can aid in the clinical integration of AI, helping with the development of AI technologies that are user-friendly and user-specific to the healthcare workers who will utilize them and ensuring they fit seamlessly into the workplace. Interdisciplinary research can foster the development of innovative solutions that address the unique challenges and requirements of lung cancer diagnosis and treatment.
4.5. Patient education and engagement
Patient education and engagement are critical components of successful AI integration in lung cancer care. Ensuring patients understand the benefits and limitations of AI-driven tools can help them make informed decisions about their care and foster trust in the technology.
5. Real-world applications and case studies
5.1. AI-driven lung cancer screening programs
AI-driven lung cancer screening programs can help identify high-risk individuals and prioritize them for further diagnostic workup, potentially reducing the number of unnecessary interventions and optimizing resource allocation.
5.2 AI-guided treatment selection and monitoring
AI algorithms can support treatment selection by identifying biomarkers associated with treatment response, allowing healthcare providers to tailor therapies for individual patients. Additionally, AI-driven monitoring tools can help track treatment progress and detect early signs of relapse, enabling timely intervention.
6. Future directions and emerging technologies
6.1. AI-assisted surgical planning and navigation
AI-assisted surgical planning and navigation systems can help surgeons perform more precise and minimally invasive procedures, potentially improving patient outcomes and reducing recovery time.
6.2 AI-powered drug discovery and development
AI-driven drug discovery and development can accelerate the identification of novel therapeutic targets and streamline the design and testing of new drug candidates, potentially leading to more effective treatments for lung cancer.
AI has immense potential to revolutionize lung cancer care by enhancing detection, diagnosis, and treatment strategies. The integration of AI technologies into clinical practice can lead to improved patient outcomes and reduced burden on healthcare systems. As we continue to advance AI in lung cancer, we hope to contribute to a future where fewer families must endure the emotional and physical toll of this devastating disease. Through interdisciplinary collaboration, innovative research, and the development of patient-centered solutions, AI can transform the landscape of lung cancer care for the better.
1. Binczyk, F., et al. (2021). Radiomics and artificial intelligence in lung cancer screening. Transl Lung Cancer Res.
2. Landy, R., et al. (2023). Deep learning algorithm for predicting lung cancer risk from low-dose CT images with presumed nonmalignant lung nodules. JAMA Network Open.
3. Pei, Q., et al. (2022). Artificial intelligence in clinical applications for lung cancer: diagnosis, treatment and prognosis. Clin Chem Lab Med.
4. Wang, L., et al. (2022) Deep Learning Techniques to Diagnose Lung Cancer. Cancers (Basel).
5. Holzinger, A., Biemann, C., Pattichis, C. S., & Kell, D. B.(2017). What do we need to build explainable AI systems for the medical domain? arXiv preprint arXiv:1712.09923.
6. Gashimova, E., et al. (2022). Non-invasive Exhaled Breath and Skin Analysis to Diagnose Lung Cancer: Study of Age Effect on Diagnostic Accuracy. ACS Omega.
7. Rajkomar, A., et al. (2018). Machine learning in medicine. New England Journal of Medicine.
8. Mooney SJ, Pejaver V. Big Data in Public Health: Terminology, Machine Learning, and Privacy. Annu Rev Public Health. 2018 Apr 1;39:95-112. doi: 10.1146/annurev-publhealth-040617-014208. Epub 2017 Dec 20. PMID: 29261408; PMCID: PMC6394411.