Attention: Restrictions on use of AUA, AUAER, and UCF content in third party applications, including artificial intelligence technologies, such as large language models and generative AI.
You are prohibited from using or uploading content you accessed through this website into external applications, bots, software, or websites, including those using artificial intelligence technologies and infrastructure, including deep learning, machine learning and large language models and generative AI.

ARTIFICIAL INTELLIGENCE Emerging Frontiers in Urology: Harnessing Big Data, Machine Learning, and Artificial Intelligence

By: Danly Omil-Lima, MD, Fox Chase Cancer Center, Philadelphia, Pennsylvania; Alberto Castro Bigalli, MD, Fox Chase Cancer Center, Philadelphia, Pennsylvania; Austin Thompson, BA, Case Western Reserve University School of Medicine, Cleveland, Ohio; Shivaram Cumarasamy, MD, Fox Chase Cancer Center, Philadelphia, Pennsylvania; David Sheyn, MD, University Hospitals Cleveland Medical Center, Ohio | Posted on: 19 Jan 2024


In recent years, artificial intelligence (AI) technologies like natural language processing (NLP) have made astonishing advances. AI systems can now hold fluent conversations, summarize complex texts, translate between languages, and generate new synthetic content.1 Fields like NLP, computer vision, robotics, and machine learning (ML) are rapidly evolving to emulate and augment human capabilities on an unprecedented scale. These breakthroughs demonstrate the transformative potential of AI across industries, including health care.

The mainstream visibility of ChatGPT, a prompt-based NLP application created by OpenAI, introduced the public to AI capabilities previously confined to the realm of science fiction. Several other AI platforms have since been introduced by major developers (eg, Google Bard, Microsoft Copilot), with the goal of integrating AI into everyday tasks. This has sparked exponentially increasing interest in AI (Figure) and how it might be applied across different domains—including health care.2

Figure. United States Google search trends for “artificial intelligence” over time.

As a field, urology has a long history of early adoption of cutting-edge developing technologies. With a wealth of patient data and complex diagnostic/treatment challenges, urology is fertile ground for applying big data, ML, and AI tools.

We stand at an exciting frontier where these tools can reshape patient care, research, surgery, and all realms of urologic practice.

Evolution of Big Data

Health care informatics has evolved significantly over the past few decades. In the 1990s, the shift from paper to electronic health records (EHRs) digitized patient information and laid the foundation for big data analytics. However, early EHR systems had limited capacity to aggregate data across health care networks.

In the 2000s, investments in health information technology infrastructure and standards (like HL7 [Health Level Seven International]) improved interoperability, allowing larger datasets to be pooled for analysis.3 The HITECH (Health Information Technology for Economic and Clinical Health) Act of 2009 further spurred EHR adoption, such that big data analytics started gaining traction in health care, with early applications in areas like predictive modeling, population health management, and genomic research.4

Today, big data in health care are derived from several sources including EHRs, insurance claims, clinical trials, genomics repositories, medical imaging, and patient-reported outcomes. Cloud computing has enabled scalable data storage and analysis.

These massive, complex datasets potentially hold invaluable insights if analyzed properly. This was never more evident than during the COVID-19 pandemic, which highlighted the value of real-time public health data surveillance.5

AI and ML

The concept of AI refers to the creation of computers capable of emulating human cognitive functions, such as problem-solving, perception, and language comprehension. ML is a subset of AI that focuses specifically on the ability of computers to learn from data across iterations. By developing ML algorithms, computers improve their performance on a task over time.

For example, one might develop an image-based ML model to detect high-grade cancer in a biopsy specimen. Such a model is first trained on specimens where the presence or absence of high-grade cancer is already known. During training, a penalty is placed on features of the model that lead to worse performance on an arbitrary metric (for example, percent of correctly classified specimens). The model’s end goal is to minimize this penalty (known among data scientists as the cost function).6 With the data (in this case, digital images of pathology slides) fed into the model in batches, the model learns to ignore features leading to worse performance. For example, a good model would learn to ignore the periphery of digitized biopsy image—which is sure to comprise empty pixels.

By performing this iterative fine-tuning and selecting features leading to the best performance, the resulting final ML model is better at detecting high-grade cancer at the end of training than it was in the beginning of training. The final model is then used to classify unknown specimens with some margin of error.

Such specialized analytic techniques are in development and can help to reveal patterns and correlations within health care data.7 This can improve clinical decision-making, identify high-risk patients, discover disease subgroups, develop personalized medicine, track population health trends, and much more to transform the health care system.

Additionally, AI language models can provide patients with health information and education in easily understandable language through interactive conversational agents. Personalized instructions, reminders, and motivational messaging can also improve patient adherence to medical treatment.8 Analyzing patient feedback and reviews with NLP can further enhance patient satisfaction.

For physicians and health care workers, AI can automate clinical documentation and note-taking, and streamline administrative tasks such as the generation of billing codes from operative note dictations.9 NLP can also allow physicians to perform faster and more targeted literature searches for evidence-based recommendations, which can support (but not replace) clinical decision-making. Furthermore, AI can potentially simplify the creation of patient education materials and discharge summaries through automated text and data extraction.

Emerging AI technologies show promise as invaluable assistants in health care. They have the potential to streamline workflows, inform decisions, personalize care, and boost outcomes. However, human judgment remains indispensable. AI should not replace physicians, but rather assist them.

Responsible development and rigorous validation are crucial to safely translate these tools to clinical care. Physicians must proactively lead the way in adopting AI, ensuring applications are thoughtfully designed and tested to avoid pitfalls like algorithmic bias, data drift, and improper automation. The data used to develop AI must be carefully assessed for quality and representativeness across diverse patient populations. With an ethical, evidence-based approach, physicians can manage risks as they reap benefits from AI’s tremendous potential. If any specialty can unleash the power of data-driven AI in medicine responsibly, it is urology. But this requires proactive efforts to validate AI systems and match their use to appropriate clinical contexts. Human oversight and stewardship are key.

  1. Dave T, Athaluri SA, Singh S. ChatGPT in medicine: an overview of its applications, advantages, limitations, future prospects, and ethical considerations. Front Artif Intell. 2023;6:1169595.
  2. Patel V, Callahan A, Andersen K, Arabandi S, Topol EJ. Implications of ChatGPT and other large language models for health care. JAMA. 2023;
  3. Bender D, Sartipi K. HL7 FHIR: an agile and RESTful approach to healthcare information exchange. In: Proceedings of the 26th IEEE International Symposium on Computer-Based Medical Systems; 2013:326-331.
  4. Mennemeyer ST, Menachemi N, Rahurkar S, Ford EW. Impact of the HITECH act on physicians’ adoption of electronic health records. J Am Med Inform Assoc. 2016;23(2):375-379.
  5. Ong E, Wong MU, Huffman A, He Y. COVID-19 coronavirus vaccine design using reverse vaccinology and machine learning. Front Immunol. 2020;11:1581.
  6. LeCun Y, Bengio Y, Hinton G. Deep learning. Nature. 2015;521(7553):436-444.
  7. Khene ZE, Kammerer-Jacquet SF, Bigot P, et al. Clinical application of digital and computational pathology in renal cell carcinoma: a systematic review. Eur Urol Oncol. 2023;10.1016/j.euo.2023.10.018.
  8. Singareddy S, Sn VP, Jaramillo AP, et al. Artificial intelligence and its role in the management of chronic medical conditions: a systematic review. Cureus. 2023;15(9):e46066.
  9. Kim JS, Vivas A, Arvind V, et al. Can natural language processing and artificial intelligence automate the generation of billing codes from operative note dictations?. Global Spine J. 2023;13(7):1946-1955.