Document Type : Editorial
Authors
1 Quantitative Magnetic Resonance Imaging and Spectroscopy Group, Research Center for Molecular and Cellular Imaging, Tehran University of Medical Sciences, Tehran, Iran.
2 Department of Medical Physics and Biomedical Engineering, School of Medicine, Tehran University of Medical Sciences, Tehran, Iran.
3 Department of Radiology Technology, School of Allied Medical Sciences, Shahid Beheshti University of Medical Sciences, Tehran, Iran
Abstract
Highlights
Keywords
Main Subjects
Editorial: Artificial Intelligence (AI) as a multidisciplinary approach is the science and engineering of building intelligent machines capable of performing tasks that typically require human intelligence. The emergence of fast-growing AI-related techniques and technologies revolutionized many businesses, including healthcare management. In other words, the practice of medicine is rapidly changing in all fields of diagnosis, treatment, and prognosis. As the number of involved parameters and factors in medical decision-making processes is vast, conclusive decision-making becomes a more complex task that needs expertise, focus, and attention. The advent of AI-based tools compensates for the lack of focus and attention in vital decision-making procedures. In addition, with the emergence of advanced radiology, genetic, proteomic, metabolic, and pathological data, AI-based computational methods are mandatory to extract meaningful and useful patterns from these complex unstructured data to support physicians in decision-making.
Furthermore, AI allows for eliminating some expensive and life-threatening procedures such as unnecessary biopsies in different stages of diseases. Urology as a broad major in medicine is not an exception. Therefore, many practices, such as urolithiasis, pediatric urology, uro-oncology, uropathology, uroradiology, and renal transplant, are potentially changing using advanced AI tools and techniques (1, 2, 3, 4). Among great AI techniques, machine learning (ML) is gaining more attention for the modeling and analysis of medical data. Moreover, Deep Learning (DL) as a branch of ML can perform modeling as an end-to-end learning machine without any feature engineering, especially in cases with complex unstructured data (5).
AI models, usually embedded in computer-aided detection/diagnosis systems, can diagnose and detect different urological disorders and diseases via the knowledge they learned from historical medical records. These systems support urologists in complex and life-changing decision-making processes, especially in diagnosing urological cancers. Prostate cancer is one of the most common urological cancer that beneficiary from AI to be precisely diagnosed. AI can analyze magnetic resonance images of the prostate gland to contour the gland, characterize its tissue, and grade cancer (6). Accurate tumor detection, grading, and staging in bladder cancer are possible by analyzing non-invasive imaging data without the need for any invasive biopsy (7, 8, 9). Histopathological images of urological organs can be automatically analyzed to improve diagnostic accuracy in cancer detection and grading (10). More interestingly, AI allows discriminating complex clinical situations, e.g., differentiating Gleason score six from 7; and distinguishing clear cell Renal Cell Carcinoma (RCC) and papillary RCC with a promising accuracy (11, 12).
AI facilitates the diagnosis of urological disease after the appearance of symptoms and allows early diagnosis before the symptoms start or the disease spreads. This allows high-risk warning people and designing different screening programs in the very early stages of the disease. Early detection of RCC (13) and kidney stones (14) are examples of urological disorders which AI can perform.
The AI techniques used in the diagnosis processes can be transferred to the treatment processes to prevent over/under treatment. In other words, accurate, personalized treatment planning based on patient-specific data can be suggested by AI models. Accurate volume measurement and localization of both urological organs and urological tissue abnormalities allow for monitoring of disease progression over time. Moreover, calculating the exact amount of radiation dosage and drug selection in the treatment procedure is determined by AI models.
AI techniques can also be employed to predict urological conditions' results accurately. Precise prediction of treatment response, recurrence, and survival allows urologists to know their practices' results in advance.
Conclusion
As a complementary technology, AI is attempting to change urology practice nowadays, thanks to health information services. AI-powered urologists can rapidly make complex and non-linear medical decisions in cases of noisy, complex, heterogeneous, and incomplete data. Therefore, false diagnoses and other medical errors will be reduced. AI techniques attempt to explain their decision (how the disease patterns and correlations are found) to support urologists more rationally. Moreover, high quality, high quantity, variety, and density of medical data improve the performance of AI models. Finally, it is highly recommended that urologists, uro-radiologists, uro-pathologists, and uro-oncologists use AI techniques to improve their performance in challenging diagnosis, prognosis, and treatment planning conditions.
Authors’ contribution:
HH designed the work and wrote the manuscript. AAA and HSR revised the manuscript critically. All authors reviewed and edited the manuscript and approved the final version of the manuscript.
Acknowledgments
None.
Conflict of interest
The authors declare that there is no conflict of interest.
Funding
There is no funding.
Ethical statement
Not applicable.
Abbreviations
AI Artificial intelligence
RCC Renal cell carcinoma