Artificial intelligence in renal pathology: Current status and future

Abstract

Renal biopsy pathology is an essential gold standard for the diagnosis of most kidney diseases. With the increase in the incidence rate of kidney diseases, the lack of renal pathologists, and an imbalance in their distribution, there is an urgent need for a new renal pathological diagnosis model. Advances in artificial intelligence (AI) along with the growing digitization of pathology slides for diagnosis are promising approach to meet the demand for more accurate detection, classification, and prediction of the outcome of renal pathology. AI has contributed substantially to a variety of clinical applications, including renal pathology. Deep learning, a subfield of AI that is highly flexible and supports automatic feature extraction, is increasingly being used in multiple areas of pathology. In this narrative review, we first provide a general description of AI methods, and then discuss the current and prospective applications of AI in the field of renal pathology. Both diagnostic and predictive prognostic applications are covered, emphasizing AI in renal pathology images, predictive models, and 3D in renal pathology. Finally, we outline the challenges associated with the implementation of AI platforms in renal pathology and provide our perspective on how these platforms might change in this field.

Urinary podocyte markers of disease activity, therapeutic efficacy, and long-term outcomes in acute and chronic kidney diseases

“…This methodology, together with machine learning and artificial intelligence, can enable the automated quantitation of kidney biopsy glomerular volume, podocyte number, size, and density, which can provide clinicians with the quantitative tools required to effectively prevent progression. [26].…”

Magnetic Resonance Imaging in Atherosclerotic Renal Artery Stenosis: The Update and Future Directions from Interventional Perspective

“…Radiomics is a recent post-processing method for computer image processing and big data mining. Driven by advances in pattern recognition and image processing techniques, radiomics provides access to more objective and quantitative features that are difficult to identify with the naked eye compared to traditional imaging phenotypic features [ 61 , 62 ]. The foundation of radiomics and deep learning based on renal functional MRI is that chronic kidney disease is a group of heterogeneous diseases that lead to changes in renal function.…”

Digital Pathology: Transforming Diagnosis in the Digital Age

“…Expedited second opinions: Telepathology demonstrates exceptional proficiency in facilitating prompt access to second opinions, a crucial component in intricate cases. In situations where the initial diagnosis is uncertain, remote consultation allows for a prompt and thorough evaluation by specialists [ 27 ]. The efficient nature of this process has the potential to significantly decrease the duration of time that patients experience uncertainty, while also simplifying the process of developing accurate treatment plans.…”

“…The initiatives undertaken by The Montefiore, the Royal Philips-AUMC partnership, and Massachusetts General Hospital exemplify the utilization of technology-driven solutions in addressing complex medical challenges. These examples showcase the transformative potential of such approaches [ 27 ].…”

“…Patient data privacy, consent, responsible AI use, regulatory approvals, and liability concerns form the ethical and legal scaffolding upon which this paradigm rests. In the backdrop of these considerations, digital pathology emerges as a transformative force that transcends geographical barriers, empowers diagnostics through AI-driven insights, and enhances patient care through more accurate diagnoses [ 27 ]. Remote consultations foster collaborations among experts worldwide, heralding a future where medical knowledge transcends borders.…”