Heart failure (HF) remains a significant global health challenge, with heart transplantation being the ultimate therapeutic option for many patients with end-stage disease [1,2]. Despite the life-saving potential of cardiac transplantation, post-transplant complications, such as cardiac allograft rejection (CAR), continue to be a critical concern for transplant recipients [3]. Endomyocardial biopsies (EMBs) currently serve as the gold standard for heart transplant (HTx) rejection surveillance [2,4]. However, limitations in diagnostic accuracy, invasiveness, and interobserver variability warrant the exploration of novel approaches to improve the diagnostic process [5], [6], [7].

Recent advancements in digital pathology (DP) suggest that machine learning (ML) holds promise in addressing some of these limitations and enhancing our understanding of CAR [8,9]. ML techniques, for instance, have showcased their capability to augment human analysis by capturing and quantifying subtle patterns in medical images that might be challenging to discern with the naked eye [8,[10], [11], [12]]. This can lead to improved diagnostic outcomes and the uncovering of novel relationships and features relevant to transplant pathology.

In this article, we offer an updated perspective on cardiac transplant diagnostics by reviewing studies that employ digital or computational pathology (CPATH) methods, with a specific emphasis on ML techniques. Our analysis underscores the potential of artificial intelligence (AI) to provide more objective evaluations and reduce interobserver variability. We incorporate findings from the latest research, especially emphasizing the advancements in machine learning and digital pathology. The advantages, limitations, and potential impact of these computer-based methods on clinical practice are discussed in detail. Furthermore, we address current methodological challenges, emphasizing the imperative of interdisciplinary collaboration for overcoming them. Overall, our comprehensive approach seeks to illuminate the continuously evolving domain of HTx and CAR diagnostics.

We conducted a literature review on MEDLINE via PubMed from January to April 2023. Our search terms included combinations of: “heart transplantation”, “cardiac transplantation”, “endomyocardial biopsy”, “rejection”, and related terms, with “machine learning”, “artificial intelligence”, “digital pathology”, “computational pathology”, and related terms. We focused on English-language human studies. Following the retrieval of initial articles, the research team thoroughly reviewed full-text articles for relevance and appropriateness. Further articles were identified through the use of the “related articles” function on MEDLINE and a manual search of the references lists of articles found through the original search. This multi-author effort ensured a comprehensive and thorough search.