Objective: Due to limitations in current imaging technologies detecting subtle cardiac microstructural changes that can lead to sudden cardiac death is a significant clinical challenge. To address this problem, we developed a forward-viewing optical coherence tomography (OCT) endoscope for the detection of relevant cardiac microstructures in the subendocardium, including Purkinje fibers, scar tissue, surviving myocytes, and adipose tissue. Methods: An endoscopic probe based on the scanning fiber principle was developed for OCT measurements in contact. The probe was evaluated in freshly excised ovine hearts exhibiting chronic myocardial infarction. Relevant regions within the cardiac chamber were measured, and distinctive microstructures were identified, characterized, and subsequently corroborated using Masson's trichrome staining. The volumetric imaging data were used to train a convolutional neural network (CNN) to detect Purkinje fibers, enabling the reconstruction of their 3D morphology. Results: We were able to distinguish between healthy myocardium, fibrotic remodeling, and critical elements of the cardiac conduction system. Our findings demonstrate the capability of this technology to provide detailed images of cardiac microstructures in large mammal hearts. Conclusion: A novel application of forward-viewing endoscopic OCT in cardiology is demonstrated by visualizing cardiac microstructures within the subendocardium at depths accessible by optical imaging modalities. Significance: By enhancing visualization at the cellular level, this method may contribute to a better understanding of cardiac physiology and pathology, potentially extending future diagnostic and therapeutic strategies.