This letter comments on Nagumo et al.'s study evaluating 1024-matrix reconstruction for intracranial perforating artery visualisation in 64-slice cerebral CTA, affirming its cost-effective value for standard CT scanners (https://doi.org/10.1002/jmrs.70055). We endorse the key finding that 1024-matrix improves small artery detection via higher sampling density, while highlighting the need for cross-vendor validation, workflow impact quantification, and clinical outcome assessment. This simple post-processing strategy is highly scalable, and further multicentre studies are warranted to confirm its universal utility. We write to express our sincere appreciation for the original article by Nagumo et al. 1 published in your journal, which investigates the impact of 1024-matrix reconstruction on the visualisation of intracranial perforating arteries in cerebral computed tomography angiography (CTA) using a conventional 64-slice CT scanner. This study addresses a critical clinical challenge in preoperative neurosurgical planning—the accurate visualisation of small perforating arteries such as the anterior choroidal artery (AChA) and posterior thalamoperforating artery (PTPA)—and provides a practical, cost-effective optimisation strategy that requires no specialised CT equipment. The findings are of great clinical value, especially for medical institutions with only standard 64-slice CT scanners, and fill an important gap in the literature regarding matrix size optimisation for cerebral CTA on conventional energy integrating detector (EID) systems. Nagumo et al.'s 1 dual-component study (phantom and clinical) design is rigorous, and the combination of physical image quality analysis (Task Transfer Function TTF, Noise Power Spectrum NPS) and quantitative/qualitative clinical evaluation provides a comprehensive assessment of the 1024-matrix reconstruction. We particularly agree with the core conclusion that increasing the matrix size from 512 to 1024 pixels enhances the visual detection of small perforating arteries via improved sampling density, despite minimal changes in physical image quality metrics. This finding aligns with previous research on matrix size optimisation in chest CT and CT colonography 2, 3, and extends this principle to cerebral CTA—a field where ultra-high-resolution CT has shown promise but remains inaccessible to many institutions 4. The significant improvement in high-quality visualisation of the PTPA (from 17.07% to 39.03%) is a notable clinical outcome, as this vessel is critical for thalamic perfusion and its injury is associated with severe postoperative neurological deficits 5. While the study is exemplary in its design and execution, it also identifies several key limitations, which we believe warrant further discussion and future investigation, and we offer three supplementary perspectives for clinical practice and subsequent research: First, the study was conducted on a single-vendor 64-slice CT scanner (GE Healthcare Revolution Ascend) with filtered back projection (FBP) reconstruction, and the results may not be generalisable to other CT platforms or reconstruction algorithms. Iterative reconstruction (IR) and deep learning-based reconstruction (DLR) are now widely used in clinical practice to reduce image noise while preserving spatial resolution 6, and combining 1024-matrix reconstruction with IR/DLR may yield additional benefits for perforating artery visualisation—especially in patients with high body mass index or motion artefacts, who are at higher risk of degraded image quality. Future multicentre studies should include CT scanners from different manufacturers and various reconstruction algorithms to validate the universality of the 1024-matrix optimisation strategy. Second, the study notes that 1024-matrix reconstruction quadruples pixel count and data volume, which may impact picture archiving and communication system (PACS) storage, network traffic and post-processing workload, but does not quantify these workflow effects. For routine clinical implementation, it is essential to assess practical parameters such as reconstruction time, MIP/VR post-processing duration and data transfer speed. In addition, for institutions with limited PACS storage capacity, a targeted 1024-matrix reconstruction strategy (applying only to clinically critical intracranial regions) may be a balanced solution, as suggested by the authors, and its clinical efficiency deserves further evaluation. Third, the study focuses on image visualisation quality but does not assess the translation of improved perforating artery visualisation into clinical outcomes. As the authors note, perforating artery injury during aneurysm clipping or other neurosurgical procedures can lead to significant postoperative deficits 5, 7, and it is critical to determine whether 1024-matrix reconstruction improves surgical planning accuracy, reduces the rate of intra-operative vessel injury, or improves patient neurological outcomes. Prospective cohort studies with long-term follow-up are needed to establish the clinical impact of this optimisation technique beyond image quality. Finally, we wish to emphasise the clinical applicability of Nagumo et al.'s findings. In many regions, conventional 64-slice CT scanners remain the primary modality for cerebral CTA, and the ability to enhance perforating artery visualisation through a simple post-processing adjustment (matrix size)—without additional radiation exposure, contrast medium, or specialised equipment—makes this strategy highly scalable. This is particularly important for preoperative evaluation of patients with intracranial aneurysms or occlusive cerebrovascular disease, where precise identification of small perforating arteries is essential for safe surgical intervention 8. In conclusion, Nagumo et al.'s study provides a valuable and practical optimisation method for cerebral CTA on conventional 64-slice CT scanners, and we commend the authors for their important contribution to the field of medical radiation sciences. We hope that subsequent research will address the aforementioned limitations and further validate the clinical utility of 1024-matrix reconstruction, and we believe this technique has the potential to become a standard post-processing approach for cerebral CTA in routine clinical practice. Yang Liu: conceptualization, writing – original draft, writing – review and editing. Wei Wu: conceptualization, writing – review and editing, supervision, project administration. The authors have nothing to report. The authors have nothing to report. The authors have nothing to report. The authors have nothing to report. The authors declare no conflicts of interest. This article is linked to Nagumo et al. papers. To view this article, visit https://doi.org/10.1002/jmrs.70055. The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.
Liu et al. (2026) studied this question.