Yoshiyuki Watanabe

The timely recognition of clinically important radiology findings is critical for patient safety, yet in Japan, no standardized multi-tiered system exists to classify the importance of radiology reports. This study aimed to propose a novel classification system for radiology reports based on the clinical importance of the findings called “Report Importance Category version 2 (RICv2)” and evaluate inter-radiologist agreement in its application. In RIC, reports are categorized into five categories; 1: “No/minor findings”, 2: “Maintain routine follow-up”, 3: “Closer follow-up required”, 4: “Action required within one month”, and 5: “Immediate action required”. Importance is defined not only by the severity of the patient, but also by the novelty and specificity of the findings. Five board-certified radiologists and five radiology residents reviewed and categorized 514 reports collected from Japan Medical Image Database (J-MID), and the inter-rater agreement was evaluated using unweighted and quadratic weighted Fleiss’ κ, and entropy analysis was applied to identify reports prone to disagreement. Among the 514 reports (408 CT, 106 MR), the distribution of final categories determined by majority consensus was RIC1 = 75 (14.6%), RIC2 = 327 (63.6%), RIC3 = 66 (12.8%), RIC4 = 34 (6.6%), and RIC5 = 12 (2.3%). Interobserver agreement was moderate, with unweighted Fleiss’ κ = 0.454, 0.478, and 0.449, for all radiologists, board-certified radiologists, and radiology residents, respectively, whereas substantial with weighted κ = 0.646, 0.669, and 0.662. Entropy analysis revealed that reports with multiple differentials or findings of uncertain clinical urgency tended to generate the greatest disagreement. RICv2 provides a structured, reproducible framework for classifying the importance of free-text radiology reports. Experience from other systems, such as Prostate Imaging Reporting and Data System (PI-RADS), suggests that refinement could further improve reproducibility, establishing RICv2 as a promising foundation for clinical and research applications.

[This corrects the article DOI: 10.1007/s13691-025-00808-4.].

To automate the notification workflow for actionable computed tomography (CT) reports, we investigated model configurations that maintained detection performance even with a very low positive rate. We assembled a dataset of 1000 head CT reports and compared the performance of an encoder model (ModernBERT-ja) with that of a decoder model (Llama-3-ELYZA-JP-8B). The encoder underwent supervised fine-tuning (SFT) using multiple loss functions, while the decoder was evaluated through few-shot prompting, zero-shot chain-of-thought (CoT), and SFT. We systematically decreased the positive rate in the training data to 14.4%, 10%, 5%, and 2% to assess model robustness. The encoder SFT with class-weighted focal loss achieved an F1 score of 0.870 at a positive rate of 14.4%, surpassing the decoder’s best few-shot configuration (10 demonstrations, label0_first; F1 = 0.717). However, encoder performance declined substantially when the positive rate dropped below 5%. The decoder improved with only a few demonstrations, and the ordering of positive demonstrations played a crucial role in few-shot prompting. Applying SFT to the decoder raised the F1 score to 0.853, approaching that of the encoder but incurring higher training expenses. Neither architecture maintained detection capability when the positive rate was 2%. Zero-shot CoT performed less effectively than few-shot prompting under all imbalanced conditions. With sufficient positive cases, encoder SFT yields the best performance. When sampling the minority class is difficult, decoder few-shot prompting without additional training is a practical alternative. These results guide the choice between encoder fine-tuning and decoder prompting for highly imbalanced radiology-report triage of actionable findings.

A controllable phantom is needed to evaluate MR measurements of slow, pulsatile cerebrospinal fluid motion. We developed a ventricular phantom in which a pulsatile pump acts on a closed cavity containing an elastic lateral ventricle model, such that wall deformation drives flow through the downstream ventricular pathway. Four-dimensional flow MRI was performed at 3T and 50 beats/min with nominal bidirectional pump displacement volumes of 0.4–2.0 mL/cycle, corresponding to 1-way syringe settings of 0.2–1.0 mL. Flow was visualized at the foramen of Monro, cerebral aqueduct, bilateral foramina of Luschka, and foramen of Magendie. Regional bidirectional stroke volume and mean velocity increased with pump setting, with the highest velocity in the aqueduct. An exploratory analysis of central-aqueduct at the middle setting, performed in 3 sessions on separate days, yielded a mean cycle-averaged speed of 3.75 ± 0.34 cm/s with a between-session coefficient of variation of 9.1%. This wall-deformation-driven phantom provides a controllable platform for studying pulsatile cerebrospinal fluid motion, with potential applications in the validation of flow-sensitive imaging techniques. Absolute measurement accuracy, physiological equivalence, and reproducibility across all pump settings remain to be established.

Purpose: Imaging studies of cerebrospinal fluid (CSF) dynamics are gaining attention. While 17O-labeled saline and gadolinium (Gd)–based contrast agents have been applied to investigate CSF dynamics using MRI in preclinical settings, studies in non-human primates remain limited. This study evaluated the feasibility of visualizing CSF kinetics in cynomolgus monkeys using these 2 tracers.Methods: Four female cynomolgus monkeys underwent dynamic MRI. In separate sessions, 17O-labeled saline and gadoterate meglumine (Gd-DOTA) were administered intrathecally into the cisterna magna. For the 17O-labeled saline protocol, T2-weighted images were acquired every minute for 20 min. For the Gd-DOTA protocol, T1-weighted images were acquired at 10 min, 30 min, 1 hr, 4 hr, and 24 hr after administration. Signal intensities (SIs) within CSF and parenchymal regions were analyzed.Results: The tracers demonstrated distinct kinetic profiles. Following 17O-labeled saline injection, the % change in SI in the cisterna magna peaked rapidly at 2–3 min and subsequently propagated caudally along the cervical spinal canal without detectable parenchymal uptake. In contrast, after Gd-DOTA administration, the % change in SI in the cisterna magna reached its maximum at 1–4 hr, with widespread distribution into ventricles, cisterns, and spinal canal. Gd-DOTA clearance via the cribriform plate was evident. Parenchymal enhancement was observed in the cerebellum, pons, and cingulate gyrus with greater signal change in superficial compared to sub-superficial layers. Marked enhancement was also observed in gray matter surrounding the spinal central canal at 4 hr.Conclusion: This pilot study demonstrated the feasibility of assessing CSF dynamics in a non-human primate model using both 17O-labeled saline and Gd-DOTA. While 17O-labeled saline enabled rapid visualization of bulk CSF flow, Gd-DOTA provided insights into slower CSF distribution and parenchymal transport. These findings provide a translational platform for future investigations into age- and disease-related alterations in CSF dynamics.

Background: Distal side‐to‐side anastomosis is an effective technique for coronary artery bypass grafting; however, restrictive suturing may adversely affect graft patency. In this study, we evaluated ex vivo side‐to‐side models through numerical simulations of fluid dynamics to compare varying degrees of anastomotic stenosis and predict their hemodynamic consequences. Methods: A carotid artery was grafted onto a porcine coronary artery in a side‐to‐side configuration, with the vessels filled with liquid silicone. Ultra‐high‐resolution computed tomography was employed to acquire the reference side‐to‐side image, and stenotic models were developed with longitudinal shortening and bilateral narrowing at 25%, 50%, and 75% stenosis, alongside 90% coronary artery stenosis. Hemodynamic analysis was conducted using computational fluid dynamics simulations to calculate streamlines, wall shear stress, and oscillatory shear index. Results: In the reference model, graft inflow impinged on the coronary artery floor, generating a recirculating vortex at the heel and blind end of the graft while maintaining laminar flow. Longitudinal stenosis resulted in a steeper graft flow angle, whereas bilateral stenosis induced helical flow along the lateral arterial wall, exacerbating with increased stenosis. At 75% longitudinal or 50% bilateral stenosis, a region of flow separation was observed, characterized by low wall shear stress and high oscillation regions distal to the toe of the anastomosis. Conclusions: Computational fluid dynamics simulations indicate that anastomotic stenosis with longitudinal length ratios less than 1.6 and bilateral width ratios less than 0.73 relative to coronary diameter is at risk of intimal hyperplasia leading to graft failure, whereas larger ratios suggest acceptable hemodynamics. Future research should investigate long‐term clinical outcomes associated with suboptimal side‐to‐side anastomotic construction.

BACKGROUND AND OBJECTIVES: The pathophysiology of obstructive hydrocephalus caused by cerebral aqueduct obstruction or stenosis, as well as communicating hydrocephalus with oscillating cerebrospinal fluid (CSF) flow and vortices in the third ventricle (3V), remains unclear. In addition, clear criteria for selecting between ventriculoperitoneal shunting (VPS) and endoscopic third ventriculostomy (ETV) have yet to be established. This study aimed to investigate the utility of 4-dimensional flow MRI for assessing CSF dynamics to enhance our understanding of hydrocephalus pathophysiology and guide treatment selection. METHODS: Using 4-dimensional flow MRI (velocity encoding = 5 cm/s), we evaluated CSF dynamics before and after VPS and ETV in patients with obstructive and communicating hydrocephalus, focusing on velocity vectors, streamlines, and vorticity. RESULTS: In obstructive hydrocephalus due to cerebral aqueduct obstruction, we initially expected a complete absence of CSF circulation in the 3V and cerebral aqueduct before ETV. However, slow flow was observed around the foramen of Monro and even within the obstructed aqueduct. After ETV, rapid CSF flow with vortex formation was observed in the 3V. In obstructive hydrocephalus, while VPS reduced ventricular size, it did not enhance CSF movement in the 3V. In communicating hydrocephalus with third ventricular floor ballooning, rapid CSF flow and abnormal vortices in the 3V persisted after ETV; however, these abnormalities were mitigated by VPS. CONCLUSION: These findings suggest that both CSF stagnation and abnormal vortices in the 3V contribute to the pathophysiology of obstructive and communicating hydrocephalus. Therefore, treatment strategies should prioritize interventions that optimize CSF flow dynamics.

Extramammary Paget’s disease (EMPD) is a rare intraepithelial adenocarcinoma. EMPD can be associated with an underlying malignancy of an adjacent organ, in which case the skin lesion is known as the pagetoid spread (PS) of the underlying tumor. PS, also known as secondary EMPD, has a worse prognosis than primary EMPD. The standard treatment is surgery, but the efficacy of nonsurgical treatments is unknown in inoperable cases. We present 2 cases of anal canal adenocarcinoma with PS that were treated with definitive concurrent chemoradiation therapy (CCRT). Although both cases experienced recurrence outside of the radiation field, and one died, anal function was preserved in both. In conclusion, CCRT may be a viable treatment option for inoperable anal canal cancer with PS.