KIKoji Iihara

Koji Iihara

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Koji Iihara has a total of 844 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Prediction of Stroke Outcome After Endovascular Thrombectomy From Clinical Characteristics: A Machine Learning Approach

    by Tomohide Yoshie, Sohei Yoshimura, Makino Sakuraba, Shinichi Wada, Kaori Miwa, Junpei Koge, Junji Miyazaki, Toshiichi Watanabe, Kazuo Kitazawa, Nobuyuki Sakai, et al.

    Background: Predicting outcomes after endovascular thrombectomy (EVT) for acute ischemic stroke is challenging. This study aimed to develop a machine learning model to predict outcomes after EVT using only clinical characteristics available shortly after hospital arrival, excluding imaging, laboratory, and angiographic data. Methods: We analyzed data from the JSDB (Japan Stroke Data Bank), a multicenter, prospective registry of patients with acute stroke. A total of 4353 patients who underwent EVT between 2016 and 2020 were included, with those from 2016 to 2019 assigned to the derivation cohort and those from 2020 to the validation cohort. The model was developed using 132 routinely collected clinical variables that can be rapidly obtained on patient arrival. Feature selection was performed using the Boruta algorithm, and the model was built using CatBoost. The primary outcome was a modified Rankin Scale (mRS) score of 0 to 2 at discharge. Results: Among 4353 patients treated with EVT, the derivation cohort included 2764 patients and the validation cohort, 1589 patients. In these cohorts, 973 (35.2%) and 547 (34.4%) achieved mRS 0 to 2, respectively. Boruta selected 47 predictors. CatBoost achieved an area under the receiver operating characteristic curve of 0.744 (95% CI, 0.710–0.776) in the derivation test set and 0.781 (95% CI, 0.755–0.803) in the validation cohort. Validation areas under the receiver operating characteristic curves were 0.792 (95% CI, 0.765–0.815) for mRS 0 to 1, 0.766 (95% CI, 0.744–0.787) for mRS 0 to 3, and 0.680 (95% CI, 0.631–0.729) for mRS 6. Conclusions: Our machine learning model predicted outcomes following EVT using only admission clinical data. This imaging‐independent machine learning approach may support early prognostication and facilitate discussions of expected clinical courses.

  • Open Access

    Impact of Aneurysm Size and Flow Diverter Use on Outcomes After Endovascular Therapy of Unruptured Paraclinoid Aneurysms

    by Kazuhiro Fujishima, Kanta Tanaka, Tetsu Satow, Toshiyuki Fujinaka, Katsunori Asai, Tomohiko Ozaki, Jun C. Takahashi, Koji Iihara, Hirotoshi Imamura, Akira Ishii, et al.

    BACKGROUND: Flow diverters (FDs) are increasingly used for unruptured paraclinoid aneurysms, especially large or giant ones. However, their benefit over conventional endovascular therapy remains unclear. This study evaluated outcomes after endovascular therapy of unruptured paraclinoid aneurysms, with aneurysm size as the primary exposure, and further examined the modifying effect of FD use. METHODS: We analyzed data of patients with unruptured paraclinoid aneurysms from the Japanese Registry of Neuroendovascular Therapy 4, a nationwide, multicenter registry enrolling patients who underwent endovascular therapy in 166 Japanese institutions from 2015 to 2019. The primary outcome was worsening in the modified Rankin Scale score ≥1 at 30 days. Secondary outcomes included ischemic and hemorrhagic complications. Outcomes were compared by aneurysm size (P =0.035), though not significant after multivariable adjustment (adjusted odds ratio, 1.26 [95% CI, 0.87–1.84]). Hemorrhagic complication rates were similarly low across groups. There was no significant difference in modified Rankin Scale score worsening by FD use, regardless of aneurysm size. FD use was associated with significantly fewer ischemic complications in aneurysms ≥10 mm (3.2% versus 10.1%, P P =0.88; P for interaction=0.03). CONCLUSIONS: Endovascular embolization of unruptured paraclinoid aneurysms ≥10 mm was associated with a higher morbidity compared with aneurysms <10 mm. FD use may reduce ischemic complications in larger aneurysms; however, this benefit was not clearly reflected in 30-day functional outcomes.

  • Device selection and clinical outcomes of adjunctive techniques for ruptured intracranial aneurysms: Insights from the JR-NET4 nationwide registry

    by Kouhei Nii, Toshio Higashi, Ritsurou Inoue, Hisatomi Arima, Hirotoshi Imamura, Nobuyuki Sakai, Koji Iihara, Akira Ishii, Yuji Matsumaru, Chiaki Sakai, et al.

    Purpose Adjunctive techniques are frequently required in endovascular treatment (EVT) for ruptured intracranial aneurysms (RIAs) with complex morphology. However, comprehensive nationwide data on technique selection and its relationship with outcomes remain limited. We aimed to clarify anatomical and clinical factors associated with technique selection and to evaluate short-term outcomes using a large nationwide registry. Methods The Japanese Registry of Neuroendovascular Therapy (JR-NET) 4 included over 4,000 RIAs treated between 2015 and 2019. After exclusions, 3,600 patients treated with balloon-assisted coil embolization (BAC), double-microcatheter coil embolization (DMC), or stent-assisted coil embolization (SAC) were analyzed. Multinomial logistic regression identified factors associated with treatment selection, and multivariable logistic regression evaluated predictors of favorable outcome (modified Rankin Scale; mRS 0–2) and mortality at 30 days. Results Wide-neck and large RIAs were associated with DMC and SAC, while bifurcation RIAs were more often treated with DMC. BAC remained the most frequently used technique. Thirty-day outcomes were primarily determined by World Federation of Neurosurgical Societies (WFNS) grade and aneurysm morphology, rather than treatment type. WFNS grade IV-V was linked to a markedly lower chance of favorable outcome (OR 0.10, 95% CI 0.07–0.14), whereas neither DMC nor SAC showed an independent effect compared with BAC. Conclusions In this nationwide registry, technique selection was chiefly determined by aneurysm morphology and location, while outcomes were influenced by patient severity and aneurysm complexity. These findings, from the largest nationwide registry of EVT for RIAs, provide real-world evidence to guide treatment planning.

  • Open Access

    Clinically Applicable Machine Learning Approach to Predict Intracerebral Hematoma Expansion

    by Shogo Watanabe, Nice Ren, Yukihiro Imaoka, Kento Morita, Syoji Kobashi, Nobutaka Mukae, Koichi Arimura, Kunihiro Nishimura, Koji Iihara

    Background: Hematoma expansion (HE) is a significant risk factor for poor prognosis in patients with intracerebral hemorrhage (ICH). Accurately predicting HE is crucial for determining optimal treatment strategies. Methods: This study enrolled 452 patients with ICH from 10 hospitals. To predict HE, 28 clinical variables available on patient arrival (including medical history, ICH location, and ICH volume) and 1142 radiomics features extracted from noncontrast computed tomography images of the ICH regions were used. Clinical variables and radiomics features were selected using gradient boosting and the least absolute shrinkage and selection operator. Three HE prediction models were built on clinical variables alone, radiomics features alone, and a third combining both. The models were compared using 5‐fold cross‐validation, and the mean area under the receiver operating characteristic curve was calculated for each. Additionally, the important features of HE prediction in the combined model were explored. Results: The combined model demonstrated the highest performance for predicting HE with a 5‐fold mean area under the receiver operating characteristic curve of 0.77±0.05, compared with 0.70±0.06 for the clinical variables alone and 0.73±0.04 for the radiomics features alone. Permutation feature importance analysis suggested that anticoagulant treatment was the most predictive of HE. Conclusions: A predictive model for HE was developed using the medical history, clinical features available on the patient’s arrival, imaging, and radiomics features extracted from computed tomography images. This prediction model will assist non–stroke care specialists in making treatment decisions for ICH in emergency settings.


  • Open Access

    Impact of anastomosis site on the regression of periventricular anastomosis in moyamoya disease

    by Saya Ozaki, Eika Hamano, Takenori Ogura, Hirotoshi Imamura, Hisae Mori, Jun C. Takahashi, Koji Iihara, Hiroharu Kataoka

    Purpose Periventricular anastomosis (PA) is a potential cause of intracranial hemorrhage in moyamoya disease. While bypass surgery can reduce PA, an effective target vessel to prevent hemorrhagic stroke remains unidentified. We investigated postoperative PA regression at different anastomosis sites. Methods Overall, 126 hemispheres of 100 patients with moyamoya disease, who underwent superficial temporal artery–middle cerebral artery single-barrel bypass between 2019 and 2022 at our institution were evaluated. Hemispheres with anastomosis anterior or posterior to the central sulcus were categorized into the anterior and posterior groups. Among the PA subtypes, lenticulostriate anastomosis (LSA) and choroidal anastomosis (ChA) were analyzed. Revascularization areas resulting from bypass surgery and PA regression were evaluated by postoperative angiography. The correlation of the anastomosis site and revascularization area with PA regression was investigated. Results In total, 126 hemispheres were analyzed individually, with 33 and 93 in the anterior and posterior groups, respectively. LSA regression was more commonly observed in the anterior group than in the posterior group. ChA regression occurred more frequently in the posterior group than in the anterior group. PA grade decreased in many large revascularization area cases in the LSA group but not in the ChA group. Multivariate analysis showed that LSA regression was associated with the anastomosis site and revascularization area. However, the anastomosis site was the only independent factor for ChA. Conclusion Postoperative PA regression correlated with anastomosis site and revascularization area. The anastomosis site may be important for ChA, and considering the target vessel based on the PA development is necessary.

  • Open Access

    Community‐Level Compact City Design, Health Care Provision, and Outcomes of Patients With Stroke

    by Yukihiro Imaoka, Nice Ren, Soshiro Ogata, Shogo Watanabe, Tomoya Itatani, Eri Kiyoshige, Hirotoshi Imamura, Kunihiro Nishimura, Syoji Kobashi, Yasuyuki Kaku, et al.

    Background: The optimal scale of urbanization for stroke health care provision and the potential impact of compact city design on stroke outcomes remain unclear. We investigated the impact of zip code area‐level compact city design using the walkability index (WI) and its mediators on stroke outcomes. Methods: This nationwide retrospective study used data from patients with stroke from the J‐ASPECT study (2017–2022). WI was calculated as the average of 3 Z ‐scored city design elements (population density, road connectivity, and variation in walkable facilities) from 113 1156 zip code areas in Japan. The association between WI and in‐hospital mortality, functional independence at discharge, and medical costs was assessed using multivariable mixed‐effects logistic regression model. Results: Overall, 555 296 patients (median age, 75 [interquartile range, 66–83] years; female, 42.5%) from 818 hospitals were included. Higher WI was significantly associated with decreased in‐hospital mortality (odds ratio [OR], 0.94 [95% CI, 0.92–0.96]) and increased functional independence (OR, 1.03 [95% CI, 1.02–1.04]). The highest WI group was associated with decreased mortality, primarily mediated by management in intensive or stroke care units (proportion mediated, 0.46 [95% CI, 0.35–0.63]), and the highest WI group was associated with increased functional independence, mediated by short road distance to the hospital (proportion mediated, 0.30 [95% CI, 0.21–0.44]). Conclusions: Zip code area‐level compact city design was associated with decreased in‐hospital mortality and increased functional independence. Compact city design at community level, even without large‐scale urbanization, may contribute to improving stroke care provision and outcomes in increasingly urbanized societies.

  • Open Access

    Quantifying the contributions of cardiovascular risk factors to cardiovascular disease trends in 21st century Japan: a microsimulation study

    by Soshiro Ogata, Eri Kiyoshige, Yusuke Yoshikawa, Koji Iihara, Hitoshi Fukuda, Masanobu Ishii, Kenichi Tsujita, Anna Head, Brendan Collins, Martin O'FLaherty, et al.


  • Open Access

    Successful endovascular occlusion of multiple fusiform aneurysms on the persistent primitive lateral basilovertebral anastomosis

    by Taichi Ikedo, Takaaki Itazu, Hirotoshi Imamura, Yuto Inoue, Shinya Tabata, Koji Shimonaga, Eika Hamano, Kiyofumi Yamada, Hisae Mori, Koji Iihara, et al.

    Background: The primitive lateral basilovertebral anastomosis (PLBA) is a transient embryonic vessel in the vertebrobasilar system that typically regresses during cerebellar artery development. Persistent PLBA (PPLBA), which forms a vertebrobasilar duplication, is a rare vascular anomaly. This is the first reported case of multiple fusiform aneurysms associated with a PPLBA. Case Description: A 14-year-old girl was diagnosed with intracranial aneurysms before undergoing coronary artery bypass grafting surgery for myocardial infarction. Digital subtraction angiography showed two fusiform aneurysms on a PPLBA, connecting the left vertebral artery (VA) and the left anterior inferior cerebellar artery (AICA). After 3 years of follow-up, a new aneurysm developed at the origin of the PPLBA, proximal to the existing two aneurysms. Due to the AICA blood flow originating mainly from the basilar artery (BA) rather than the PPLBA, endovascular parent artery occlusion of the PPLBA was planned to prevent aneurysmal rupture and subarachnoid hemorrhage. Complete occlusion of all three aneurysms was achieved without complications. Conclusion: Understanding the embryological anatomy of this rare vertebrobasilar duplication involving the PPLBA, AICA, BA, and VA facilitated the successful development of a therapeutic strategy. Aneurysms associated with PPLBA exhibit various vascular structures and can be treated effectively with tailored endovascular approaches.


  • Open Access

    Early neurological deterioration after endovascular treatment for acute stroke with mild symptoms

    by Satoru Fujiwara, Kazutaka Uchida, Tsuyoshi Ohta, Nobuyuki Ohara, Michi Kawamoto, Hiroshi Yamagami, Mikito Hayakawa, Akira Ishii, Koji Iihara, Hirotoshi Imamura, et al.

    Purpose Endovascular treatment (EVT) for acute ischemic stroke (AIS) with mild symptoms (National Institute of Health Stroke Scale [NIHSS] score of ≤ 5) has a specific concern about early neurological deterioration (END) after the procedure, but its significance remains unclear. This study aims to examine the prognostic impact and characteristics of END in this patient population. Methods We conducted a post-hoc analysis of the Japan Registry of NeuroEndovascular Therapy 4, enrolling 13,479 patients who underwent EVT for AIS at 166 participating centers between 2015 and 2019. We included patients with AIS and NIHSS scores ≤ 5 and compared outcomes (modified Rankin Scale [mRS] score at 30 days) between patients with and without END after EVT. Multivariate logistic regression models were used to explore associations between END and outcomes and risk factors for END. Results Of 503 patients analyzed, 66 (13%) developed END. Patients who developed END had a lower proportion of mRS scores of 0–2 at 30 days than the No-END group (17% vs. 80%, P < 0.0001), with an adjusted odds ratio (OR) of 0.04 (95% confidence interval [CI], 0.02–0.11). Mortality within 30 days was higher among patients with END (9.1% vs. 0.7%, P < 0.0001). A multivariable logistic regression model identified that factors associated with END included age ≥ 75 years (adjusted OR 3.67; 95% CI, 1.78–7.55), ASPECTS (adjusted OR 0.82; 95% CI, 0.69–0.97), NIHSS scores before EVT of 0–2 compared with those of 3–5 (adjusted OR 2.11; 95% CI, 1.04–4.25), and unsuccessful recanalization (adjusted OR 8.39; 95% CI, 3.90–18.0). Conclusion END after EVT for AIS with mild symptoms was relatively common and associated with worse outcomes. Risk stratification for END and successful recanalization would be crucial, particularly for EVT candidates with mild symptoms.


  • Open Access

    Recurrent prediction within 1, 3, and 5 years after acute ischemic stroke based on machine learning using 10 years J-ASPECT study

    by Shogo Watanabe, Nice Ren, Soshiro Ogata, Yuriko Nakaoku, Akihito Hagihara, Syoji Kobashi, Haruhiko Hiramatsu, Tsuyoshi Ohta, Teruo Noguchi, Hiroharu Kataoka, et al.

    【背景および目的】本研究の目的は急性期脳梗塞発症患者の退院後1,3,5年以内の再発を予測することである.【方法】対象は2010–2019年度にJ-ASPECT Study参加施設に急性期脳梗塞で入院した患者とした.117個の臨床情報を変数として機械学習モデルを作成し,Stroke Prognosis Instrument II (SPI-II),Essen Stroke Risk Score(ESRS)と比較した.【結果】機械学習による再発予測の平均受信者操作特性曲線のArea Under the Curve(ROC AUC)は1,3,5年以内でそれぞれ0.62,0.63,0.63となり,SPI-IIは0.54,0.54,0.54,ESRSは0.55,0.54,0.53となった.【結論】機械学習を用いた脳梗塞再発予測は既存のリスクスコアよりも高い予測精度が得られた.