Koji Yamashita
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Koji Yamashita has a total of 400 co-authors. Here are the top co-authors by number of shared publications:
Hiroshi Honda
Affiliation Department of Clinical Radiology, Graduate School of Medical Sciences, Kyushu University, Fukuoka, Japan
62 shared publicationsOsamu Togao
Department of Molecular Imaging and Diagnosis, Graduate School of Medical Sciences Kyushu University Fukuoka Japan
55 shared publicationsAkio Hiwatashi
Affiliation Department of Molecular Imaging & Diagnosis, Graduate School of Medical Sciences, Kyushu University, Fukuoka, Japan
51 shared publicationsKazufumi Kikuchi
Kyushu University
42 shared publications- Akio Hiwatashi
Department of Radiology, Nagoya City University Graduate School of Medical Sciences , 1 Kawasumi, Mizuho-cho, Mizuho-ku, Nagoya, Aichi 467-8601
39 shared publications - Osamu Togao
Kyushu University
27 shared publications
Latest Publications
- 11 November 2026
- Published by Springer Nature in Radiological Physics and Technology
- p. 1-16
Monte Carlo-based three-dimensional dose evaluation for rotational cerebral angiography: peak dose to brain and skin and Dose–Volume Histogram (DVH) analysis of the brain
Background and purpose Rotational cerebral angiography produces complex, nonuniform dose distributions; the mean brain dose may not sufficiently represent localized high-dose regions. This study aimed to visualize three-dimensional dose distributions in the brain and skin using PHITS, estimate peak brain dose (PBD) and peak skin dose (PSD), and evaluate the brain dose distribution using dose–volume histogram (DVH) analysis. Materials and methods X-ray spectra for 3D digital subtraction angiography (3D-DSA, 70 kV) and cone-beam computed tomography (CBCT, 109 kV) were generated using SpekPy and implemented in PHITS. A rotational irradiation model covering approximately 200° was constructed using a multisource approach based on projection-dependent tube current–time product values. The model was validated using a PMMA phantom. Three-dimensional dose distributions were calculated using the adult male mesh-type reference computational phantom; PBD, PSD, and brain DVH indices were evaluated. Results Calculated doses agreed with measured values within 5.2%. Dose concentrations were mainly observed in the occipital region. The PBD and PSD were 25.1 and 44.1 mGy for 3D-DSA and 101.3 and 139.1 mGy for CBCT, respectively. The PBD was 3.3 times higher than the mean brain dose for 3D-DSA and 2.3 times higher for CBCT. DVH analysis showed that CBCT shifted the brain-dose distribution toward higher-dose regions, with D10% values of 14.9 and 70.0 mGy for 3D-DSA and CBCT, respectively. Conclusions PHITS-based three-dimensional dose evaluation allowed localized high-dose regions in the brain and skin to be visualized and quantified. DVH analysis complemented the peak dose assessment by providing volumetric information on the brain-dose distribution.
- 11 November 2026
- Published by Springer Nature in Radiological Physics and Technology
- p. 1-16
Monte Carlo-based three-dimensional dose evaluation for rotational cerebral angiography: peak dose to brain and skin and Dose–Volume Histogram (DVH) analysis of the brain
Background and purpose Rotational cerebral angiography produces complex, nonuniform dose distributions; the mean brain dose may not sufficiently represent localized high-dose regions. This study aimed to visualize three-dimensional dose distributions in the brain and skin using PHITS, estimate peak brain dose (PBD) and peak skin dose (PSD), and evaluate the brain dose distribution using dose–volume histogram (DVH) analysis. Materials and methods X-ray spectra for 3D digital subtraction angiography (3D-DSA, 70 kV) and cone-beam computed tomography (CBCT, 109 kV) were generated using SpekPy and implemented in PHITS. A rotational irradiation model covering approximately 200° was constructed using a multisource approach based on projection-dependent tube current–time product values. The model was validated using a PMMA phantom. Three-dimensional dose distributions were calculated using the adult male mesh-type reference computational phantom; PBD, PSD, and brain DVH indices were evaluated. Results Calculated doses agreed with measured values within 5.2%. Dose concentrations were mainly observed in the occipital region. The PBD and PSD were 25.1 and 44.1 mGy for 3D-DSA and 101.3 and 139.1 mGy for CBCT, respectively. The PBD was 3.3 times higher than the mean brain dose for 3D-DSA and 2.3 times higher for CBCT. DVH analysis showed that CBCT shifted the brain-dose distribution toward higher-dose regions, with D10% values of 14.9 and 70.0 mGy for 3D-DSA and CBCT, respectively. Conclusions PHITS-based three-dimensional dose evaluation allowed localized high-dose regions in the brain and skin to be visualized and quantified. DVH analysis complemented the peak dose assessment by providing volumetric information on the brain-dose distribution.
- Open Access
- 1 October 2026
- Published by SAGE Publications in Multiple Sclerosis Journal – Experimental, Translational and Clinical
- Vol. 12 (4)
Clinical and radiological features of fifteen patients presenting with concentric lesions like Baló's concentric sclerosis
Background: Baló's concentric sclerosis (BCS) is a rare demyelinating disorder defined by alternating rings of demyelinated and preserved myelin. Methods: We retrospectively reviewed the medical records of 15 Japanese patients with BCS lesions seen at our institution over a 15-year period. Results: The median age at onset was 29 years, with a female predominance (67%). All patients fulfilled the 2024 McDonald criteria for multiple sclerosis (MS), and 83% exhibited cerebrospinal fluid oligoclonal bands (OCBs). In patients who underwent susceptibility-weighted imaging, peripheral hypointensity was frequently observed, whereas classic lamellar architecture was not clearly visualized. Over long-term follow-up, new BCS lesions were detected during treatment with interferon beta, dimethyl fumarate, fingolimod, or siponimod, whereas no new BCS lesions were observed during natalizumab or ofatumumab exposure, except shortly after switching to natalizumab. Conclusions: These findings support the concept that BCS constitutes a clinical variant within the MS spectrum. Some patients appeared to benefit from natalizumab or ofatumumab, although the optimal DMT remains uncertain. The frequent presence of OCBs and MS-like radiological features may indicate overlap with classical MS. BCS may thus represent a common and distinct immune-mediated demyelinating phenotype at the interface between classical MS and atypical demyelination.
- 1 October 2026
- Published by SAGE Publications in Multiple Sclerosis Journal – Experimental, Translational and Clinical
- Vol. 12 (4)
Open AccessClinical and radiological features of fifteen patients presenting with concentric lesions like Baló's concentric sclerosis
Background: Baló's concentric sclerosis (BCS) is a rare demyelinating disorder defined by alternating rings of demyelinated and preserved myelin. Methods: We retrospectively reviewed the medical records of 15 Japanese patients with BCS lesions seen at our institution over a 15-year period. Results: The median age at onset was 29 years, with a female predominance (67%). All patients fulfilled the 2024 McDonald criteria for multiple sclerosis (MS), and 83% exhibited cerebrospinal fluid oligoclonal bands (OCBs). In patients who underwent susceptibility-weighted imaging, peripheral hypointensity was frequently observed, whereas classic lamellar architecture was not clearly visualized. Over long-term follow-up, new BCS lesions were detected during treatment with interferon beta, dimethyl fumarate, fingolimod, or siponimod, whereas no new BCS lesions were observed during natalizumab or ofatumumab exposure, except shortly after switching to natalizumab. Conclusions: These findings support the concept that BCS constitutes a clinical variant within the MS spectrum. Some patients appeared to benefit from natalizumab or ofatumumab, although the optimal DMT remains uncertain. The frequent presence of OCBs and MS-like radiological features may indicate overlap with classical MS. BCS may thus represent a common and distinct immune-mediated demyelinating phenotype at the interface between classical MS and atypical demyelination.
- Open Access
- 8 September 2026
- Published by Elsevier in Academic Radiology
Feasibility of Using Short Repetition Time for Brain Tumors in 23Na-MRI
- 8 September 2026
- Published by Elsevier in Academic Radiology
Open AccessFeasibility of Using Short Repetition Time for Brain Tumors in 23Na-MRI
- 31 July 2026
- Published by Wiley in Neurology and Clinical Neuroscience
Recurrent Severe Drug‐Induced Aseptic Meningitis due to Trimethoprim–Sulfamethoxazole Mimicking Ventriculitis‐Like MRI Findings in a Patient Receiving Anti–PD‐1 Therapy
Drug‐induced aseptic meningitis is a rare and usually self‐limiting condition. Although clinical and cerebrospinal fluid findings may resemble those of bacterial meningitis, brain magnetic resonance imaging findings are typically normal. An 80‐year‐old woman receiving anti–PD‐1 antibody therapy for bladder cancer developed severe meningitis with high fever and disturbed consciousness. Brain magnetic resonance imaging revealed findings suggestive of ventriculitis, strongly suggesting bacterial meningitis; however, blood and cerebrospinal fluid cultures obtained before the initiation of antibiotic therapy were negative. Two months earlier, the patient had experienced a similar episode of severe meningitis of unknown etiology. Careful review of the clinical course revealed that trimethoprim‐sulfamethoxazole had been administered shortly before the onset of both episodes. The patient was diagnosed with recurrent drug‐induced aseptic meningitis caused by trimethoprim‐sulfamethoxazole. Patients with drug‐induced aseptic meningitis rarely present with severe clinical and radiological features mimicking bacterial meningitis. Therefore, it should be considered in patients with culture‐negative meningitis, particularly those receiving immune checkpoint inhibitors.
- 31 July 2026
- Published by Wiley in Neurology and Clinical Neuroscience
Recurrent Severe Drug‐Induced Aseptic Meningitis due to Trimethoprim–Sulfamethoxazole Mimicking Ventriculitis‐Like MRI Findings in a Patient Receiving Anti–PD‐1 Therapy
Drug‐induced aseptic meningitis is a rare and usually self‐limiting condition. Although clinical and cerebrospinal fluid findings may resemble those of bacterial meningitis, brain magnetic resonance imaging findings are typically normal. An 80‐year‐old woman receiving anti–PD‐1 antibody therapy for bladder cancer developed severe meningitis with high fever and disturbed consciousness. Brain magnetic resonance imaging revealed findings suggestive of ventriculitis, strongly suggesting bacterial meningitis; however, blood and cerebrospinal fluid cultures obtained before the initiation of antibiotic therapy were negative. Two months earlier, the patient had experienced a similar episode of severe meningitis of unknown etiology. Careful review of the clinical course revealed that trimethoprim‐sulfamethoxazole had been administered shortly before the onset of both episodes. The patient was diagnosed with recurrent drug‐induced aseptic meningitis caused by trimethoprim‐sulfamethoxazole. Patients with drug‐induced aseptic meningitis rarely present with severe clinical and radiological features mimicking bacterial meningitis. Therefore, it should be considered in patients with culture‐negative meningitis, particularly those receiving immune checkpoint inhibitors.
- 1 May 2026
- Published by Radiological Society of North America (RSNA) in Radiology
- Vol. 319 (2) , e254090
The Utility of Sodium Inversion Recovery MRI to Identify Astrocytoma, IDH -Mutant: A Prospective Pilot Study
Although further reduction of acquisition time is needed, 3-T sodium inversion recovery MRI may enhance current diagnostic approaches for detecting isocitrate dehydrogenase–mutant astrocytoma.
- 1 May 2026
- Published by Radiological Society of North America (RSNA) in Radiology
- Vol. 319 (2) , e254090
The Utility of Sodium Inversion Recovery MRI to Identify Astrocytoma, IDH -Mutant: A Prospective Pilot Study
Although further reduction of acquisition time is needed, 3-T sodium inversion recovery MRI may enhance current diagnostic approaches for detecting isocitrate dehydrogenase–mutant astrocytoma.
- 17 March 2026
- Published by Wiley in NMR in Biomedicine
- Vol. 39 (4) , e70271
Effect of the Temporal Sampling Strategy for the Assessment of Cerebral Perfusion in Multidelay Arterial Spin Labeling
Multidelay arterial spin labeling (ASL) MRI requires balancing temporal resolution with scan duration. The goal of the current study is to evaluate the influence of the temporal sampling strategy, implemented through temporal downsampling, on the quantitative accuracy of cerebral perfusion parameters obtained using multidelay ASL. In this prospective single‐center study, 12 healthy volunteers (mean age, 37 ± 9 years; seven men and five women) underwent multidelay pseudo‐continuous ASL on a 3‐T MRI system. Data were acquired using a variable time repetition scheme with 61 time points at a 75‐ms sampling interval. Cerebral blood flow (CBF) and arterial transit time (ATT) maps were calculated using seven temporally downsampled datasets with sampling intervals of 75 (gold standard), 150, 300, 450, 600, 750, and 900 ms. The normalized root mean square error (RMSE) between each sampling interval and the gold standard (75 ms) was calculated for both CBF and ATT. Additionally, subgroup analyses were performed for regions with ATT shorter or longer than the median, as determined from the 75‐ms interval dataset. The normalized RMSE of the CBF and ATT increased with more aggressive temporal downsampling. Across all sampling schemes, the normalized RMSE of CBF was significantly lower in regions with shorter ATT than in those with longer ATT ( p < 0.05). Temporal sampling configuration influences the quantitative reliability of CBF and ATT estimation. CBF bias is particularly pronounced in regions with prolonged ATT.
- 17 March 2026
- Published by Wiley in NMR in Biomedicine
- Vol. 39 (4) , e70271
Effect of the Temporal Sampling Strategy for the Assessment of Cerebral Perfusion in Multidelay Arterial Spin Labeling
Multidelay arterial spin labeling (ASL) MRI requires balancing temporal resolution with scan duration. The goal of the current study is to evaluate the influence of the temporal sampling strategy, implemented through temporal downsampling, on the quantitative accuracy of cerebral perfusion parameters obtained using multidelay ASL. In this prospective single‐center study, 12 healthy volunteers (mean age, 37 ± 9 years; seven men and five women) underwent multidelay pseudo‐continuous ASL on a 3‐T MRI system. Data were acquired using a variable time repetition scheme with 61 time points at a 75‐ms sampling interval. Cerebral blood flow (CBF) and arterial transit time (ATT) maps were calculated using seven temporally downsampled datasets with sampling intervals of 75 (gold standard), 150, 300, 450, 600, 750, and 900 ms. The normalized root mean square error (RMSE) between each sampling interval and the gold standard (75 ms) was calculated for both CBF and ATT. Additionally, subgroup analyses were performed for regions with ATT shorter or longer than the median, as determined from the 75‐ms interval dataset. The normalized RMSE of the CBF and ATT increased with more aggressive temporal downsampling. Across all sampling schemes, the normalized RMSE of CBF was significantly lower in regions with shorter ATT than in those with longer ATT ( p < 0.05). Temporal sampling configuration influences the quantitative reliability of CBF and ATT estimation. CBF bias is particularly pronounced in regions with prolonged ATT.
- 26 February 2026
- Published by SAGE Publications in Multiple Sclerosis Journal
- Vol. 32 (4) , 385-395
Diagnostic performance of central vein sign, paramagnetic rim lesion and cortical lesion in Asian patients with multiple sclerosis
Background: The 2024 revised McDonald criteria for multiple sclerosis (MS) incorporate central vein sign (CVS) and paramagnetic rim lesion (PRL), but their utility in Asian patients remains unclear. Objective: To evaluate the performance of CVS, PRL and cortical lesion (CL) in distinguishing MS from neuromyelitis optica spectrum disorders (NMOSD) and myelin-oligodendrocyte glycoprotein antibody-associated disease (MOGAD) in Japanese patients. Methods: We conducted a cross-sectional study of 139 MS, 41 NMOSD and 12 MOGAD patients who underwent 3-Tesla magnetic resonance imaging (MRI), including susceptibility-weighted imaging. Results: At least one CVS was identified in 95% of MS, 59% of NMOSD and 42% of MOGAD, while PRL occurred in 63% of MS only. The CVS number (cutoff ⩾ 2) yielded 85.6% sensitivity and 75.5% specificity (area under the curve (AUC) 0.856), which improved in patients <50 (AUC 0.920) or <70 years (AUC 0.907), and when combined with PRL and CL (AUC 0.926). The Select 6 algorithm achieved 69.1% sensitivity and 98.1% specificity (AUC 0.836), which were comparable in patients <50 (AUC 0.879) or <70 years (AUC 0.839), but improved when combined with PRL and CL (AUC 0.925). Conclusions: Combining Select 6 or CVS number with PRL and CL, particularly in patients <50 or <70 years, provides robust diagnostic performance for MS in the Asian population.
- 26 February 2026
- Published by SAGE Publications in Multiple Sclerosis Journal
- Vol. 32 (4) , 385-395
Diagnostic performance of central vein sign, paramagnetic rim lesion and cortical lesion in Asian patients with multiple sclerosis
Background: The 2024 revised McDonald criteria for multiple sclerosis (MS) incorporate central vein sign (CVS) and paramagnetic rim lesion (PRL), but their utility in Asian patients remains unclear. Objective: To evaluate the performance of CVS, PRL and cortical lesion (CL) in distinguishing MS from neuromyelitis optica spectrum disorders (NMOSD) and myelin-oligodendrocyte glycoprotein antibody-associated disease (MOGAD) in Japanese patients. Methods: We conducted a cross-sectional study of 139 MS, 41 NMOSD and 12 MOGAD patients who underwent 3-Tesla magnetic resonance imaging (MRI), including susceptibility-weighted imaging. Results: At least one CVS was identified in 95% of MS, 59% of NMOSD and 42% of MOGAD, while PRL occurred in 63% of MS only. The CVS number (cutoff ⩾ 2) yielded 85.6% sensitivity and 75.5% specificity (area under the curve (AUC) 0.856), which improved in patients <50 (AUC 0.920) or <70 years (AUC 0.907), and when combined with PRL and CL (AUC 0.926). The Select 6 algorithm achieved 69.1% sensitivity and 98.1% specificity (AUC 0.836), which were comparable in patients <50 (AUC 0.879) or <70 years (AUC 0.839), but improved when combined with PRL and CL (AUC 0.925). Conclusions: Combining Select 6 or CVS number with PRL and CL, particularly in patients <50 or <70 years, provides robust diagnostic performance for MS in the Asian population.
- Open Access
- 20 February 2026
- Published by Scientific Scholar in Surgical Neurology International
- Vol. 17, 108
Ventriculoperitoneal shunt-associated cystic lesion with perilesional cerebral edema 45 years after placement for posterior fossa tumor: A case report
Background: Ventriculoperitoneal (VP) shunts are widely used to treat hydrocephalus in children with posterior fossa tumors. While many patients gain shunt independence after tumor control, indwelling shunts are often not removed. Shunt-related intracranial cysts are extremely rare and typically develop along the intracranial catheter tract within days to several years after shunt insertion; to date, the longest reported interval has been 19 years. Here, we report a case a cystic lesion with marked perilesional edema developing 45 years after childhood VP shunt placement for tumor-related hydrocephalus. Case Description: We report the case of a 58-year-old man who presented with acute severe left-sided hemiparesis. At the age of 13 years, he underwent resection of a posterior fossa ependymoma, VP shunt placement, and whole-brain radiotherapy. Neuroimaging revealed a large right frontal low-density lesion with vasogenic edema and a cystic cavity surrounding the intracranial shunt catheter. Retrospective review of magnetic resonance imaging performed 3 months earlier showed a smaller preexisting cyst at the same site. The shunt catheter was removed without resistance, and the membranous material adhering to the distal tip caused partial intraluminal obstruction. Postoperatively, the hemiparesis improved markedly; follow-up imaging demonstrated resolution of the edema and reduction in the cyst size. As hydrocephalus did not recur, shunt reinsertion was not required. Conclusion: This case represents the longest latency reported in the literature for VP shunt-related intracranial cyst formation, occurring 45 years after shunt implantation, exceeding the previously reported maximum interval of 19 years. The acute symptoms closely resemble stroke or infection, underscoring the diagnostic challenge. Clinicians should be vigilant of such delayed complications even decades after shunt placement.
- 20 February 2026
- Published by Scientific Scholar in Surgical Neurology International
- Vol. 17, 108
Open AccessVentriculoperitoneal shunt-associated cystic lesion with perilesional cerebral edema 45 years after placement for posterior fossa tumor: A case report
Background: Ventriculoperitoneal (VP) shunts are widely used to treat hydrocephalus in children with posterior fossa tumors. While many patients gain shunt independence after tumor control, indwelling shunts are often not removed. Shunt-related intracranial cysts are extremely rare and typically develop along the intracranial catheter tract within days to several years after shunt insertion; to date, the longest reported interval has been 19 years. Here, we report a case a cystic lesion with marked perilesional edema developing 45 years after childhood VP shunt placement for tumor-related hydrocephalus. Case Description: We report the case of a 58-year-old man who presented with acute severe left-sided hemiparesis. At the age of 13 years, he underwent resection of a posterior fossa ependymoma, VP shunt placement, and whole-brain radiotherapy. Neuroimaging revealed a large right frontal low-density lesion with vasogenic edema and a cystic cavity surrounding the intracranial shunt catheter. Retrospective review of magnetic resonance imaging performed 3 months earlier showed a smaller preexisting cyst at the same site. The shunt catheter was removed without resistance, and the membranous material adhering to the distal tip caused partial intraluminal obstruction. Postoperatively, the hemiparesis improved markedly; follow-up imaging demonstrated resolution of the edema and reduction in the cyst size. As hydrocephalus did not recur, shunt reinsertion was not required. Conclusion: This case represents the longest latency reported in the literature for VP shunt-related intracranial cyst formation, occurring 45 years after shunt implantation, exceeding the previously reported maximum interval of 19 years. The acute symptoms closely resemble stroke or infection, underscoring the diagnostic challenge. Clinicians should be vigilant of such delayed complications even decades after shunt placement.
- 6 February 2026
- Published by American Society of Neuroradiology (ASNR) in American Journal of Neuroradiology
- Vol. 47 (7) , 1869-1878
The Utility of Cystathionine Assessment Using Proton MRS for the Preoperative Differential Diagnosis of Adult-Type Diffuse Gliomas
Adult-type diffuse gliomas-astrocytoma, IDH-mutant; oligodendroglioma, IDH-mutant and 1p/19q-codeleted; and glioblastoma, IDH-wildtype-exhibit distinct prognoses and treatment responses. Accurate preoperative subtype estimation is therefore important for clinical decision-making. Proton MR spectroscopy (1H-MRS) enables noninvasive assessment of tumor metabolism. Cystathionine, detectable at 2.72 ppm, has been proposed as a metabolic marker of oligodendroglioma, but its diagnostic performance across adult-type diffuse gliomas remains incompletely defined. This study evaluated the utility of cystathionine quantification by 1H-MRS for differentiating glioma subtypes and assessed whether combined analysis with 2-hydroxyglutarate (2HG) improves diagnostic performance. Eighty-five patients with histologically and molecularly confirmed adult-type diffuse gliomas (25 oligodendrogliomas, 28 astrocytomas, 32 glioblastomas) underwent preoperative 3T MRI including single-voxel PRESS 1H-MRS (TE = 97 ms). Spectra with severe artifacts were excluded; no cases were excluded based on full width at half maximum (FWHM < 12.8 Hz, 0.1 ppm at 3T). Metabolites were quantified using LCModel, with concentrations normalized to the unsuppressed water signal and relaxation-corrected. Group comparisons of cystathionine levels were performed using Kruskal-Wallis and Bonferroni-corrected pairwise tests. The ROC analysis evaluated diagnostic performance for differentiating oligodendrogliomas from astrocytomas and glioblastomas. Supplementary analyses excluding spectra with cystathionine CRLB ≥ 50% and combined cystathionine-2HG ROC analyses were also performed. Cystathionine levels were highest in oligodendrogliomas (1.040 ± 0.908 mM), intermediate in glioblastomas, and lowest in astrocytomas (0.437 ± 0.403 mM). Oligodendrogliomas showed significantly higher levels than astrocytomas (P = 0.003), whereas no significant difference was observed between oligodendrogliomas and glioblastomas. ROC analysis showed moderate diagnostic performance (AUC = 0.69 for oligodendroglioma vs astrocytoma; AUC = 0.56 for oligodendroglioma vs glioblastoma). After CRLB-based exclusion, sensitivity increased but specificity decreased (AUC = 0.83 for oligodendroglioma vs astrocytoma). Combining cystathionine with 2HG modestly improved AUCs (0.72 and 0.61, respectively). Cystathionine quantification by 1H-MRS reflects biologically meaningful metabolic differences among adult-type diffuse gliomas, with higher levels characteristic of oligodendrogliomas compared with astrocytomas. However, overlap with glioblastomas limits its role as a stand-alone discriminator. When interpreted alongside 2HG and conventional imaging features, cystathionine may serve as a supportive metabolic marker to enhance preoperative glioma subtype classification.
- 6 February 2026
- Published by American Society of Neuroradiology (ASNR) in American Journal of Neuroradiology
- Vol. 47 (7) , 1869-1878
The Utility of Cystathionine Assessment Using Proton MRS for the Preoperative Differential Diagnosis of Adult-Type Diffuse Gliomas
Adult-type diffuse gliomas-astrocytoma, IDH-mutant; oligodendroglioma, IDH-mutant and 1p/19q-codeleted; and glioblastoma, IDH-wildtype-exhibit distinct prognoses and treatment responses. Accurate preoperative subtype estimation is therefore important for clinical decision-making. Proton MR spectroscopy (1H-MRS) enables noninvasive assessment of tumor metabolism. Cystathionine, detectable at 2.72 ppm, has been proposed as a metabolic marker of oligodendroglioma, but its diagnostic performance across adult-type diffuse gliomas remains incompletely defined. This study evaluated the utility of cystathionine quantification by 1H-MRS for differentiating glioma subtypes and assessed whether combined analysis with 2-hydroxyglutarate (2HG) improves diagnostic performance. Eighty-five patients with histologically and molecularly confirmed adult-type diffuse gliomas (25 oligodendrogliomas, 28 astrocytomas, 32 glioblastomas) underwent preoperative 3T MRI including single-voxel PRESS 1H-MRS (TE = 97 ms). Spectra with severe artifacts were excluded; no cases were excluded based on full width at half maximum (FWHM < 12.8 Hz, 0.1 ppm at 3T). Metabolites were quantified using LCModel, with concentrations normalized to the unsuppressed water signal and relaxation-corrected. Group comparisons of cystathionine levels were performed using Kruskal-Wallis and Bonferroni-corrected pairwise tests. The ROC analysis evaluated diagnostic performance for differentiating oligodendrogliomas from astrocytomas and glioblastomas. Supplementary analyses excluding spectra with cystathionine CRLB ≥ 50% and combined cystathionine-2HG ROC analyses were also performed. Cystathionine levels were highest in oligodendrogliomas (1.040 ± 0.908 mM), intermediate in glioblastomas, and lowest in astrocytomas (0.437 ± 0.403 mM). Oligodendrogliomas showed significantly higher levels than astrocytomas (P = 0.003), whereas no significant difference was observed between oligodendrogliomas and glioblastomas. ROC analysis showed moderate diagnostic performance (AUC = 0.69 for oligodendroglioma vs astrocytoma; AUC = 0.56 for oligodendroglioma vs glioblastoma). After CRLB-based exclusion, sensitivity increased but specificity decreased (AUC = 0.83 for oligodendroglioma vs astrocytoma). Combining cystathionine with 2HG modestly improved AUCs (0.72 and 0.61, respectively). Cystathionine quantification by 1H-MRS reflects biologically meaningful metabolic differences among adult-type diffuse gliomas, with higher levels characteristic of oligodendrogliomas compared with astrocytomas. However, overlap with glioblastomas limits its role as a stand-alone discriminator. When interpreted alongside 2HG and conventional imaging features, cystathionine may serve as a supportive metabolic marker to enhance preoperative glioma subtype classification.
- 24 January 2026
- Published by Springer Nature in Neuroradiology
- Vol. 68 (2) , 533-539
The diagnostic value of cortical high-flow sign combined with cystathionine on 1H-MRS for prediction of 1p/19q-codeletion status in IDH-mutant adult-type diffuse glioma
Objectives To leverage a combination of cortical high-flow sign on arterial spin labeling (ASL) and cystathionine measurement using 1H-MR spectroscopy (1H-MRS) to distinguish oligodendroglioma, IDH-mutant and 1p/19q-codeleted (O_IDHm-codel) from astrocytoma, IDH-mutant (A_IDHm-noncodel). Materials and methods We implemented pseudo-continuous ASL technique (post-labeling delay = 2000 ms) using a 3.0-T MRI scanner. Relative perfusion maps were generated by subtracting paired labeled images from control images. 1H-MRS data were acquired by using the single-voxel point-resolved spectroscopy sequence (PRESS) sequence (TR = 2000 ms, TE = 97 ms, NEX = 128, volume of interest = 203 mm3). Our study included cases with a Cramér-Rao Lower Bound of cystathionine levels at 2.7 ppm that were below 50%. The presence or absence of cortical high-flow sign and the estimated concentration of cystathionine were compared between A_IDHm-noncodel and O_IDHm-codel. The receiver operating characteristic curves were used to evaluate the diagnostic performance of each parameter, as well as the combination of both. Results The cortical high-flow sign was identified more frequent in O_IDHm-codel (7/12, 58.3%) than in A_IDHm-noncodel (3/18, 16.7%; p = 0.018). The cystathionine levels in O_IDHm-codel (1.50 ± 0.63 mM) was significantly higher than in A_IDHm-noncodel (0.85 ± 0.24 mM; p = 0.001). The area under the curve for distinguishing O_IDHm-codel from A_IDHm-noncodel using the presence of cortical high-flow sign, cystathionine levels, and their combination was 0.708 (0.509–0.908), 0.750 (0.553–0.947), and 0.875 (0.737–1.000), respectively. Conclusion The presence of cortical high-flow sign on ASL, along with elevated cystathionine levels measured by 1H-MRS, could differentiate O_IDHm-codel from A_IDHm-noncodel.
- 24 January 2026
- Published by Springer Nature in Neuroradiology
- Vol. 68 (2) , 533-539
The diagnostic value of cortical high-flow sign combined with cystathionine on 1H-MRS for prediction of 1p/19q-codeletion status in IDH-mutant adult-type diffuse glioma
Objectives To leverage a combination of cortical high-flow sign on arterial spin labeling (ASL) and cystathionine measurement using 1H-MR spectroscopy (1H-MRS) to distinguish oligodendroglioma, IDH-mutant and 1p/19q-codeleted (O_IDHm-codel) from astrocytoma, IDH-mutant (A_IDHm-noncodel). Materials and methods We implemented pseudo-continuous ASL technique (post-labeling delay = 2000 ms) using a 3.0-T MRI scanner. Relative perfusion maps were generated by subtracting paired labeled images from control images. 1H-MRS data were acquired by using the single-voxel point-resolved spectroscopy sequence (PRESS) sequence (TR = 2000 ms, TE = 97 ms, NEX = 128, volume of interest = 203 mm3). Our study included cases with a Cramér-Rao Lower Bound of cystathionine levels at 2.7 ppm that were below 50%. The presence or absence of cortical high-flow sign and the estimated concentration of cystathionine were compared between A_IDHm-noncodel and O_IDHm-codel. The receiver operating characteristic curves were used to evaluate the diagnostic performance of each parameter, as well as the combination of both. Results The cortical high-flow sign was identified more frequent in O_IDHm-codel (7/12, 58.3%) than in A_IDHm-noncodel (3/18, 16.7%; p = 0.018). The cystathionine levels in O_IDHm-codel (1.50 ± 0.63 mM) was significantly higher than in A_IDHm-noncodel (0.85 ± 0.24 mM; p = 0.001). The area under the curve for distinguishing O_IDHm-codel from A_IDHm-noncodel using the presence of cortical high-flow sign, cystathionine levels, and their combination was 0.708 (0.509–0.908), 0.750 (0.553–0.947), and 0.875 (0.737–1.000), respectively. Conclusion The presence of cortical high-flow sign on ASL, along with elevated cystathionine levels measured by 1H-MRS, could differentiate O_IDHm-codel from A_IDHm-noncodel.