Fumiyasu Tsushima
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Fumiyasu Tsushima has a total of 47 co-authors. Here are the top co-authors by number of shared publications:
Shingo Kakeda
Department of Radiology, Hirosaki University Graduate School of Medicine, Hirosaki, Japan
3 shared publications- Shingo Kakeda
Department of Radiology, Hirosaki University Graduate School of Medicine, Hirosaki 036-8562, Japan
3 shared publications Shinya Kakehata
Department of Radiology, Hirosaki University Graduate School of Medicine, Hirosaki, Japan
3 shared publicationsSho Maruyama
Department of Radiology, Hirosaki University Graduate School of Medicine, Hirosaki, Japan
3 shared publicationsHiroyuki Miura
Department of Radiology, Hirosaki University Graduate School of Medicine, Hirosaki, Japan
2 shared publicationsTaiki Koshiishi
Department of Radiology, Hirosaki University Graduate School of Medicine, Hirosaki, Japan
2 shared publications
Latest Publications
- 1 June 2026
- Published by Springer Nature in Annals of Nuclear Medicine
- Vol. 40 (10) , 1-8
Utility of preoperative [¹⁸F]FDG PET/CT for CT-guided biopsy planning of large tumors: target adjustment and malignant yield with hotspot targeting
Objective To assess the clinical utility of preoperative [¹⁸F]FDG PET/CT for CT-guided biopsy planning of large tumors, including predictors of PET/CT-informed target adjustment and malignant yield with FDG hotspot targeting in very large tumors. Methods In this retrospective single-center study (January 2015–December 2025), we analyzed 82 patients who underwent pre-biopsy [¹⁸F]FDG PET/CT at our institution within 90 days. Target adjustment was retrospectively adjudicated as the presence of a spatial difference between the contrast-enhanced CT-based target and the PET/CT-informed target selected after considering intratumoral FDG uptake. ROC analysis was used to evaluate whether tumor size predicted target adjustment and to determine the optimal cutoff. Multivariable logistic regression included age, sex, tumor size, PET system (analog vs. digital), ΔSUV, lesion location (chest vs. other), and lymphoma (vs. other). In tumors ≥ 52 mm, malignant yield was compared between hotspot (highest uptake) and non-hotspot targeting. Results Target adjustment was performed in 28/82 cases (34.1%). Interobserver agreement was 89% with Cohen’s κ = 0.74 (95% CI, 0.58–0.89). Tumor size predicted target adjustment (AUC 0.847; 95% CI 0.761–0.934), and the Youden-optimal cutoff was 52 mm (sensitivity 0.82, specificity 0.81). Target adjustment rates were 5/49 (10.2%) for < 52 mm and 23/33 (69.7%) for ≥ 52 mm (p < 0.001). In multivariable analysis, tumor size (per 10 mm) was independently associated with target adjustment (OR 2.33; 95% CI 1.56–3.50; p < 0.001), while female sex was inversely associated (OR 0.20; 95% CI 0.04–0.96; p = 0.045). Lymphoma showed a trend toward an inverse association that did not reach statistical significance (OR 0.11; 95% CI 0.01–1.11; p = 0.062). The multivariable model showed good discrimination (AUC 0.90; 95% CI 0.82–0.98; DeLong). Among tumors ≥ 52 mm, malignant pathology was obtained in 14/14 cases (100%) with hotspot targeting versus 5/9 (55.6%) with non-hotspot targeting (p = 0.014). Conclusions Preoperative [¹⁸F]FDG PET/CT supports CT-guided biopsy planning of large tumors by identifying cases likely to require target adjustment and was associated with higher malignant yield when the FDG hotspot was targeted in tumors ≥ 52 mm.
- 1 June 2026
- Published by Springer Nature in Annals of Nuclear Medicine
- Vol. 40 (10) , 1-8
Utility of preoperative [¹⁸F]FDG PET/CT for CT-guided biopsy planning of large tumors: target adjustment and malignant yield with hotspot targeting
Objective To assess the clinical utility of preoperative [¹⁸F]FDG PET/CT for CT-guided biopsy planning of large tumors, including predictors of PET/CT-informed target adjustment and malignant yield with FDG hotspot targeting in very large tumors. Methods In this retrospective single-center study (January 2015–December 2025), we analyzed 82 patients who underwent pre-biopsy [¹⁸F]FDG PET/CT at our institution within 90 days. Target adjustment was retrospectively adjudicated as the presence of a spatial difference between the contrast-enhanced CT-based target and the PET/CT-informed target selected after considering intratumoral FDG uptake. ROC analysis was used to evaluate whether tumor size predicted target adjustment and to determine the optimal cutoff. Multivariable logistic regression included age, sex, tumor size, PET system (analog vs. digital), ΔSUV, lesion location (chest vs. other), and lymphoma (vs. other). In tumors ≥ 52 mm, malignant yield was compared between hotspot (highest uptake) and non-hotspot targeting. Results Target adjustment was performed in 28/82 cases (34.1%). Interobserver agreement was 89% with Cohen’s κ = 0.74 (95% CI, 0.58–0.89). Tumor size predicted target adjustment (AUC 0.847; 95% CI 0.761–0.934), and the Youden-optimal cutoff was 52 mm (sensitivity 0.82, specificity 0.81). Target adjustment rates were 5/49 (10.2%) for < 52 mm and 23/33 (69.7%) for ≥ 52 mm (p < 0.001). In multivariable analysis, tumor size (per 10 mm) was independently associated with target adjustment (OR 2.33; 95% CI 1.56–3.50; p < 0.001), while female sex was inversely associated (OR 0.20; 95% CI 0.04–0.96; p = 0.045). Lymphoma showed a trend toward an inverse association that did not reach statistical significance (OR 0.11; 95% CI 0.01–1.11; p = 0.062). The multivariable model showed good discrimination (AUC 0.90; 95% CI 0.82–0.98; DeLong). Among tumors ≥ 52 mm, malignant pathology was obtained in 14/14 cases (100%) with hotspot targeting versus 5/9 (55.6%) with non-hotspot targeting (p = 0.014). Conclusions Preoperative [¹⁸F]FDG PET/CT supports CT-guided biopsy planning of large tumors by identifying cases likely to require target adjustment and was associated with higher malignant yield when the FDG hotspot was targeted in tumors ≥ 52 mm.
- Open Access
- 19 January 2026
- Published by Springer Nature in Research Square
Preoperative 18F-FDG PET/CT for CT-Guided Biopsy Planning: Predicting Target Adjustment and Improving Malignant Yield in Large Tumors
Objective To assess the clinical utility of preoperative 18F-FDG PET/CT for CT-guided biopsy planning of large tumors, including predictors of PET-driven target adjustment and malignant yield with FDG hotspot targeting in very large tumors. Methods In this retrospective single-center study, we analyzed 82 patients who underwent CT-guided biopsy between January 2015 and December 2025 with preoperative 18F-FDG PET/CT performed within 90 days before biopsy. Target adjustment was defined as changing the planned intratumoral biopsy target from the contrast-enhanced CT–based target to an alternative intratumoral site based on FDG uptake. Tumor size was evaluated by ROC analysis. Multivariable logistic regression included age, sex, tumor size, PET system (analog vs digital), ΔSUVmax, lesion location (chest vs other), and lymphoma (vs other). In tumors ≥ 52 mm, malignant yield was compared between hotspot (highest uptake) and non-hotspot targeting. Results Target adjustment was performed in 28/82 cases (34.1%). Interobserver agreement was 89% with Cohen’s κ = 0.74 (95% CI, 0.58–0.89). Tumor size predicted target adjustment (AUC 0.847; 95% CI 0.761–0.934), and the Youden-optimal cutoff was 52 mm (sensitivity 0.82, specificity 0.81). Target adjustment rates were 5/49 (10.2%) for < 52 mm and 23/33 (69.7%) for ≥ 52 mm (p < 0.001). In multivariable analysis, tumor size (per 10 mm) was independently associated with target adjustment (OR 2.33; 95% CI 1.56–3.50; p < 0.001), while female sex (OR 0.20; 95% CI 0.03–0.91; p = 0.049) and lymphoma (OR 0.082; 95% CI 0.0058–0.746; p = 0.041) were inversely associated. The multivariable model showed good discrimination (AUC 0.90; 95% CI 0.82–0.98; DeLong). Among tumors ≥ 52 mm, malignant pathology was obtained in 14/14 cases (100%) with hotspot targeting versus 5/9 (55.6%) with non-hotspot targeting (p = 0.014). Conclusions Preoperative 18F-FDG PET/CT supports CT-guided biopsy planning of large tumors by identifying cases likely to require target adjustment and by improving malignant yield when the FDG hotspot is targeted in tumors ≥ 52 mm.
- 19 January 2026
- Published by Springer Nature in Research Square
Open AccessPreoperative 18F-FDG PET/CT for CT-Guided Biopsy Planning: Predicting Target Adjustment and Improving Malignant Yield in Large Tumors
Objective To assess the clinical utility of preoperative 18F-FDG PET/CT for CT-guided biopsy planning of large tumors, including predictors of PET-driven target adjustment and malignant yield with FDG hotspot targeting in very large tumors. Methods In this retrospective single-center study, we analyzed 82 patients who underwent CT-guided biopsy between January 2015 and December 2025 with preoperative 18F-FDG PET/CT performed within 90 days before biopsy. Target adjustment was defined as changing the planned intratumoral biopsy target from the contrast-enhanced CT–based target to an alternative intratumoral site based on FDG uptake. Tumor size was evaluated by ROC analysis. Multivariable logistic regression included age, sex, tumor size, PET system (analog vs digital), ΔSUVmax, lesion location (chest vs other), and lymphoma (vs other). In tumors ≥ 52 mm, malignant yield was compared between hotspot (highest uptake) and non-hotspot targeting. Results Target adjustment was performed in 28/82 cases (34.1%). Interobserver agreement was 89% with Cohen’s κ = 0.74 (95% CI, 0.58–0.89). Tumor size predicted target adjustment (AUC 0.847; 95% CI 0.761–0.934), and the Youden-optimal cutoff was 52 mm (sensitivity 0.82, specificity 0.81). Target adjustment rates were 5/49 (10.2%) for < 52 mm and 23/33 (69.7%) for ≥ 52 mm (p < 0.001). In multivariable analysis, tumor size (per 10 mm) was independently associated with target adjustment (OR 2.33; 95% CI 1.56–3.50; p < 0.001), while female sex (OR 0.20; 95% CI 0.03–0.91; p = 0.049) and lymphoma (OR 0.082; 95% CI 0.0058–0.746; p = 0.041) were inversely associated. The multivariable model showed good discrimination (AUC 0.90; 95% CI 0.82–0.98; DeLong). Among tumors ≥ 52 mm, malignant pathology was obtained in 14/14 cases (100%) with hotspot targeting versus 5/9 (55.6%) with non-hotspot targeting (p = 0.014). Conclusions Preoperative 18F-FDG PET/CT supports CT-guided biopsy planning of large tumors by identifying cases likely to require target adjustment and by improving malignant yield when the FDG hotspot is targeted in tumors ≥ 52 mm.
- 22 November 2025
- Published by Springer Nature in Neuroradiology
- Vol. 68 (5) , 1377-1385
Deep-learning reconstruction enables about one minute 3D T1-weighted MRI: quantitative evaluation of Acceleration-quality Trade-offs and motion reduction
Purpose Deep learning-based reconstruction techniques, such as the prototype DL-Speed, have been developed to accelerate 3D T1-weighted imaging, but their clinical utility and impact on quantitative analysis have not been systematically validated. The purpose of this study is to evaluate the utility of a deep learning-based reconstruction method for substantially accelerating image acquisition without compromising quantitative image quality and morphometry. Methods Six healthy volunteers were scanned with 3D MPRAGE using acceleration factors ranging from 2 to 16. Image quality was assessed using the CAT12 Image Quality Rating (IQR), and morphometry with CAT12 global cortical thickness and total gray matter (GM) volume. In a clinical cohort of 40 patients, DL-Speed with 11-fold acceleration (1 min 10 s) was compared to conventional imaging (4 min 59 s). Head motion was quantified via total vector change (TVC) using volumetric navigators. Results In healthy volunteers, image quality remained within CAT12 Rank A at all acceleration levels; IQR followed a quadratic trend (R² = 0.98). Global cortical thickness decreased approximately linearly with acceleration (e.g., DLS2→DLS11: −0.068 mm [− 2.8%]; DLS2→DLS16: −0.140 mm [− 5.7%]). Total GM volume showed a shallow, monotonic decline (group means: 748.8 mL at DLS2 → 729.5 mL at DLS11 [− 2.6%] → 726.3 mL at DLS16 [− 3.0%]). In clinical cohort, DL-Speed significantly reduced head motion (TVC: 52.3 ± 9.4 mm vs. 140.4 ± 32.8 mm, p < 0.001) while maintaining acceptable image quality (IQR: 93.9 ± 1.0% vs. 94.8 ± 1.2%, p < 0.001). Bland–Altman analysis showed narrow limits of agreement (–2.49% to + 0.75%). Regarding cortical thickness and GM volume, we found significant correlations between conventional images and DL-Speed (thickness r = 0.97; GM r = 0.99; both p < 0.001). Conclusion DL-Speed enables about-one-minute 3D T1-weighted imaging that markedly reduces motion while preserving quantitative image integrity suitable for morphometry. Because image quality (IQR) follows a non-linear (quadratic) dependency whereas cortical thickness and GM volume decline nearly linearly, the data support a practical operating range (e.g., 6–11-fold acceleration) rather than a single optimal factor. Trial registration Not applicable.
- 22 November 2025
- Published by Springer Nature in Neuroradiology
- Vol. 68 (5) , 1377-1385
Deep-learning reconstruction enables about one minute 3D T1-weighted MRI: quantitative evaluation of Acceleration-quality Trade-offs and motion reduction
Purpose Deep learning-based reconstruction techniques, such as the prototype DL-Speed, have been developed to accelerate 3D T1-weighted imaging, but their clinical utility and impact on quantitative analysis have not been systematically validated. The purpose of this study is to evaluate the utility of a deep learning-based reconstruction method for substantially accelerating image acquisition without compromising quantitative image quality and morphometry. Methods Six healthy volunteers were scanned with 3D MPRAGE using acceleration factors ranging from 2 to 16. Image quality was assessed using the CAT12 Image Quality Rating (IQR), and morphometry with CAT12 global cortical thickness and total gray matter (GM) volume. In a clinical cohort of 40 patients, DL-Speed with 11-fold acceleration (1 min 10 s) was compared to conventional imaging (4 min 59 s). Head motion was quantified via total vector change (TVC) using volumetric navigators. Results In healthy volunteers, image quality remained within CAT12 Rank A at all acceleration levels; IQR followed a quadratic trend (R² = 0.98). Global cortical thickness decreased approximately linearly with acceleration (e.g., DLS2→DLS11: −0.068 mm [− 2.8%]; DLS2→DLS16: −0.140 mm [− 5.7%]). Total GM volume showed a shallow, monotonic decline (group means: 748.8 mL at DLS2 → 729.5 mL at DLS11 [− 2.6%] → 726.3 mL at DLS16 [− 3.0%]). In clinical cohort, DL-Speed significantly reduced head motion (TVC: 52.3 ± 9.4 mm vs. 140.4 ± 32.8 mm, p < 0.001) while maintaining acceptable image quality (IQR: 93.9 ± 1.0% vs. 94.8 ± 1.2%, p < 0.001). Bland–Altman analysis showed narrow limits of agreement (–2.49% to + 0.75%). Regarding cortical thickness and GM volume, we found significant correlations between conventional images and DL-Speed (thickness r = 0.97; GM r = 0.99; both p < 0.001). Conclusion DL-Speed enables about-one-minute 3D T1-weighted imaging that markedly reduces motion while preserving quantitative image integrity suitable for morphometry. Because image quality (IQR) follows a non-linear (quadratic) dependency whereas cortical thickness and GM volume decline nearly linearly, the data support a practical operating range (e.g., 6–11-fold acceleration) rather than a single optimal factor. Trial registration Not applicable.
- 1 March 2022
- Published by Elsevier in Academic Radiology
- Vol. 29 (3) , 388-394
Effectiveness of Cytological Diagnosis with Outer Cannula Washing Solution for Computed Tomography-Guided Needle Biopsy
- 1 March 2022
- Published by Elsevier in Academic Radiology
- Vol. 29 (3) , 388-394
Effectiveness of Cytological Diagnosis with Outer Cannula Washing Solution for Computed Tomography-Guided Needle Biopsy
- 16 February 2022
- Published by SAGE Publications in Acta Radiologica
- Vol. 64 (2) , 690-697
Improved visualization of the subthalamic nucleus on synthetic MRI with optimized parameters: initial study
Background: Synthetic magnetic resonance imaging (SyMRI) enables to reformat various images by adjusting the MR parameters. Purpose: To investigate whether customization of the repetition time (TR), echo time (TE), and inversion time (TI) in SyMRI could improve the visualization of subthalamic nucleus (STN). Material and Methods: We examined five healthy volunteers using both coronal SyMRI and quantitative susceptibility mapping (QSM), seven patients with Parkinson's disease (PD) using coronal SyMRI, and 15 patients with PD using coronal QSM. Two neuroradiologists reformatted SyMRI (optimized SyMRI) by adjusting TR, TE, and TI to achieve maximum tissue contrast between the STN and the adjacent brain parenchyma. The optimized MR parameters in the PD patients varied according to the individual. For regular SyMRI (T2-weighted imaging [T2WI] and STIR), optimized SyMRI, and QSM, qualitative visualization scores of the STN (STN score) were recorded. The contrast-to-noise ratio (CNR) of the STN was also measured. Results: For the STN scores in both groups, the optimized SyMRI were significantly higher than the regular SyMRI ( P < 0.05), and there were no significant differences between optimized SyMRI and QSM. For the CNR of differentiation of the STN from the substantia nigra, the optimized SyMRI was higher than the regular SyMRI (volunteer: T2WI P = 0.10 and STIR P = 0.26; PD patient: T2WI P = 0.43 and STIR P = 0.25), but the optimized SyMRI was lower than the QSM (volunteer: P = 0.26; PD patient: P = 0.03). Conclusions: On SyMRI, optimization of MR parameters (TR, TE, and TI) on an individual basis may be useful to increase the conspicuity of the STN.
- 16 February 2022
- Published by SAGE Publications in Acta Radiologica
- Vol. 64 (2) , 690-697
Improved visualization of the subthalamic nucleus on synthetic MRI with optimized parameters: initial study
Background: Synthetic magnetic resonance imaging (SyMRI) enables to reformat various images by adjusting the MR parameters. Purpose: To investigate whether customization of the repetition time (TR), echo time (TE), and inversion time (TI) in SyMRI could improve the visualization of subthalamic nucleus (STN). Material and Methods: We examined five healthy volunteers using both coronal SyMRI and quantitative susceptibility mapping (QSM), seven patients with Parkinson's disease (PD) using coronal SyMRI, and 15 patients with PD using coronal QSM. Two neuroradiologists reformatted SyMRI (optimized SyMRI) by adjusting TR, TE, and TI to achieve maximum tissue contrast between the STN and the adjacent brain parenchyma. The optimized MR parameters in the PD patients varied according to the individual. For regular SyMRI (T2-weighted imaging [T2WI] and STIR), optimized SyMRI, and QSM, qualitative visualization scores of the STN (STN score) were recorded. The contrast-to-noise ratio (CNR) of the STN was also measured. Results: For the STN scores in both groups, the optimized SyMRI were significantly higher than the regular SyMRI ( P < 0.05), and there were no significant differences between optimized SyMRI and QSM. For the CNR of differentiation of the STN from the substantia nigra, the optimized SyMRI was higher than the regular SyMRI (volunteer: T2WI P = 0.10 and STIR P = 0.26; PD patient: T2WI P = 0.43 and STIR P = 0.25), but the optimized SyMRI was lower than the QSM (volunteer: P = 0.26; PD patient: P = 0.03). Conclusions: On SyMRI, optimization of MR parameters (TR, TE, and TI) on an individual basis may be useful to increase the conspicuity of the STN.
- Open Access
- 21 September 2021
- Published by Wiley in The Prostate
- Vol. 81 (16) , 1411-1427
Characteristics of α2,3‐sialyl N‐glycosylated PSA as a biomarker for clinically significant prostate cancer in men with elevated PSA level
Background: The presence of glycosylated isoforms of prostate‐specific antigen (PSA) in prostate cancer (PC) cells is a potential marker of their aggressiveness. We characterized the origin of α2,3‐sialylated prostate‐specific antigen (S23PSA) by tissue‐based sialylation‐related gene expression and studied the performance of S23PSA density (S23PSAD) alone and in combination with multiparametric magnetic resonance imaging (MRI) for the detection of clinically significant prostate cancer in men with elevated PSA.Methods: Tissue‐based quantification of S23PSA and sialyltransferase and sialidase gene expression was evaluated in 71 radical prostatectomy specimens. The diagnostic performance of S23PSAD was studied in 1099 men retrospectively enrolled in a multicenter systematic biopsy (SBx) cohort. We correlated the S23PSAD with Prostate Imaging Reporting and Data System (PI‐RADS) scores in 98 men prospectively enrolled in a single‐center MRI‐targeted biopsy (MRI–TBx) cohort. The primary outcome was the PC‐diagnostic performance of the S23PSAD, the secondary outcome was the avoidable biopsy rate of S23PSAD combined with DRE and total PSA (tPSA), and with or without PI‐RADS.Results: S23PSA was significantly higher in Gleason pattern 4 and 5 compared with benign prostate tissue. In the retrospective cohort, the performance of S23PSAD for detecting PC was superior to tPSA or PSA density (PSAD) (AUC: 0.7758 vs. 0.6360 and 0.7509, respectively). In the prospective cohort, S23PSAD was superior to tPSA, PSAD, and PI‐RADS (AUC: 0.7725 vs. 0.5901, 0.7439 and 0.7305, respectively), and S23PSAD + PI‐RADS + DRE + tPSA was superior to DRE + tPSA+PI‐RADS with avoidance rate of MRI–TBx (13% vs. 1%) at 30% risk threshold.Conclusions: The diagnostic performance of S23PSAD was superior to conventional strategies but comparable to mpMRI.
- 21 September 2021
- Published by Wiley in The Prostate
- Vol. 81 (16) , 1411-1427
Open AccessCharacteristics of α2,3‐sialyl N‐glycosylated PSA as a biomarker for clinically significant prostate cancer in men with elevated PSA level
Background: The presence of glycosylated isoforms of prostate‐specific antigen (PSA) in prostate cancer (PC) cells is a potential marker of their aggressiveness. We characterized the origin of α2,3‐sialylated prostate‐specific antigen (S23PSA) by tissue‐based sialylation‐related gene expression and studied the performance of S23PSA density (S23PSAD) alone and in combination with multiparametric magnetic resonance imaging (MRI) for the detection of clinically significant prostate cancer in men with elevated PSA.Methods: Tissue‐based quantification of S23PSA and sialyltransferase and sialidase gene expression was evaluated in 71 radical prostatectomy specimens. The diagnostic performance of S23PSAD was studied in 1099 men retrospectively enrolled in a multicenter systematic biopsy (SBx) cohort. We correlated the S23PSAD with Prostate Imaging Reporting and Data System (PI‐RADS) scores in 98 men prospectively enrolled in a single‐center MRI‐targeted biopsy (MRI–TBx) cohort. The primary outcome was the PC‐diagnostic performance of the S23PSAD, the secondary outcome was the avoidable biopsy rate of S23PSAD combined with DRE and total PSA (tPSA), and with or without PI‐RADS.Results: S23PSA was significantly higher in Gleason pattern 4 and 5 compared with benign prostate tissue. In the retrospective cohort, the performance of S23PSAD for detecting PC was superior to tPSA or PSA density (PSAD) (AUC: 0.7758 vs. 0.6360 and 0.7509, respectively). In the prospective cohort, S23PSAD was superior to tPSA, PSAD, and PI‐RADS (AUC: 0.7725 vs. 0.5901, 0.7439 and 0.7305, respectively), and S23PSAD + PI‐RADS + DRE + tPSA was superior to DRE + tPSA+PI‐RADS with avoidance rate of MRI–TBx (13% vs. 1%) at 30% risk threshold.Conclusions: The diagnostic performance of S23PSAD was superior to conventional strategies but comparable to mpMRI.