Automated Detection and Segmentation of Bone Metastases on Spine MRI Using U-Net : A Multicenter Study
Copyright © 2024 The Korean Society of Radiology..
OBJECTIVE: To develop and evaluate a deep learning model for automated segmentation and detection of bone metastasis on spinal MRI.
MATERIALS AND METHODS: We included whole spine MRI scans of adult patients with bone metastasis: 662 MRI series from 302 patients (63.5 ± 11.5 years; male:female, 151:151) from three study centers obtained between January 2015 and August 2021 for training and internal testing (random split into 536 and 126 series, respectively) and 49 MRI series from 20 patients (65.9 ± 11.5 years; male:female, 11:9) from another center obtained between January 2018 and August 2020 for external testing. Three sagittal MRI sequences, including non-contrast T1-weighted image (T1), contrast-enhanced T1-weighted Dixon fat-only image (FO), and contrast-enhanced fat-suppressed T1-weighted image (CE), were used. Seven models trained using the 2D and 3D U-Nets were developed with different combinations (T1, FO, CE, T1 + FO, T1 + CE, FO + CE, and T1 + FO + CE). The segmentation performance was evaluated using Dice coefficient, pixel-wise recall, and pixel-wise precision. The detection performance was analyzed using per-lesion sensitivity and a free-response receiver operating characteristic curve. The performance of the model was compared with that of five radiologists using the external test set.
RESULTS: The 2D U-Net T1 + CE model exhibited superior segmentation performance in the external test compared to the other models, with a Dice coefficient of 0.699 and pixel-wise recall of 0.653. The T1 + CE model achieved per-lesion sensitivities of 0.828 (497/600) and 0.857 (150/175) for metastases in the internal and external tests, respectively. The radiologists demonstrated a mean per-lesion sensitivity of 0.746 and a mean per-lesion positive predictive value of 0.701 in the external test.
CONCLUSION: The deep learning models proposed for automated segmentation and detection of bone metastases on spinal MRI demonstrated high diagnostic performance.
| Medienart: |
E-Artikel |
|---|
| Erscheinungsjahr: |
2024 |
|---|---|
| Erschienen: |
2024 |
| Enthalten in: |
Zur Gesamtaufnahme - volume:25 |
|---|---|
| Enthalten in: |
Korean journal of radiology - 25(2024), 4 vom: 25. Apr., Seite 363-373 |
| Sprache: |
Englisch |
|---|
| Beteiligte Personen: |
Kim, Dong Hyun [VerfasserIn] |
|---|
| Links: |
|---|
| Themen: |
Bone neoplasms |
|---|
| Anmerkungen: |
Date Completed 27.03.2024 Date Revised 02.04.2024 published: Print Citation Status MEDLINE |
|---|
| doi: |
10.3348/kjr.2023.0671 |
|---|
| funding: |
|
|---|---|
| Förderinstitution / Projekttitel: |
|
| PPN (Katalog-ID): |
NLM370182553 |
|---|
| LEADER | 01000caa a22002652 4500 | ||
|---|---|---|---|
| 001 | NLM370182553 | ||
| 003 | DE-627 | ||
| 005 | 20240403000627.0 | ||
| 007 | cr uuu---uuuuu | ||
| 008 | 240327s2024 xx |||||o 00| ||eng c | ||
| 024 | 7 | |a 10.3348/kjr.2023.0671 |2 doi | |
| 028 | 5 | 2 | |a pubmed24n1361.xml |
| 035 | |a (DE-627)NLM370182553 | ||
| 035 | |a (NLM)38528694 | ||
| 040 | |a DE-627 |b ger |c DE-627 |e rakwb | ||
| 041 | |a eng | ||
| 100 | 1 | |a Kim, Dong Hyun |e verfasserin |4 aut | |
| 245 | 1 | 0 | |a Automated Detection and Segmentation of Bone Metastases on Spine MRI Using U-Net |b A Multicenter Study |
| 264 | 1 | |c 2024 | |
| 336 | |a Text |b txt |2 rdacontent | ||
| 337 | |a ƒaComputermedien |b c |2 rdamedia | ||
| 338 | |a ƒa Online-Ressource |b cr |2 rdacarrier | ||
| 500 | |a Date Completed 27.03.2024 | ||
| 500 | |a Date Revised 02.04.2024 | ||
| 500 | |a published: Print | ||
| 500 | |a Citation Status MEDLINE | ||
| 520 | |a Copyright © 2024 The Korean Society of Radiology. | ||
| 520 | |a OBJECTIVE: To develop and evaluate a deep learning model for automated segmentation and detection of bone metastasis on spinal MRI | ||
| 520 | |a MATERIALS AND METHODS: We included whole spine MRI scans of adult patients with bone metastasis: 662 MRI series from 302 patients (63.5 ± 11.5 years; male:female, 151:151) from three study centers obtained between January 2015 and August 2021 for training and internal testing (random split into 536 and 126 series, respectively) and 49 MRI series from 20 patients (65.9 ± 11.5 years; male:female, 11:9) from another center obtained between January 2018 and August 2020 for external testing. Three sagittal MRI sequences, including non-contrast T1-weighted image (T1), contrast-enhanced T1-weighted Dixon fat-only image (FO), and contrast-enhanced fat-suppressed T1-weighted image (CE), were used. Seven models trained using the 2D and 3D U-Nets were developed with different combinations (T1, FO, CE, T1 + FO, T1 + CE, FO + CE, and T1 + FO + CE). The segmentation performance was evaluated using Dice coefficient, pixel-wise recall, and pixel-wise precision. The detection performance was analyzed using per-lesion sensitivity and a free-response receiver operating characteristic curve. The performance of the model was compared with that of five radiologists using the external test set | ||
| 520 | |a RESULTS: The 2D U-Net T1 + CE model exhibited superior segmentation performance in the external test compared to the other models, with a Dice coefficient of 0.699 and pixel-wise recall of 0.653. The T1 + CE model achieved per-lesion sensitivities of 0.828 (497/600) and 0.857 (150/175) for metastases in the internal and external tests, respectively. The radiologists demonstrated a mean per-lesion sensitivity of 0.746 and a mean per-lesion positive predictive value of 0.701 in the external test | ||
| 520 | |a CONCLUSION: The deep learning models proposed for automated segmentation and detection of bone metastases on spinal MRI demonstrated high diagnostic performance | ||
| 650 | 4 | |a Multicenter Study | |
| 650 | 4 | |a Journal Article | |
| 650 | 4 | |a Bone neoplasms | |
| 650 | 4 | |a Deep learning | |
| 650 | 4 | |a Magnetic resonance imaging | |
| 650 | 4 | |a Metastasis | |
| 650 | 4 | |a Spine | |
| 700 | 1 | |a Seo, Jiwoon |e verfasserin |4 aut | |
| 700 | 1 | |a Lee, Ji Hyun |e verfasserin |4 aut | |
| 700 | 1 | |a Jeon, Eun-Tae |e verfasserin |4 aut | |
| 700 | 1 | |a Jeong, DongYoung |e verfasserin |4 aut | |
| 700 | 1 | |a Chae, Hee Dong |e verfasserin |4 aut | |
| 700 | 1 | |a Lee, Eugene |e verfasserin |4 aut | |
| 700 | 1 | |a Kang, Ji Hee |e verfasserin |4 aut | |
| 700 | 1 | |a Choi, Yoon-Hee |e verfasserin |4 aut | |
| 700 | 1 | |a Kim, Hyo Jin |e verfasserin |4 aut | |
| 700 | 1 | |a Chai, Jee Won |e verfasserin |4 aut | |
| 773 | 0 | 8 | |i Enthalten in |t Korean journal of radiology |d 2000 |g 25(2024), 4 vom: 25. Apr., Seite 363-373 |w (DE-627)NLM11642950X |x 2005-8330 |7 nnns |
| 773 | 1 | 8 | |g volume:25 |g year:2024 |g number:4 |g day:25 |g month:04 |g pages:363-373 |
| 856 | 4 | 0 | |u http://dx.doi.org/10.3348/kjr.2023.0671 |3 Volltext |
| 912 | |a GBV_USEFLAG_A | ||
| 912 | |a GBV_NLM | ||
| 951 | |a AR | ||
| 952 | |d 25 |j 2024 |e 4 |b 25 |c 04 |h 363-373 | ||