Deep Learning-Assisted Quantitative Measurement of Thoracolumbar Fracture Features on Lateral Radiographs
OBJECTIVE: This study aimed to develop and validate a deep learning (DL) algorithm for the quantitative measurement of thoracolumbar (TL) fracture features, and to evaluate its efficacy across varying levels of clinical expertise.
METHODS: Using the pretrained Mask Region-Based Convolutional Neural Networks model, originally developed for vertebral body segmentation and fracture detection, we fine-tuned the model and added a new module for measuring fracture metrics-compression rate (CR), Cobb angle (CA), Gardner angle (GA), and sagittal index (SI)-from lumbar spine lateral radiographs. These metrics were derived from six-point labeling by 3 radiologists, forming the ground truth (GT). Training utilized 1,000 nonfractured and 318 fractured radiographs, while validations employed 213 internal and 200 external fractured radiographs. The accuracy of the DL algorithm in quantifying fracture features was evaluated against GT using the intraclass correlation coefficient. Additionally, 4 readers with varying expertise levels, including trainees and an attending spine surgeon, performed measurements with and without DL assistance, and their results were compared to GT and the DL model.
RESULTS: The DL algorithm demonstrated good to excellent agreement with GT for CR, CA, GA, and SI in both internal (0.860, 0.944, 0.932, and 0.779, respectively) and external (0.836, 0.940, 0.916, and 0.815, respectively) validations. DL-assisted measurements significantly improved most measurement values, particularly for trainees.
CONCLUSION: The DL algorithm was validated as an accurate tool for quantifying TL fracture features using radiographs. DL-assisted measurement is expected to expedite the diagnostic process and enhance reliability, particularly benefiting less experienced clinicians.
Errataetall: |
CommentIn: Neurospine. 2024 Mar;21(1):44-45. - PMID 38569630 |
---|---|
Medienart: |
E-Artikel |
Erscheinungsjahr: |
2024 |
---|---|
Erschienen: |
2024 |
Enthalten in: |
Zur Gesamtaufnahme - volume:21 |
---|---|
Enthalten in: |
Neurospine - 21(2024), 1 vom: 04. März, Seite 30-43 |
Sprache: |
Englisch |
---|
Beteiligte Personen: |
Yuh, Woon Tak [VerfasserIn] |
---|
Links: |
---|
Themen: |
Artificial intelligence |
---|
Anmerkungen: |
Date Revised 06.04.2024 published: Print-Electronic CommentIn: Neurospine. 2024 Mar;21(1):44-45. - PMID 38569630 Citation Status PubMed-not-MEDLINE |
---|
doi: |
10.14245/ns.2347366.683 |
---|
funding: |
|
---|---|
Förderinstitution / Projekttitel: |
|
PPN (Katalog-ID): |
NLM370590554 |
---|
LEADER | 01000caa a22002652 4500 | ||
---|---|---|---|
001 | NLM370590554 | ||
003 | DE-627 | ||
005 | 20240406232455.0 | ||
007 | cr uuu---uuuuu | ||
008 | 240404s2024 xx |||||o 00| ||eng c | ||
024 | 7 | |a 10.14245/ns.2347366.683 |2 doi | |
028 | 5 | 2 | |a pubmed24n1367.xml |
035 | |a (DE-627)NLM370590554 | ||
035 | |a (NLM)38569629 | ||
040 | |a DE-627 |b ger |c DE-627 |e rakwb | ||
041 | |a eng | ||
100 | 1 | |a Yuh, Woon Tak |e verfasserin |4 aut | |
245 | 1 | 0 | |a Deep Learning-Assisted Quantitative Measurement of Thoracolumbar Fracture Features on Lateral Radiographs |
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 Revised 06.04.2024 | ||
500 | |a published: Print-Electronic | ||
500 | |a CommentIn: Neurospine. 2024 Mar;21(1):44-45. - PMID 38569630 | ||
500 | |a Citation Status PubMed-not-MEDLINE | ||
520 | |a OBJECTIVE: This study aimed to develop and validate a deep learning (DL) algorithm for the quantitative measurement of thoracolumbar (TL) fracture features, and to evaluate its efficacy across varying levels of clinical expertise | ||
520 | |a METHODS: Using the pretrained Mask Region-Based Convolutional Neural Networks model, originally developed for vertebral body segmentation and fracture detection, we fine-tuned the model and added a new module for measuring fracture metrics-compression rate (CR), Cobb angle (CA), Gardner angle (GA), and sagittal index (SI)-from lumbar spine lateral radiographs. These metrics were derived from six-point labeling by 3 radiologists, forming the ground truth (GT). Training utilized 1,000 nonfractured and 318 fractured radiographs, while validations employed 213 internal and 200 external fractured radiographs. The accuracy of the DL algorithm in quantifying fracture features was evaluated against GT using the intraclass correlation coefficient. Additionally, 4 readers with varying expertise levels, including trainees and an attending spine surgeon, performed measurements with and without DL assistance, and their results were compared to GT and the DL model | ||
520 | |a RESULTS: The DL algorithm demonstrated good to excellent agreement with GT for CR, CA, GA, and SI in both internal (0.860, 0.944, 0.932, and 0.779, respectively) and external (0.836, 0.940, 0.916, and 0.815, respectively) validations. DL-assisted measurements significantly improved most measurement values, particularly for trainees | ||
520 | |a CONCLUSION: The DL algorithm was validated as an accurate tool for quantifying TL fracture features using radiographs. DL-assisted measurement is expected to expedite the diagnostic process and enhance reliability, particularly benefiting less experienced clinicians | ||
650 | 4 | |a Journal Article | |
650 | 4 | |a Artificial intelligence | |
650 | 4 | |a Deep learning | |
650 | 4 | |a Radiography | |
650 | 4 | |a Spinal curvatures | |
650 | 4 | |a Spinal fractures | |
650 | 4 | |a Spinal injuries | |
700 | 1 | |a Khil, Eun Kyung |e verfasserin |4 aut | |
700 | 1 | |a Yoon, Yu Sung |e verfasserin |4 aut | |
700 | 1 | |a Kim, Burnyoung |e verfasserin |4 aut | |
700 | 1 | |a Yoon, Hongjun |e verfasserin |4 aut | |
700 | 1 | |a Lim, Jihe |e verfasserin |4 aut | |
700 | 1 | |a Lee, Kyoung Yeon |e verfasserin |4 aut | |
700 | 1 | |a Yoo, Yeong Seo |e verfasserin |4 aut | |
700 | 1 | |a An, Kyeong Deuk |e verfasserin |4 aut | |
773 | 0 | 8 | |i Enthalten in |t Neurospine |d 2018 |g 21(2024), 1 vom: 04. März, Seite 30-43 |w (DE-627)NLM283050233 |x 2586-6583 |7 nnns |
773 | 1 | 8 | |g volume:21 |g year:2024 |g number:1 |g day:04 |g month:03 |g pages:30-43 |
856 | 4 | 0 | |u http://dx.doi.org/10.14245/ns.2347366.683 |3 Volltext |
912 | |a GBV_USEFLAG_A | ||
912 | |a GBV_NLM | ||
951 | |a AR | ||
952 | |d 21 |j 2024 |e 1 |b 04 |c 03 |h 30-43 |