Update Figure 3 and its caption; current manuscript title in the citation
Browse filesFigure 3 regenerated in the submitted 2x2 layout from the final patient-wise results, external early arterial phase omitted (one examination). Caption follows the new layout. Citation title updated to the revised manuscript title. Field-strength note now names both institutions included in that analysis.
README.md
CHANGED
|
@@ -99,16 +99,16 @@ External test set, LiverSegMRI vs comparators:
|
|
| 99 |
| HD95 (mm) | 6.6 | 15.3 | 27.0 |
|
| 100 |
| Volume error (%) | 5.0 | 13.9 | 12.5 |
|
| 101 |
|
| 102 |
-
- **By field strength:** in
|
| 103 |
- **Comparisons with the other models:** all paired comparisons were significant with Holm-adjusted P < .001.
|
| 104 |
|
| 105 |
### Performance across sequence types and liver morphology
|
| 106 |
|
| 107 |

|
| 108 |
|
| 109 |
-
Radar plots of patient-wise mean Dice for the three models, by MRI sequence type (
|
| 110 |
|
| 111 |
-
Relative to conventional T1-weighted phases, Dice falls by 0.011 on DWI and 0.016 on ADC for LiverSegMRI, against 0.071 and 0.085 for TotalSegmentator MRI and 0.181 and 0.145 for MRAnnotator. Subgroup performance (
|
| 112 |
|
| 113 |
ADC = apparent diffusion coefficient, AP = arterial phase, DWI = diffusion-weighted imaging, EAP = early arterial phase, HBP = hepatobiliary phase, PVP = portal venous phase, T1W = T1-weighted, T2W = T2-weighted, TP/DP = transitional/delayed phase.
|
| 114 |
|
|
@@ -141,7 +141,7 @@ will be completed after publication.
|
|
| 141 |
|
| 142 |
```bibtex
|
| 143 |
@misc{liversegmri,
|
| 144 |
-
title = {LiverSegMRI: Development and External Validation of a Deep Learning Model for Automated Liver Segmentation across Multiparametric MRI Sequences and Comparison with
|
| 145 |
author = {Yuce, Murat and Tordjman, Mickael and Meribout, Anis and Ozkaya, Efe and Lee, Jung-Oh and Lee, Jeong Min and Akinci D'Antonoli, Tugba and Wasserthal, Jakob and Mei, Xueyan and Taouli, Bachir},
|
| 146 |
note = {Manuscript under review; journal, year and DOI to be added after publication},
|
| 147 |
url = {https://huggingface.co/TaouliLab/LiverSegMRI}
|
|
|
|
| 99 |
| HD95 (mm) | 6.6 | 15.3 | 27.0 |
|
| 100 |
| Volume error (%) | 5.0 | 13.9 | 12.5 |
|
| 101 |
|
| 102 |
+
- **By field strength:** in internal patients with DICOM metadata (Mount Sinai and SNUH), performance was similar at 1.5 T and 3.0 T.
|
| 103 |
- **Comparisons with the other models:** all paired comparisons were significant with Holm-adjusted P < .001.
|
| 104 |
|
| 105 |
### Performance across sequence types and liver morphology
|
| 106 |
|
| 107 |

|
| 108 |
|
| 109 |
+
Radar plots of patient-wise mean Dice for the three models, by MRI sequence type (top) and by morphologic subgroup (bottom), in the internal (left) and external (right) test sets. Each axis is one sequence type or subgroup. LiverSegMRI encloses both comparators on every axis. The external early arterial phase is omitted because it contained a single examination.
|
| 110 |
|
| 111 |
+
Relative to conventional T1-weighted phases, Dice falls by 0.011 on DWI and 0.016 on ADC for LiverSegMRI, against 0.071 and 0.085 for TotalSegmentator MRI and 0.181 and 0.145 for MRAnnotator. Subgroup performance (bottom row) stays close to the whole-cohort level for LiverSegMRI in cirrhosis, irregular borders, prior resection, left lobe extension, exophytic lesions and ascites.
|
| 112 |
|
| 113 |
ADC = apparent diffusion coefficient, AP = arterial phase, DWI = diffusion-weighted imaging, EAP = early arterial phase, HBP = hepatobiliary phase, PVP = portal venous phase, T1W = T1-weighted, T2W = T2-weighted, TP/DP = transitional/delayed phase.
|
| 114 |
|
|
|
|
| 141 |
|
| 142 |
```bibtex
|
| 143 |
@misc{liversegmri,
|
| 144 |
+
title = {LiverSegMRI: Development and External Validation of a Deep Learning Model for Automated Liver Segmentation across Multiparametric MRI Sequences and Comparison with Public Multi-Organ Segmentation Models},
|
| 145 |
author = {Yuce, Murat and Tordjman, Mickael and Meribout, Anis and Ozkaya, Efe and Lee, Jung-Oh and Lee, Jeong Min and Akinci D'Antonoli, Tugba and Wasserthal, Jakob and Mei, Xueyan and Taouli, Bachir},
|
| 146 |
note = {Manuscript under review; journal, year and DOI to be added after publication},
|
| 147 |
url = {https://huggingface.co/TaouliLab/LiverSegMRI}
|
figures/figure3_performance_by_sequence_and_subgroup.png
CHANGED
|
Git LFS Details
|
|
Git LFS Details
|