Incremental retraining, clinical implementation, and acceptance rate of deep learning auto-segmentation for male pelvis in a multiuser environment
© 2023 American Association of Physicists in Medicine..
BACKGROUND: Deep learning auto-segmentation (DLAS) models have been adopted in the clinic; however, they suffer from performance deterioration owing to the clinical practice variability. Some commercial DLAS software provide an incremental retraining function that enables users to train a custom model using their institutional data to account for clinical practice variability.
PURPOSE: This study was performed to evaluate and implement the commercial DLAS software with the incremental retraining function for definitive treatment of patients with prostate cancer in a multi-user environment.
METHODS: CT-based target organs and organs-at-risk (OAR) delineation of 215 prostate cancer patients were utilized. The performance of three commercial DLAS software built-in models was validated with 20 patients. A retrained custom model was developed using 100 patients and evaluated on the remaining data (n = 115). Dice similarity coefficient (DSC), Hausdorff distance (HD), mean surface distance (MSD), and surface DSC (SDSC) were utilized for quantitative evaluation. A multi-rater qualitative evaluation was blindly performed with a five-level scale. Visual inspection was performed in consensus and non-consensus unacceptable cases to identify the failure modes.
RESULTS: Three commercial DLAS vendor built-in models achieved sub-optimal performance in 20 patients. The retrained custom model had a mean DSC of 0.82 for prostate, 0.48 for seminal vesicles (SV), and 0.92 for rectum, respectively. This represents a significant improvement over the built-in model with DSC of 0.73, 0.37, and 0.81 for the corresponding structures. Compared to the acceptance rate of 96.5% and consensus unacceptable rate (i.e., both reviewers rated as unacceptable) of 3.5% achieved by manual contours, the custom model achieved a 91.3% acceptance rate and 8.7% consensus unacceptable rate. The failure modes of retrained custom model were attributed to the following: cystogram (n = 2), hip prosthesis (n = 2), low dose rate brachytherapy seeds (n = 2), air in endorectal balloon(n = 1), non-iodinated spacer (n = 2), and giant bladder(n = 1).
CONCLUSION: The commercial DLAS software with the incremental retraining function was validated and clinically adopted for prostate patients in a multi-user environment. AI-based auto-delineation of the prostate and OARs is shown to achieve improved physician acceptance, overall clinical utility, and accuracy.
Medienart: |
E-Artikel |
---|
Erscheinungsjahr: |
2023 |
---|---|
Erschienen: |
2023 |
Enthalten in: |
Zur Gesamtaufnahme - volume:50 |
---|---|
Enthalten in: |
Medical physics - 50(2023), 7 vom: 09. Juli, Seite 4079-4091 |
Sprache: |
Englisch |
---|
Beteiligte Personen: |
Duan, Jingwei [VerfasserIn] |
---|
Links: |
---|
Themen: |
Deep learning auto-segmentation |
---|
Anmerkungen: |
Date Completed 11.07.2023 Date Revised 18.07.2023 published: Print-Electronic Citation Status MEDLINE |
---|
doi: |
10.1002/mp.16537 |
---|
funding: |
|
---|---|
Förderinstitution / Projekttitel: |
|
PPN (Katalog-ID): |
NLM357891651 |
---|
LEADER | 01000naa a22002652 4500 | ||
---|---|---|---|
001 | NLM357891651 | ||
003 | DE-627 | ||
005 | 20231226073552.0 | ||
007 | cr uuu---uuuuu | ||
008 | 231226s2023 xx |||||o 00| ||eng c | ||
024 | 7 | |a 10.1002/mp.16537 |2 doi | |
028 | 5 | 2 | |a pubmed24n1192.xml |
035 | |a (DE-627)NLM357891651 | ||
035 | |a (NLM)37287322 | ||
040 | |a DE-627 |b ger |c DE-627 |e rakwb | ||
041 | |a eng | ||
100 | 1 | |a Duan, Jingwei |e verfasserin |4 aut | |
245 | 1 | 0 | |a Incremental retraining, clinical implementation, and acceptance rate of deep learning auto-segmentation for male pelvis in a multiuser environment |
264 | 1 | |c 2023 | |
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 11.07.2023 | ||
500 | |a Date Revised 18.07.2023 | ||
500 | |a published: Print-Electronic | ||
500 | |a Citation Status MEDLINE | ||
520 | |a © 2023 American Association of Physicists in Medicine. | ||
520 | |a BACKGROUND: Deep learning auto-segmentation (DLAS) models have been adopted in the clinic; however, they suffer from performance deterioration owing to the clinical practice variability. Some commercial DLAS software provide an incremental retraining function that enables users to train a custom model using their institutional data to account for clinical practice variability | ||
520 | |a PURPOSE: This study was performed to evaluate and implement the commercial DLAS software with the incremental retraining function for definitive treatment of patients with prostate cancer in a multi-user environment | ||
520 | |a METHODS: CT-based target organs and organs-at-risk (OAR) delineation of 215 prostate cancer patients were utilized. The performance of three commercial DLAS software built-in models was validated with 20 patients. A retrained custom model was developed using 100 patients and evaluated on the remaining data (n = 115). Dice similarity coefficient (DSC), Hausdorff distance (HD), mean surface distance (MSD), and surface DSC (SDSC) were utilized for quantitative evaluation. A multi-rater qualitative evaluation was blindly performed with a five-level scale. Visual inspection was performed in consensus and non-consensus unacceptable cases to identify the failure modes | ||
520 | |a RESULTS: Three commercial DLAS vendor built-in models achieved sub-optimal performance in 20 patients. The retrained custom model had a mean DSC of 0.82 for prostate, 0.48 for seminal vesicles (SV), and 0.92 for rectum, respectively. This represents a significant improvement over the built-in model with DSC of 0.73, 0.37, and 0.81 for the corresponding structures. Compared to the acceptance rate of 96.5% and consensus unacceptable rate (i.e., both reviewers rated as unacceptable) of 3.5% achieved by manual contours, the custom model achieved a 91.3% acceptance rate and 8.7% consensus unacceptable rate. The failure modes of retrained custom model were attributed to the following: cystogram (n = 2), hip prosthesis (n = 2), low dose rate brachytherapy seeds (n = 2), air in endorectal balloon(n = 1), non-iodinated spacer (n = 2), and giant bladder(n = 1) | ||
520 | |a CONCLUSION: The commercial DLAS software with the incremental retraining function was validated and clinically adopted for prostate patients in a multi-user environment. AI-based auto-delineation of the prostate and OARs is shown to achieve improved physician acceptance, overall clinical utility, and accuracy | ||
650 | 4 | |a Journal Article | |
650 | 4 | |a deep learning auto-segmentation | |
650 | 4 | |a inter-observer contour variation | |
650 | 4 | |a prostate radiotherapy | |
700 | 1 | |a Vargas, Carlos E |e verfasserin |4 aut | |
700 | 1 | |a Yu, Nathan Y |e verfasserin |4 aut | |
700 | 1 | |a Laughlin, Brady S |e verfasserin |4 aut | |
700 | 1 | |a Toesca, Diego Santos |e verfasserin |4 aut | |
700 | 1 | |a Keole, Sameer |e verfasserin |4 aut | |
700 | 1 | |a Rwigema, Jean Claude M |e verfasserin |4 aut | |
700 | 1 | |a Wong, William W |e verfasserin |4 aut | |
700 | 1 | |a Schild, Steven E |e verfasserin |4 aut | |
700 | 1 | |a Feng, Xue |e verfasserin |4 aut | |
700 | 1 | |a Chen, Quan |e verfasserin |4 aut | |
700 | 1 | |a Rong, Yi |e verfasserin |4 aut | |
773 | 0 | 8 | |i Enthalten in |t Medical physics |d 1974 |g 50(2023), 7 vom: 09. Juli, Seite 4079-4091 |w (DE-627)NLM000294748 |x 2473-4209 |7 nnns |
773 | 1 | 8 | |g volume:50 |g year:2023 |g number:7 |g day:09 |g month:07 |g pages:4079-4091 |
856 | 4 | 0 | |u http://dx.doi.org/10.1002/mp.16537 |3 Volltext |
912 | |a GBV_USEFLAG_A | ||
912 | |a GBV_NLM | ||
951 | |a AR | ||
952 | |d 50 |j 2023 |e 7 |b 09 |c 07 |h 4079-4091 |