Low-Rank Adaptation of Pre-Trained Large Vision Models for Improved Lung Nodule Malignancy Classification
<italic>Goal:</italic> This paper investigates using Low-Rank Adaptation (LoRA) to adapt large vision models (LVMs) pretrained with self-supervised learning (SSL) for lung nodule malignancy classification. Inspired by LoRA's success in the field of Natural Language Processing,...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Open Journal of Engineering in Medicine and Biology |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10843806/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1825207041604452352 |
|---|---|
| author | Benjamin P. Veasey Amir A. Amini |
| author_facet | Benjamin P. Veasey Amir A. Amini |
| author_sort | Benjamin P. Veasey |
| collection | DOAJ |
| description | <italic>Goal:</italic> This paper investigates using Low-Rank Adaptation (LoRA) to adapt large vision models (LVMs) pretrained with self-supervised learning (SSL) for lung nodule malignancy classification. Inspired by LoRA's success in the field of Natural Language Processing, we hypothesized that such an adaptation technique can significantly improve classification performance, parameter efficiency, and training speed for the novel application of lung image cancer diagnostic. <italic>Methods:</italic> Utilizing two comprehensive lung nodule datasets, NLSTx and LIDC, which together encompass a diverse array of biopsy- and radiologist-confirmed lung CT scans, our rigorous experimental setup demonstrates that LoRA-adapted models markedly surpass traditional fine-tuning methods. <italic>Results:</italic> The best LoRA-adapted model achieved a 3% increase in ROC AUC over the state-of-the-art model, utilized 89.9% fewer parameters, and reduced training times by 36.5%. <italic>Conclusions:</italic> Integrating LoRA with out-of-domain pretrained LVMs offers a promising avenue for enhancing performance of lung nodule malignancy classification. The annotations for the NLSTx dataset are also released with this paper on GitHub at <uri>https://github.com/benVZ/NLSTx</uri>. |
| format | Article |
| id | doaj-art-aec14f0a479f4c7c9f32667a362bb30e |
| institution | Kabale University |
| issn | 2644-1276 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Open Journal of Engineering in Medicine and Biology |
| spelling | doaj-art-aec14f0a479f4c7c9f32667a362bb30e2025-02-07T00:01:59ZengIEEEIEEE Open Journal of Engineering in Medicine and Biology2644-12762025-01-01629630410.1109/OJEMB.2025.353084110843806Low-Rank Adaptation of Pre-Trained Large Vision Models for Improved Lung Nodule Malignancy ClassificationBenjamin P. Veasey0https://orcid.org/0000-0002-7012-4493Amir A. Amini1https://orcid.org/0000-0001-9698-8989Medical Imaging Laboratory, University of Louisville, Louisville, KY, USAMedical Imaging Laboratory, University of Louisville, Louisville, KY, USA<italic>Goal:</italic> This paper investigates using Low-Rank Adaptation (LoRA) to adapt large vision models (LVMs) pretrained with self-supervised learning (SSL) for lung nodule malignancy classification. Inspired by LoRA's success in the field of Natural Language Processing, we hypothesized that such an adaptation technique can significantly improve classification performance, parameter efficiency, and training speed for the novel application of lung image cancer diagnostic. <italic>Methods:</italic> Utilizing two comprehensive lung nodule datasets, NLSTx and LIDC, which together encompass a diverse array of biopsy- and radiologist-confirmed lung CT scans, our rigorous experimental setup demonstrates that LoRA-adapted models markedly surpass traditional fine-tuning methods. <italic>Results:</italic> The best LoRA-adapted model achieved a 3% increase in ROC AUC over the state-of-the-art model, utilized 89.9% fewer parameters, and reduced training times by 36.5%. <italic>Conclusions:</italic> Integrating LoRA with out-of-domain pretrained LVMs offers a promising avenue for enhancing performance of lung nodule malignancy classification. The annotations for the NLSTx dataset are also released with this paper on GitHub at <uri>https://github.com/benVZ/NLSTx</uri>.https://ieeexplore.ieee.org/document/10843806/Low-rank adaptationlung cancernodule classificationparameter-efficient fine-tuningvision transformers |
| spellingShingle | Benjamin P. Veasey Amir A. Amini Low-Rank Adaptation of Pre-Trained Large Vision Models for Improved Lung Nodule Malignancy Classification IEEE Open Journal of Engineering in Medicine and Biology Low-rank adaptation lung cancer nodule classification parameter-efficient fine-tuning vision transformers |
| title | Low-Rank Adaptation of Pre-Trained Large Vision Models for Improved Lung Nodule Malignancy Classification |
| title_full | Low-Rank Adaptation of Pre-Trained Large Vision Models for Improved Lung Nodule Malignancy Classification |
| title_fullStr | Low-Rank Adaptation of Pre-Trained Large Vision Models for Improved Lung Nodule Malignancy Classification |
| title_full_unstemmed | Low-Rank Adaptation of Pre-Trained Large Vision Models for Improved Lung Nodule Malignancy Classification |
| title_short | Low-Rank Adaptation of Pre-Trained Large Vision Models for Improved Lung Nodule Malignancy Classification |
| title_sort | low rank adaptation of pre trained large vision models for improved lung nodule malignancy classification |
| topic | Low-rank adaptation lung cancer nodule classification parameter-efficient fine-tuning vision transformers |
| url | https://ieeexplore.ieee.org/document/10843806/ |
| work_keys_str_mv | AT benjaminpveasey lowrankadaptationofpretrainedlargevisionmodelsforimprovedlungnodulemalignancyclassification AT amiraamini lowrankadaptationofpretrainedlargevisionmodelsforimprovedlungnodulemalignancyclassification |