RadDenseNet: An Explainable AI Model for Detection of Hepatic Steatosis from Abdominal CT Images
Abstract
Hepatic steatosis affects approximately 30.2% of the global population and, if undetected, can silently progress to non-alcoholic steatohepatitis, cirrhosis, and hepatocellular carcinoma. Manual CT interpretation for fatty liver assessment is time-consuming and prone to inter-observer variability, motivating research into automated screening tools. We propose RadDenseNet, a DenseNet121 model initialized with RadImageNet weights pretrained on 1.35 million medical images spanning CT, MRI, and ultrasound, and evaluate it as a research prototype. The first 397 of 427 backbone layers were frozen, and the remaining 30 layers were fine-tuned alongside a classification head comprising global average pooling, a dense layer of 256 units, batch normalization, and dropout of 0.5. The model was trained on the Kaggle Liver CT Image Dataset - 3,557 abdominal CT images across Healthy and Hepatic Steatosis classes - using a stratified 70/20/10 split with 5-fold cross-validation. RadDenseNet was compared against six baseline models spanning ImageNet and RadImageNet pretraining across ResNet50, InceptionV3, and DenseNet121 architectures. The decision threshold was optimized on the validation set, yielding τ = 0.48. On the held-out test set, RadDenseNet achieved 96.35% accuracy, AUC of 0.9966, sensitivity of 0.9538, specificity of 0.9752, precision of 0.9789, and F1-score of 0.9662, with 5-fold cross-validation confirming stable performance (mean accuracy 96.46% ± 0.68%, mean AUC 0.9962 ± 0.0014). RadImageNet pretraining consistently outperformed ImageNet pretraining across all architectures, with accuracy improvements of 2.81 to 5.90 percentage points. Grad-CAM analysis covering all four prediction categories revealed that false positives arose from extra-hepatic attention and false negatives from weak parenchymal activation - clinically meaningful failure modes. External validation on multi-center datasets, extension to multi-class steatosis grading, and liver ROI segmentation as preprocessing are recommended as future directions
Downloads
References
Z. M. Younossi, P. Golabi, J. M. Paik, A. Henry, C. Van Natta, and L. Younossi, "The global epidemiology of nonalcoholic fatty liver disease (NAFLD) and nonalcoholic steatohepatitis (NASH): a systematic review," Hepatology, vol. 77, no. 4, pp. 1335–1347, 2023, doi: 10.1097/HEP.0000000000000004.
K. Riazi, H. Azhari, J. H. Charette, F. E. Underwood, J. A. King, E. E. Afshar, M. G. Swain, S. S. Congly, G. G. Kaplan, and F. Shaheen, "The prevalence and incidence of NAFLD worldwide: a systematic review and meta-analysis," Lancet Gastroenterol. Hepatol., vol. 7, no. 9, pp. 851–861, 2022, doi: 10.1016/S2468-1253(22)00165-0.
D. Georgescu, D. F. Lighezan, C. I. Rosca, O. M. Bratu, and C. Diaconu, "NASH/NAFLD-related hepatocellular carcinoma: an added burden," Life (Basel), vol. 14, no. 1, p. 25, 2023, doi: 10.3390/life14010025.
M. J. Zhu, X. H. Sun, and X. M. Li, "Pathogenesis of NAFLD-related hepatocellular carcinoma: an up-to-date review," J. Hepatocell. Carcinoma, vol. 10, pp. 347–356, 2023, doi: 10.2147/JHC.S400231.
J. Starekova, D. Hernando, P. J. Pickhardt, and S. B. Reeder, "Quantification of liver fat content with CT and MRI: state of the art," Radiology, vol. 301, no. 2, pp. 250–262, 2021, doi: 10.1148/radiol.2021204288.
H. N. Kim, H. J. Jeon, H. G. Choi, I. S. Kwon, W. S. Ryu, T. H. Lee, S. H. Kim, and B. S. Lee, "CT-based Hounsfield unit values reflect the degree of steatohepatitis in patients with low-grade fatty liver disease," BMC Gastroenterol., vol. 23, no. 1, p. 77, 2023, doi: 10.1186/s12876-023-02706-4.
S. Survarachakan, R. Prasad, Z. S. A. Rahman, J. de Frutos, F. Lindseth, and T. L. A. van Walsum, "Deep learning for image-based liver analysis — a comprehensive review focusing on malignant lesions," Artif. Intell. Med., vol. 130, p. 102331, 2022, doi: 10.1016/j.artmed.2022.102331.
K. Xiang, B. Jiang, and D. Shang, "The overview of the deep learning integrated into the medical imaging of liver," Hepatol. Int., vol. 15, no. 4, pp. 868–880, 2021, doi: 10.1007/s12072-021-10229-z.
G. Litjens, T. Kooi, B. E. Bejnordi, A. A. A. Setio, F. Ciompi, M. Ghafoorian, J. A. W. M. van der Laak, B. van Ginneken, and C. I. Sánchez, "A survey on deep learning in medical image analysis," Med. Image Anal., vol. 42, pp. 60–88, 2017, doi: 10.1016/j.media.2017.07.005.
D. Ardila, A. P. Kiraly, S. Bharadwaj, B. Choi, J. J. Reicher, L. Peng, D. Tse, M. Etemadi, W. Ye, G. Corrado, D. P. Naidich, and S. Shetty, "End-to-end lung cancer screening with three-dimensional deep learning on low-dose chest computed tomography," Nat. Med., vol. 25, no. 6, pp. 954–961, 2019, doi: 10.1038/s41591-019-0447-x.
S. Deepak and P. M. Ameer, "Brain tumor classification using deep CNN features via transfer learning," Comput. Biol. Med., vol. 111, p. 103345, 2019, doi: 10.1016/j.compbiomed.2019.103345.
D. Alzu'bi, A. Abdullah, M. Hmeidi, R. AlAzab, M. Gharaibeh, A. El-Heis, A. Almotairi, A. Forestiero, C. Yuan, and G. Merigó, "Kidney tumor detection and classification based on deep learning approaches: a new dataset in CT scans," J. Healthc. Eng., vol. 2022, p. 3861161, 2022, doi: 10.1155/2022/3861161.
Z. Zhang, G. Li, Z. Wang, F. Xia, N. Zhao, H. Nie, Z. Ye, J. S. Lin, Y. Hui, and X. Liu, "Deep-learning segmentation to select liver parenchyma for categorizing hepatic steatosis on multinational chest CT," Sci. Rep., vol. 14, p. 11987, 2024, doi: 10.1038/s41598-024-62887-2.
H. Zhang, J. Liu, D. Su, Z. Bai, Y. Wu, Y. Ma, Q. Miao, M. Wang, and X. Yang, "Diagnostic of fatty liver using radiomics and deep learning models on non-contrast abdominal CT," PLoS ONE, vol. 20, no. 2, p. e0310938, 2025, doi: 10.1371/journal.pone.0310938.
P. M. Graffy, V. Sandfort, R. M. Summers, and P. J. Pickhardt, "Automated liver fat quantification at nonenhanced abdominal CT for population-based steatosis assessment," Radiology, vol. 293, no. 2, pp. 334–342, 2019, doi: 10.1148/radiol.2019190512.
J. S. Leiby, M. E. Lee, E. K. Choe, and D. Kim, "Data-driven classification of fatty liver from 3D unenhanced abdominal CT scans," in Proc. Machine Learning in Medical Imaging (MLMI 2023), Lecture Notes in Computer Science, vol. 14349, Springer, Cham, pp. 264–272, 2024, doi: 10.1007/978-3-031-45676-3_27.
X. Mei, Z. Liu, P. M. Robson, B. Marinelli, M. Huang, A. Doshi, A. Jacobi, C. Cao, K. E. Link, T. Yang, Y. Wang, H. Greenspan, T. Deyer, Z. A. Fayad, and Y. Yang, "RadImageNet: an open radiologic deep learning research dataset for effective transfer learning," Radiol. Artif. Intell., vol. 4, no. 5, p. e210315, 2022, doi: 10.1148/ryai.210315.
UCI Machine Learning, "Liver CT Image Dataset," Kaggle, 2024. [Online]. Available: https://www.kaggle.com/datasets/ucimachinelearning/liver-ct-image-dataset.
J. D. Kelleher, B. Mac Namee, and A. D'Arcy, Fundamentals of Machine Learning for Predictive Data Analytics: Algorithms, Worked Examples, and Case Studies, 2nd ed. MIT Press, Cambridge, MA, USA, 2020, ISBN: 978-0262044691.
C. Shorten and T. M. Khoshgoftaar, "A survey on image data augmentation for deep learning," J. Big Data, vol. 6, no. 1, p. 60, 2019, doi: 10.1186/s40537-019-0197-0.
J. Deng, W. Dong, R. Socher, L. J. Li, K. Li, and L. Fei-Fei, "ImageNet: a large-scale hierarchical image database," in Proc. IEEE Conf. Computer Vision and Pattern Recognition (CVPR), Miami, FL, USA, pp. 248–255, 2009, doi: 10.1109/CVPR.2009.5206848.
K. He, X. Zhang, S. Ren, and J. Sun, "Deep residual learning for image recognition," in Proc. IEEE Conf. Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA, pp. 770–778, 2016, doi: 10.1109/CVPR.2016.90.
C. Szegedy, V. Vanhoucke, S. Ioffe, J. Shlens, and Z. Wojna, "Rethinking the inception architecture for computer vision," in Proc. IEEE Conf. Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA, pp. 2818–2826, 2016, doi: 10.1109/CVPR.2016.308.
G. Huang, Z. Liu, L. Van Der Maaten, and K. Q. Weinberger, "Densely connected convolutional networks," in Proc. IEEE Conf. Computer Vision and Pattern Recognition (CVPR), Honolulu, HI, USA, pp. 4700–4708, 2017, doi: 10.1109/CVPR.2017.243.
M. Raghu, C. Zhang, J. Kleinberg, and S. Bengio, "Transfusion: understanding transfer learning for medical imaging," in Proc. Advances in Neural Information Processing Systems (NeurIPS), vol. 32, Vancouver, Canada, pp. 3347–3357, 2019. [Online]. Available: https://proceedings.neurips.cc/paper/2019/hash/eb1e78328c46506b46a4ac4a1e378b91-Abstract.html.
D. P. Kingma and J. Ba, "Adam: a method for stochastic optimization," in Proc. Int. Conf. Learning Representations (ICLR), San Diego, CA, USA, 2015. [Online]. Available: https://openreview.net/forum?id=8gmWwjFyLj.
I. Goodfellow, Y. Bengio, and A. Courville, Deep Learning. MIT Press, Cambridge, MA, USA, 2016. [Online]. Available: https://www.deeplearningbook.org.
M. Buda, A. Maki, and M. A. Mazurowski, "A systematic study of the class imbalance problem in convolutional neural networks," Neural Netw., vol. 106, pp. 249–259, 2018, doi: 10.1016/j.neunet.2018.07.011.
L. N. Smith, "Cyclical learning rates for training neural networks," in Proc. IEEE Winter Conf. Applications of Computer Vision (WACV), Santa Rosa, CA, USA, pp. 464–472, 2017, doi: 10.1109/WACV.2017.45.
R. R. Selvaraju, M. Cogswell, A. Das, R. Vedantam, D. Parikh, and D. Batra, "Grad-CAM: visual explanations from deep networks via gradient-based localization," Int. J. Comput. Vis., vol. 128, no. 2, pp. 336–359, 2020, doi: 10.1007/s11263-019-01228-7.
A. B. Arrieta, N. Díaz-Rodríguez, J. Del Ser, A. Bennetot, S. Tabik, A. Barbado, S. García, S. Gil-López, D. Molina, R. Benjamins, R. Chatila, and F. Herrera, "Explainable artificial intelligence (XAI): concepts, taxonomies, opportunities and challenges toward responsible AI," Inf. Fusion, vol. 58, pp. 82–115, 2020, doi: 10.1016/j.inffus.2019.12.012.
T. Saito and M. Rehmsmeier, "The precision-recall plot is more informative than the ROC plot when evaluating binary classifiers on imbalanced datasets," PLoS ONE, vol. 10, no. 3, p. e0118432, 2015, doi: 10.1371/journal.pone.0118432.
T. Fawcett, "An introduction to ROC analysis," Pattern Recognit. Lett., vol. 27, no. 8, pp. 861–874, 2006, doi: 10.1016/j.patrec.2005.10.010.
S. J. Pan and Q. Yang, "A survey on transfer learning," IEEE Trans. Knowl. Data Eng., vol. 22, no. 10, pp. 1345–1359, 2010, doi: 10.1109/TKDE.2009.191.
A. Raza, M. S. Alshehri, S. Almakdi, A. A. Siddique, M. Alsulami, and M. Alhaisoni, "Enhancing brain tumor classification with transfer learning: leveraging DenseNet121 for accurate and efficient detection," Int. J. Imaging Syst. Technol., vol. 34, no. 1, p. e22957, 2024, doi: 10.1002/ima.22957.
S. A. Harrison and C. P. Day, "Benefits of lifestyle modification in NAFLD," Gut, vol. 56, no. 12, pp. 1760–1769, 2007, doi: 10.1136/gut.2006.112094.
P. J. Pickhardt, P. M. Graffy, M. G. Zea, A. T. Sandfort, and R. M. Summers, "Utilizing fully automated abdominal CT-based biomarkers for opportunistic screening for metabolic syndrome," AJR Am. J. Roentgenol., vol. 215, no. 5, pp. 1173–1179, 2020, doi: 10.2214/AJR.19.22537.
R. Benjamens, P. Dhunnoo, and B. Meijboom, "The state of artificial intelligence-based FDA-approved medical devices and algorithms: an online database," NPJ Digit. Med., vol. 3, no. 1, p. 118, 2020, doi: 10.1038/s41746-020-00324-0.
P. Bilic, P. Christ, H. B. Li et al., "The liver tumor segmentation benchmark (LiTS)," Med. Image Anal., vol. 84, p. 102680, 2023, doi: 10.1016/j.media.2022.102680.
A. L. Rinella, M. E. Lazarus, V. Ratziu, F. M. Francque, A. Sanyal, L. Kanwal, et al., "A multisociety Delphi consensus statement on new fatty liver disease nomenclature," Hepatology, vol. 78, no. 6, pp. 1966–1986, 2023, doi: 10.1097/HEP.0000000000000520.
M. Ekstedt, H. Hagstrom, P. Nasr, M. Fredrikson, P. Stal, R. Kechagias, and S. Hultcrantz, "Fibrosis stage is the strongest predictor for disease-specific mortality in NAFLD after up to 33 years of follow-up," Hepatology, vol. 61, no. 5, pp. 1547–1554, 2015, doi: 10.1002/hep.27368.
J. R. Zech, M. A. Badgeley, M. Liu, A. B. Costa, J. J. Titano, and E. K. Oermann, "Variable generalization performance of a deep learning model to detect pneumonia in chest radiographs: a cross-sectional study," PLOS Med., vol. 15, no. 11, p. e1002686, 2018, doi: 10.1371/journal.pmed.1002686.
M. Q. Pembury Smith and G. D. Ruxton, "Effective use of the McNemar test," Behav. Ecol. Sociobiol., vol. 74, no. 11, p. 133, 2020, doi: 10.1007/s00265-020-02916-z.
S. Ioffe and C. Szegedy, "Batch normalization: accelerating deep network training by reducing internal covariate shift," in Proc. 32nd Int. Conf. Machine Learning (ICML), Lille, France, vol. 37, pp. 448–456, 2015. [Online]. Available: http://proceedings.mlr.press/v37/ioffe15.html.
N. Srivastava, G. Hinton, A. Krizhevsky, I. Sutskever, and R. Salakhutdinov, "Dropout: a simple way to prevent neural networks from overfitting," J. Mach. Learn. Res., vol. 15, no. 1, pp. 1929–1958, 2014. [Online]. Available: https://jmlr.org/papers/v15/srivastava14a.html.
C. Guo, G. Pleiss, Y. Sun, and K. Q. Weinberger, "On calibration of modern neural networks," in Proc. 34th Int. Conf. Machine Learning (ICML), Sydney, Australia, vol. 70, pp. 1321–1330, 2017. [Online].Available:http://proceedings.mlr.press/v70/guo17a.html.
R. C. Gonzalez and R. E. Woods, Digital Image Processing, 4th ed. Pearson, New York, USA, 2018, ISBN: 978-0133356724.
S. Vabalas, E. Gowen, E. Poliakoff, and A. J. Casson, "Machine learning algorithm validation with a limited sample size," PLOS ONE, vol. 14, no. 11, p. e0224365, 2019, doi: 10.1371/journal.pone.0224365.
B. Norgeot et al., "Minimum information about clinical artificial intelligence modeling: the MI-CLAIM checklist," Nat. Med., vol. 26, no. 9, pp. 1320–1324, 2020, doi: 10.1038/s41591-020-1041-y.
H. Y. Kim, K. J. Lee, S. S. Lee, S. J. Choi, D. H. Kim, S. Heo, H. J. Jang, and S. H. Choi, “Diagnosis of moderate-to-severe hepatic steatosis using deep learning-based automated attenuation measurements on contrast-enhanced CT”. Abdom. Radiol., vol. 50, no. 9, pp. 4139-4147, 2025, doi: 10.1007/s00261-025-04872-5.
V. Nivethitha, R. A. Daniel, A. Debnath, V. Dwarakanathan, G. Jeer, and G. Kavipriya, “Diagnostic accuracy of artificial intelligence models for imaging detection of hepatic steatosis through systematic review and meta-analysis”. Sci. Rep., vol. 15, no. 1, p. 34408, 2025, doi: 10.1038/s41598-025-17386-3.
Copyright (c) 2026 Rahul LR, Satya Prasad Gaddamedi
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution-ShareAlikel 4.0 International (CC BY-SA 4.0) that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).
.png)
.png)
.png)
.png)
.png)
