Comparison of LSTM and RNN model performance in predicting F-18 NaF kinetics in prostate cancer bone metastasis based on PET/CT
DOI:
https://doi.org/10.61511/crsusf.v2i2.1951Keywords:
long short term memory (LSTM), NaF, prostate, recurrent neural network (RNN)Abstract
Background: Positron Emission Tomography-Computed Tomography (PET/CT) imaging using F-18 NaF is an important modality for evaluating bone metastasis in prostate cancer. The accuracy of this tracer kinetics prediction can improve monitoring of therapeutic response. Although Recurrent Neural Network (RNN) and Long Short Term Memory (LSTM) have been used for modeling sequential data, no comprehensive study has specifically compared their performance for predicting F-18 NaF uptake in prostate cancer, despite the clinical importance of cost and PET/CT service availability. Methods: This study analyzed data from nine patients in the NaF PROSTATE dataset The Cancer Imaging Archive (TCIA). SUVmean was extracted from PET/CT imaging series. Bidirectional LSTM and RNN models with dropout layers were developed. Evaluation was performed using R², RMSE, MAE metrics, a nd kinetic analysis via biexponential curve fitting to assess the biological plausibility of predictions. Findings: Evaluation of model LSTM demonstrated superior performance than RNN. Kinetic curve analysis confirmed that LSTM was able to reproduce uptake and clearance patterns more stably and physiologically than RNN, which showed fluctuations. These findings are consistent with the theoretical advantage of LSTM in handling long - term dependencies. Conclusion: LSTM is proven to be superior to RNN in predicting the kinetics of F-18 NaF in prostate cancer bone metastases, both statistically and clinically. Its accuracy and stability support its potential application in molecular imaging and therapy monitoring. Novelty/Originality of this article: This study provides quantitative evidence of LSTM's superiority over RNN for predicting F-18 NaF kinetics, using an innovative validation approach through kinetic curve analysis that enriches clinical assessment beyond conventional statistical metrics.
References
Carson, R. E., & Feng, D. (2021). Compartmental modeling in dynamic PET: Limitations and challenges in clinical translation. Annals of Nuclear Medicine, 35(8), 833–844. https://doi.org/10.1007/s12149-021-01664-5
Clark, K., Vendt, B., Smith, K., Freymann, J., Kirby, J., Koppel, P., Moore, S., Phillips, S., Maffitt, D., Pringle, M., Tarbox, L., & Prior, F. (2013). The Cancer Imaging Archive (TCIA): Maintaining and operating a public information repository. Journal of Digital Imaging, 26(6), 1045–1057. https://doi.org/10.1007/s10278-013-9622-7
Coleman, R. E. (2006). Clinical features of metastatic bone disease and risk of skeletal morbidity. Clinical Cancer Research, 12(20), 6243s–6249s. https://doi.org/10.1158/1078-0432.CCR-06-0931
Fedorov, A., Beichel, R., Kalpathy-Cramer, J., Finet, J., Fillion-Robin, J. C., Pujol, S., Bauer, C., Jennings, D., Fennessy, F., Sonka, M., Buatti, J., Aylward, S., Miller, J. V., Pieper, S., & Kikinis, R. (2012). 3D Slicer as an image computing platform for the Quantitative Imaging Network. Magnetic Resonance Imaging, 30(9), 1323–1341. https://doi.org/10.1016/j.mri.2012.05.001
Grant, F. D., Fahey, F. H., Packard, A. B., Davis, R. T., Alavi, A., & Treves, S. T. (2008). Skeletal PET with 18F-NaF: Applying new technology to an old tracer. Journal of Nuclear Medicine, 49(1), 68–78. https://doi.org/10.2967/jnumed.107.045328
Gunn, R. N., Gunn, S. R., & Cunningham, V. J. (2001). Positron emission tomography compartmental models. Journal of Cerebral Blood Flow & Metabolism, 21(6), 635–652. https://doi.org/10.1097/00004647-200106000-00002
Herzog, H., Lassen, M. L., & Ahn, S. (2020). Reducing scan duration in dynamic PET using AI-based reconstruction of missing time frames. European Journal of Nuclear Medicine and Molecular Imaging, 47(13), 3051–3061. https://doi.org/10.1007/s00259-020-04888-7
Hospodarskyy, O., Martsenyuk, V., Kukharska, N., Hospodarskyy, A., & Sverstiuk, S. (2024, June). Understanding the Adam Optimization Algorithm in Machine Learning. 2nd International Workshop on Computer Information Technologies in Industry 4.0 . https://doi.org/10.1016/j.ejmp.2023.103192
Iagaru, A., & Mittra, E. (2012). Advances in PET detection of bone metastases in prostate cancer. Seminars in Nuclear Medicine, 42(6), 465–480. https://doi.org/10.1053/j.semnuclmed.2012.07.005
Ivashchenko, O. V., O’Doherty, J., Hardiansyah, D., Cremonesi, M., Tran-Gia, J., Hippeläinen, E., Stokke, C., Grassi, E., Sandström, M., & Glatting, G. (2024). Time-activity data fitting in molecular radiotherapy: Methodology and pitfalls. Physica Medica, 117, 103192. https://doi.org/10.1016/j.ejmp.2023.103192
Jin, J. K., Dayyani, F., & Gallick, G. E. (2011). Steps in prostate cancer progression that led to bone metastasis. International Journal of Cancer, 128(11), 2545–2561. https://doi.org/10.1002/ijc.26024
Kaalep, A., Sera, T., Rijnsdorp, S., Yaqub, M., Tohka, J., Boellaard, R., & Simplified Kinetic Modeling Taskforce. (2021). Variability and reproducibility of compartmental model outcome parameters in dynamic PET. EJNMMI Physics, 8(1), 77. https://doi.org/10.1186/s40658-019-0265-8
Karakatsanis, N. A., Lodge, M. A., Tahari, A. K., Zhou, Y., Wahl, R. L., & Rahmim, A. (2013). Dynamic whole-body PET parametric imaging: Concept, acquisition protocol optimization and clinical application. Physics in Medicine and Biology, 58(20), 7391–7418. https://doi.org/10.1088/0031-9155/58/20/7391
Kinahan, P. E., & Fletcher, J. W. (2010). Positron emission tomography-computed tomography standardized uptake values in clinical practice and assessing response to therapy. Seminars in Ultrasound, CT and MRI, 31(6), 496–505. https://doi.org/10.1053/j.sult.2010.10.001
Kurdziel, K. A. , Apolo, A. B. , Lindenberg, L. , Mena, E. , McKinney, Y. Y. , Adler, S. S. , Turkbey, B. , Dahut, W. , Gulley, J. L. , Madan, R. A. , Landgren, O. , & Choyke, P. L. (2015). Data From NaF PROSTATE (Version 1) [Data set]. The Cancer Imaging Archive. https://doi.org/https://doi.org/10.7937/K9/TCIA.2015.ISOQTHKO
Kurdziel, K. A., Shih, J. H., Apolo, A. B., Lindenberg, L., Mena, E., McKinney, Y. Y., Adler, S. S., Turkbey, B., Dahut, W., Gulley, J. L., Madan, R. A., Landgren, O., & Choyke, P. L. (2012). The kinetics and reproducibility of 18F-sodium fluoride for oncology using current PET camera technology. Journal of Nuclear Medicine, 53(8), 1175–1184. https://doi.org/10.2967/jnumed.111.100883
Lei, Y., Dong, X., Higgins, K., Liu, T., Curran, W. J., Mao, H., & Yang, X. (2022). Deep learning for dynamic PET parametric imaging using recurrent networks. Physics in Medicine & Biology, 67(6), 065002. https://doi.org/10.1088/1361-6560/ac4b6d
Liu, G., Li, Y., Zhang, X., Chen, W., & Cherry, S. R. (2023). Total-body PET kinetic modeling: Opportunities and challenges for multi-organ quantification. European Journal of Nuclear Medicine and Molecular Imaging, 50(9), 2345–2358. https://doi.org/10.1007/s00259-023-06299-w
Muraina, & Olaniyi, I. (2022). Ideal Dataset Splitting Ratios In Machine Learning Algorithms: General Concerns For Data Scientists And Data Analysts. 7th International Mardin Artuklu Scientific Researches Conference, 496–504. https://www.researchgate.net/publication/358284895
Muzi, M., Mankoff, D. A., Grierson, J. R., Wells, J. M., Vesselle, H., & Krohn, K. A. (2012). Kinetic modeling of 18F-NaF PET images of bone metastases. Journal of Nuclear Medicine, 53(12), 1904–1910. https://doi.org/10.2967/jnumed.112.105080
Muzic, R. F., & Cornelius, S. (2001). COMKAT: Compartment model kinetic analysis tool. Journal of Nuclear Medicine, 42(4), 636–645. https://jnm.snmjournals.org/content/42/4/636
Muzic, R. F., Lodge, M. A., & Rahmim, A. (2025). Advances in PET tracer kinetic modeling in the era of total-body scanners. Annals of Nuclear Medicine, 39(2), 215–227. https://doi.org/10.1007/s12149-025-02014-x
Rahmim, A., & Zaidi, H. (2008). PET versus SPECT: Strengths, limitations and challenges. Nuclear Medicine Communications, 29(3), 193–207. https://doi.org/10.1097/MNM.0b013e3282f3a515
Rajkomar, A., Dean, J., & Kohane, I. (2018). Machine learning in medicine. New England Journal of Medicine, 380(14), 1347–1358. https://doi.org/10.1056/NEJMra1814259
Schwartz, J., Humm, J. L., Gonen, M., Kalaigian, H., Schoder, H., Larson, S. M., & Nehmeh, S. A. (2011). Repeatability of SUV measurements in serial PET. Medical Physics, 38(5), 2629–2638. https://doi.org/10.1118/1.3578604
Shen, W., Zhou, M., Yang, F., & Yang, C. (2020). Predicting 18F-FDG PET dynamic images using deep learning for dose reduction. Physics in Medicine & Biology, 65(11), 115005. https://doi.org/10.1088/1361-6560/ab896d
Tang, J., & Rahmim, A. (2021). Emerging approaches beyond compartmental models: Machine learning for PET kinetic analysis. International Journal of Imaging Systems and Technology, 31(3), 1122–1135. https://doi.org/10.1002/ird3.66
Thie, J. A. (2004). Understanding the Standardized Uptake Value, Its Methods, and Implications for Usage. Journal of Nuclear Medicine, 45(9), 1431–1434. https://jnm.snmjournals.org/content/45/9/1431
Townsend, D. W. (2020). Positron emission tomography/computed tomography. Seminars in Nuclear Medicine, 50(2), 123–131. https://doi.org/10.1053/j.semnuclmed.2019.11.003
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, Ł., & Polosukhin, I. (2017). Attention is all you need. Advances in Neural Information Processing Systems (NeurIPS), 30, 5998–6008. https://doi.org/10.48550/arXiv.1706.03762
Wang, G., Qi, J., & Chen, K. (2021). Deep learning for PET parametric imaging: Accurate kinetic modeling from reduced acquisition protocols. Medical Physics, 48(6), 3204–3214. https://doi.org/10.1002/mp.14873
Downloads
Published
Issue
Section
Citation Check
License
Copyright (c) 2025 Sani Inuwa Sani, Mohammad Sidik Cahyana, Haryanto
This work is licensed under a Creative Commons Attribution 4.0 International License.
