CNN-based prediction using early post-radiotherapy MRI as a proxy for toxicity in the murine head and neck
DOI:
https://doi.org/10.2340/1651-226X.2025.44020Keywords:
head and neck cancer, mice, radiotherapy, deep learning, convolutional neural network, toxicity early detection, magnetic resonance imagingAbstract
Background and purpose: Radiotherapy (RT) of head and neck cancer can cause severe toxicities. Early identification of individuals at risk could enable personalized treatment. This study evaluated whether convolutional neural networks (CNNs) applied to Magnetic Resonance (MR) images acquired early after irradiation can predict radiation-induced tissue changes associated with toxicity in mice.
Patient/material and methods: Twenty-nine C57BL/6JRj mice were included (irradiated: n = 14; control:
n = 15). Irradiated mice received 65 Gy of fractionated RT to the oral cavity, swallowing muscles and salivary glands. T2-weighted MR images were acquired 3–5 days post-irradiation. CNN models (VGG, MobileNet, ResNet, EfficientNet) were trained to classify sagittal slices as irradiated or control (n = 586 slices). Predicted class probabilities were correlated with five toxicity endpoints assessed 8–105 days post-irradiation. Model explainability was assessed with VarGrad heatmaps, to verify that predictions relied on clinically relevant image regions.
Results: The best-performing model (EfficientNet B3) achieved 83% slice-level accuracy (ACC) and correctly classified 28 of 29 mice. Higher predicted probabilities of the irradiated class were strongly associated with oral mucositis, dermatitis, reduced saliva production, late submandibular gland fibrosis and atrophy of salivary gland acinar cells. Explainability heatmaps confirmed that CNNs focused on irradiated regions.
Interpretation: The high CNN classification ACC, the regions highlighted by the explainability analysis and the strong correlations between model predictions and toxicity suggest that CNNs, together with post-irradiation magnetic resonance imaging, may identify individuals at risk of developing toxicity.
Downloads
References
Van den Bosch L, van der Schaaf A, van der Laan HP, Hoebers FJP, Wijers OB, van den Hoek JGM, et al. Comprehensive toxicity risk profiling in radiation therapy for head and neck cancer: a new concept for individually optimised treatment. Radiother Oncol. 2021;157:147–54.
https://doi.org/10.1016/j.radonc.2021.01.024 DOI: https://doi.org/10.1016/j.radonc.2021.01.024
van der Laan HP, Van den Bosch L, Schuit E, Steenbakkers RJHM, van der Schaaf A, Langendijk JA. Impact of radiation-induced toxicities on quality of life of patients treated for head and neck cancer. Radiother Oncol. 2021;160:47–53.
https://doi.org/10.1016/j.radonc.2021.04.011 DOI: https://doi.org/10.1016/j.radonc.2021.04.011
Hunter M, Kellett J, Toohey K, D’Cunha NM, Isbel S, Naumovski N. Toxicities caused by head and neck cancer treatments and their influence on the development of malnutrition: review of the literature. Eur J Investig Health Psychol Educ. 2020;10(4):935–49.
https://doi.org/10.3390/ejihpe10040066 DOI: https://doi.org/10.3390/ejihpe10040066
Isaksson LJ, Pepa M, Zaffaroni M, Marvaso G, Alterio D, Volpe S, et al. Machine learning-based models for prediction of toxicity outcomes in radiotherapy. Front Oncol. 2020;10:790.
https://doi.org/10.3389/fonc.2020.00790 DOI: https://doi.org/10.3389/fonc.2020.00790
Pota M, Scalco E, Sanguineti G, Farneti A, Cattaneo GM, Rizzo G, et al. Early prediction of radiotherapy-induced parotid shrinkage and toxicity based on CT radiomics and fuzzy classification. Artif Intell Med. 2017;81:41–53.
https://doi.org/10.1016/j.artmed.2017.03.004 DOI: https://doi.org/10.1016/j.artmed.2017.03.004
Kawamura M, Yoshimura M, Asada H, Nakamura M, Matsuo Y, Mizowaki T. A scoring system predicting acute radiation dermatitis in patients with head and neck cancer treated with intensity-modulated radiotherapy. Radiat Oncol. 2019;14(1):14.
https://doi.org/10.1186/s13014-019-1215-2 DOI: https://doi.org/10.1186/s13014-019-1215-2
Samant P, de Ruysscher D, Hoebers F, Canters R, Hall E, Nutting C, et al. Machine learning for normal tissue complication probability prediction: predictive power with versatility and easy implementation. Clin Transl Radiat Oncol. 2023;39:100595.
https://doi.org/10.1016/j.ctro.2023.100595 DOI: https://doi.org/10.1016/j.ctro.2023.100595
Carbonara R, Bonomo P, Di Rito A, Didonna V, Gregucci F, Ciliberti MP, et al. Investigation of radiation-induced toxicity in head and neck cancer patients through radiomics and machine learning: a systematic review. J Oncol. 2021;2021:1–9.
https://doi.org/10.1155/2021/5566508 DOI: https://doi.org/10.1155/2021/5566508
Khajetash B, Mahdavi SR, Nikoofar A, Johnson L, Tavakoli M. Ensemble learning approach for prediction of early complications after radiotherapy for head and neck cancer using CT and MRI radiomic features. Sci Rep. 2025;15(1):14229.
https://doi.org/10.1038/s41598-025-93676-0 DOI: https://doi.org/10.1038/s41598-025-93676-0
Lambin P, Leijenaar RTH, Deist TM, Peerlings J, de Jong EEC, van Timmeren J, et al. Radiomics: the bridge between medical imaging and personalized medicine. Nat Rev Clin Oncol. 2017;14(12):749–62.
https://doi.org/10.1038/nrclinonc.2017.141 DOI: https://doi.org/10.1038/nrclinonc.2017.141
Zwanenburg A. Radiomics in nuclear medicine: robustness, reproducibility, standardization, and how to avoid data analysis traps and replication crisis. Eur J Nucl Med Mol Imaging. 2019;46(13):2638–55.
https://doi.org/10.1007/s00259-019-04391-8 DOI: https://doi.org/10.1007/s00259-019-04391-8
Afshar P, Mohammadi A, Plataniotis KN, Oikonomou A, Benali H. From handcrafted to deep-learning-based cancer radiomics: challenges and opportunities. IEEE Sig Process Mag. 2019;36(4):132–60.
https://doi.org/10.1109/msp.2019.2900993 DOI: https://doi.org/10.1109/MSP.2019.2900993
Reyes M, Meier R, Pereira S, Silva CA, Dahlweid F-M, von Tengg-Kobligk H, et al. On the interpretability of artificial intelligence in radiology: challenges and opportunities. Radiol Artif Intell. 2020;2(3):e190043.
https://doi.org/10.1148/ryai.2020190043 DOI: https://doi.org/10.1148/ryai.2020190043
Hooker S, Erhan D, Kindermans PJ, Kim B. A benchmark for interpretability methods in deep neural networks. Advances in neural information processing systems. 2019;32:9737-9748.
Schlaak RA, SenthilKumar G, Boerma M, Bergom C. Advances in preclinical research models of radiation-induced cardiac toxicity. Cancers. 2020;12(2):415.
https://doi.org/10.3390/cancers12020415 DOI: https://doi.org/10.3390/cancers12020415
Tillner F, Thute P, Bütof R, Krause M, Enghardt W. Pre-clinical research in small animals using radiotherapy technology – a bidirectional translational approach. Z Med Phys. 2014;24(4):335–51.
https://doi.org/10.1016/j.zemedi.2014.07.004 DOI: https://doi.org/10.1016/j.zemedi.2014.07.004
Kakar M, Huynh BN, Zlygosteva O, Juvkam IS, Edin N, Tomic O, et al. Attention-based vision transformer enables early detection of radiotherapy-induced toxicity in magnetic resonance images of a preclinical model. Technol Cancer Res Treat. 2025;24:15330338251333018.
https://doi.org/10.1177/15330338251333018 DOI: https://doi.org/10.1177/15330338251333018
Belgioia L, Morbelli SD, Corvò R. Prediction of response in head and neck tumor: focus on main hot topics in research. Front Oncol. 2021;10:604965.
https://doi.org/10.3389/fonc.2020.604965 DOI: https://doi.org/10.3389/fonc.2020.604965
Juvkam IS, Zlygosteva O, Arous D, Galtung HK, Malinen E, Søland TM, et al. A preclinical model to investigate normal tissue damage following fractionated radiotherapy to the head and neck. J Radiat Res. 2023;64(1):44–52.
https://doi.org/10.1093/jrr/rrac066 DOI: https://doi.org/10.1093/jrr/rrac066
Zlygosteva O, Juvkam IS, Arous D, Sitarz M, Sørensen BS, Ankjærgaard C, et al. Acute normal tissue responses in a murine model following fractionated irradiation of the head and neck with protons or X-rays. Acta Oncol. 2023;62(11):1574–80.
https://doi.org/10.1080/0284186x.2023.2254481 DOI: https://doi.org/10.1080/0284186X.2023.2254481
Juvkam IS, Zlygosteva O, Sitarz M, Sørensen BS, Aass HCD, Edin NJ, et al. Proton- compared to X-irradiation leads to more acinar atrophy and greater hyposalivation accompanied by a differential cytokine response. Sci Rep. 2024;14(1):22311.
https://doi.org/10.1038/s41598-024-73110-7 DOI: https://doi.org/10.1038/s41598-024-73110-7
Zlygosteva O, Juvkam IS, Aass HCD, Galtung HK, Søland TM, Malinen E, et al. Cytokine levels in saliva are associated with salivary gland fibrosis and hyposalivation in mice after fractionated radiotherapy of the head and neck. Int J Mol Sci. 2023;24(20):15218.
https://doi.org/10.3390/ijms242015218 DOI: https://doi.org/10.3390/ijms242015218
Percie du Sert N, Hurst V, Ahluwalia A, Alam S, Avey MT, Baker M, et al. The ARRIVE guidelines 2.0: updated guidelines for reporting animal research. BMC Vet Res. 2020;16(1):242.
https://doi.org/10.1186/s12917-020-02451-y DOI: https://doi.org/10.1186/s12917-020-02451-y
Huang L-K, Wang M-JJ. Image thresholding by minimizing the measures of fuzziness. Pattern Recognit. 1995;28(1):41–51.
https://doi.org/10.1016/0031-3203(94)e0043-k DOI: https://doi.org/10.1016/0031-3203(94)E0043-K
Beare R. Histogram-based thresholding – some missing methods. Insight J. 2011 Jan-Jun:3279-3284.
https://doi.org/10.54294/efycla DOI: https://doi.org/10.54294/efycla
Diehl PU, Neil D, Binas J, Cook M, Liu S-C, Pfeiffer M. Fast-classifying, high-accuracy spiking deep networks through weight and threshold balancing. In: 2015 International Joint Conference on Neural Networks (IJCNN); 2015. p. 1–8.
https://doi.org/10.1109/ijcnn.2015.7280696 DOI: https://doi.org/10.1109/IJCNN.2015.7280696
Howard AG, Zhu M, Chen B, Kalenichenko D, Wang W, Weyand T, et al. Mobilenets: efficient convolutional neural networks for mobile vision applications. arXiv preprint arXiv:170404861. 2017.
Wu Z, Shen C, van den Hengel A. Wider or deeper: revisiting the ResNet model for visual recognition. Pattern Recognit. 2019;90:119–33.
https://doi.org/10.1016/j.patcog.2019.01.006 DOI: https://doi.org/10.1016/j.patcog.2019.01.006
Tan M, Le Q. Efficientnet: Rethinking model scaling for convolutional neural networks. In: International conference on machine learning 2019 May 24 held in Long Beach, California, USA. pp. 6105-6114. PMLR 97.
Deng J, Dong W, Socher R, Li L-J, Kai L, Li F-F. ImageNet: a large-scale hierarchical image database. In: 2009 IEEE Conference on Computer Vision and Pattern Recognition; 2009. p. 248–55.
https://doi.org/10.1109/cvpr.2009.5206848 DOI: https://doi.org/10.1109/CVPR.2009.5206848
Cawley GC, Talbot NL. On over-fitting in model selection and subsequent selection bias in performance evaluation. J Machine Learn Res. 2010;11:2079–107.
Huynh BN, Groendahl AR, Tomic O, Liland KH, Knudtsen IS, Hoebers F, et al. Head and neck cancer treatment outcome prediction: a comparison between machine learning with conventional radiomics features and deep learning radiomics. Front Med. 2023;10:1217037.
https://doi.org/10.3389/fmed.2023.1217037 DOI: https://doi.org/10.3389/fmed.2023.1217037
Adebayo J, Gilmer J, Muelly M, Goodfellow I, Hardt M, Kim B. Sanity checks for saliency maps. Adv Neural Inform Process Syst. 2018;31:9505-9515.
Spearman C. The proof and measurement of association between two things. Am J Psychol. 1904;15(1):72–101.
https://doi.org/10.2307/1412159 DOI: https://doi.org/10.2307/1412159
Baak M, Koopman R, Snoek H, Klous S. A new correlation coefficient between categorical, ordinal and interval variables with Pearson characteristics. Comput Statist Data Anal. 2020;152:107043.
https://doi.org/10.1016/j.csda.2020.107043 DOI: https://doi.org/10.1016/j.csda.2020.107043
McKnight PE, Najab J. Mann‐whitney U test. The Corsini encyclopedia of psychology. 2010 Jan 30:1-1.
https://doi.org/10.1002/9780470479216.corpsy0524 DOI: https://doi.org/10.1002/9780470479216.corpsy0524
Sagi O, Rokach L. Ensemble learning: a survey. Wiley Interdiscipl Rev Data Mining Knowl Discov. 2018;8(5):e1249.
https://doi.org/10.1002/widm.1249 DOI: https://doi.org/10.1002/widm.1249
Zhen X, Chen J, Zhong Z, Hrycushko B, Zhou L, Jiang S, et al. Deep convolutional neural network with transfer learning for rectum toxicity prediction in cervical cancer radiotherapy: a feasibility study. Phys Med Biol. 2017;62(21):8246–63.
https://doi.org/10.1088/1361-6560/aa8d09 DOI: https://doi.org/10.1088/1361-6560/aa8d09
Araújo ALD, Moraes MC, Pérez-de-Oliveira ME, da Silva VM, Saldivia-Siracusa C, Pedroso CM, et al. Machine learning for the prediction of toxicities from head and neck cancer treatment: a systematic review with meta-analysis. Oral Oncol. 2023;140:106386.
https://doi.org/10.1016/j.oraloncology.2023.106386 DOI: https://doi.org/10.1016/j.oraloncology.2023.106386
van Dijk LV, Noordzij W, Brouwer CL, Boellaard R, Burgerhof JGM, Langendijk JA, et al. 18F-FDG PET image biomarkers improve prediction of late radiation-induced xerostomia. Radiother Oncol. 2018;126(1):89–95.
https://doi.org/10.1016/j.radonc.2017.08.024 DOI: https://doi.org/10.1016/j.radonc.2017.08.024
Sheikh K, Lee SH, Cheng Z, Lakshminarayanan P, Peng L, Han P, et al. Predicting acute radiation induced xerostomia in head and neck cancer using MR and CT radiomics of parotid and submandibular glands. Radiat Oncol. 2019;14(1):131.
https://doi.org/10.1186/s13014-019-1339-4 DOI: https://doi.org/10.1186/s13014-019-1339-4
van Dijk LV, Thor M, Steenbakkers RJHM, Apte A, Zhai T-T, Borra R, et al. Parotid gland fat related magnetic resonance image biomarkers improve prediction of late radiation-induced xerostomia. Radiother Oncol. 2018;128(3):459–66.
https://doi.org/10.1016/j.radonc.2018.06.012 DOI: https://doi.org/10.1016/j.radonc.2018.06.012
van Dijk LV, Langendijk JA, Zhai T-T, Vedelaar TA, Noordzij W, Steenbakkers RJHM, et al. Delta-radiomics features during radiotherapy improve the prediction of late xerostomia. Sci Rep. 2019;9(1):12483.
https://doi.org/10.1038/s41598-019-48184-3 DOI: https://doi.org/10.1038/s41598-019-48184-3
Liu Y, Shi H, Huang S, Chen X, Zhou H, Chang H, et al. Early prediction of acute xerostomia during radiation therapy for nasopharyngeal cancer based on delta radiomics from CT images. Quant Imaging Med Surg. 2019;9(7):1288–302.
https://doi.org/10.21037/qims.2019.07.08 DOI: https://doi.org/10.21037/qims.2019.07.08
Kapoor R, Sleeman W, Palta J, Weiss E. 3D deep convolution neural network for radiation pneumonitis prediction following stereotactic body radiotherapy. J Appl Clin Med Phys. 2022;24(3):e13875.
https://doi.org/10.1002/acm2.13875 DOI: https://doi.org/10.1002/acm2.13875
Ribeiro MT, Singh S, Guestrin C. Why should I trust you? In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining; 2016. p. 1135–44.
https://doi.org/10.1145/2939672.2939778 DOI: https://doi.org/10.1145/2939672.2939778
Lundberg SM, Lee S-I. A unified approach to interpreting model predictions. Adv Neural Inform Process Syst. 2017;30:4765-4774.
Kirillov A, Mintun E, Ravi N, Mao H, Rolland C, Gustafson L, et al., editors. Segment anything. In: Proceedings of the IEEE/CVF international conference on computer vision Paris, France, on October 2-3, 2023. pp. 4015-4026. IEEE DOI: https://doi.org/10.1109/ICCV51070.2023.00371
Zhang Y, Shen Z, Jiao R. Segment anything model for medical image segmentation: current applications and future directions. Comput Biol Med. 2024;171:108238.
https://doi.org/10.1016/j.compbiomed.2024.108238 DOI: https://doi.org/10.1016/j.compbiomed.2024.108238
Additional Files
Published
How to Cite
Issue
Section
Categories
License
Copyright (c) 2025 Bao Ngoc Huynh, Manish Kakar, Olga Zlygosteva, Inga Solgård Juvkam, Nina Edin, Oliver Tomic, Cecilia Marie Futsaether, Eirik Malinen
This work is licensed under a Creative Commons Attribution 4.0 International License.
