Accurate early detection of Parkinson’s disease from single photon emission computed tomography imaging through convolutional neural networks

Abstract

Early and accurate detection of Parkinson’s disease (PD) remains a crucial diagnostic challenge with substantial clinical implications, particularly for ensuring effective treatment and patient management. For instance, a group of subjects with scans without evidence of dopaminergic deficit (SWEDD) who are initially diagnosed as PD but exhibit normal single photon emission computed tomography (SPECT) scans. Over time, follow-up assessments often lead to a revised diagnosis of non-PD. In the meantime, these subjects may receive PD-specific medications that can cause more harm than benefit. In this paper, a case study is presented in which machine learning models are developed and trained on SPECT images to distinguish early PD from healthy controls, as well as to differentiate SWEDD cases from early PD. The case study utilizes a well-known, publicly available dataset and explores several machine learning classifiers, including support vector machines, logistic regression, feed forward neural networks, and convolutional neural networks (CNNs). The CNN model gave the best performance in differentiating PD from healthy subjects. All these models demonstrated strong potential for early differentiation of SWEDD cases from PD. These results suggest that the proposed approach could support clinicians in making more accurate and timely diagnostic decisions.

Keywords

Computer-aided diagnosis / Machine learning / Deep learning / Parkinson’s disease / Medical imaging

Funding

PPMI, a public-private partnership, is funded by the Michael J. Fox Foundation for Parkinson’s Research and other funding partners include AbbVie, Allergan, Amathus Therapeutics, Avid Radiopharmaceuticals, Biogen Idec, BioLegend, Bristol-Myers Squibb, Celgene, Denali Therapeutics, GE Healthcare, Genentech, GlaxoSmithKline, Eli Lilly and Company, Lundbeck, Merck & Co., Meso Scale Discovery, Pfizer, Piramal, Prevail Therapeutics, Hoffmann-La Roche, Sanofi Genzyme, Servier, Takeda Pharmaceutical Company, Teva, Verily Life Sciences, Voyager Therapeutics, and UCB (Union ChimiqueBelge).

Conflict of interest

The author declares no competing interests.

References

[1] Jankovic J. Parkinson’s disease: Clinical features and diagnosis. J Neurol Neurosurg Psychiatry. 2008; 79(4):368-376. doi: 10.1136/jnnp.2007.131045

[2] Fahn S. Description of Parkinson’s disease as a clinical syndrome. Ann N Y Acad Sci. 2003; 991:1-14. doi: 10.1111/j.1749-6632.2003.tb07458.x

[3] Tysnes OB, Storstein A. Epidemiology of Parkinson’s disease. J Neural Transm (Vienna). 2017; 124(8):901-905. doi: 10.1007/s00702-017-1686-y

[4] Marek K, Jennings D, Lasch S, et al. The Parkinson progression marker initiative (PPMI). Prog Neurobiol. 2011; 95(4):629-635. doi: 10.1016/j.pneurobio.2011.09.005

[5] Groveman BR, Orrù CD, Hughson AG, et al. Rapid and ultra-sensitive quantitation of disease-associated α-synuclein seeds in brain and cerebrospinal fluid by αSyn RT-QuIC. Acta Neuropathol Commun. 2018; 6:7. doi: 10.1186/s40478-018-0508-2

[6] Booij J, Tissingh G, Boer GJ, et al. [123I]FP-CIT SPECT shows a pronounced decline of striatal dopamine transporter labelling in early and advanced Parkinson’s disease. J Neurol Neurosurg Psychiatry. 1997; 62(2):133-140. doi: 10.1136/jnnp.62.2.133

[7] Booth TC, Nathan M, Waldman AD, Quigley AM, Schapira AH, Buscombe J. The role of functional dopamine-transporter SPECT imaging in Parkinsonian syndromes, part 1. AJNR Am J Neuroradiol. 2015; 36(2):236-244. doi: 10.3174/ajnr.A3971

[8] Cummings JL, Henchcliffe C, Schaier S, Simuni T, Waxman A, Kemp P. The role of dopaminergic imaging in patients with symptoms of dopaminergic system neurodegeneration. Brain. 2011; 134(11):3146-3166. doi: 10.1093/brain/awr177

[9] Seibyl J, Jennings D, Grachev I, Coffey C, Marek K. 123-I Ioflupane SPECT measures of Parkinson disease progression in the Parkinson progression marker initiative (PPMI) trial. J Nucl Med. 2013; 54(1):57-58.

[10] Beach TG, Adler CH. Importance of low diagnostic accuracy for early Parkinson’s disease. Mov Disord. 2018; 33(10):1551-1554. doi: 10.1002/mds.27485

[11] Choi H, Ha S, Im HJ, Paek SH, Lee DS. Refining diagnosis of Parkinson’s disease with deep learning-based interpretation of dopamine transporter imaging. Neuroimage Clin. 2017; 16:586-594. doi: 10.1016/j.nicl.2017.09.010

[12] Coarelli G, Garcin B, Roze E, Vidailhet M, Degos B. Invalidation of Parkinson’s disease diagnosis after years of follow-up based on clinical, radiological and neurophysiological examination. J Neurol Sci. 2019; 406:116454. doi: 10.1016/j.jns.2019.116454

[13] Catafau AM, Tolosa E, DaTscan Clinically Uncertain Parkinsonian Syndromes Study Group. Impact of dopamine transporter SPECT using 123I-Ioflupane on diagnosis and management of patients with clinically uncertain Parkinsonian syndromes. Mov Disord. 2004; 19(10):1175-1182. doi: 10.1002/mds.20112

[14] Kupsch AR, Bajaj N, Weiland F, et al. Impact of DaTscan SPECT imaging on clinical management, diagnosis, confidence of diagnosis, quality of life, health resource use and safety in patients with clinically uncertain parkinsonian syndromes: A prospective 1-year follow-up of an open-label controlled study. J Neurol Neurosurg Psychiatry. 2012; 83(6):620-628. doi: 10.1136/jnnp-2011-301695

[15] Marek K, Jennings D, Seibyl JP. Long-term follow-up of patients with scans without evidence of dopaminergic deficit (SWEDD) in the ELLDOPA study. Neurology. 2005; 64:A274.

[16] Benamer TS, Patterson J, Grosset DG, et al. Accurate differentiation of parkinsonism and essential tremor using visual assessment of [123I]-FP-CIT SPECT imaging: The [123I]-FP-CIT study group. Mov Disord. 2000; 15(3):503-510.

[17] Scherfler C, Nocker M. Dopamine transporter SPECT: How to remove subjectivity? Mov Disord. 2009; 24(Suppl 2):S721-S724. doi: 10.1002/mds.22590

[18] Staff RT, Ahearn TS, Wilson K, et al. Shape analysis of 123I-N-omega-fluoropropyl-2-beta-carbomethoxy- 3beta-(4-iodophenyl) nortropane single-photon emission computed tomography images in the assessment of patients with parkinsonian syndromes. Nucl Med Commun. 2009; 30(3):194-201. doi: 10.1097/MNM.0b013e328314b863

[19] Hirschauer TJ, Adeli H, Buford JA. Computer-aided diagnosis of Parkinson’s disease using enhanced probabilistic neural network. J Med Syst. 2015; 39(11):179. doi: 10.1007/s10916-015-0353-9

[20] Kim DH, Wit H, Thurston M. Artificial intelligence in the diagnosis of Parkinson’s disease from ioflupane-123 single-photon emission computed tomography dopamine transporter scans using transfer learning. Nucl Med Commun. 2018; 39(10):887-893. doi: 10.1097/MNM.0000000000000890

[21] Martínez-Murcia FJ, Górriz JM, Ramírez J, Illán IA, Ortiz A, Parkinson’s Progression Markers Initiative. Automatic detection of Parkinsonism using significance measures and component analysis in DaTSCAN imaging. Neurocomputing. 2014; 126:58-70. doi: 10.1016/j.neucom.2013.01.054

[22] Martínez-Murcia FJ, Ortiz A, Górriz JM, et al. A 3D convolutional neural network approach for the diagnosis of Parkinson’s disease. In: International Work-Conference on the Interplay Between Natural and Artificial Computation. Cham: Springer; 2017. p. 324-333. doi: 10.1007/978-3-319-59740-9_32

[23] Oliveira FP, Castelo-Branco M. Computer-aided diagnosis of Parkinson’s disease based on [(123)I]FP-CIT SPECT binding potential images, using the voxels-as-features approach and support vector machines. J Neural Eng. 2015; 12(2):026008. doi: 10.1088/1741-2560/12/2/026008

[24] Ortiz A, Munilla J, Martínez-Ibañez M, Górriz JM, Ramírez J, Salas-Gonzalez D. Parkinson’s disease detection using isosurfaces-based features and convolutional neural networks. Front Neuroinform. 2019; 13:48. doi: 10.3389/fninf.2019.00048

[25] Prashanth R, Dutta Roy S, Mandal PK, Ghosh S. Automatic classification and prediction models for early Parkinson’s disease diagnosis from SPECT imaging. Expert Syst Appl. 2014; 41(8):3333-3342. doi: 10.1016/j.eswa.2013.11.031

[26] Prashanth R, Dutta Roy S, Mandal PK, Ghosh S. High-accuracy detection of early Parkinson’s disease through multimodal features and machine learning. Int J Med Inform. 2016; 90:13-21. doi: 10.1016/j.ijmedinf.2016.03.001

[27] Prashanth R, Roy SD, Mandal PK, Ghosh S. High-accuracy classification of Parkinson’s disease through shape analysis and surface fitting in 123I-Ioflupane SPECT imaging. IEEE J Biomed Health Inform. 2017; 21(2):794-802. doi: 10.1109/JBHI.2016.2547901

[28] Segovia F, Gorriz JM, Ramirez J, Alvarez I, Jimenez- Hoyuela JM, Ortega SJ. Improved parkinsonism diagnosis using a partial least squares-based approach. Med Phys. 2012; 39(7):4395-4403. doi: 10.1118/1.4730289

[29] Illan IA, Gorrz JM, Ramirez J, Segovia F, Jimenez-Hoyuela JM, Ortega Lozano SJ. Automatic assistance to Parkinson’s disease diagnosis in DaTSCAN SPECT imaging. Med Phys. 2012; 39(10):5971-5980. doi: 10.1118/1.4742055

[30] Rojas A, Górriz JM, Ramírez J, et al. Application of empirical mode decomposition (EMD) on DaTSCAN SPECT images to explore Parkinson disease. Expert Syst Appl. 2013; 40(7):2756-2766. doi: 10.1016/j.eswa.2012.11.017

[31] Towey DJ, Bain PG, Nijran KS. Automatic classification of 123I-FP-CIT (DaTSCAN) SPECT images. Nucl Med Commun. 2011; 32(8):699-707. doi: 10.1097/MNM.0b013e328347cd09

[32] Huertas-Fernandez I, Garcia-Gomez F, Garcia-Solis D, et al. Machine learning models for the differential diagnosis of vascular parkinsonism and Parkinson’s disease using [(123) I] FP-CIT SPECT. Eur J Nucl Med Mol Imaging. 2015; 42(1):112-119. doi: 10.1007/s00259-014-2882-8

[33] Oliveira FP, Faria DB, Costa DC, Castelo-Branco M, Tavares JMR. Extraction, selection and comparison of features for an effective automated computer-aided diagnosis of Parkinson’s disease based on [123I] FP-CIT SPECT images. Eur J Nucl Med Mol Imaging. 2018; 45(6):1052-1062. doi: 10.1007/s00259-017-3918-7

[34] Zhang X, Chou J, Liang J, et al. Data-driven subtyping of Parkinson’s disease using longitudinal clinical records: A cohort study. Sci Rep. 2019; 9(1):797. doi: 10.1038/s41598-018-37545-z

[35] Marek K, Chowdhury S, Siderowf A, et al. The Parkinson’s progression markers initiative (PPMI) - establishing a PD biomarker cohort. Ann Clin Transl Neurol. 2018; 5(12):1460-1477. doi: 10.1002/acn3.644

[36] Hoehn MM, Yahr MD. Parkinsonism: Onset, progression and mortality. Neurology. 1967; 17(5):427-442. doi: 10.1212/wnl.17.5.427

[37] LeCun Y, Bengio Y, Hinton G. Deep learning. Nature. 2015; 521(7553):436-444. doi: 10.1038/nature14539

[38] Cortes C, Vapnik V. Support-vector networks. Mach Learn. 1995; 20(3):273-297. doi: 10.1007/BF00994018

[39] Lee SI, Lee H, Abbeel P, Ng AY. Efficient l-1 regularized logistic regression. In: Proceedings of the 21st National Conference on Artificial Intelligence (AAAI). 2006. p. 401-408.

[40] Bergstra JS, Bardenet R, Bengio Y, Kégl B. Algorithms for hyper-parameter optimization. In: Advances in Neural Information Processing Systems 24. 2011. p. 2546-2554.

[41] Srivastava N, Hinton G, Krizhevsky A, Sutskever I, Salakhutdinov R. Dropout: A simple way to prevent neural networks from overfitting. J Mach Learn Res. 2014; 15(1):1929-1958.

[42] Perez L, Wang J. The Effectiveness of Data Augmentation in Image Classification using Deep Learning. [arXiv Preprint]; 2017.

[43] Goodfellow I, Pouget-Abadie J, Mirza M, et al. Generative adversarial nets. In: Advances in Neural Information Processing Systems 27. [Preprint]; 2014. p. 2672-2680

[44] Muller R, Kornblith S, Hinton GE. When does label smoothing help? In: Advances in Neural Information Processing Systems 32. 2019. p. 4696-4705.

[45] Shiiba T, Takano K, Takaki A, Suwazono S. Dopamine transporter single-photon emission computed tomography-derived radiomics signature for detecting Parkinson’s disease. EJNMMI Res. 2022; 12(1):39. doi: 10.1186/s13550-022-00910-1

[46] Tufail AB, Ma YK, Zhang QN, et al. 3D convolutional neural networks-based multiclass classification of Alzheimer’s and Parkinson’s diseases using PET and SPECT neuroimaging modalities. Brain Inform. 2021; 8(1):23. doi: 10.1186/s40708-021-00144-2

[47] Majhi B, Kashyap A, Mohanty SS, et al. An improved method for diagnosis of Parkinson’s disease using deep learning models enhanced with metaheuristic algorithm. BMC Med Imaging. 2024; 24(1):156. doi: 10.1186/s12880-024-01335-z

[48] Khachnaoui H, Chikhaoui B, Khlifa N, Mabrouk R. Enhanced Parkinson’s disease diagnosis through convolutional neural network models applied to spect datscan images. IEEE Access. 2023; 11:91157-91172. doi: 10.1109/ACCESS.2023.3308075