Early diagnosis of Idiopathic Pulmonary Fibrosis disease using Community and Deep learning techniques

Abstract

Idiopathic Pulmonary Fibrosis (IPF); is a chronic and progressive lung disease of unknown etiology and poor prognosis, characterized by advanced fibrosis. Histologically, it is characterized by the usual pattern of interstitial pneumonia. Predicting the progression of IPF disease is yet to be possible with known techniques. However, early diagnosis of IPF is very important to start treatment early. In this research study, a limited number of High-Resolution Computed Tomography (HRCT) images from open sources were used for this study in the diagnosis of IPF. The study aims to assist in the early diagnosis of IPF disease by using high-resolution CT scan images. Idiopathic Pulmonary Fibrosis (IPF) is a chronic and progressive lung disease with a poor prognosis. It is characterized by advanced fibrosis and is diagnosed histologically by the usual interstitial pneumonia pattern. Predicting the progression of IPF is currently not possible, but early diagnosis is vital for early treatment. This study used a limited number of high-resolution computed tomography (HRCT) images from open sources to investigate the use of HRCT images for early diagnosis of IPF. The images were preprocessed, and 25 relevant features were selected from 502 features for 2310 patients. The CT image dataset was divided into 80% training and 20% test sets. Random oversampling was applied to the training dataset. The prepared data was then trained with machine, ensemble, and deep learning techniques. The results showed that community learning at the region of interest (ROI) level achieved an accuracy of 96.52%, sensitivity of 86.45%, and specificity of 92.14%. These results suggest that HRCT images can assist in the early diagnosis of IPF. Moreover, the results of this study are promising, but further research is needed to validate these findings and to develop a clinical decision-support tool for the early diagnosis of IPF. This tool could help pulmonologists identify patients with IPF at an early stage when treatment is most effective. There is a need for a more extensive and diverse dataset of HRCT images. Despite such challenges, the potential benefits of using HRCT images for early diagnosis of IPF are significant. Identifying patients with IPF early can improve their chances of survival and quality of life.

Keywords

• Deep Learning
• ensemble learning
• convolutional neural networks
• IPF

Topluluk ve Derin öğrenme teknikleri kullanılarak İdiyopatik Pulmoner Fibrozis hastalığının erken teşhisi

Abstract

İdiyopatik Pulmoner Fibrozis (IPF); hâlihazırda etyolojisi bilinmeyen, kötü prognozlu, ileri derecede fibroz ile karakterize, kronik ve progresif olan bir akciğer hastalığıdır. Histolojik olarak olağan interstisyel pnömoni paterni ile karakterizedir. IPF hastalığının ilerlemesinin öngörülmesi bilinen tekniklerle henüz mümkün değildir. Fakat IPF’nin erken teşhisi, tedaviye erken başlamak için oldukça önemlidir. Bu araştırma çalışmasında, açık kaynaklardan alınan sınırlı sayıda Yüksek Çözünürlüklü Bilgisayarlı Tomografi (YÇBT) imajı IPF tanısında bu çalışma için kullanılmıştır. Çalışmanın amacı, yüksek çözünürlüklü Bilgisayarlı Tomografi (BT) tarama imajlarından faydalanarak IPF hastalığının erken teşhisine yardımcı olmaktır. Öncelikle, bu araştırmada kullanılan BT imaj verileri bir dizi ön işleme tekniklerine tabi tutulmuştur. 2310 hasta için toplam 502 özellik arasından, Özyinelemeli Öznitelik Eleme yöntemi (Recursive Feature Elimination) kullanılarak 25 alakalı özellik seçilmiştir. Ön işleme sürecinden sonra, BT imaj veri seti %80 eğitim ve %20 test kümelerine ayrılmıştır. Eğitim veri kümesine Üst Örnekleme (Random Oversampling) uygulanmıştır. Bu işlemden sonra, hazırlanan veri, Makine Öğrenmesi (ML), Topluluk Öğrenmesi (Ensemble Learning) ve Derin Öğrenme (Deep Learning) teknikleri ile eğitilmiştir. Yapılan çalışmada sonuç olarak İlgi Alanı (Region of Interest-ROI) düzeyinde Topluluk Öğrenmesi performansı sırasıyla %96,52 doğruluk, %86,45 hassasiyet ve %92.14 özgüllük olarak elde edilmiştir. Öncelikle, bu araştırmada kullanılan BT imaj verileri bir dizi ön işleme tekniklerine tabi tutulmuştur. 2310 hasta için toplam 502 özellik arasından, Özyinelemeli Öznitelik Eleme yöntemi (Recursive Feature Elimination) kullanılarak 25 alakalı özellik seçilmiştir. Ön işleme sürecinden sonra, BT imaj veri seti %80 eğitim ve %20 test kümelerine ayrılmıştır. Eğitim veri kümesine Üst Örnekleme (Random Oversampling) uygulanmıştır. Bu işlemden sonra, hazırlanan veri, Makine Öğrenmesi, Topluluk Öğrenmesi (Ensemble Learning) ve Derin Öğrenme (Deep Learning) teknikleri ile eğitilmiştir. Yapılan çalışmada sonuç olarak İlgi Alanı (Region of Interest-ROI) düzeyinde Topluluk Öğrenmesi performansı sırasıyla %96,52 doğruluk, %86,45 hassasiyet ve %92.14 özgüllük olarak elde edilmiştir.

Keywords

• Derin Öğrenme
• topluluk öğrenmesi
• evrişimli sinir ağları
• İdiyopatik Pulmoner Fibrozis

References

1. Novak, C., Ballinger, M. N., & Ghadiali, S. (2021). Mechanobiology of Pulmonary Diseases: A Review of Engineering Tools to Understand Lung Mechanotransduction. Journal of Biomechanical Engineering, 143(11)
2. Grutters, J. C., & Du Bois, R. M. (2005). Genetics of fibrosing lung diseases. European Respiratory Journal, 25(5), 915-927.
3. Kuwana, M., Gil-Vila, A., & Selva-O’Callaghan, A. (2021). Role of autoantibodies in the diagnosis and prognosis of interstitial lung disease in autoimmune rheumatic disorders. Therapeutic Advances in Musculoskeletal Disease, 13, 1759720X211032457.
4. Ley, B., Elicker, B. M., Hartman, T. E., Ryerson, C. J., Vittinghoff, E., Ryu, J. H., & Collard, H. R. (2014). Idiopathic pulmonary fibrosis: CT and risk of death. Radiology, 273(2), 570.
5. Shi, Y., Wong, W. K., Goldin, J. G., Brown, M. S., & Kim, G. H. J. (2019). Prediction of progression in idiopathic pulmonary fibrosis using CT scans at baseline: A quantum particle swarm optimization-Random Forest approach. Artificial intelligence in medicine, 100, 101709.
6. Refaee, T., Bondue, B., Van Simaeys, G., Wu, G., Yan, C., Woodruff, H. C., ... & Lambin, P. (2022). A Handcrafted Radiomics-Based Model for the Diagnosis of Usual Interstitial Pneumonia in Patients with Idiopathic Pulmonary Fibrosis. Journal of Personalized Medicine, 12(3), 373.
7. Christe, A., Peters, A. A., Drakopoulos, D., Heverhagen, J. T., Geiser, T., Stathopoulou, T., ... & Ebner, L. (2019). Computer-aided diagnosis of pulmonary fibrosis using deep learning and CT images. Investigative radiology, 54(10), 627.
8. Anthimopoulos, M., Christodoulidis, S., Ebner, L., Christe, A., & Mougiakakou, S. (2016). Lung pattern classification for interstitial lung diseases using a deep convolutional neural network. IEEE transactions on medical imaging, 35(5), 1207-1216.

9. Peters, u. (2016). assessment of respiratory system mechanics in adults: effect of weight loss, posture, bronchodilation and artefacts on respiratory impedance and its repeatability (doctoral dissertation).
10. Nallanthighal, V. S., Mostaani, Z., Härmä, A., Strik, H., & Magimai-Doss, M. (2021). Deep learning architectures for estimating breathing signal and respiratory parameters from speech recordings. Neural Networks, 141, 211-224.
11. Rehman, A., & Khan, F. G. (2021). A deep learning based review on abdominal images. Multimedia Tools and Applications, 80(20), 30321-30352.
12. Walsh, S. L., Calandriello, L., Silva, M., & Sverzellati, N. (2018). Deep learning for classifying fibrotic lung disease on high-resolution computed tomography: a case-cohort study. The Lancet Respiratory Medicine, 6(11), 837-845.
13. Comelli, A., Coronnello, C., Dahiya, N., Benfante, V., Palmucci, S., Basile, A., ... & Stefano, A. (2020). Lung segmentation on high-resolution computerized tomography images using deep learning: a preliminary step for radiomics studies. Journal of Imaging, 6(11), 125.
14. Hofmanninger, J., Prayer, F., Pan, J., Röhrich, S., Prosch, H., & Langs, G. (2020). Automatic lung segmentation in routine imaging is primarily a data diversity problem, not a methodology problem. European Radiology Experimental, 4(1), 1-13.
15. 15 Soffer, S., Morgenthau, A. S., Shimon, O., Barash, Y., Konen, E., Glicksberg, B. S., & Klang, E. (2021). Artificial intelligence for interstitial lung disease analysis on chest computed tomography: a systematic review. Academic Radiology.
16. Christe, A., Peters, A. A., Drakopoulos, D., Heverhagen, J. T., Geiser, T., Stathopoulou, T., ... & Ebner, L. (2019). Computer-aided diagnosis of pulmonary fibrosis using deep learning and CT images. Investigative radiology, 54(10), 627.
17. Salahuddin, Z., Frix, A. N., Yan, C., Wu, G., Woodruff, H. C., Gietema, H., ... & Lambin, P. (2022). Diagnosis of Idiopathic Pulmonary Fibrosis in High-Resolution Computed Tomography Scans Using a Combination of Handcrafted Radiomics and Deep Learning. Frontiers in medicine, 9.
18. Trusculescu, A. A., Manolescu, D., Tudorache, E., & Oancea, C. (2020). Deep learning in interstitial lung disease—how long until daily practice. European radiology, 30(11), 6285-6292.
19. Furukawa, T., Oyama, S., Yokota, H., Kondoh, Y., Kataoka, K., Johkoh, T., ... & Hasegawa, Y. (2022). A comprehensible machine learning tool to differentially diagnose idiopathic pulmonary fibrosis from other chronic interstitial lung diseases. Respirology, 27(9), 739-746.
20. Comelli, A., Coronnello, C., Dahiya, N., Benfante, V., Palmucci, S., Basile, A., ... & Stefano, A. (2020). Lung segmentation on high-resolution computerized tomography images using deep learning: a preliminary step for radiomics studies. Journal of Imaging, 6(11), 125.
21. Dai, D., Weigt, S., Goldin, J., Song, J. W., Pourzand, L., Oh, J. H., ... & Kim, G. H. H. (2021). Imaging Signatures in Idiopathic Pulmonary Fibrosis (IS-IPF) Study from Multi-Center Multidisciplinary Experiences in ILD. In TP26. TP026 dıagnosis, assessment, and prognosis of fıbrotıc ild (pp. A1849-A1849). American Thoracic Society.
22. Sharma, B., Lobato, B., Rao, S., Daga, M. K., & Janota, B. Deep learning using a convolutional neural network to differentiate between CT lung images of pulmonary fibrosis and nonspecific interstitial pneumonia.
23. Anthimopoulos, M., Christodoulidis, S., Ebner, L., Geiser, T., Christe, A., & Mougiakakou, S. (2018). Semantic segmentation of pathological lung tissue with dilated fully convolutional networks. IEEE journal of biomedical and health informatics, 23(2), 714-722.
24. Jacob, J., Bartholmai, B. J., Rajagopalan, S., Van Moorsel, C. H., Van Es, H. W., Van Beek, F. T., ... & Wells, A. U. (2018). Predicting outcomes in idiopathic pulmonary fibrosis using automated computed tomographic analysis. American journal of respiratory and critical care medicine, 198(6), 767-776.
25. Bratt, A., Williams, J. M., Liu, G., Panda, A., Patel, P. P., Walkoff, L., ... & Koo, C. W. (2022). Predicting usual interstitial pneumonia histopathology from chest CT imaging with deep learning. Chest.
26. Gerard, S. E., Herrmann, J., Kaczka, D. W., Musch, G., Fernandez-Bustamante, A., & Reinhardt, J. M. (2020). Multi-resolution convolutional neural networks for fully automated segmentation of acutely injured lungs in multiple species. Medical image analysis, 60, 101592.
27. Hu, Q., Souza, L. F. D. F., Holanda, G. B., Alves, S. S., Silva, F. H. D. S., Han, T., & Reboucas Filho, P. P. (2020). An effective approach for CT lung segmentation using mask region-based convolutional neural networks. Artificial intelligence in medicine, 103, 101792.
28. Lee, S. H., Kim, S. Y., Kim, D. S., Kim, Y. W., Chung, M. P., Uh, S. T., ... & Park, M. S. (2016). Predicting survival of patients with idiopathic pulmonary fibrosis using GAP score: a nationwide cohort study. Respiratory research, 17(1), 1-9.
29. Trusculescu, A. A., Manolescu, D., Tudorache, E., & Oancea, C. (2020). Deep learning in interstitial lung disease—how long until daily practice. European radiology, 30(11), 6285-6292.
30. Zelaya, C.V.G. Towards explaining the effects of data preprocessing on machine learning. In Proceedings of the 2019 IEEE 35th International Conference on Data Engineering (ICDE), Macao, China, 8–11 April 2019.
31. O’Brien, R., & Ishwaran, H. (2019). A random forests quantile classifier for class imbalanced data. Pattern recognition, 90, 232-249.
32. Kimber, T. B., Chen, Y., & Volkamer, A. (2021). Deep learning in virtual screening: recent applications and developments. International Journal of Molecular Sciences, 22(9), 4435.
33. Chaabane, I., Guermazi, R., & Hammami, M. (2020). Enhancing techniques for learning decision trees from imbalanced data. Advances in Data Analysis and Classification, 14(3), 677-745.
34. Bae, S. Y., Lee, J., Jeong, J., Lim, C., & Choi, J. (2021). Effective data-balancing methods for class-imbalanced genotoxicity datasets using machine learning algorithms and molecular fingerprints. Computational Toxicology, 20, 100178.
35. Sáez, J. A., Luengo, J., Stefanowski, J., & Herrera, F. (2015). SMOTE–IPF: Addressing the noisy and borderline examples problem in imbalanced classification by a re-sampling method with filtering. Information Sciences, 291, 184-203.
36. Li, Q., Li, M., Zheng, K., Li, H., Yang, H., Ma, S., & Zhong, M. (2020). Detection of microRNA expression levels based on microarray analysis for classification of idiopathic pulmonary fibrosis. Experimental and therapeutic medicine, 20(4), 3096-3103.
37. Puri, A., Gupta, M. K., & Sachdev, K. (2022). An ensemble-based approach using structural feature extraction method with class imbalance handling technique for drug-target interaction prediction. Multimedia Tools and Applications, 1-19.
38. Gal, Y., & Ghahramani, Z. (2016, June). Dropout as a bayesian approximation: Representing model uncertainty in deep learning. In international conference on machine learning (pp. 1050-1059). PMLR.