Kafa İçi Kanamalarda Yapay Zeka ile Acil Tanı ve Triaj

Abstract

Akut kafa içi kanamalar, hangi türden olursa olsun, mortalitesi yüksek, hızlı tanısı ve tedavisi yüksek önem arzeden patolojilerdir, ancak erken operasyondan en fazla fayda görecek hasta grubu, gürültülü bir tabloyla gelmediği için, fayda görmeyecek olan hasta grubuna göre daha geç opere edilmektedir. Bu çalışmada, küçük bir veri setinde , kafa içi kanamanın varlığını ayırt edebilen bir derin öğrenme modelini değerlendirmeyi amaçladık. Materyal Metod: Çalışmaya qure.ai beyin BT veritabanından, 5 intrakraniyal kanamalı 3 sağlıklı hasta rastgele olarak dahil edildi. 100 adet kanamalı 100 adet de sağlıklı olmak üzere toplamda 200 adet BT kesit görüntüsü ile veri seti oluşturuldu, eğitim, doğrulama ve test seti olarak üçe bölündü. Yapay sinir ağı eğitim setinde eğitilerek doğrulama setinde hassasiyeti test edildi, hassasiyet %80 dolaylarına çıktıktan sonra sabitlendi ve yapay sinir ağının eğitimi durduruldu. Daha sonra bu yapay sinir ağı, test setinde değerlendirildi. Sonuçlar: Derin öğrenme modeli test seti üzerinde çalıştırıldı. Sensitivite %90.0 , Spesifite: %70.0 ,Pozitif Prediktif Değer: %75.0 , Negatif Prediktif Değer: %87.5 Toplam Doğruluk: %80.0 olarak geldi. Derin öğrenme modeli, daha önce hiç görmediği 20 kesitte, yalnızca 1 defa yanlış negatif değerlendirme yaptı. Neticede, bir derin öğrenme modelinin küçük bir veri setinde bile oldukça yüksek doğrulukta sonuçlar çıkarabileceği ve potansiyel olarak acil servislerde hızlı triaj amacıyla kullanılabileceğini düşünmekteyiz.

Keywords

• kafa içi kanama
• beyin cerrahisi
• yapay zeka
• derin öğrenme

Urgent diagnosis and triage in Intracranial Haemorrages with Machine Learning

Abstract

Acute intracranial hemorrhages, regardless of their type, are pathologies with high mortality and require rapid diagnosis and treatment, however the patient group who will benefit most from early operation is operated later than the patient group with nless favorable outcome, because they do not admit with a severe clinical presentation. In this study, we aimed to evaluate a deep learning model that can distinguish the presence of intracranial hemorrhage in a small data set. Material Method: 3 healthy patients and 5 patients with intracranial hemorrhages were randomly seleceted for the study from the qure.ai Cranial CT database. The data set was created with a total of 200 CT cross-section images, 100 of which were hemorrhagic and 100 were healthy, and it was divided into three groups as training, validation and test set. The artificial neural network was trained in the training set and its accuracy was tested in the validation set, the accuracy did not improve after reaching around 80% and the training of the artificial neural network was stopped. Later, this artificial neural network was evaluated in the test set. Results: The deep learning model was run on the test set. Results were as follows; Sensitivity 90.0%, Specificity: 70.0%, Positive Predictive Value: 75.0%, Negative Predictive Value: 87.5% Total Accuracy: 80.0%. The deep learning model made only one false-negative assessment in 20 crosss-sections that it had never seen before. As a result, we think that a deep learning model can produce highly accurate results even if they are trained in a small data set and potentially be used for rapid triage in emergency departments.

Keywords

• intracranial haemmorage
• neurosurgery
• artificial intelligence
• deep learning

References

1. 1. Alagoz F, Yildirim AE, Sahinoglu M, et al. Traumatic acute subdural hematomas: Analysis of outcomes and predictive factors at a single center. Turk Neurosurg. 2017;27(2):187-191. doi:10.5137/1019-5149.JTN.15177-15.2
2. 2. Solaroǧlu I, Kaptanoǧlu E, Okutan Ö, Beşkonakli E, Taşkin Y. Prognostic value of initial computed tomography findings in patients with traumatic acute subdural hematoma. Turk Neurosurg. 2002.
3. 3. Haselsberger K, Pucher R, Auer LM. Prognosis after acute subdural or epidural haemorrhage. Acta Neurochir (Wien). 1988. doi:10.1007/BF01560563
4. 4. Karnjanasavitree W, Phuenpathom N, Tunthanathip T. The optimal operative timing of traumatic intracranial acute subdural hematoma correlated with outcome. Asian J Neurosurg. 2018;13(4):1158. doi:10.4103/ajns.ajns_199_18
5. 5. Koza JR, Bennett FH, Andre D, Keane MA. Automated Design of Both the Topology and Sizing of Analog Electrical Circuits Using Genetic Programming. In: Artificial Intelligence in Design ’96. Springer Netherlands; 1996:151-170. doi:10.1007/978-94-009-0279-4_9
6. 6. Lakhani P, Sundaram B. Deep learning at chest radiography: Automated classification of pulmonary tuberculosis by using convolutional neural networks. Radiology. 2017. doi:10.1148/radiol.2017162326
7. 7. Bar A, Wolf L, Amitai OB, Toledano E, Elnekave E. Compression Fractures Detection on CT. June 2017. http://arxiv.org/abs/1706.01671. Accessed February 6, 2020.
8. 8. Yasaka K, Akai H, Abe O, Kiryu S. Deep learning with convolutional neural network for differentiation of liver masses at dynamic contrast-enhanced CT: A preliminary study. Radiology. 2018;286(3):887-896. doi:10.1148/radiol.2017170706

9. 9. Liu F, Xie L, Xia Y, Fishman EK, Yuille AL. Joint Shape Representation and Classification for Detecting PDAC. April 2018. http://arxiv.org/abs/1804.10684. Accessed February 6, 2020.
10. 10. Shadmi R, Mazo V, Bregman-Amitai O, Elnekave E. Fully-convolutional deep-learning based system for coronary calcium score prediction from non-contrast chest CT. In: Proceedings - International Symposium on Biomedical Imaging. Vol 2018-April. IEEE Computer Society; 2018:24-28. doi:10.1109/ISBI.2018.8363515
11. 11. Topol EJ. High-performance medicine: the convergence of human and artificial intelligence. Nat Med. 2019;25(1):44-56. doi:10.1038/s41591-018-0300-7
12. 12. Organisation for Economic Co-operation and Development O. OECD - Computed Tomography (CT) Exams (indicator). Computed tomography (CT) exams (indicator). doi:10.1787/3c994537-en
13. 13. Chilamkurthy S, Ghosh R, Tanamala S, et al. Deep learning algorithms for detection of critical findings in head CT scans: a retrospective study. Lancet. 2018;392(10162):2388-2396. doi:10.1016/S0140-6736(18)31645-3
14. 14. Shafer G, Vovk V. A Tutorial on Conformal Prediction. Vol 9.; 2008.