Deep learning models for osteoporosis diagnosis and future bone fracture prediction
Osteoporoz tanısı ve gelecekteki kemik kırığı tahmini için derin öğrenme modelleri
- Tez No: 966791
- Danışmanlar: DOÇ. DR. OĞUZ ATA
- Tez Türü: Doktora
- Konular: Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve Kontrol, Computer Engineering and Computer Science and Control
- Anahtar Kelimeler: Belirtilmemiş.
- Yıl: 2025
- Dil: İngilizce
- Üniversite: Altınbaş Üniversitesi
- Enstitü: Lisansüstü Eğitim Enstitüsü
- Ana Bilim Dalı: Elektrik ve Bilgisayar Mühendisliği Ana Bilim Dalı
- Bilim Dalı: Belirtilmemiş.
- Sayfa Sayısı: Belirtilmemiş.
Özet
The research develops deep learning-based models for diagnosing osteoporosis and the prediction of future fracture risk by incorporating medical imaging with clinical data. Osteoporosis is one of the very common bone diseases that result in a reduction in density within bones, increasing their susceptibility to fracture, especially within the elderly population. The currently used clinical diagnostic tools like DEXA have specific shortcomings in early diagnosis and risk prediction. This study has provided a hybrid deep learning framework by CNNs for integrative image analyses with clinical metadata in overall risk assessment. The model of osteoporosis prediction divides bone health into three subgroups: normal, osteopenia, and osteoporosis. In the fracture prediction model, using a number of clinical features, a random forest classifier was employed to find individuals who are at a high risk of fracture. Both models were evaluated by cross-validation with very good performance for both, with an F1-score higher than 0.88 in all cases. Advanced methods were proposed and used to improve model interpretability, such as SMOTEENN balancing to deal with class imbalance. The results expose the big potential of deep learning methods to enhance diagnostic precision by reducing false negatives and providing practical insights into clinical decision-making. This study sheds light on the importance and the impact of different factors of lifestyle on the prediction and prevention of Osteoporosis and future fractures. By visualizing and interpreting these factors, this study contributes to the growing body of knowledge on osteoporosis and its risk factors, paving the way for more effective disease management strategies. These results constitute a great step toward improving the diagnosis of osteoporosis and fracture risk by advanced machine learning techniques.
Özet (Çeviri)
The research develops deep learning-based models for diagnosing osteoporosis and the prediction of future fracture risk by incorporating medical imaging with clinical data. Osteoporosis is one of the very common bone diseases that result in a reduction in density within bones, increasing their susceptibility to fracture, especially within the elderly population. The currently used clinical diagnostic tools like DEXA have specific shortcomings in early diagnosis and risk prediction. This study has provided a hybrid deep learning framework by CNNs for integrative image analyses with clinical metadata in overall risk assessment. The model of osteoporosis prediction divides bone health into three subgroups: normal, osteopenia, and osteoporosis. In the fracture prediction model, using a number of clinical features, a random forest classifier was employed to find individuals who are at a high risk of fracture. Both models were evaluated by cross-validation with very good performance for both, with an F1-score higher than 0.88 in all cases. Advanced methods were proposed and used to improve model interpretability, such as SMOTEENN balancing to deal with class imbalance. The results expose the big potential of deep learning methods to enhance diagnostic precision by reducing false negatives and providing practical insights into clinical decision-making. This study sheds light on the importance and the impact of different factors of lifestyle on the prediction and prevention of Osteoporosis and future fractures. By visualizing and interpreting these factors, this study contributes to the growing body of knowledge on osteoporosis and its risk factors, paving the way for more effective disease management strategies. These results constitute a great step toward improving the diagnosis of osteoporosis and fracture risk by advanced machine learning techniques.
Benzer Tezler
- Tibia plato kırıklarının yazılım algoritması kullanılarak yorumlanma başarısı
Success in interpreting tibia plateau fractures using software algorithm
BURAK ACAR
Tıpta Uzmanlık
Türkçe
2025
Acil Tıpİzmir Katip Çelebi ÜniversitesiAcil Tıp Ana Bilim Dalı
DOÇ. DR. UMUT PAYZA
DOÇ. DR. MEHMET GÖKTUĞ EFGAN
- Dual enerji x-ışını absorbsiyometri ve lomber grafi görüntülerinin derin öğrenme ile karşılaştırılması: Osteoporoz tanısında yeni bir yaklaşım
Comparison of dual-energy x-ray absorptiometry and lumbar radiograph images using deep learning: A novel approach for osteoporosis diagnosis
DENİZ APALAN
Tıpta Uzmanlık
Türkçe
2025
Radyoloji ve Nükleer TıpSivas Cumhuriyet ÜniversitesiRadyoloji Ana Bilim Dalı
DR. ÖĞR. ÜYESİ BÜLENT YILDIZ
- Üretici yapay zeka ile medikal veri çoğaltma: Farklı gan mimarilerinin derin öğrenme tabanlı osteoporoz teşhisindeki başarılarının kıyaslanması
Generator artificial intelligence for medical data replica: Comparison of the success of different ganarchitectures in deep learning-based osteoporosis diagnosis
SEFA KARTAL
Yüksek Lisans
Türkçe
2026
Elektrik ve Elektronik MühendisliğiÇankırı Karatekin ÜniversitesiElektrik-Elektronik Mühendisliği Ana Bilim Dalı
DR. ÖĞR. ÜYESİ ZAFER CİVELEK
- Dental panoramik radyografilerde derin öğrenme tabanlı segmentasyon yaklaşımları
Deep learning based segmentation approaches on dental panoramic radiographies
NİDA KUMBASAR
Doktora
Türkçe
2024
Elektrik ve Elektronik MühendisliğiAtatürk ÜniversitesiElektrik-Elektronik Ana Bilim Dalı
PROF. İBRAHİM YÜCEL ÖZBEK
- Radyolojik görüntülerde yoğunluk tabanlı kemik segmentasyonu v eosteoporoz değerlendirmesi
Ensity-based bone segmentation and osteoporosisassessment in radiological images
EMİNE REYHAN ŞENTÜRK
Yüksek Lisans
Türkçe
2025
Elektrik ve Elektronik MühendisliğiÇankırı Karatekin ÜniversitesiElektrik-Elektronik Mühendisliği Ana Bilim Dalı
DR. ÖĞR. ÜYESİ ZAFER CİVELEK