Hacettepe Üniversitesi Açık Erişim Sistemi (HÜAES)
- HÜAES, Hacettepe Üniversitesi bünyesinde üretilen kitap, makale, tez, bildiri, rapor gibi tüm akademik kaynakları uluslararası standartlarda dijital ortamda depolar, etkisini artırmak için telif haklarına uygun olarak Açık Erişime sunar.
Communities in DSpace
Select a community to browse its collections.
- Ankara State Conservatory
- Directorates
- Faculty of Dentistry
- Faculty of Pharmacy
- Faculty of Letters
Recent Submissions
Kemik İliği Biyopsisi ile İlk Tanı Alan Solid Tümörlerde Klinik ve Laboratuvar Özellikler
(Tıp Fakültesi, 2026-06-05) Dilan Kayaalp; İç Hastalıkları
Kayaalp, D. Clinical and Laboratory Features of Solid Tumors Initially Diagnosed by
Bone Marrow Biopsy. Hacettepe University Faculty of Medicine, Department of Internal
Medicine, Thesis in Internal Medicine Specialty, Ankara, 2026.
Objective: Bone marrow metastasis is an uncommon but clinically significant manifestation
of advanced solid malignancies and is associated with poor prognosis. This study aimed to
evaluate the clinical, hematological, and biochemical characteristics of patients with solid
tumors initially diagnosed by bone marrow biopsy and to investigate their association with
overall survival.
Materials and Methods: A total of 30 patients with bone marrow metastasis from solid
tumors were retrospectively analyzed. Demographic data, primary tumor types, hematological
and biochemical findings, peripheral blood smear results, and survival data were recorded.
Overall survival was analyzed using the Kaplan–Meier method. Survival comparisons
between groups were performed using the log-rank test. Cox proportional hazards regression
analysis was conducted for selected variables.
Results: The median age at diagnosis was 57.5 years, and 73.3% of patients were male. The
most common primary tumors were lung (20%), breast (16.7%), gastric (13.3%), and prostate
cancer (13.3%); in 23.3% of patients, the primary tumor could not be identified. Anemia and
thrombocytopenia were each observed in 75% of patients, while bicytopenia was present in
60.7%. Among patients with available peripheral smear data, leukoerythroblastic changes
were observed in 50%, and features of microangiopathic hemolytic anemia in 25%.
Biochemical abnormalities were frequent, including elevated LDH (80.8%), ALP (85.7%),
CRP (72%), and ESR (65.4%), while hypoalbuminemia was detected in 28.6% of patients.
The median overall survival was 64 days (95% CI: 43.8–84.2 days). No statistically significant
association was found between hematological or biochemical parameters and overall survival.
Conclusion: Patients with bone marrow metastasis from solid tumors have a markedly poor
prognosis with a short median survival. Primary tumors are frequently unidentified, and
anemia and thrombocytopenia are the most common hematological abnormalities.
Leukoerythroblastosis may serve as an important peripheral blood finding suggestive of bone
marrow infiltration. However, conventional hematological and biochemical parameters alone
appear to have limited prognostic value in predicting survival.
Keywords: Bone marrow metastasis, solid tumors, myelophthisis, thrombocytopenia, anemia,
leukoerythroblastosis
GASTROENTEROLOJİ’YE ÖZGÜ YAPAY ZEKA DİL MODELİNİN KLİNİK PRATİKTE KULLANIMI
(Tıp Fakültesi, 2026-04-20) Ali Berkcan Bozdoğan; İç Hastalıkları
Bozdoğan Ali B.: Clinical Use of a Gastroenterology-Specific Artificial
Intelligence Language Model; Hacettepe University Adult Hospital Residency
Thesis, Department of Internal Medicine, Hacettepe University Faculty of
Medicine, Ankara, 2026.
Objective: This study aimed to evaluate the clinical decision support capacity of
GastroGPT, a retrieval-augmented generation (RAG)-based artificial intelligence
language model developed specifically for gastroenterology, and to perform a
comparative analysis with general-purpose large language models.
Materials and Methods: In this retrospective, multicenter comparative model
evaluation study, 200 standardized clinical cases selected from the archives of
Hacettepe University Faculty of Medicine, Department of Gastroenterology were
utilized. Cases were stratified to represent the following subspecialty areas: general
gastroenterology (50%), hepatology (26%), pancreatic diseases (9%), inflammatory
bowel diseases (6%), gastrointestinal oncology (5%), and endoscopy (4%). A total of
eight AI language models were evaluated, comprising four GastroGPT variants
(Deeplake, Faiss, Chroma, and 16K) and four general-purpose models (GPT-4,
Claude, Bard/Gemini, You.com). Model responses were scored by two independent
expert gastroenterologists in a blinded fashion using a 5-point Likert scale across five
criteria: diagnostic accuracy, appropriateness of recommended investigations,
guideline concordance of treatment recommendations, applicability of patient
management strategies, and overall performance. Additionally, hallucination rates,
source attribution quality, and response consistency were assessed.
Results: GastroGPT-Deeplake achieved the highest overall performance score
(53.8±5.9; maximum 60 points), significantly outperforming GPT-4 (50.1±6.8,
p=0.028) and Claude (49.3±7.2, p=0.012). GastroGPT variants demonstrated marked
superiority over general-purpose models in hallucination rates (4.2–6.8% vs. 8.3–
12.1%, p=0.018) and source attribution quality (62.8–78.3% vs. 23.5–34.7%). In the
vector database comparison, Deeplake significantly outperformed Faiss and Chroma
(p<0.001). GastroGPT's mean scores across the five evaluation criteria ranged from 4.00 to 4.21, with the highest performance observed in guideline concordance
(4.21±0.63). Case complexity had a significant effect on diagnostic accuracy
(F=7.361, p=0.0008); diagnostic accuracy scores in high-complexity cases
(3.81±0.83) were significantly lower compared to low-complexity cases (4.43±0.70;
Cohen's d=0.803). No significant performance differences were observed across
disease prevalence categories or subspecialty areas. Inter-rater reliability was excellent
(ICC>0.99, Pearson r=0.814, Cronbach's α=0.802). ROC analysis yielded an AUC of
0.908 for the overall performance criterion.
Conclusion: GastroGPT-Deeplake demonstrated superiority over general-purpose
large language models in both overall performance and safety parameters through
domain-specific RAG integration. The advantages of RAG technology in reducing
hallucination risk and ensuring source traceability are of critical importance for the
safety of medical AI applications. Our findings support the preferential use of domainspecific approaches over general-purpose models in developing AI-assisted clinical
decision support systems for gastroenterology practice.
Keywords: Artificial intelligence, large language model, gastroenterology,
GastroGPT, retrieval-augmented generation, clinical decision support
system, hallucination, vector database
Pediatrik Kriptojenik Nörolojik Hastalıklarda İnsizyonel Beyin Biyopsisinin Tanıya Katkısı
(2026-05-22) Eylül Nazlı Sungur; Çocuk Sağlığı ve Hastalıkları
SUNGUR, E. The Diagnostic Contribution of Incisional Brain Biopsy in Pediatric Cryptogenic Neurological Diseases. Hacettepe University Faculty of Medicine, Department of Pediatrics, Medical Specialty Thesis, Ankara-2026.
Cryptogenic neurological diseases pose significant diagnostic challenges in childhood due to their broad differential diagnosis spectrum and heterogeneous clinical course. In cases where diagnostic uncertainty persists despite clinical evaluation, laboratory investigations, cerebrospinal fluid analysis, and neuroimaging studies, brain biopsy may contribute to both the diagnostic process and clinical management. The aim of this study was to evaluate the diagnostic contribution and impact on clinical management of brain biopsy in pediatric cryptogenic neurological diseases, and to identify factors that may influence its diagnostic yield. In this study, pediatric patients evaluated by the Division of Pediatric Neurology at Hacettepe University İhsan Doğramacı Children’s Hospital between January 1, 2000 and December 1, 2024, whose brain tissue specimens were examined by the Department of Medical Pathology at Hacettepe University Faculty of Medicine, were retrospectively reviewed. A total of 892 pathology reports were screened, and 41 patients who underwent 45 brain biopsies for cryptogenic neurological disease were included in the study. Brain biopsy contributed to the diagnosis in 51.1% of cases by confirming the presumed diagnosis or leading to a diagnostic change, while diagnostic change alone was observed in 6.7% of biopsies. Its impact on management was higher than its diagnostic contribution, influencing management in 68.9% of biopsies. Diagnostic contribution differed significantly across final diagnostic groups; diagnostic yield was higher in neoplastic, infectious, and vasculitic processes, whereas it was more limited in metabolic/neurodegenerative diseases and immune-mediated encephalitis (p=0.002). Diagnostic change was more frequent in vasculitic processes; in these cases, vasculitis had not been included among the pre-biopsy differential diagnoses (p=0.008). The effect on clinical management also differed significantly across final diagnostic groups (p<0.001); clinical impact was particularly high in neoplastic, infectious, vasculitic, and immune-mediated processes, while it was limited in metabolic/neurodegenerative diseases. The presence of contrast enhancement on brain magnetic resonance imaging (p=0.005), the dominant involved compartment of the lesion (p=0.012), and a shorter interval between symptom onset and biopsy (p=0.005) were significantly associated with the diagnostic contribution of biopsy. No statistically significant association was found between diagnostic contribution and pre-biopsy treatment, biopsy technique (open/stereotactic), or biopsy type (incisional/excisional). The biopsy-related complication rate was 6.7%, and no biopsy-related mortality was observed. A borderline significant association was found between complication development and the anatomical biopsy site (p=0.050), whereas biopsy technique, biopsy type, and the patient’s clinical status at the time of biopsy were not significantly associated with complications. In conclusion, brain biopsy in pediatric cryptogenic neurological diseases is a feasible procedure with an acceptable safety profile that can provide substantial diagnostic and clinical benefit, particularly in the presence of targetable and contrast-enhancing lesions and when performed in a timely manner.
Keywords: cryptogenic neurological disease, brain biopsy, diagnostic contribution, clinical management, complication.
AKUT KORONER SENDROMLU HASTALARDA HASTANE İÇİ MORTALİTENİN MAKİNE ÖĞRENME YÖNTEMLERİ KULLANILARAK KESTİRİMİ
(Sağlık Bilimleri Enstitüsü, 2026) Aktürk Sevinç; Biyoistatistik
Akturk, S., Prediction of In-Hospital Mortality in Patients with Acute Coronary Syndrome Using Machine Learning Methods, Hacettepe University Graduate School of Health Sciences, Biostatistics Program, Master's Thesis, Ankara, 2026. This study aimed to compare the performance of different machine learning algorithms in predicting in-hospital mortality in patients with acute coronary syndrome. The retrospective study included 657 patients, and the dataset was stratified according to mortality status, dividing it into 80% training (n=525) and 20% test (n=132) sets. Decision tree, random forest, XGBoost, logistic regression, Naive Bayes, support vector machines, and artificial neural networks methods were used to predict in-hospital mortality. Five-fold cross-validation, hyperparameter and Youden index-based threshold optimization were applied in model development; class imbalance was addressed with SMOTE. Variable significance was evaluated using a model-independent permutation-based approach. The median age of the patients was 62 years (IQR: 53–71), 72.8% were male, and 41.4% had STEMI, with an in-hospital mortality rate of 5.6%. Model performance varied depending on the applied modeling approach. In the baseline evaluation, logistic regression; following hyperparameter and threshold optimization, random forest; among the baseline models with SMOTE, the decision tree; and in the final stage, in which SMOTE, hyperparameter optimization, and threshold optimization were applied together, XGBoost were the models that stood out.In variable significance analyses, acute heart failure, acute kidney injury, leukocytes, and glucose were prominent variables. In conclusion, no single model consistently superior across all performance measures was identified; model performance varied depending on the modeling approach and evaluation criteria. The findings highlight the importance of evaluating model performance in clinical data with class imbalances using different criteria together.
Ankara’daki Bir Üniversite İdari Personelinin Kanser Taramalarına Yönelik Bilgi ve Tutumlarının Değerlendirilmesi
(Tıp Fakültesi, 2026-08-25) Kaan Aksu
Aksu K., Evaluation of Knowledge and Attitudes of Administrative Staff at a University in Ankara Regarding Cancer Screenings, Hacettepe University Faculty of Medicine Department of Public Health, Public Health Specialty Thesis, Ankara, 2026. This research was conducted to determine the knowledge levels, attitudes, and the variables influencing these factors among administrative staff working at the Beytepe Campus of a state university in Ankara regarding breast, cervical, and colorectal cancer screenings within the scope of national cancer screening programs. Cancer is recognized as one of the most significant public health problems both globally and in Türkiye due to high mortality rates, disability, and treatment costs. The success of early diagnosis and organized screening programs (such as KETEM, Family Health Centers, etc.) depends on the adoption of these programs by the target population and a participation rate exceeding 70%. This descriptive epidemiological study was carried out on 398 volunteers accessible within the university administrative staff population. Data were collected using a comprehensive questionnaire including sociodemographic characteristics, history of chronic diseases (most commonly thyroid, diabetes, and hypertension), healthy lifestyle habits such as tobacco and alcohol use, as well as the validated and reliable "Attitude Scale Towards Cancer Screenings" (24 items, maximum 120 points) and "Knowledge Scale Towards Cancer Screenings" (25 items, maximum 25 points). According to the findings, 63.6% of the participants were female, the mean age was 45.5, and 82.7% of the staff had a university or postgraduate education level. While 53.8% of the participants stated they were informed about national cancer screening programs, family physicians (48.6%) and social media platforms (36.9%) emerged as the primary sources of information. Only 34.8% of the staff had undergone at least one cancer screening in their lifetime; among these, breast cancer (81.2%) and cervical cancer (66.9%) screenings were common, while colorectal cancer screening remained significantly behind targets at 18.7%. The fact that a vast majority (80.4%) of those who did not undergo screening cited "not having any complaints" reveals an insufficient awareness that screenings are intended for asymptomatic individuals. Statistical analyses showed that women's median scores for both attitude (104.0) and knowledge (18.0) were significantly higher than those of men (p < 0.05). According to the knowledge scale results, 44.4% of the participants had an adequate level of knowledge (17.5 points and above). Statistical analyses revealed a positive and low correlation between knowledge levels and attitude scores (r=0.380; p<0.001). According to multivariate regression results, having children under the age of 18 negatively affected attitudes, whereas having knowledge and experience had a positive impact. Meanwhile, female gender, 55–64 age group, physical activity, and KETEM awareness were identified as the strongest predictors of the knowledge score. In conclusion, screening participation is low even among a highly educated group of university staff; therefore, it is recommended to develop strategies to increase participation in colorectal screenings—especially among male staff—to promote screening behavior as a preventive routine rather than a symptom-oriented action, and to expand institutional awareness training.