Talha Üstüntaş

Talha Üstüntaş

Auto-generated profile, based on publicly available information

Talha Üstüntaş

Faculty of Medicine, Departments of Pediatrics, Necmettin Erbakan University, Konya, Turkey

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
22 December 2025 in European Journal of Pediatrics
Verified
Journal Article
Closed access
PDF file
MGMustafa Gençeli
Mustafa Gençeli
GSGonca Başak Soran
Gonca Başak Soran
ÖAÖzge Metin Akcan
Özge Metin Akcan
MYMehtap Yücel
Mehtap Yücel
TÜTalha Üstüntaş
Talha Üstüntaş
SŞSüleyman Şahin
Süleyman Şahin

24 March 2026 in European Journal of Pediatrics
Verified
Journal Article
Open access
USUğur Saraç
Uğur Saraç
APAyşe Büşra Paydaş
Ayşe Büşra Paydaş
MGMustafa Gençeli
Mustafa Gençeli
TÜTalha Üstüntaş
Talha Üstüntaş
MYMehtap Yücel
Mehtap Yücel
MOMehmet Burhan Oflaz
Mehmet Burhan Oflaz
We evaluated three multimodal LLMs, ChatGPT (GPT-5.2), Gemini 3, and Microsoft Copilot, in pediatric ECG interpretation, focusing on clinically significant abnormalities and emergency arrhythmias with likelihood ratios as primary outcome measures. This prospective comparative diagnostic accuracy study (STARD/STARD-AI) included 264 pediatric patients with 12-lead ECGs (November 2024–November 2025). De-identified images were submitted via standardized zero-shot prompt. Three blinded pediatric cardiologists established the reference diagnosis by majority-vote consensus. Cases were classified as Tier 1 (normal), Tier 2 (abnormal, non-urgent), or Tier 3 (urgent). Two binary endpoints were assessed: clinically significant abnormality (Tier 2 + 3 vs Tier 1) and emergency abnormality (Tier 3 vs Tier 1 + 2). Clinically significant abnormalities were present in 54.5% of patients. AUC values ranged from 0.550 to 0.623, reflecting modest discrimination. For the clinically significant endpoint, + LR values were 2.05 (ChatGPT), 1.26 (Gemini), and 1.21 (Copilot); − LR values were 0.68, 0.55, and 0.81, indicating limited rule-in and insufficient rule-out utility. For the emergency endpoint, Gemini achieved 100% sensitivity (95% CI = 85.1–100.0) with − LR 0.07 (95% CI = 0.00–1.12) in a small subgroup ( n = 22); however, specificity of 30.2% and + LR of 1.40 indicate overcalling rather than diagnostic precision. No model achieved clinically meaningful rule-in utility for either endpoint. Conclusions : Current multimodal LLMs showed limited diagnostic utility in pediatric ECG interpretation, with + LR values near 1.0 across both endpoints. Standalone deployment is not supported; these tools may at most serve as adjunctive screening aids under clinician oversight. What is Known: • Deep learning algorithms trained on large ECG datasets perform well in adult populations, but evidence in pediatric ECG interpretation is limited. • General-purpose LLMs show variable accuracy in medical examinations; reliability in subspecialty domains such as pediatric cardiology remains unproven. What is New: • This is the[FCA1] first head-to-head comparative diagnostic accuracy study of multimodal LLMs in pediatric ECG evaluation, using likelihood ratios as primary outcome measures. • All three LLMs showed limited rule-in utility (+LR near 1.0); Gemini achieved potentially meaningful rule-out performance for emergency arrhythmias (−LR = 0.07), but with wide confidence intervals reflecting the small emergency subgroup (n = 22). • Gemini’s 100% sensitivity in the emergency subgroup reflects overcalling (specificity 30.2%) consistent with a triage/screening behavior rather than diagnostic precision.

8 May 2026 in European Journal of Pediatrics
Verified
Journal Article
Open access
PDF file
MGMustafa Gençeli
Mustafa Gençeli
ÖAÖzge Metin Akcan
Özge Metin Akcan
GSGonca Başak Soran
Gonca Başak Soran
AÇAbdulkerim Çokbiçer
Abdulkerim Çokbiçer
USUğur Saraç
Uğur Saraç
AAAbdullah Akkuş
Abdullah Akkuş
Pediatric exanthematous diseases pose diagnostic challenges because clinical presentations overlap. To determine whether current artificial intelligence (AI) models achieve diagnostic accuracy within or above the performance distribution of pediatric residents and specialists for common rash-associated diseases. Participants and AI models were evaluated against definitive diagnoses confirmed by clinical features, laboratory findings, and consensus of two pediatric infectious disease specialists. A volunteer sample of 263 pediatric clinicians: 107 residents (years 1 through 4) and 156 specialists. Each clinician completed a blinded multiple-choice questionnaire with a clinical photograph and accompanying clinical data per case. The same cases were presented to three AI models: ChatGPT, Gemini, and Copilot. Among 263 clinicians (107 residents, 156 specialists), specialists scored higher than residents (median, 46 [IQR, 42—50] vs 41 [IQR, 36—46]; P < .001; r = 0.32). ChatGPT correctly diagnosed 53 of 61 cases (86.9%), Gemini 50 (82.0%), and Copilot 44 (72.1%). Both ChatGPT and Gemini exceeded the upper bound of the specialist population median 95% CI (47.17). All three AI models scored above the resident 95% CI upper bound (42.76). Disease-level accuracy ranged from 0% (insect bites, all models) to 100% (9 conditions, all models). Fourth-year residents scored higher than first- and second-year residents ( P = .001; ε 2 = 0.13). Conclusions : AI models given clinical data alongside images matched or exceeded specialist-level performance for pediatric exanthems. Accuracy varied by disease; failures clustered in conditions that require contextual reasoning. Physician oversight remains necessary where AI accuracy is lowest. What is Known • Childhood exanthematous diseases pose significant diagnostic challenges due to their overlapping clinical presentations. • Artificial intelligence models are becoming increasingly proficient in accurately diagnosing these conditions. What is New • In this diagnostic accuracy study of 61 cases evaluated by 263 clinicians and 3 artificial intelligence models, ChatGPT (86.9%) and Gemini (82.0%) exceeded the 95% CI upper bound of the specialist population median. Disease-level accuracy ranged from 0 to 100% across models • Artificial intelligence models given clinical data alongside images can match or exceed specialist-level accuracy for common pediatric exanthems, but failures in context-dependent diagnoses require physician oversight

10 December 2024 in TURKISH JOURNAL OF MEDICAL SCIENCES
Verified
Journal Article
Open access
PDF file
MGMustafa Gençeli̇
Mustafa Gençeli̇
TÜTalha Üstüntaş
Talha Üstüntaş
ÖAÖzge Akcan
Özge Akcan
SSSi̇nAn Saylik
Si̇nAn Saylik
FEFati̇h Ercan
Fati̇h Ercan
MAMustafa Argun
Mustafa Argun
Background/aim: Differentiating multisystem inflammatory syndrome in children (MIS-C) from adenovirus infection (AI) can be challenging due to similar clinical and laboratory findings. This study aimed to identify distinguishing characteristics and develop a scoring system to facilitate accurate diagnosis.Materials and methods: A comprehensive review of medical records was undertaken for 108 MIS-C cases and 259 patients with confirmed AI. A comparison of laboratory data and clinical findings was conducted across patient groups through appropriate statistical tests.Results: MIS-C patients were significantly older than AI patients (p<0.001). Diarrhea, rash, abdominal pain, vomiting, non-exudative conjunctivitis, lymphadenopathy tachycardia, bradycardia, hypotension, hypoxia seizures, agitation, headache, and altered consciousness symptoms were more frequently associated with MIS-C (p<0.001), while cough and runny nose were significantly more common in AI (p<0.001). Lymphopenia and thrombocytopenia were more prevalent in MIS-C patients (p<0.001). AI and MIS-C were scored by giving 1 point each to the parameters that created the difference. For AI, ≤60 months, presence of cough, runny nose and absence of diarrhea, rash, abdominal pain, vomiting, non-exudative conjunctivitis, lymphadenopathy, tachycardia, bradycardia, hypotension, hypoxia seizures, agitation, headache, and altered consciousness, lymphopenia, thrombocytopenia and CRP <124.5 mg/L were determined as each parameter plus one point. A total score above 14 could predict AI with a high degree of accuracy sensitivity values stood at around 97.5% and specificity at 92.6%.Conclusion: The proposed inpatient scoring system, when used in conjunction with polymerase chain reaction testing, may improve the early differentiation of AI and MIS-C. This approach could help reduce unnecessary testing and optimize resource allocation. Further research with larger samples should investigate this novel scoring system to establish its reliability and generalizability.

3 March 2025 in Turkish Archives of Pediatrics
Verified
Journal Article
Open access
TÜTalha Üstüntaş
Talha Üstüntaş
MAMehmet Emre Atabek
Mehmet Emre Atabek
BEBeray Selver Eklioğlu
Beray Selver Eklioğlu
SDSaime Ergen Dibeklioğlu
Saime Ergen Dibeklioğlu
Objective: Type 1 diabetes mellitus (T1DM) is a prevalent chronic disease affecting children, necessitating ongoing care from multidisciplinary clinics. This study aimed to determine the prevalence of thyroid dysfunction in newly diagnosed T1DM patients, investigating long-term thyroid function through periodic monitoring. Materials and Methods: A retrospective cohort study utilized the medical records of pediatric patients with newly diagnosed T1DM. Extracted data included age at initial diagnosis, presenting clinical manifestations, standard laboratory findings, and thyroid function panel results. Results: The sample included 208 T1DM patients, with a higher proportion of males (54.8%) than females (45.2%). Initial assessment revealed that 69.2% of patients were euthyroid, 24.5% exhibited euthyroid sick syndrome (ESS), 5.8% compensated hypothyroidism, and 0.5% primary hypothyroidism. About 76.5% of patients initially diagnosed with ESS were euthyroid at the second visit. Analysis confirmed a significant difference in the distribution of euthyroid and ESS patients between the first and second visits (P < .001). Temporal analysis indicated that thyroid function tended to normalize within 1-2 weeks of the initial presentation. Conclusion: Tests should be performed within 1-2 weeks following clinical stabilization to evaluate thyroid function in newly diagnosed T1DM patients. However, patients with a history of thyroid disease, or those with clinical signs indicative of thyroid dysfunction, should undergo immediate testing at the time of diagnosis. Cite this article as: Üstüntaş T, Atabek ME, Selver Eklioğlu B. Transient thyroid dysfunction in new-onset type 1 diabetes: Reassessing test timing for accurate results. Turk Arch Pediatr. 2025;60(2):153-158.

16 April 2026 in European Journal of Pediatrics
Verified
Journal Article
Open access
PDF file
KEKübra Nur Erdoğan
Kübra Nur Erdoğan
MGMustafa Gençeli
Mustafa Gençeli
EBEsra Babayiğit
Esra Babayiğit
AŞAyşe Şimşek
Ayşe Şimşek
TÜTalha Üstüntaş
Talha Üstüntaş
SPSevgi Pekcan
Sevgi Pekcan
To examine factors associated with severe respiratory disease at initial presentation in pediatric hMPV-associated LRTIs. We analyzed patients aged 1 month to 18 years with PCR-confirmed hMPV-associated LRTI (January 2018–January 2024). Clinical severity was stratified via the Modified Tal Score at admission (mild ≤ 5, moderate 6–10, severe ≥ 11). Each patient contributed one episode. Multivariable binary logistic regression identified risk factors for severe respiratory disease; ROC analysis evaluated biomarker diagnostic accuracy. A total of 676 hMPV-positive patients were identified; after restricting to one episode per patient, 421 unique episodes were analyzed. Of these, 54.9% were mild, 38.5% moderate, and 6.7% ( n = 28) severe. Comorbidities were present in 67.9% of severe versus 35.5% of mild cases ( p = .004). Viral co-detections occurred in 41.3% but were unrelated to severity ( p = .235). Median CRP was 4.0, 10.0, and 32.0 mg/L in mild, moderate, and severe groups, respectively (p < .001). In multivariable regression, comorbidity was the strongest risk factor (OR 2.96, 95% CI 1.17–7.49, p = 0.022), followed by CRP (OR 1.55 per 1 SD [≈32 mg/L], p = 0.018). No age group retained significance after adjustment (model AUC 0.733). Conclusions : Severe respiratory disease was infrequent (6.7%) and associated primarily with underlying comorbidities, which increased the odds approximately threefold. Age was not independently associated with severity. CRP showed modest diagnostic accuracy (AUC 0.678, negative likelihood ratio 0.59); low levels alone could not reliably exclude severe respiratory disease. What is Known: • hMPV is an established cause of lower respiratory tract infections in children, Ranging from mild illness to severe pneumonia requiring hospitalization and respiratory support. • Infants younger than 12 months and children with comorbidities such as prematurity, chronic lung disease, congenital heart disease, or neurological disorders carry increased risk for severe hMPV infection. What is New: • Multivariable logistic regression identifies underlying comorbidity as the dominant independent risk factor for severe hMPV respiratory disease (OR 2.96, p = 0.022). No age group retains significance after adjustment, indicating that the apparent severity signal in children older than 5 years reflects a comorbidity-driven population shift rather than older age acting as an independent risk factor. • Elevated CRP, decreased lymphocyte counts, and decreased monocyte counts at admission are associated with severe respiratory disease. CRP retained independent association in multivariable analysis, though diagnostic accuracy for all three biomarkers was modest.