Ali Salbas

Dr. Ali Salbas

orcid

Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

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

Representative Publications

View All Publications
30 July 2025 in Diagnostics
Verified
Journal Article
Open access
PDF file
ASAli Salbas
Ali Salbas
RBRaşit Eren Büyüktoka
Raşit Eren Büyüktoka
Background/Objectives: Multimodal large language models (LLMs) are increasingly used in radiology. However, their ability to recognize fundamental imaging features, including modality, anatomical region, imaging plane, contrast-enhancement status, and particularly specific magnetic resonance imaging (MRI) sequences, remains underexplored. This study aims to evaluate and compare the performance of three advanced multimodal LLMs (ChatGPT-4o, Claude 4 Opus, and Gemini 2.5 Pro) in classifying brain MRI sequences. Methods: A total of 130 brain MRI images from adult patients without pathological findings were used, representing 13 standard MRI series. Models were tested using zero-shot prompts for identifying modality, anatomical region, imaging plane, contrast-enhancement status, and MRI sequence. Accuracy was calculated, and differences among models were analyzed using Cochran’s Q test and McNemar test with Bonferroni correction. Results: ChatGPT-4o and Gemini 2.5 Pro achieved 100% accuracy in identifying the imaging plane and 98.46% in identifying contrast-enhancement status. MRI sequence classification accuracy was 97.7% for ChatGPT-4o, 93.1% for Gemini 2.5 Pro, and 73.1% for Claude 4 Opus (p < 0.001). The most frequent misclassifications involved fluid-attenuated inversion recovery (FLAIR) sequences, often misclassified as T1-weighted or diffusion-weighted sequences. Claude 4 Opus showed lower accuracy in susceptibility-weighted imaging (SWI) and apparent diffusion coefficient (ADC) sequences. Gemini 2.5 Pro exhibited occasional hallucinations, including irrelevant clinical details such as “hypoglycemia” and “Susac syndrome.” Conclusions: Multimodal LLMs demonstrate high accuracy in basic MRI recognition tasks but vary significantly in specific sequence classification tasks. Hallucinations emphasize caution in clinical use, underlining the need for validation, transparency, and expert oversight.

13 November 2025 in Academic Radiology
Verified
Journal Article
Closed access
ASAli Salbas
Ali Salbas
MYMurat Yogurtcu
Murat Yogurtcu

16 February 2026 in Insights into Imaging
Verified
Journal Article
Open access
ASAli Salbas
Ali Salbas
RBRaşit Eren Büyüktoka
Raşit Eren Büyüktoka
MYMurat Yogurtcu
Murat Yogurtcu
MKMünevver İlKe Kaya
Münevver İlKe Kaya
BEBinnur Emre
Binnur Emre
AKAli Murat Koc
Ali Murat Koc
Objectives: To evaluate the publication outcomes of oral presentations delivered at the European Congress of Radiology (ECR) 2019 and examine factors influencing conversion to full-text articles; findings were also compared with ECR 2010. Materials and methods: A total of 1817 oral presentations from ECR 2019 were analyzed. Publication status was determined by searching PubMed/MEDLINE up to December 2023. For each matched article, Journal Impact Factor (JIF) and Google Scholar citations/year were recorded. Additional variables included country of origin, collaboration type, imaging modality, and study design. Statistical analyses used chi-square and Kruskal–Wallis, with p < 0.05 considered significant. Results: Of 1817 oral presentations, 844 (46.5%) were published, with no significant difference from ECR 2010 (43%, p = 0.091). Abstracts originated from 71 countries, with Italy (16.5%) and China (15.5%) contributing the most. Publications appeared in 254 journals. Publication rates varied significantly by country ( p < 0.001), with Switzerland (74.4%) and the Netherlands (68.8%) achieving the highest rates. When analyzed by continent, abstracts from Asia showed a significantly higher publication rate than those from Europe (52.3% vs. 43.6%, p = 0.001). Publication outcomes also varied significantly by imaging modality ( p = 0.002) and subspecialty ( p < 0.001). Breast imaging achieved the highest median JIF (4.9), whereas Artificial Intelligence/Machine Learning (AI/ML) demonstrated the highest median annual citation count (10.5). Conclusions: Nearly half of the ECR 2019 oral presentations achieved peer-reviewed publication, maintaining rates from 2010. The congress’s contributor landscape has become more global, with greater participation from Asia. While traditional radiological fields remain prevalent, AI/ML abstracts demonstrated high citation rates. These findings reflect contemporary trends in radiological research. Critical relevance statement: By analyzing the publication outcomes of ECR 2019, with comparisons to 2010 data, this study examines evolving global patterns in publication outcomes, offering insights to enhance the dissemination of radiological research. Key Points: Converting oral presentations to publications remains challenging in radiological research. Nearly half of the ECR 2019 oral presentations were published, showing a modest, non-significant increase from ECR 2010. The congress has become increasingly global, with notable growth in participation from Asia. This study reveals radiology’s evolving scientific landscape and current research priorities. Graphical

13 March 2026 in Neuroradiology
Verified
Journal Article
Open access
PDF file
AKAli Murat Koc
Ali Murat Koc
ASAli Salbas
Ali Salbas
Purpose: This study aimed to evaluate the publication outcomes of oral abstracts presented at the European Society of Neuroradiology (ESNR) congresses between 2018 and 2022 and to identify factors associated with subsequent publication in peer-reviewed journals. Methods: All oral abstracts presented at ESNR congresses from 2018 to 2022 were identified from the official Books of Abstracts. Publication status was determined through a systematic search of the PubMed/MEDLINE database. For each abstract, subspecialty, study design, and type of institutional collaboration were recorded. Publication rates, time to publication, journal impact factors, and citation metrics were analyzed. Statistical comparisons were performed using chi-square, Kruskal–Wallis, and Mann–Whitney U tests with Bonferroni correction. Results: Among 241 oral presentations, 114 (47.3%) were subsequently published as full-text articles. The mean time to publication was 16 months, and 90.4% of papers were published within three years after presentation. The most frequent journal of publication was Neuroradiology . Publication rates differed significantly by subspecialty, study design, and collaboration type. Studies involving multi-institutional collaboration showed higher publication success, and internationally collaborative studies also received higher citation counts. Conclusion: Approximately half of the oral presentations at ESNR congresses were published in peer-reviewed journals. Multi-institutional collaboration substantially enhances publication success and academic impact. Encouraging collaboration and supporting research development within the ESNR framework may further improve the visibility and strengthen the academic influence of future studies.

23 September 2025 in Joint Diseases and Related Surgery
Verified
Journal Article
Open access
ASAli Salbas
Ali Salbas
EBEbru Kul Baysan
Ebru Kul Baysan
Joint Diseases and Related Surgery (JDRS), the official journal of the Turkish Joint Diseases Foundation, is an international, peer-reviewed, immediate open access journal. All published articles in JDRS has been free for everyone, without embargo, since 1990.

15 September 2025 in Diagnostic and Interventional Radiology
Verified
Journal Article
Open access
PDF file
ASAli Salbas
Ali Salbas
AKALİ MURAT KOÇ
ALİ MURAT KOÇ
Bibliometric analysis of radiology residency theses in Türkiye: publication metrics and trends - Diagnostic and Interventional Radiology