Denisa Oana Zelinschi

Dr. Denisa Oana Zelinschi


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
Verified
Journal Article
Open access
PDF file
Theodor Florin PantilimonescuTheodor Florin Pantilimonescu
Theodor Florin Pantilimonescu
CDCostin Dămian
Costin Dămian
VRViorel Dragos Radu
Viorel Dragos Radu
MHMaximilian Hogea
Maximilian Hogea
OCOana Andreea Costachescu
Oana Andreea Costachescu
Ionela Lacramioara SerbanIonela Lacramioara Serban
Ionela Lacramioara Serban
Urinary tract infections (UTIs) are a common pathology worldwide, frequently associated with kidney stones. We aimed to determine how artificial intelligence (AI) could assist and enhance human medical activities in this field. We performed a search in PubMed using different sets of keywords. When using the keywords “AI, artificial intelligence, urinary tract infections, Escherichia coli (E. coli)”, we identified 16 papers, 12 of which fulfilled our research criteria. When using the keywords “urolithiasis, AI, artificial intelligence”, we identified 72 results, 30 of which were suitable for analysis. We identified that AI/machine learning can be used to detect Gram-negative bacilli involved in UTIs in a fast and accurate way and to detect antibiotic-resistant genes in E. coli. The most frequent AI applications for urolithiasis can be summarized into three categories: The first category relates to patient follow-up, trying to improve physical and medical conditions after specific urologic surgical procedures. The second refers to urinary stone disease (USD), focused on stone evaluation, using different AI and machine learning systems, regarding the stone’s composition in terms of uric acid, its dimensions, its volume, and its speed of detection. The third category comprises the comparison of the ChatGPT-4, Bing AI, Grok, Claude, and Perplexity chatbots in different applications for urolithiasis. ChatGPT-4 has received the most positive evaluations. In conclusion, the impressive number of papers published on different applications of AI in UTIs and urology suggest that machine learning will be exploited effectively in the near future to optimize patient follow-up, diagnosis, and treatment.

30 July 2025 in Preprints.org
Verified
Preprint
Open access
PDF file
DZDenisa Oana Zelinschi
Denisa Oana Zelinschi
AUAlexandra Ursache
Alexandra Ursache
EMEDUARD CRISTIAN MIHOCI
EDUARD CRISTIAN MIHOCI
EAEmil Anton
Emil Anton
DNDan Navolan
Dan Navolan
DNDragos Nemescu
Dragos Nemescu
Generative artificial intelligence (GenAI), particularly large language models (LLMs) like ChatGPT, is emerging as a tool to address declining birth rates and rising infertility. It enhances reproductive education, preventive interventions, and doctor-patient communication, providing personalized healthcare information discreetly. Sexual and reproductive health is a fundamental human right, yet many lack access to quality care. GenAI offers 24/7 access to reproductive health information through chatbots, aiding patients with diagnoses and lifestyle advice. While useful, AI tools sometimes lack depth for sensitive topics. AI can create educational content tailored to various literacy levels and cultural contexts, helping individuals understand complex medical information. AI tools can destigmatize menstruation and provide personalized insights through cycle tracking, improving health management and education. AI enhances in vitro fertilization (IVF) success through better diagnostics and personalized risk profiles, though it requires careful validation and regulation. AI can improve the accessibility and quality of sexual health education, offering personalized guidance on contraceptive methods. AI chatbots can bridge gaps in reproductive healthcare for marginalized groups by providing confidential, multilingual support. AI tools can enhance communication, helping patients prepare for consultations and engage in shared decision-making. Despite its potential, GenAI faces challenges such as accuracy, bias, privacy concerns, and accessibility issues. Regulatory oversight is still developing. GenAI is transforming reproductive health by providing personalized support and education. However, it must be implemented with a focus on safety, equity, and human connection, ideally through a hybrid approach that combines AI tools with human oversight.

26 July 2026 in Diagnostics
Verified
Journal Article
Open access
PDF file
Alexandra Elena CristoforAlexandra Elena Cristofor
Alexandra Elena Cristofor
OOOriana-Maria Onicescu
Oriana-Maria Onicescu
DZDenisa Oana Zelinschi
Denisa Oana Zelinschi
AUAlexandra Ursache
Alexandra Ursache
ACAlexandru Carauleanu
Alexandru Carauleanu
DNDragos Nemescu
Dragos Nemescu
Background: First-trimester screening for preterm preeclampsia relies on validated competing-risks models integrating maternal characteristics with biochemical and biophysical markers to generate individualized risk estimates. Large language models (LLMs) are increasingly used for pregnancy-related health information, yet their alignment with established clinical risk models remains unclear. Objective: to perform an exploratory, local perturbation-based assessment of how LLM-generated numerical risk estimates reproduce the direction and relative magnitude of established preeclampsia predictor effects compared with the Fetal Medicine Foundation (FMF) competing-risks model. Methods: A total of 129 synthetic clinical scenarios were generated from a low-risk reference pregnancy using a one-factor-at-a-time perturbation approach across 16 risk factors represented as 22 predictors. Eight LLMs (Claude, GPT, DeepSeek, Gemini, Copilot, Meta, Mistral, and Grok) estimated the probability of preeclampsia requiring delivery before 37 weeks. Corresponding risks were calculated using the FMF model. Model behavior was analyzed in logit space using a local perturbation-based modeling approach inspired by Local Interpretable Model-Agnostic Explanations (LIME) to derive feature-effect coefficients. Agreement with the FMF model was assessed using correlation, directional concordance, cosine similarity, and normalized root mean squared error, summarized using an exploratory composite score. Results: The FMF model identified mean arterial pressure, placental growth factor, parity, uterine artery pulsatility index, and chronic hypertension as dominant predictors. Alignment between LLM outputs and the FMF model was heterogeneous, with composite scores ranging from 0.59 to 0.82. Models with higher descriptive scores preserved predictor directionality (up to 90.9%) and rank ordering, but agreement in magnitude and scaling was limited (R2: 0.44–0.59). Intermediate models showed preserved directionality with reduced magnitude agreement, while lower-scoring models demonstrated more frequent sign inconsistencies and minimal variance explained. Conclusions: LLMs demonstrated partial, prompt-specific alignment with the FMF model in this local perturbation analysis, particularly for predictor direction and relative importance, but did not consistently reproduce quantitative effect sizes. This approach was intended to characterize local model behavior around a predefined reference case rather than evaluate clinically realistic combinations of interacting risk factors or global clinical prediction performance. Given the evolving nature of LLMs, ongoing reassessment using standardized approaches is required.