Background Large language models (LLMs) are increasingly being studied as potentially valuable support tools in oncology practice including clinical decision support. Yet, their real-world utility in melanoma treatment decision-making is still not sufficiently considered, especially in resource-limited settings. Accordingly, this study evaluated the performance of four LLMs against real-world treatment decisions of a melanoma multidisciplinary tumor board (MDT). Methods This retrospective single-center study included 151 consecutive patients with newly diagnosed cutaneous melanoma discussed at the MDT at the University Clinical Center Tuzla, Bosnia and Herzegovina, between 2020 and 2024. Melanoma treatment recommendations generated by four LLMs, ChatGPT-4o, ChatGPT-5 Thinking, Gemini 2.5 Pro and DeepSeek-V3.2, were evaluated by four board-certified oncologists against the actual MDT treatment decisions. Additionally, the LLM-generated recommendations were also rated across five pre-specified domains: clarity, clinical applicability, coverage, explanation and support with evidence, and guideline concordance. Results In this study, inter-rater reliability was acceptable to good, supporting the consistency of expert evaluation. ChatGPT-5 Thinking showed the strongest and most consistent overall performance, followed by ChatGPT-4o, while Gemini 2.5 Pro and DeepSeek-V3.2 were rated less favorably. Differences between LLMs were statistically significant across all evaluated domains. Performance differences appeared most clinically relevant in more complex scenarios, particularly when consideration of adjuvant or systemic treatment strategies was required. Conclusion The findings of this study suggest that selected LLMs may have a supportive role in everyday melanoma MDT practice particularly in oncology centers with limited resources. However, the current results do not support the use of LLM-generated recommendations as independent treatment decisions, and further prospective studies are required before LLM-assisted treatment recommendations can be safely integrated into the MDT workflow.

Chimeric Antigen Receptors (CAR) T-cell therapy is a ground-breaking discovery in immunotherapy, mainly known for its exceptional results in treating haematological malignancies. The latest research has revealed that the potential of CAR T-cell therapy extends far beyond its current capabilities and could represent a novel therapeutic approach for treating various cancers. This review aims to summarize the latest innovations in CAR T-cell therapy applied in cancer treatment, including multiple myeloma, osteosarcoma, glioblastoma, melanoma and various childhood malignancies. However, several challenges limit success of CAR T-cell therapy, including the antigen escape phenomenon, 'on-target off-tumour' toxicity, penetration into solid tumour tissue, alongside the cost-effectiveness concerns. The improvement of cancer immunotherapies currently available requires an increase in the effectiveness of CAR T-cells in managing refractory and solid cancers. This could be achieved by using CAR T-cells to target various antigens, enhancing their local delivery and tumour infiltration capabilities and utilizing CAR T-cells in combination with checkpoint blockade and immunotherapy, such as PD-1 blockade and CD19 CAR T-cell combined therapy. Although CAR T-cell treatment offers a lot of promise, its cost needs to be taken into account, especially in healthcare systems with limited funding. More importantly, frameworks for Health Technology Assessment (HTA) must adapt to incorporate ethical, sociological and psychological aspects. Reducing CAR T-cell toxicity is also essential, as it remains among biggest obstacles to their widespread application in clinical practice. Future research should therefore focus on enhancing our understanding of CAR T-cell therapy and expanding the application of immunotherapy in treatment.

During the early months of the COVID-19 pandemic, treatment protocols varied substantially among countries and even between hospitals. This study compared the clinical characteristics, management strategies, and outcomes of hospitalized COVID-19 patients treated in tertiary centers in Türkiye and Bosnia and Herzegovina. We retrospectively analyzed 1338 adults hospitalized with laboratory-confirmed SARS-CoV-2 infection: 657 patients in Tuzla (Bosnia and Herzegovina, June–December 2020) and 681 in İstanbul (Turkiye, April–May 2020). Demographic, clinical, and laboratory data, treatment details (including favipiravir use), need for invasive or non-invasive mechanical ventilation, and in-hospital mortality were extracted from medical records. Patients in Bosnia and Herzegovina were older (61.6 ± 14.4 vs. 56.9 ± 15.8 years; p < 0.001) and had longer hospital stays (9.0 ± 5.5 vs. 7.7 ± 6.1 days; p < 0.001). In the Bosnian cohort, leukocyte, neutrophil, platelet, ferritin, CRP, troponin, creatinine, AST, and ALT levels were significantly higher, whereas hemoglobin and D-dimer levels were lower. The need for ventilatory support was greater in Bosnia and Herzegovina (15.1% vs. 12.2%, p < 0.001), and overall mortality was higher (25.7% vs. 9.3%, p < 0.001). No mortality difference was observed between patients treated and not treated with favipiravir. Despite similar inclusion criteria, patients in Bosnia and Herzegovina exhibited more severe disease, greater organ involvement, and higher mortality than those in Turkiye. Favipiravir use did not influence survival. Inter-country comparisons highlight how differing patient profiles and treatment protocols may impact COVID-19 outcomes; however, interpretation should consider that the two centers contributed data from different phases of the 2020 pandemic, and that country-level differences in circulating variants, healthcare capacity, hospital strain, and evolving clinical guidelines may also have influenced the observed patterns.

Background: Breast cancer remains the most common cancer in women worldwide. Treatment has evolved into multimodal approaches, with pathologic complete response (pCR) after neoadjuvant chemotherapy (NAC) serving as a key prognostic marker. The aim of this study was to evaluate the value of inflammatory markers in predicting pCR to NAC in breast cancer. Methods: This cross-sectional study of 74 patients with breast cancer who underwent NAC followed by surgery included demographic, tumor, and immune-inflammatory marker data. Receiver operating characteristic curve analysis and the Youden index were used to determine optimal cutoff values. Univariate and multivariate logistic regression assessed associations between markers and pCR, adjusting for tumor stage, human epidermal growth factor receptor 2 (HER2), and estrogen receptor (ER) status. Results: Our multivariate analysis identified the pan-immune-inflammation value (PIV), HER2 status, and ER status as significant independent predictors of pCR. PIV (OR, 4.28; 95% CI, 1.59–16.88) remained significant among inflammatory markers, while the neutrophil-to-lymphocyte ratio (NLR), monocyte-to-lymphocyte ratio (MLR), and platelet-to-lymphocyte ratio (PLR) did not. HER2-positive (OR, 7.45; 95% CI, 2.30–24.15) and hormone receptor (HR)–negative (OR, 7.02; 95% CI, 2.63–18.70) statuses were also strongly associated with pCR. Conclusion: PIV is a robust predictor of pCR in patients with breast cancer receiving NAC, offering a comprehensive reflection of the immune-inflammatory state. Incorporating PIV with tumor-specific markers (e.g., receptor status, Ki-67, grade) may enhance treatment stratification. Further validation in diverse cohorts is warranted.

Background/Objectives: In many low- and middle-income countries (LMICs), including Bosnia and Herzegovina, oncology services are constrained by a limited number of specialists and uneven access to evidence-based care. Artificial intelligence (AI), particularly large language models (LLMs) such as ChatGPT, may provide clinical decision support to help standardize treatment and assist clinicians where oncology expertise is scarce. This study aimed to evaluate the concordance, safety, and clinical appropriateness of ChatGPT-generated treatment recommendations compared to decisions made by a multidisciplinary team (MDT) in the management of newly diagnosed breast cancer patients. Methods: This retrospective study included 91 patients with newly diagnosed, treatment-naïve breast cancer, presented to an MDT in Bosnia and Herzegovina in 2023. Patient data were entered into ChatGPT-4.0 to generate treatment recommendations. Four board-certified oncologists, two internal and two external, evaluated ChatGPT’s suggestions against MDT decisions using a 4-point Likert scale. Agreement was analyzed using descriptive statistics, Cronbach’s alpha, and Fleiss’ kappa. Results: The mean agreement score between ChatGPT and MDT decisions was 3.31 (SD = 0.10), with high consistency across oncologist ratings (Cronbach’s alpha = 0.86). Fleiss’ kappa indicated moderate inter-rater reliability (κ = 0.31, p < 0.001). Higher agreement was observed in patients with hormone receptor-negative tumors and those treated with standard chemotherapy regimens. Lower agreement occurred in cases requiring individualized decisions, such as low-grade tumors or uncertain indications for surgery or endocrine therapy. Conclusions: ChatGPT showed high concordance with MDT treatment plans, especially in standardized clinical scenarios. In resource-limited settings, AI tools may support oncology decision-making and help bridge gaps in clinical expertise. However, careful validation and expert oversight remain essential for safe and effective use in practice.

Sex cord-stromal tumors are rare ovarian neoplasms, with fibromas comprising approximately 4% and thecomas accounting for 0.5%–1% of all ovarian tumors. The occurrence of these tumors outside the ovaries is exceptionally rare and diagnostically challenging, often mimicking malignancy when associated with ascites, elevated CA-125 levels, or Meigs-like syndrome. This review aims to synthesize current knowledge on the histopathological, immunohistochemical, radiological, and molecular features of ovarian fibroma-thecoma group tumors and highlight their clinical relevance. We report two postmenopausal women with large abdominal masses located extraovarian: one in the broad ligament and the other adherent to the omentum and intestines. In the first case, markedly elevated CA-125, ascites, and pleural effusion initially suggested Meigs syndrome. The second case presented with an abdominal mass and ascites. Imaging studies indicated the possibility of malignant ovarian tumors in both patients, leading to surgical excision. Histopathological examination revealed spindle-to-oval tumor cells arranged in fascicular or storiform patterns, with focal lipid-rich theca-like cells. Immunohistochemical analysis showed that the tumors were positive for vimentin, WT1, progesterone receptor (PR), and variably for estrogen receptor (ER), CD56, inhibin, and calretinin, while being negative for markers of epithelial, melanocytic, and gastrointestinal stromal tumors. A review of the literature identified only 11 well-documented cases of extraovarian fibroma-thecoma group tumors, which most commonly arise in the broad ligament or pelvic cavity. These cases are frequently associated with ascites and elevated CA-125 levels and are often misdiagnosed preoperatively as malignant disease. Our cases underscore the importance of considering extraovarian fibromas and thecomas in the differential diagnosis of pelvic and abdominal masses presenting with similar features. Accurate pathological assessment can prevent unnecessary radical surgeries and promote more favorable patient outcomes.

Introduction. Sentinel lymph node biopsy (SLNB) has significantly advanced axillary staging in clinically node-negative breast cancer, offering lower morbidity compared to traditional axillary lymph node dissection (ALND). Nonetheless, precise prediction of non-sentinel lymph node (non-SLN) involvement remains essential for optimizing surgical decisions and preventing unnecessary ALND. Methods. A retrospective cohort analysis was performed on 176 patients with clinically node-negative breast cancer who underwent SLNB. Clinicopathological data were reviewed to evaluate associations between various predictive factors and non-SLN involvement. Variables analyzed included tumor size, histological grade, lymphovascular invasion (LVI), Ki-67 proliferation index, and sentinel lymph node characteristics. Results. Multivariable logistic regression identified the type of SLN metastasis (OR=21.4; 95% CI 1.7–43.6; p=0.01), the number of positive SLNs (OR=5.66; 95% CI 1.18–36.6; p=0.03), and the number of negative SLNs (OR=0.04; 95% CI 0.006–0.27; p=0.001) as independent predictors of non-SLN metastases. The predictive model demonstrated excellent discriminatory power, with an area under the receiver operating characteristic curve (AUC) of 0.91. Conclusion. Specific clinical and histopathological variables reliably predict non-SLN involvement in SLN-positive breast cancer patients. Incorporation of these predictors into clinical practice may enhance individualized axillary management and reduce unnecessary ALND procedures. Further validation through larger prospective studies is warranted. Key words: Breast Neoplasms, Sentinel Lymph Node Biopsy, Axillary Lymph Nodes, Lymph Node Dissection, Neoplasm Staging.

DIALOG + is a low-cost intervention proven to improve the subjective quality of life in patients with psychosis and anxiety disorders in low- and middle-income countries. In a recent study, DIALOG + was shown to be feasible for patients in primary care settings with long-term physical conditions and to result in an improvement in patient outcomes. The aim of this qualitative study was to explore the experiences of patients and clinicians using DIALOG + in Bosnia and Herzegovina to gain a better understanding of its impact in this setting. In-depth semi-structured interviews were conducted with 11 patients and 4 physicians, as well as two focus groups with 5 patients in each, all of whom participated in the intervention. Specific life and treatment domains discussed during the sessions between patients and clinicians were also analysed to determine which domains were most frequently addressed and where patients needed the most support. The interviews were audio-recorded, transcribed, and analysed using thematic analysis. Four qualitative themes were identified: (1) DIALOG + structure and solution-oriented approach are helpful; (2) DIALOG + allows space for conversation; (3) Therapeutic relationship is improved, and (4) The intervention has its limitations. DIALOG + is a novel primary care intervention with positive effects on patients’ lives, which enhance primary care. Nevertheless, it presents a new challenge in this setting. It is necessary to make adjustments in primary care, such as providing clinicians with more extensive training and ongoing support, as well as providing more time for the intervention’s implementation. Study was registered prospectively within the ISRCTN Registry: ISRCTN17003451, 02/12/2020.

INTRODUCTION Acute bacterial meningitis (ABM) is a serious infectious disease and medical emergency. Given the fact of its high mortality and morbidity, detecting prognostic factors is potentially useful in improving treatment strategies. This paper aims to determine prognostic factors of short-term outcomes of patients with ABM in a developing country - Bosnia and Herzegovina, measured by the Modified Rankin Scale (MRS) at discharge. METHODOLOGY In this retrospective cohort study, 56 patients treated at the Clinic of Infectious Diseases, Clinical Center University of Sarajevo, for 11 years (2012-2022) were included. Statistical analysis was performed using IBM SPSS Statistics version 29.0.1.0. RESULTS The subjects' average age was 31.6 ± 27.7 years, with the youngest patient being 4 months and the oldest 75 years old. Among those, 31 were male and 25 were female. Unfavorable outcomes had 16 (28.6%) patients, including fatal outcomes in four patients and severe disabilities in 12 patients (MRS: 2-6), while 40 (71.4%) patients had favorable outcomes (MRS: 0-1). In the multivariate analysis, predictors of unfavorable outcomes included age older than 60 years, duration of symptoms longer than 24 hours, presence of neurological defects at admission, impaired consciousness, respiratory distress, and no corticosteroid use during treatment. CONCLUSIONS There is certainly a window of opportunity for patients with ABM: The shorter the time between disease onset and treatment initiation, the better the disease outcome.

Background/Objectives: Squamous cell lung cancer (SCC) presents a significant treatment challenge due to its poor prognosis and limited therapeutic options. In many resource-limited countries, access to advanced molecular testing is often unavailable, making the identification of novel and reliable prognostic markers crucial for improving patient selection for systemic treatments. Methods: This single-center, retrospective study investigated the prognostic value of inflammatory biomarkers, including the systemic immune-inflammation index (SII), neutrophil-to-lymphocyte ratio (NLR), and platelet-to-lymphocyte ratio (PLR), in 134 patients diagnosed with SCC. Patients were stratified into groups based on optimal cut-off values determined by ROC analysis for each biomarker. Results: Elevated levels of the SII, NLR, and PLR were significantly associated with shorter overall survival in patients with SCC (all p < 0.05). Conclusions: These easily accessible and cost-effective laboratory parameters are particularly valuable in settings where molecular testing is not available, aiding in the identification of high-risk patients and optimizing treatment selection for chemotherapy.

AIM The position of science and scientific research in Bosnia and Herzegovina (B&H) is unfavourable. University education lacks research programs, hindering students from developing critical evaluation skills. METHODS This cross-sectional study examined knowledge, attitudes, experience and perceived barriers towards scientific research among biomedical students and recent graduates at the University of Sarajevo in B&H. A self-reporting questionnaire was distributed via social networks and among students from all years of the Medical, Dentistry, Pharmaceutical faculties, the Faculty of Health Sciences, and the Faculty of Science, as well as recent graduates. RESULTS The survey was completed by 243 participants. The mean knowledge score of 4.3/11 was noticed; 216 (96%) believed research was necessary for healthcare improvement. Although 147 participants (65.3%) attended courses on research methodology, only 63 (28%) engaged in scientific research at their faculties. Only 5 (8.5%) employed graduates participated in research at their jobs. Almost half, 99 (44%) participants consulted scientific papers only upon recommendation. The majority of participants, 199 (82.2%), reported they experienced barriers to conducting research in B&H. CONCLUSION Biomedical students and graduates demonstrated limited research knowledge, but had positive attitudes and acknowledged significant barriers. Improved faculty curricula in research areas and more opportunities are needed.

Introduction The management of long-term physical conditions is a challenge worldwide, absorbing a majority resources despite the importance of acute care. The management of these conditions is done largely in primary care and so interventions to improve primary care could have an enormous impact. However, very little data exist on how to do this. Mental distress is frequently comorbid with long term physical conditions, and can impact on health behaviour and adherence, leading to poorer outcomes. DIALOG+ is a low-cost, patient-centred and solution-focused intervention, which is used in routine patient-clinician meetings and has been shown to improve outcomes in mental health care. The question arises as to whether it could also be used in primary care to improve the quality of life and mental health of patients with long-term physical conditions. This is particularly important for low- and middle-income countries with limited health care resources. Methods An exploratory non-controlled multi-site trial was conducted in Bosnia and Herzegovina, Colombia, and Uganda. Feasibility was determined by recruitment, retention, and session completion. Patient outcomes (quality of life, anxiety and depression symptoms, objective social situation) were assessed at baseline and after three approximately monthly DIALOG+ sessions. Results A total of 117 patients were enrolled in the study, 25 in Bosnia and Herzegovina, 32 in Colombia, and 60 in Uganda. In each country, more than 75% of anticipated participants were recruited, with retention rates over 90% and completion of the intervention exceeding 92%. Patients had significantly higher quality of life and fewer anxiety and depression symptoms at post-intervention follow-up, with moderate to large effect sizes. There were no significant improvements in objective social situation. Conclusion The findings from this exploratory trial suggest that DIALOG+ is feasible in primary care settings for patients with long-term physical conditions and may substantially improve patient outcomes. Future research may test implementation and effectiveness of DIALOG+ in randomized controlled trials in wider primary care settings in low- and middle-income countries. Trial registration All studies were registered prospectively within the ISRCTN Registry. ISRCTN17003451, 02/12/2020 (Bosnia and Herzegovina), ISRCTN14018729, 01/12/2020 (Colombia) and ISRCTN50335796, 02/12/2020 (Uganda).

Background: Chest pain represents a prevalent complaint in emergency departments (EDs), where the precise differentiation between acute coronary syndrome and alternative conditions assumes paramount significance. This pilot study aimed to assess the HEART score’s implementation in West Balkan EDs. Methods: A retrospective analysis was performed on a prospective cohort comprising patients presenting with chest pain admitted to EDs in Sarajevo, Zenica, and Belgrade between July and December 2022. Results: A total of 303 patients were included, with 128 classified as low-risk based on the HEART score and 175 classified as moderate-to-high-risk. The low-risk patients exhibited younger age and a lower prevalence of cardiovascular risk factors. Laboratory and anamnestic findings revealed higher levels of C-reactive protein, ALT, and creatinine, higher rates of moderately to highly suspicious chest pain history, a greater number of cardiovascular risk factors, and elevated troponin levels in moderate-to-high-risk patients. Comparatively, among patients with a low HEART score, 2.3% experienced MACE, whereas those with a moderate-to high-risk HEART score had a MACE rate of 10.2%. A moderate-to-high-risk HEART score demonstrated a sensitivity of 91.2% (95%CI 90.2–93.4%) and specificity of 46.5% (95%CI 39.9–48.3%) for predicting MACE. Conclusion: This pilot study offers preliminary insights into the integration of the HEART score within the emergency departments of the West Balkan region.

Abstract Gut microbiota is known as unique collection of microorganisms (including bacteria, archaea, eukaryotes and viruses) that exist in a complex environment of the gut. Recently, this has become one of the most popular areas of research in medicine because this plays not only an important role in disease development, but gut microbiota also influences drug pharmacokinetics. These alterations in drug pharmacokinetic pathways and drug concentration in plasma and blood often lead to an increase in the incidence of toxicological events in patients. This review aims to present current knowledge of the most commonly used drugs in clinical practice and their dynamic interplay with the host’s gut microbiota as well as the mechanisms underlying these metabolic processes and the consequent effect on their therapeutic efficacy and safety. These new findings set a foundation for the development of personalized treatments specific to each metabolism, maximizing drugs’ therapeutic effects and minimizing the side effects because they are one of the major limiting factors in treating patients.

Summary Background Congenital anomalies are the fifth leading cause of mortality in children younger than 5 years globally. Many gastrointestinal congenital anomalies are fatal without timely access to neonatal surgical care, but few studies have been done on these conditions in low-income and middle-income countries (LMICs). We compared outcomes of the seven most common gastrointestinal congenital anomalies in low-income, middle-income, and high-income countries globally, and identified factors associated with mortality. Methods We did a multicentre, international prospective cohort study of patients younger than 16 years, presenting to hospital for the first time with oesophageal atresia, congenital diaphragmatic hernia, intestinal atresia, gastroschisis, exomphalos, anorectal malformation, and Hirschsprung's disease. Recruitment was of consecutive patients for a minimum of 1 month between October, 2018, and April, 2019. We collected data on patient demographics, clinical status, interventions, and outcomes using the REDCap platform. Patients were followed up for 30 days after primary intervention, or 30 days after admission if they did not receive an intervention. The primary outcome was all-cause, in-hospital mortality for all conditions combined and each condition individually, stratified by country income status. We did a complete case analysis. Findings We included 3849 patients with 3975 study conditions (560 with oesophageal atresia, 448 with congenital diaphragmatic hernia, 681 with intestinal atresia, 453 with gastroschisis, 325 with exomphalos, 991 with anorectal malformation, and 517 with Hirschsprung's disease) from 264 hospitals (89 in high-income countries, 166 in middle-income countries, and nine in low-income countries) in 74 countries. Of the 3849 patients, 2231 (58·0%) were male. Median gestational age at birth was 38 weeks (IQR 36–39) and median bodyweight at presentation was 2·8 kg (2·3–3·3). Mortality among all patients was 37 (39·8%) of 93 in low-income countries, 583 (20·4%) of 2860 in middle-income countries, and 50 (5·6%) of 896 in high-income countries (p<0·0001 between all country income groups). Gastroschisis had the greatest difference in mortality between country income strata (nine [90·0%] of ten in low-income countries, 97 [31·9%] of 304 in middle-income countries, and two [1·4%] of 139 in high-income countries; p≤0·0001 between all country income groups). Factors significantly associated with higher mortality for all patients combined included country income status (low-income vs high-income countries, risk ratio 2·78 [95% CI 1·88–4·11], p<0·0001; middle-income vs high-income countries, 2·11 [1·59–2·79], p<0·0001), sepsis at presentation (1·20 [1·04–1·40], p=0·016), higher American Society of Anesthesiologists (ASA) score at primary intervention (ASA 4–5 vs ASA 1–2, 1·82 [1·40–2·35], p<0·0001; ASA 3 vs ASA 1–2, 1·58, [1·30–1·92], p<0·0001]), surgical safety checklist not used (1·39 [1·02–1·90], p=0·035), and ventilation or parenteral nutrition unavailable when needed (ventilation 1·96, [1·41–2·71], p=0·0001; parenteral nutrition 1·35, [1·05–1·74], p=0·018). Administration of parenteral nutrition (0·61, [0·47–0·79], p=0·0002) and use of a peripherally inserted central catheter (0·65 [0·50–0·86], p=0·0024) or percutaneous central line (0·69 [0·48–1·00], p=0·049) were associated with lower mortality. Interpretation Unacceptable differences in mortality exist for gastrointestinal congenital anomalies between low-income, middle-income, and high-income countries. Improving access to quality neonatal surgical care in LMICs will be vital to achieve Sustainable Development Goal 3.2 of ending preventable deaths in neonates and children younger than 5 years by 2030. Funding Wellcome Trust.