Building Trustworthy Generative Artificial Intelligence for Diabetes Care and Limb Preservation: A Medical Knowledge Extraction Case.

BACKGROUND: Large language models (LLMs) offer significant potential in medical information extraction but carry risks of generating incorrect information. This study aims to develop and validate a retriever-augmented generation (RAG) model that provides accurate medical knowledge about diabetes and diabetic foot care to laypersons with an eighth-grade literacy level. Improving health literacy through patient education is paramount to addressing the problem of limb loss in the diabetic population. In addition to affecting patient well-being through improved outcomes, improved physician well-being is an important outcome of a self-management model for patient health education. METHODS: We used an RAG architecture and built a question-and-answer artificial intelligence (AI) model to extract knowledge in response to questions pertaining to diabetes and diabetic foot care. We utilized GPT-4 by OpenAI, with Pinecone as a vector database. The NIH National Standards for Diabetes Self-Management Education served as the basis for our knowledge base. The model's outputs were validated through expert review against established guidelines and literature. Fifty-eight keywords were used to select 295 articles and the model was tested against 175 questions across topics. RESULTS: The study demonstrated that with appropriate content volume and few-shot learning prompts, the RAG model achieved 98% accuracy, confirming its capability to offer user-friendly and comprehensible medical information. CONCLUSION: The RAG model represents a promising tool for delivering reliable medical knowledge to the public which can be used for self-education and self-management for diabetes, highlighting the importance of content validation and innovative prompt engineering in AI applications.

CYP2C8 rs11572080 and CYP3A4 rs2740574 risk genotypes in paclitaxel-treated premenopausal breast cancer patients.

Breast cancer (BC) is the most prevalent malignancy in women globally. At time of diagnosis, premenopausal BC is considered more aggressive and harder to treat than postmenopausal cases. Cytochrome P450 (CYP) enzymes are responsible for phase I of estrogen metabolism and thus, they are prominently involved in the pathogenesis of BC. Moreover, CYP subfamily 2C and 3A play a pivotal role in the metabolism of taxane anticancer agents. To understand genetic risk factors that may have a role in pre-menopausal BC we studied the genotypic variants of CYP2C8, rs11572080 and CYP3A4, rs2740574 in female BC patients on taxane-based therapy and their association with menopausal status. Our study comprised 105 female patients with histologically proven BC on paclitaxel-therapy. They were stratified into pre-menopausal (n = 52, 49.5%) and post-menopausal (n = 53, 50.5%) groups. Genotyping was done using TaqMan assays and employed on Quantstudio 12 K flex real-time platform. Significant increased frequencies of rs11572080 heterozygous CT genotype and variant T allele were established in pre-menopausal group compared to post-menopausal group (p = 0.023, 0.01, respectively). Moreover, logistic regression analysis revealed a significant association between rs11572080 CT genotype and premenopausal BC. However, regarding rs2740574, no significant differences in genotypes and allele frequencies between both groups were detected. We reported a significant association between CYP2C8 genotypic variants and premenopausal BC risk in Egyptian females. Further studies on larger sample sizes are still needed to evaluate its importance in early prediction of BC in young women and its effect on treatment outcome.

Assessment of vitamin D status and vitamin D receptor polymorphism in Egyptian children with Type 1 diabetes.

BACKGROUND: The endocrine system of vitamin D regulates about 3 % of the human genome. Vitamin D exerts its actions via a nuclear vitamin D receptor (VDR) which in turn regulates insulin secretion from the pancreas. VDR gene polymorphisms could have an impact on how autoimmune illnesses like Type 1 diabetes mellitus (T1DM) develop. We aimed to explore the relation between T1DM and VDR gene polymorphisms in Egyptian diabetic children and their siblings. METHODS: Enzyme-linked immunosorbent assay was used to quantify 25(OH) vitamin D in the study, which had 179 participants (group 1 = 85 diabetic children, group 2 = 57 siblings of the patients, group 3 = 37 healthy controls). Real-time polymerase chain reaction (RT-PCR) was used to analyze the genotyping of the VDR gene polymorphisms Apa-I (rs7975232), Fok-I (rs2228570), Taq-I (rs731236) and Bsm-I (rs1544410). RESULTS: The mean serum 25(OH) vitamin D levels was significantly lower in T1DM patients (14.99 ± 9.24 ng/mL) and siblings (16.31 ± 7.96 ng/mL) compared to the controls (19.48 ± 7.42 ng/mL) (p = 0.031). The genotypes distribution of VDR Fok-I (rs2228570) and Bsm-I (rs1544410) polymorphisms showed a significant difference between patients, siblings and controls as P = 0.001 and 0.026 respectively, while the VDR ApaI and TaqI polymorphisms did not. FokI-A allele frequency was significantly lower in T1DM patients and siblings than in controls (p < 0.001). FokI-AA genotype had a statistical significant higher vitamin D levels than other genotypes with p value of 0.024. CONCLUSION: Our study found that T1DM children had lower vitamin D levels, and VDR FokI and BsmI gene polymorphisms were linked to T1DM in Egyptian children. Determining the relationship between vitamin D levels and VDR polymorphisms, particularly the FokI and other genetic analyses may aid in the early diagnosis of T1DM in children.

Study of CD4+ T-lymphocytes in chronic kidney disease patients with COVID-19 infection.

Chronic kidney disease (CKD) is a functional and/or structural kidney damage that lasts more than three months duration. This study aimed to analyze CD4+ T-lymphocytes levels in chronic CKD patients specifically, during the coronavirus disease 2019 (COVID-19) pandemic to assess the adaptive cell-mediated immunity. The study measured absolute CD4+ T-lymphocytes counts by flowcytometry among participating individuals. The study included 146 subjects, 40 CKD patients and tested positive for COVID-19, 44 CKD patients and tested negative for COVID-19 and 62 normal individuals as controls. There was a significant impact of COVID-19 infection in CKD patients showing lower absolute CD4+ T-lymphocytes values to more than six folds compared to the control individuals (Odds Ratio: 72.63, p= 0.0001). Also, there was a significant correlation between the decrease in absolute CD4+ T-lymphocytes counts and the advanced stages of CKD. Therefore, the study indicated that CKD causes an obvious alteration in the body immune system as decreased CD4+ T-lymphocytes levels alongside with the advanced CKD stages. While COVID-19 infection exposes CKD patients to be 50% more likely to express lower values of CD4+ T-lymphocytes levels compared to the negative tested CKD patients. In conclusion, poor immune response and increased morbidity and mortality could be correlated with CKD patients especially when associated with COVID-19 infection as comorbidity.

Evaluation of prognostic variables in chronic lymphocytic leukemia and association with disease stage.

The aim of the present study was to investigate different biological prognostic markers to identify high-risk patients with chronic lymphocytic leukemia (CLL) with a higher tumor burden, in order to ensure appropriate management. A total of 81 Egyptian patients with CLL were enrolled in the present study, with 75 healthy subjects serving as the control group. The expression of CD49d, CD38 and ZAP-70 in CLL cells was assessed using flow cytometry. The fluorescence in situ hybridization technique was employed to evaluate TP53 (del17p), ataxia-telangiectasia (del11q) and 13q14 (del13q14) genes and the presence of trisomy 12. The serological markers ß2 microglobulin (B2M) and sCD23 were measured by ELISA. The CD49d gene was highly expressed in 25.9% and cytogenetic aberrations were observed in 66.6% of all recruited CLL patients. The patients were categorized according to the Binet staging system and a significant increase in the expression of sCD23, CD49d and ZAP-70 was detected in group C (P=0.008, 0.034 and 0.017, respectively) when compared to groups A and B. CD49d+ patients exhibited significantly higher expression of CD38 (P=0.002) and trisomy 12 (P=0.015) and lower expression of del13q14 (P=0.001). Patients who were CD49d+ with B2M>3.5 µg/ml exhibited higher total leukocyte count (P=0.048), higher absolute lymphocyte count (P=0.036), higher expression of CD38 (P=0.002) and trisomy 12 (P=0.034) and lower expression of del13q14 (P=0.002). Therefore, sCD23, CD49d and ZAP-70 may be considered as an optimal prognostic marker combination to be evaluated in the early stages of CLL and throughout disease management. Integrating both serological markers and CD49d expression by flow cytometry may add to the prognostic value of each marker alone and help identify high-risk patients with a higher tumor burden.

Risk factors for severity of thrombocytopenia in full term infants: a single center study.

BACKGROUND: Neonatal thrombocytopenia (NT) (platelet count < 150 × 109/L) is a common finding in the neonatal intensive care unit (NICU). The main aim of this study was to assess the prevalence, risk factors, and outcomes of severe NT in full term (FT) infants. METHODS: During the study period, all FT infants who met the inclusion criteria for NT on two occasions were included. Maternal data, such as maternal age, weight, gestational age, mode of delivery, and history of systemic diseases, including diabetes mellitus, pre-eclampsia, systemic lupus erythematosus, and immune thrombocytopenic purpura, were recorded. Furthermore, neonatal data, such as gender, neonatal weight, causes/duration of admission, types of respiratory support used, complete blood count measurements, and outcomes for neonates admitted to the NICU, were recorded. RESULTS: In total, 55 FT infants with NT met the inclusion criteria, and 29 (52.73%) cases had severe NT. The most common cause of NT was neonatal sepsis (20 cases, 36.35%), followed by a postoperative state (5 cases, 9.09%). Moreover, in cases of positive blood cultures, the most commonly isolated organism was Escherichia coli (6 cases, 10.90%), followed by Klebsiella (5 cases, 9.09%). Cases of severe NT needed more platelet transfusions (P = 0.001) and had higher rates of mortality (P = 0.001) when compared to cases of mild/moderate NT associated with signs of bleeding and pulmonary/intraventricular hemorrhage (IVH) (P = 0.001). CONCLUSION: Severe NT compared to mild/moderate NT, associated with signs of bleeding and pulmonary/IVH, needed more platelet transfusions, and had increased mortality. Further research is needed to explain which of these complications related to severity of thrombocytopenia or were associated with original disease of the babies.

ITGA4 gene methylation status in chronic lymphocytic leukemia.

BACKGROUND: We aimed to investigate ITGA4 gene expression pattern and to explore its methylation heterogeneity in chronic lymphocytic leukemia (CLL). PATIENTS & METHODS: Eighty one CLL patients and 75 healthy subjects were enrolled and prognostic evaluation of patients was assessed. ITGA4 q-realtime PCR was performed using Applied Biosystems, TaqMan gene expression assay. ITGA4 gene-specific CpG methylation was investigated in real time using pyrosequencing technology. RESULTS: ITGA4 was differentially expressed in CLL patients. The CpG sites-1, 2 and 3 showed significantly higher mean levels than healthy controls (p = <0.001, 0.007 and 0.009). Significant association between CpG site-1 and CLL has been detected using age-adjusted logistic regression (p < 0.001). CONCLUSION: Hypermethylation at ITGA4 gene CpG sites (1,2,3) is a characteristic feature in CLL.

Osteopontin for Early Detection of Microvascular and Macrovascular Type 1 Diabetic Complication.

AIM: To evaluate the relationship between osteopontin and diabetes complication in type 1 diabetic patient. PATIENTS AND METHODS: Seventy types 1 diabetic and 60 healthy volunteers were studied. Full history, examination, laboratory tests of glycosylated haemoglobin (HbA1c), serum lipids {cholesterol, triglyceride (Tg), high-density lipoprotein-cholesterol (HDL-c), low-density lipoprotein - cholesterol (LDL-c)}, oxidised low-density lipoprotein (OxLDL), Osteopontin and urinary microalbuminuria (albumin/creatinine ratio) were done. Image study in the form of a carotid intimal medial thickness (cIMT) and aortic intimal medial thickness (aIMT), renal doppler for resistivity index was also done for all participant included in the study. RESULTS: Urinary albumin/creatinine ratio, lipid profile, osteopontin, cIMT and aIMT were higher in people with diabetes. Osteopontin was higher in people with diabetes with positive microalbuminuria and cIMT. Systolic blood pressure, microalbuminuria and cIMT had a positive correlation with osteopontin in people with diabetes. Stepwise multiple regression analysis showed that osteopontin had a significant correlation with cIMT. Receiver operating characteristic (ROC) curve showed that the cut off value of Osteopontin for detection of cIMT was > 60 with a specificity of 100% and sensitivity 80.5%, while that of albumin/creatinine ratio was > 64 with a specificity of 66.7 and sensitivity of 92.3. CONCLUSION: Osteopontin is higher in type 1 diabetics and is useful for early detection of diabetic microvascular and macrovascular complication.

Association of STAT4 rs7574865 and PTPN22 rs2476601 polymorphisms with rheumatoid arthritis and non-systemically reacting antibodies in Egyptian patients.

The aim of this study was to investigate association of protein tyrosine phosphatase non-receptor type 22 (PTPN22) rs2476601 and signal transducer and activator of transcription 4 (STAT4) rs7574865 polymorphisms with rheumatoid arthritis (RA) susceptibility and to assess potential association with the status of rheumatoid factor (RF) and anti-cyclic citrullinated peptide (anti-CCP) antibodies, serum neopterin, and disease activity. RF, anti-CCP antibodies, and neopterin were assayed in serum of 100 unrelated RA patients and 114 controls. STAT4 rs7574865 G/T and PTPN22 rs2476601 C/T polymorphisms were genotyped by the TaqMan allelic discrimination method. The frequency of STAT4 variant allele was significantly higher in RA patients than in controls (p = 0.01), while the variant allele of PTPN22 was identified in only two RA patients, in a heterozygous form and in none of control subjects. The frequency of STAT4 variant allele carrier genotypes (GT+TT) was significantly higher among RA patients than in controls (43.7 vs. 10.5%, p = 0.02) and associated with RA under additive and dominant models. The frequency of RF and anti-CCP positivity was significantly higher among RA patients carrying T allele genotypes compared to patients carrying wild genotype (P = 0.02 and 0.04, respectively). No significant associations between STAT4 variant and serum neopterin or disease activity parameters were identified. Our study confirmed the association of STAT4 rs7574865 polymorphism with RA and was the first to indicate an association with RF and anti-CCP antibodies positivity. We also found PTPN22 rs2476601 has no role in susceptibility to RA in Egyptian patients.

Genetic polymorphisms of human cytochrome P450 CYP1A1 in an Egyptian population and tobacco-induced lung cancer.

BACKGROUND: Cytochrome P450 CYP1A1 helps detoxify the potential carcinogens in tobacco smoke, it was reported that polymorphisms in the coding gene result in variation in the expression and activity levels which alter metabolism and clearance of carcinogens and therefore modify cancer risk. In this work, we aimed to identify CYP1A1 gene polymorphisms associated with lung cancer in Egyptian population and to examine the interaction effect with Tobacco smoking in modulating disease risk. METHODS: A case-control study was conducted on 150 unrelated lung cancer patients and 150 unrelated control subjects. Genomic DNA was extracted and sequencing analysis of CYP1A1 gene was performed on ABI PRISM 3100 genetic analyzer. RESULTS: Three variants in CYP1A1 gene were identified in heterozygous forms in lung cancer patients I462V, T461N and I286T. A combined variant T461N/ I462V associated with lung cancer and those who carried this variant were 2-times more likely to develop lung cancer (OR = 2.03, 95% CI = 1.81-2.29, P = 0.04), specially the non-small cell type (NSCLC) (OR = 2.20, 95% CI = 1.93-2.50, P = 0.02). Wild type was more frequent among smoker controls (83.3%) compared to smoker lung cancer patients (54.8%), P = 0.03. Association studies to examine the interaction effect of identified variants with Tobacco smoking in modulating disease risk showed no significant associations. Identified polymorphisms showed no significant implication on the stage or the prognosis of the disease. CONCLUSION: Our findings support that CYP1A1 polymorphisms play a role in the pathogenesis of lung cancer. In Egyptian population, CYP1A1 I462V, T461N and I286T variants were identified among lung cancer patients and combined T461N/ I462V was a risk variant for NSCLC in non smokers.