GFR decline predicts total mortality and mediates the effect of tryptophan metabolism on death risk in type 2 diabetes.

CONTEXT: The independent role of glomerular filtration rate (GFR) decline in shaping the risk of mortality in people with type 2 diabetes has only been partially addressed. OBJECTIVE: The objective of the study was twofold: i) to investigate the association between all-cause mortality and eGFR changes over time; ii) to understand whether renal dysfunction mediates the effect of tryptophan metabolism on death risk. DESIGN: Prospective study with an average follow-up of 14.8 years. SETTING: Research Hospital. PATIENTS: The aggregate Gargano Mortality Study included 962 patients with type 2 diabetes who had at least three eGFR recordings and at least 1.5 years of follow-up. INTERVENTIONS: This was an observational study, with no intervention. MAIN OUTCOME MEASURES: Rate of all-cause mortality. RESULTS: Age and sex adjusted annual incident rate of mortality was 2.75 events per 100 person-years. The median annual rate of decline of eGFR was 1.3 ml/min per 1.73 m2 per year (range -3.7; 7.8). The decline of kidney function was strongly and independently associated with the risk of death. Serum kynurenine-to-tryptophan ratio (KTR) was associated with both eGFR decline and all-cause mortality. Causal mediation analysis showed that 24.3% of the association between KTR and mortality was mediated by eGFR decline. CONCLUSIONS: In patients with type 2 diabetes, eGFR decline is independently associated with the risk of all-cause mortality and mediates a significant proportion of the association between tryptophan metabolism and death.

Inflammation and prediction of death in type 2 diabetes. Evidence of an intertwined link with tryptophan metabolism.

OBJECTIVE: To study whether inflammation is associated with and helps predict mortality risk in patients with type 2 diabetes. To explore the intertwined link between inflammation and tryptophan metabolism on death risk. DESIGN: Two prospective cohorts: the aggregate Gargano Mortality Study (1,731 individuals; 872 all-cause deaths) as discovery sample, the Foggia Mortality Study (490 individuals; 256 deaths) as validation sample. Twenty-seven inflammatory markers were measured. Causal mediation analysis and in vitro studies were carried out to explore the link between inflammatory markers and the kynurenine-to-tryptophan ratio (KTR) in shaping mortality risk. RESULTS: Using multivariable stepwise Cox regression analysis, IL-4, IL-6, IL-8, IL-13, RANTES and IP-10, were independently associated with death. An inflammation score (I-score) comprising these six molecules is strongly associated with death in both the discovery and the validation cohorts HR (95%CI) = 2.13 (1.91-2.37) and 2.20 (1.79-2.72), respectively. The I-score improved discrimination and reclassification measures (all P<0.01) of two mortality prediction models based on clinical variables. The causal mediation analysis showed that 28% of the KTR effect on mortality was mediated by IP-10. Studies in cultured endothelial cells showed that 5-Methoxy-tryptophan, an anti-inflammatory metabolite derived from tryptophan, reduces the expression of IP-10, thus providing a functional basis for the observed causal mediation. CONCLUSIONS: Adding the I-score to clinical prediction models may help identify individuals who are at greater risk of death. Deeply addressing the intertwined relationship between low-grade inflammation and imbalanced tryptophan metabolism in shaping mortality risk may help discover new therapies targeting patients characterized by these abnormalities.

GAME-SCORE predicts pathological and radiological response to chemotherapy in patients with colorectal liver metastases.

BACKGROUND: Genetic And Morphological Evaluation (GAME) score is the newest prognostic model for patient with colorectal liver metastases (CRLMs). Pathological and radiological responses to neoadjuvant chemotherapy (NAC) are key factors for prognostic stratification of these patients. The present study aims to evaluate the GAME-score's ability to predict pathological and radiologic responses to NAC. METHODS: CRLM patients who underwent liver resection after NAC from January 2010 to December 2021 were categorized by GAME scores: low risk (LR, 0-1), moderate risk (MR, 2-3), and high risk (HR, ≥4). Correlations between groups and radiological/pathological features were analyzed. Poor pathological response was defined as Tumor Regression Grade 4-5. RESULTS: Of 1054 liver resections for CRLMs, 448 were included. GAME scores were LR: 80 (18 %), MR: 228 (51 %), and HR: 140 (31 %). In this cohort, HR-GAME scores were associated with lower pathological response (LR: 67.1 %, MR: 74.9 %, HR: 82.6 %; p = 0.010). Radiologic progression occurred in 10 % of HR patients, significantly more than in LR (3.8 %) and MR (3.5 %) groups (p = 0.011). Multivariable analysis for independent predictors of pathological response confirmed HR-GAME (RR 1.843, p=0.025) along with age higher than 70 years (RR 2.111, p=0.022) and irinotecan-based NAC (RR 3.066, p < 0.001). For radiological progression disease after NAC, the HR-GAME score (RR 2.77, p=0.016) was the only independent predictor. HR-GAME scores were also associated with higher rates of mucinous differentiation (p = 0.021), satellitosis (p = 0.001), vascular invasion (p = 0.011), and perineural invasion (p = 0.010). CONCLUSIONS: GAME score category should be considered into planning of therapeutic strategy of patients with CRLMs.

Comparative study of 155Tb production via 155Dy precursor with p, d and α beams on natural targets for medical applications.

Three different reactions with the use of natural targets are investigated to produce 155Tb for medical applications from the decay of its precursor 155Dy. The TALYS code has been exploited to optimize the cross section description and to improve the agreement with the full set of available data. The study is completed by a theoretical model for the two radio-chemical separations: optimal solutions are presented for the production of high quality 155Tb samples, guaranteed by the absence of the main contaminant, 156Tb.

Vitamin D Supplementation in Functional Recovery of Subjects with Severe Acquired Brain Injury: A Pilot Controlled Randomized Study.

Low vitamin D (VD) has been associated with poor clinical course in several neurological diseases. Supplementation has been suggested to improve outcomes. Severe acquired brain injury (sABI) subjects have low VD levels and disabling conditions requiring rehabilitation. The aim of the present study was to evaluate if VD supplementation produced a better clinical course and a better functional outcome in sABI during rehabilitation. A randomized single-blind study was performed. sABI subjects were randomized to the VD supplementation group (VDsG) (initial dose of 50.000 UI and 1.000 daily) and usual care control group (CG). Disability Rating Scale (DRS), Glasgow Outcome Scale (GOS), and Level of Cognitive Functioning (LCF) were used in assessing disability. A total of 73 subjects (42 M and 31 F; mean age 53.2 ± 15.7) were randomized: 36 (21 M and 15 F; mean age 57.52 ± 14.88) to VDsG and 37 (20 M and 17 F; mean age 48.28 ± 17.47) to CG. Both groups significantly improved after rehabilitation, and no between-group difference was observed. The mean score values for DRS, GOS, and LCF in VDsG were 18.83 ± 4.27 and 9.42 ± 5.83; 2.89 ± 0.32 and 3.78 ± 0.80; and 4.81 ± 1.70 and 7.53 ± 1.28, at admission and discharge, respectively. Likewise, mean values for DRS, GOS, and LCF in CG were 18.57 ± 4.80 and 9.84 ± 6.34; 2.84 ± 0.37 and 3.81 ± 0.94; and 4.97 ± 2.01 and 7.41 ± 1.32, respectively. VD supplementation did not improve functional outcomes in sABI during rehabilitation treatment.

Association between R2CHA2DS2-VASc score and three-vessel coronary artery disease in a large population at high cardiovascular risk.

Coronary artery disease (CAD), particularly three-vessel coronary disease (3VD), is the main cause of death in industrialized countries. Chronic kidney disease is an independent risk factor for CAD. The CHA2DS2-VASc score shows a good ability to predict CV events in high-risk population independently from atrial fibrillation. The aim of the present study was to evaluate the association between the R2CHA2DS2-VASc score and 3VD in a population of patients at high cardiovascular risk. Monocentric prospective study evaluated 1017 patients undergoing coronary angiography. The R2CHA2DS2-VASc score was obtained by adding 2 points to the CHA2DS2-VASc score in case of eGFR < 60 ml/min/1.73m2. Coronary lesions causing ≥ 50% reduction of a major epicardial vessel diameter were considered significant. Patients were grouped based on R2CHA2DS2-VASc tertiles and according to the severity of CAD: 3VD vs No-3VD. The 3VD group showed significantly higher R2CHA2DS2-VASc score than the No-3VD group (4.20 ± 2.18 vs 3.36 ± 2.06, p < 0.001). The risk of 3VD increased by 21% for every 1-point increase in the score (OR 1.21; 95% CI 1.13-1.28, p < 0.001). The prevalence of 3VD was higher among patients belonging to higher tertiles of R2CHA2DS2-VASc (17.2% vs 26.7% vs 33.6% for first, second, and third tertile respectively, p < 0.001) with a risk more than doubled for the third tertile compared to the first one (OR 2.45; 95% CI 1.71-3.49, p < 0.001). The R2CHA2DS2-VASc score is independently associated with 3VD in patients at high cardiovascular risk. The score could be considered a useful tool for clinicians to identify patients who are at high risk of 3VD.

Efficacy of a Multistrain Synbiotic Treatment in Acute and Post-Acute COVID-19 Patients: A Double-Blind, Placebo-Controlled Randomized Trial.

BACKGROUND AND AIMS: Several studies reported the effect of COVID-19 on inducing gut dysbiosis, which is also correlated with disease severity. This study aims to investigate the effect of a nutraceutical formula on the shift of microbiota profiles and, secondly, on the clinical-pathological parameters of acute and post-acute COVID-19 patients. METHODS: In this randomised, double-blind, placebo-controlled trial conducted at National Institute for Infectious diseases (INMI) Lazzaro Spallanzani (Italy), 52 patients were randomly assigned (1:1) to receive a multistrain synbiotic formula (Kebirah®) or placebo orally for 35 days at COVID-19 diagnosis. Health professionals, investigators, and patients were masked to group assignments. The V3-V4 hypervariable region of 16S rRNA gene sequencing was employed to study the gut microbiota composition in the two groups. RESULTS: Supplementation with Kebirah® prevented the decrease in the Shannon diversity index of gut microbiota, which was instead observed in patients receiving the placebo. In addition, decreases in lymphocyte count and haemoglobin levels were observed only in the placebo group and not in the treated group, which was also characterised by an amelioration of the gut microbial profile, with an enrichment in beneficial bacteria and a preservation of species diversity. CONCLUSIONS: Our data suggest that modulating the gut microbiota in acute disease through administration of a specific symbiotic formula could be a useful strategy in the frame of SARS-CoV-2 infections.

Interplay of Dyadic Consensus, Reflective Functioning, and Perinatal Affective Difficulties in Modulating Fear of COVID-19 among First-Time Mothers: A Mediation Analysis.

BACKGROUND: The COVID-19 pandemic has exacerbated fears and anxieties, potentially influencing maternal perinatal mental health. This study addresses a gap in the literature on fear of COVID-19 in pregnant women, aiming to identify contributing factors. METHOD: Participants were 401 primiparous women with an average age of 34 years (SD = 4.56) who were recruited through the National Health System during birth support courses. They completed a series of self-reported instruments via an online survey, providing information on their levels of reflective functioning, dyadic consensus, perinatal maternal affectivity, and fear of COVID-19. Pearson's correlation and mediation analysis via a generalized linear model were implemented to analyze the collected data. RESULTS: The relationship between dyadic consensus and fear of COVID-19 was significant and negative. Furthermore, a significant parallel mediation involving perinatal maternal affectivity and reflective functioning was found to the extent that, when these factors were inserted into the model, the direct association between dyadic consensus and fear of COVID-19 became non-significant (total mediation). CONCLUSIONS: The results highlight the importance of dyadic adjustment in alleviating maternal COVID-19 fear, emphasizing interventions promoting couple functioning, mentalization, and addressing affective difficulties. Such approaches are vital for supporting expecting mothers during challenging times like the COVID-19 pandemic.

Polymer physics models reveal structural folding features of single-molecule gene chromatin conformations.

Here, we employ polymer physics models of chromatin to investigate the 3D folding of a 2Mb wide genomic region encompassing the human LTN1 gene, a crucial DNA locus involved in key cellular functions. Through extensive Molecular Dynamics simulations, we reconstruct in-silico the ensemble of single-molecule LTN1 3D structures, which we benchmark against recent in-situ Hi-C 2.0 data. The model-derived single molecules are then used to predict structural folding features at the single-cell level, providing testable predictions for super-resolution microscopy experiments.

A powerful machine learning approach to identify interactions of differentially abundant gut microbial subsets in patients with metastatic and non-metastatic pancreatic cancer.

Pancreatic cancer has a dismal prognosis, as it is often diagnosed at stage IV of the disease and is characterized by metastatic spread. Gut microbiota and its metabolites have been suggested to influence the metastatic spread by modulating the host immune system or by promoting angiogenesis. To date, the gut microbial profiles of metastatic and non-metastatic patients need to be explored. Taking advantage of the 16S metagenomic sequencing and the PEnalized LOgistic Regression Analysis (PELORA) we identified clusters of bacteria with differential abundances between metastatic and non-metastatic patients. An overall increase in Gram-negative bacteria in metastatic patients compared to non-metastatic ones was identified using this method. Furthermore, to gain more insight into how gut microbes can predict metastases, a machine learning approach (iterative Random Forest) was performed. Iterative Random Forest analysis revealed which microorganisms were characterized by a different level of relative abundance between metastatic and non-metastatic patients and established a functional relationship between the relative abundance and the probability of having metastases. At the species level, the following bacteria were found to have the highest discriminatory power: Anaerostipes hadrus, Coprobacter secundus, Clostridium sp. 619, Roseburia inulinivorans, Porphyromonas and Odoribacter at the genus level, and Rhodospirillaceae, Clostridiaceae and Peptococcaceae at the family level. Finally, these data were intertwined with those from a metabolomics analysis on fecal samples of patients with or without metastasis to better understand the role of gut microbiota in the metastatic process. Artificial intelligence has been applied in different areas of the medical field. Translating its application in the field of gut microbiota analysis may help fully exploit the potential information contained in such a large amount of data aiming to open up new supportive areas of intervention in the management of cancer.

Liver Resection vs Nonsurgical Treatments for Patients With Early Multinodular Hepatocellular Carcinoma.

Importance: The 2022 Barcelona Clinic Liver Cancer algorithm currently discourages liver resection (LR) for patients with multinodular hepatocellular carcinoma (HCC) presenting with 2 or 3 nodules that are each 3 cm or smaller. Objective: To compare the efficacy of liver resection (LR), percutaneous radiofrequency ablation (PRFA), and transarterial chemoembolization (TACE) in patients with multinodular HCC. Design, Setting, and Participants: This cohort study is a retrospective analysis conducted using data from the HE.RC.O.LE.S register (n = 5331) for LR patients and the ITA.LI.CA database (n = 7056) for PRFA and TACE patients. A matching-adjusted indirect comparison (MAIC) method was applied to balance data and potential confounding factors between the 3 groups. Included were patients from multiple centers from 2008 to 2020; data were analyzed from January to December 2023. Interventions: LR, PRFA, or TACE. Main Outcomes and Measures: Survival rates at 1, 3, and 5 years were calculated. Cox MAIC-weighted multivariable analysis and competing risk analysis were used to assess outcomes. Results: A total of 720 patients with early multinodular HCC were included, 543 males (75.4%), 177 females (24.6%), and 350 individuals older than 70 years (48.6%). There were 296 patients in the LR group, 240 who underwent PRFA, and 184 who underwent TACE. After MAIC, LR exhibited 1-, 3-, and 5-year survival rates of 89.11%, 70.98%, and 56.44%, respectively. PRFA showed rates of 94.01%, 65.20%, and 39.93%, while TACE displayed rates of 90.88%, 48.95%, and 29.24%. Multivariable Cox survival analysis in the weighted population showed a survival benefit over alternative treatments (PRFA vs LR: hazard ratio [HR], 1.41; 95% CI, 1.07-1.86; P = .01; TACE vs LR: HR, 1.86; 95% CI, 1.29-2.68; P = .001). Competing risk analysis confirmed a lower risk of cancer-related death in LR compared with PRFA and TACE. Conclusions and Relevance: For patients with early multinodular HCC who are ineligible for transplant, LR should be prioritized as the primary therapeutic option, followed by PRFA and TACE when LR is not feasible. These findings provide valuable insights for clinical decision-making in this patient population.

Multiscale modelling of chromatin 4D organization in SARS-CoV-2 infected cells.

SARS-CoV-2 can re-structure chromatin organization and alter the epigenomic landscape of the host genome, but the mechanisms that produce such changes remain unclear. Here, we use polymer physics to investigate how the chromatin of the host genome is re-organized upon infection with SARS-CoV-2. We show that re-structuring of A/B compartments can be explained by a re-modulation of intra-compartment homo-typic affinities, which leads to the weakening of A-A interactions and the enhancement of A-B mixing. At the TAD level, re-arrangements are physically described by a reduction in the loop extrusion activity coupled with an alteration of chromatin phase-separation properties, resulting in more intermingling between different TADs and a spread in space of the TADs themselves. In addition, the architecture of loci relevant to the antiviral interferon response, such as DDX58 or IFIT, becomes more variable within the 3D single-molecule population of the infected model, suggesting that viral infection leads to a loss of chromatin structural specificity. Analysing the time trajectories of pairwise gene-enhancer and higher-order contacts reveals that this variability derives from increased fluctuations in the chromatin dynamics of infected cells. This suggests that SARS-CoV-2 alters gene regulation by impacting the stability of the contact network in time.

Epidemiological analysis to identify predictors of X-linked hypophosphatemia (XLH) diagnosis in an Italian pediatric population: the EPIX project.

PURPOSE: X-linked hypophosphatemia (XLH) is a rare multi-systemic disease characterized by low plasma phosphate levels. The aim of this study was to investigate the annual XLH prevalence and internally evaluate predictive algorithms' application performance for the early diagnosis of XLH. METHODS: The PediaNet database, containing data on more than 400,000 children aged up to 14 years, was used to identify a cohort of XLH patients, which were matched with up to 10 controls by date of birth and gender. The annual prevalence of XLH cases per 100,000 patients registered in PediaNet database was estimated. To identify possible predictors associated with XLH diagnosis, a logistic regression model and two machine learning algorithms were applied. Predictive analyses were separately carried out including patients with at least 1 or 2 years of database history in PediaNet. RESULTS: Among 431,021 patients registered in the PediaNet database between 2007-2020, a total of 12 cases were identified with a mean annual prevalence of 1.78 cases per 100,000 patients registered in PediaNet database. Overall, 8 cases and 60 matched controls were included in the analysis. The random forest algorithm achieved the highest area under the receiver operating characteristic curve (AUC) value both in the one-year prior ID (AUC = 0.99, 95% CI = 0.99-1.00) and the two-year prior ID (AUC = 1.00, 95% CI = 1.00-1.00) analysis. Overall, the XLH predictors selected by the three predictive methods were: the number of vitamin D prescriptions, the number of recorded diagnoses of acute respiratory infections, the number of prescriptions of antihistamine for systemic use, the number of prescriptions of X-ray of the lower limbs and pelvis and the number of allergology visits. CONCLUSION: Findings showed that data-driven machine learning models may play a prominent role for the prediction of the diagnosis of rare diseases such as XLH.

Laparoscopic versus open parenchymal sparing liver resections for high tumour burden colorectal liver metastases: a propensity score matched analysis.

BACKGROUND: Laparoscopic liver resection (LLR) has proved effective in the treatment of oligometastatic disease (1 or 2 colorectal liver metastases CRLM) with similar long-term outcomes and improved short-term results compared to open liver resection (OLR). Feasibility of parenchymal sparing LLR for high tumour burden diseases is largely unknown. Aim of the study was to compare short and long-term results of LLR and OLR in patients with ≥ 3 CRLM. METHODS: Patients who underwent first LR of at least two different segments for ≥ 3 CRLM between 01/2012 and 12/2021 were analysed. Propensity score nearest-neighbour 1:1 matching was based on relevant prognostic factors. RESULTS: 277 out of 673 patients fulfilled inclusion criteria (47 LLR and 230 OLR). After match two balanced groups of 47 patients with a similar mean number of CRLM (5 in LLR vs 6.5 in OLR, p = 0.170) were analysed. The rate of major hepatectomy was similar between the two group (10.6% OLR vs. 12.8% LLR). Mortality (2.1% OLR vs 0 LLR) and overall morbidity rates (34% OLR vs 23.4% LLR) were comparable. Length of stay (LOS) was shorter in the LLR group (5 vs 9 days, p = 0.001). No differences were observed in median overall (41.1 months OLR vs median not reached LLR) and disease-free survival (18.3 OLR vs 27.9 months LLR). CONCLUSION: Laparoscopic approach should be considered in selected patients scheduled to parenchymal sparing LR for high tumour burden disease as associated to shorter LOS and similar postoperative and long-term outcomes compared to the open approach.

Colorectal metastases with intrabiliary growth: incidence, treatment, and outcomes.

Intrabiliary growth (IG) is an unusual modality for colorectal metastases to spread. Relatively little is known about this condition because large series are lacking. The aim of the study was to compare the surgical and oncological outcomes of patients with or without IG. From 01/2010 to 12/2020, 999 patients underwent hepatectomy for colorectal metastases. Clinicopathological variables were retrospectively analyzed from a prospective-collected database of patients with or without IG. A propensity score matched (PSM) analysis to compare OS and DFS was performed. At first hepatectomy, 29 patients (2.9%) had IG: 7 isolated IG and 22 mixed-type (mass-forming lesion with IG). 4 patients presented IG at repeat hepatectomy for recurrence, of whom 3 had no biliary invasion at initial surgery. IG resulted to be more common in older patients (median age 70 in IG vs 60 years of no-IG, p = 0.004). Mean time from colorectal tumor was longer in IG (20.4 months) than no-IG (12.9 months), p = 0.038. Major hepatectomies (55.2% IG vs 29.7% no-IG, p = 0.003) and anatomic resections (89.7% vs 58.2%, p = 0.001) were more frequently required to treat IG. In 5 (17%) of IG, a resection of main bile duct was performed. Overall postoperative mortality and complications were similar in the two groups, while bile leak was 17.2% IG vs 5.6% no-IG (p = 0.024). Median margin width was comparable in IG (1.4 mm) and no-IG (2 mm). Five-year overall survival (IG 45.9% vs no-IG 44.5%) and Disease-Free Survival (IG 35.9% vs no-IG 36.6%) were similar in the two groups. According to PSM, 145 patients with no-IG were compared to 29 of IG group. After PSM, OS and DFS did not show any statistically significant difference. IG has similar oncological outcomes of resected colorectal metastases without IG, although it affects surgical management.

A wearable system for visual cueing gait rehabilitation in Parkinson's disease: a randomized non-inferiority trial.

BACKGROUND: Gait disturbances represent one of the most disabling features of Parkinson's disease (PD). AIM: The aim of this study was to evaluate the non-inferiority of a new wearable visual cueing system (Q-Walk) for gait rehabilitation in PD subjects, compared to traditional visual cues (stripes on the floor). DESIGN: Open-label, monocentric, randomized controlled non-inferiority trial. SETTING: Outpatients. POPULATION: Patients affected by idiopathic PD without cognitive impairment, Hoehn and Yahr stage II-IV, Unified Parkinson's Disease Rating Scale motor section III ≥2, stable drug usage since at least 3 weeks. METHODS: At the enrollment (T0), all subjects underwent a clinical/functional evaluation and the instrumental gait and postural analysis; then they were randomly assigned to the Study Group (SG) or Control Group (CG). Rehabilitation program consisted in 10 consecutive individual sessions (5 sessions/week for 2 consecutive weeks). Each session included 60 minutes of conventional physiotherapy plus 30 minutes of gait training by Q-Walk (SG) or by traditional visual cues (CG). Follow-up visits were scheduled at the end of the treatment (T1) and after 3 months (T2). RESULTS: Fifty-two subjects were enrolled in the study, 26 in each group. The within-groups analysis showed a significant improvement in clinical scales and instrumental data at T1 and at T2, compared to baseline, in both groups. According to the between-group analysis, Q-Walk cueing system was not-inferior to the traditional cues for gait rehabilitation. The satisfaction questionnaire revealed that most subjects described the Q-Walk cueing system as simple, motivating and easily usable, possibly suitable for home use. CONCLUSIONS: Data showed that motor rehabilitation of PD subjects performed by means of the new wearable Q-Walk cueing system was feasible and as effective as traditional cues in improving gait parameters and balance. CLINICAL REHABILITATION IMPACT: Wearable devices can act as an additional rehabilitation strategy for long-term and continuous care, allowing patients to train intensively and extensively in household settings, favoring a tailor-made and personalized approach as well as remote monitoring.

Machine learning-based algorithms applied to drug prescriptions and other healthcare services in the Sicilian claims database to identify acromegaly as a model for the earlier diagnosis of rare diseases.

Acromegaly is a rare disease characterized by a diagnostic delay ranging from 5 to 10 years from the symptoms' onset. The aim of this study was to develop and internally validate machine-learning algorithms to identify a combination of variables for the early diagnosis of acromegaly. This retrospective population-based study was conducted between 2011 and 2018 using data from the claims databases of Sicily Region, in Southern Italy. To identify combinations of potential predictors of acromegaly diagnosis, conditional and unconditional penalized multivariable logistic regression models and three machine learning algorithms (i.e., the Recursive Partitioning and Regression Tree, the Random Forest and the Support Vector Machine) were used, and their performance was evaluated. The random forest (RF) algorithm achieved the highest Area under the ROC Curve value of 0.83 (95% CI 0.79-0.87). The sensitivity in the test set, computed at the optimal threshold of predicted probabilities, ranged from 28% for the unconditional logistic regression model to 69% for the RF. Overall, the only diagnosis predictor selected by all five models and algorithms was the number of immunosuppressants-related pharmacy claims. The other predictors selected by at least two models were eventually combined in an unconditional logistic regression to develop a meta-score that achieved an acceptable discrimination accuracy (AUC = 0.71, 95% CI 0.66-0.75). Findings of this study showed that data-driven machine learning algorithms may play a role in supporting the early diagnosis of rare diseases such as acromegaly.