ObjectiveTo assess whether a DFUC2022-derived soft ulcer-region prior improves four-class infection/ischaemia label prediction in ulcer-centred DFUC2021 patches, using comparators matched for classifier architecture or base-model outputs.MethodsWe analysed 5, 955 labelled patches using five fixed image-similarity-grouped outer folds and five training seeds. CNN checkpoint selection and stacker fitting and selection excluded each outer evaluation fold. Three incremental comparisons assessed Full lesion-prior transfer (LPT) against an architecture-matched neutral-prior control, reliability-aware LPT (RA-LPT) against its output-only stack, and strong RGB stacks with versus without explicit prior features. Pointwise 95% confidence intervals and approximate two-sided p values were obtained from 5, 000 paired crossed seed × similarity-group bootstrap resamples; p values were Holm-adjusted across the three comparisons.ResultsFull LPT achieved Macro-F1 of 0.8921 ± 0.0059 versus 0.8777 ± 0.0085 for neutral Full LPT (mean ± sample SD across seeds). The mean paired difference was +0.0144 (95% CI, + 0.0048 to +0.0244; Holm-adjusted approximate p = 0.0132), positive in all five seeds. RA-LPT achieved Macro-F1 of 0.9181 ± 0.0038 versus 0.9206 ± 0.0028 for its corresponding output-only stack, without demonstrating an incremental improvement (difference, −0.0026; 95% CI, −0.0066 to +0.0012; adjusted p = 0.3079). Adding explicit prior features to the strong RGB stack also did not demonstrate an improvement in Macro-F1 (−0.0016; 95% CI, −0.0057 to +0.0025; adjusted p = 0.4223). The strong RGB stack achieved 0.9370 ± 0.0050, above Full LPT. In a later exploratory training-size analysis, Full LPT and its neutral control improved with increasing labelled-training budgets; mean Full LPT-minus-neutral Macro-F1 differences were positive at all four budgets, but all pointwise 95% confidence intervals included zero.ConclusionIn the main five-seed evaluation, sample-specific prior information improved Full LPT relative to its architecture-matched neutral control. Its incremental value was comparator-dependent: additional explicit prior features showed no confirmed improvement beyond the evaluated model-output stacks. Patient identifiers and target-domain lesion masks were unavailable; these internal development findings require patient-linked, multicentre external validation before clinical use.
The 2026 American Thyroid Association guidelines no longer use thyroid-antibody status to guide levothyroxine treatment in subclinical hypothyroidism (SCH), yet the relevance of SCH and thyroid autoimmunity to obstetric outcomes beyond early pregnancy has rarely been examined at population scale. We evaluated the association between maternal thyroid function, thyroid autoimmunity, and a composite of late obstetric complications in a population-based cohort. We conducted a retrospective, population-based cohort study of 23,499 singleton pregnancies among Clalit Health Services members in Israel (2013–2022) with at least one gestational thyroid-stimulating hormone (TSH) measurement and documented thyroid-autoantibody testing; pregnancies with persistently suppressed TSH (maximum <0.4 mIU/L) were excluded. Pregnancies were classified by gestational TSH (euthyroid, maximum 0.4–4 mIU/L; SCH, mean ≥4 and maximum <10 mIU/L) and antibody status, yielding four groups, and assessed for a prespecified composite of 14 obstetric complications. Inverse-probability weighting (IPW) balanced baseline covariates, and rate ratios (RRs; ratio of mean complication counts) were estimated using IPW-weighted quasi-Poisson regression; “any complication” was analyzed as a secondary binary outcome. After weighting, neither SCH nor thyroid autoimmunity was associated with an increased rate of composite complications. Antibody-positive euthyroid women had a modestly lower composite rate than antibody-negative euthyroid women (RR 0.92; 95% CI 0.88-0.96), as did SCH antibody-positive women versus euthyroid antibody-negative women (RR 0.93; 95% CI 0.87-0.98), although the latter attenuated after additional adjustment for parity and obstetric history. Within SCH, antibody status was not associated with the composite (RR 0.99; 95% CI 0.88-1.11), a finding stable after adjustment for recorded levothyroxine treatment, and among antibody-negative women SCH was not associated with an increased composite rate (RR 0.93; 95% CI 0.87-0.99). Healthcare-utilization patterns were consistent with, but did not establish, a surveillance effect. In this clinically detected and managed cohort, we did not observe an increased burden of late composite obstetric complications associated with SCH or thyroid autoimmunity; the modestly lower rates observed in some comparisons may reflect surveillance, survivor effects or residual confounding. These findings are consistent with current TSH-guided, guideline-based management rather than antibody-driven treatment decisions.
BackgroundThis study aimed to develop and validate machine learning (ML) models for diagnosing endometriosis based on common clinical characteristics, aiming to offer guidance for clinical screening and early diagnosis of endometriosis.MethodsWe retrospectively collected data on female patients aged 18 years and older from a tertiary hospital in China. The data were randomly split into a training set and a validation set in a 7:3 ratio. Patients were divided into an endometriosis group and a non-endometriosis group. In the training set, nivariable and multivariable logistic regression analyses were performed to identify independent risk factors for endometriosis. (logistic regression [LR, a traditional statistical inference model with a linear baseline], random forest [RF, bagging ensemble of decision trees], support vector machine [SVM, radial-basis-function kernel], and XGBoost [gradient-boosted trees]) were constructed. These four algorithms were selected because they cover complementary learning paradigms, encompassing interpretable linear models and nonlinear ensemble methods.The optimal Youden index was used to select the cutoff probability, thereby determining the sensitivity and specificity of each model.ResultsA total of 233 patients were included. Among them, 120 patients had endometriosis. Our analysis showed that a history of uterine fibroids, menopausal status, dysmenorrhea, white blood cell count, CA125, and CA199 were independent influencing factors for endometriosis. Results revealed that in the validation set, the AUC, sensitivity, and specificity were 0.878 (95% confidence interval [CI]: 0.793–0.962), 0.944, and 0.706 for LR; 0.878 (95% CI: 0.795–0.960), 0.694, and 0.912 for RF; 0.853 (95% CI: 0.764–0.941), 0.667, and 0.853 for SVM; 0.860 (95% CI: 0.777–0.943), 0.722, and 0.794 for XGBoost. The detection performance of the LR model was not significantly inferior to that of the other models.ConclusionBuilding ML models to detect endometriosis based on basic demographic information, clinical characteristics, and laboratory indicators is a feasible approach with desirable performance. The sample size in this study was limited, and internal validation was performed using a random split of the dataset. Therefore, caution is still warranted regarding the validation performance and interpretation of this model.
The diagnosis of rare bone diseases (RBDs) is challenged by their heterogeneity and rare occurrence. However, correct and timely diagnosis is highly relevant for the best therapeutic outcome of patients. We regarded two AI-approaches developed by the Bone2Gene project that aim to support the diagnostic process by assigning hand radiographs to RBDs. We noticed a predominant misclassification of several RBDs into the group of SHOX-related short stature. To find potential genetic explanations for this overlap, we performed a literature research for SHOX-related signaling in endochondral ossification and connections to the signaling underlying other RBDs. We found that SHOX activates the expression of the structural protein aggrecan (ACAN) and represses FGFR3-induced RAS/RAF/MAPK signaling, displaying signaling connections to ACAN-related growth disorder as well as Hypochondroplasia and Achondroplasia, respectively. Additionally, we conclude that the high phenotypic variety of SHOX-related short stature also contributes to the misclassification of RBDs that do not show a direct connection to SHOX-related signaling. Overall, we deepened our understanding of the occurrence of misclassification in both Bone2Gene approaches. Complemented with genetic testing, Bone2Gene can be a valuable tool in the diagnostic process of RBDs.
The tripeptide Arg-Gly-Asp (RGD) is a minimal integrin-binding motif found in multiple extracellular matrix (ECM) proteins. We previously determined that immature ovarian fragments cultured in the RGD-immobilized dextran hydrogel significantly promoted antral follicle development compared to those cultured in the dextran hydrogel control. In this study, using neonatal mouse ovaries, we examined the effects of RGD peptide and ovarian fragment size on early-stage follicle development in our 3D culture system. One neonatal ovary was cut into 1/2, 1/4, or 1/8 fragments and cultured in the dextran hydrogel with or without RGD. Follicle development progressively increased as fragment size decreased (intact < 1/2 < 1/4 < 1/8). Notably, RGD treatment further enhanced follicle development in a size-dependent manner. In particular, the RGD-treated 1/8 fragments significantly promoted follicle development compared to all other groups with or without RGD. Previous studies have shown that ovarian fragmentation stimulates early-stage follicle development through disruption of the Hippo pathway. We found that the RGD treatment further increased the Hippo pathway disruption in the 1/8 fragments compared to those cultured without RGD. As a ligand, RGD binds to integrin receptors to induce various cell functions. In the neonatal ovary, ovary-specific RGD-binding integrins αvβ3 and αvβ5 were predominantly localized in the ovarian interstitial/theca cells. To evaluate the role of the RGD-integrin interaction, we blocked integrins αvβ3 and αvβ5 using Cilengitide (Ci). In the presence of Ci, RGD-induced follicle development and the Hippo pathway disruption were completely suppressed in the 1/8 fragments in 3D culture. These results suggest that the RGD-integrin interaction is crucial for promoting early-stage follicle development in 3D culture. Overall, our study demonstrates that ovarian fragment size significantly influences early-stage follicle development in culture. Fragmentation exposes the ovarian stromal compartment expressing RGD-binding integrins to the surrounding hydrogel. Moreover, the effect of RGD treatment on follicle development depends on ovarian fragment size in our 3D culture system, with 1/8 fragments yielding the greatest enhancement.
