Multiplexed Serum Steroid Profiling Reveals Metabolic Signatures of Subtypes in Congenital Adrenal Hyperplasia

Context

Altered metabolic signatures on steroidogenesis may characterize individual subtypes of congenital adrenal hyperplasia (CAH), but conventional diagnostic approaches are limited to differentiate subtypes.

Objective

We explored metabolic characterizations and identified multiple diagnostic biomarkers specific to individual subtypes of CAH.

Methods

Liquid chromatography-mass spectrometry-based profiling of 33 adrenal steroids was developed and applied to serum samples obtained from 67 CAH patients and 38 healthy volunteers.

Results

Within- and between-run precisions were 95.4% to 108.3% and 94.1% to 110.0%, respectively, while all accuracies were <12% and the correlation coefficients (r2) were > 0.910. Metabolic ratios corresponding to 21-hydroxylase characterized 21-hydroxylase deficiency (21-OHD; n = 63) from healthy controls (area under the curve = 1.0, P < 1 × 10-18 for all) and other patients with CAH in addition to significantly increased serum 17α-hydroxyprogesterone (P < 1 × 10-16) and 21-deoxycortisol (P < 1 × 10-15) levels. Higher levels of mineralocorticoids, such as corticosterone (B) and 18-hydroxyB, were observed in 17α-hydroxylase deficiency (17α-OHD; N = 3), while metabolic ratio of dehydroepiandrosterone sulfate to pregnenolone sulfate was remarkably decreased against all subjects. A patient with 11β-hydroxylase deficiency (11β-OHD) demonstrated significantly elevated 11-deoxycortisol and its metabolite tetrahydroxy-11-deoxyF, with reduced metabolic ratios of 11β-hydroxytestosterone/testosterone and 11β-hydroxyandrostenedione/androstenedione. The steroid profiles resulted in significantly decreased cortisol metabolism in both 21-OHD and 17α-OHD but not in 11β-OHD.

Conclusion

The metabolic signatures with specific steroids and their corresponding metabolic ratios may reveal individual CAH subtypes. Further investigations with more substantial sample sizes should be explored to enhance the clinical validity.

Longitudinal clinical, cognitive and biomarker profiles in dominantly inherited versus sporadic early-onset Alzheimer's disease

Approximately 5% of Alzheimer's disease cases have an early age at onset (<65 years), with 5-10% of these cases attributed to dominantly inherited mutations and the remainder considered as sporadic. The extent to which dominantly inherited and sporadic early-onset Alzheimer's disease overlap is unknown. In this study, we explored the clinical, cognitive and biomarker profiles of early-onset Alzheimer's disease, focusing on commonalities and distinctions between dominantly inherited and sporadic cases. Our analysis included 117 participants with dominantly inherited Alzheimer's disease enrolled in the Dominantly Inherited Alzheimer Network and 118 individuals with sporadic early-onset Alzheimer's disease enrolled at the University of California San Francisco Alzheimer's Disease Research Center. Baseline differences in clinical and biomarker profiles between both groups were compared using t-tests. Differences in the rates of decline were compared using linear mixed-effects models. Individuals with dominantly inherited Alzheimer's disease exhibited an earlier age-at-symptom onset compared with the sporadic group [43.4 (SD ± 8.5) years versus 54.8 (SD ± 5.0) years, respectively, P < 0.001]. Sporadic cases showed a higher frequency of atypical clinical presentations relative to dominantly inherited (56.8% versus 8.5%, respectively) and a higher frequency of APOE-ε4 (50.0% versus 28.2%, P = 0.001). Compared with sporadic early onset, motor manifestations were higher in the dominantly inherited cohort [32.5% versus 16.9% at baseline (P = 0.006) and 46.1% versus 25.4% at last visit (P = 0.001)]. At baseline, the sporadic early-onset group performed worse on category fluency (P < 0.001), Trail Making Test Part B (P < 0.001) and digit span (P < 0.001). Longitudinally, both groups demonstrated similar rates of cognitive and functional decline in the early stages. After 10 years from symptom onset, dominantly inherited participants experienced a greater decline as measured by Clinical Dementia Rating Sum of Boxes [3.63 versus 1.82 points (P = 0.035)]. CSF amyloid beta-42 levels were comparable [244 (SD ± 39.3) pg/ml dominantly inherited versus 296 (SD ± 24.8) pg/ml sporadic early onset, P = 0.06]. CSF phosphorylated tau at threonine 181 levels were higher in the dominantly inherited Alzheimer's disease cohort (87.3 versus 59.7 pg/ml, P = 0.005), but no significant differences were found for t-tau levels (P = 0.35). In summary, sporadic and inherited Alzheimer's disease differed in baseline profiles; sporadic early onset is best distinguished from dominantly inherited by later age at onset, high frequency of atypical clinical presentations and worse executive performance at baseline. Despite these differences, shared pathways in longitudinal clinical decline and CSF biomarkers suggest potential common therapeutic targets for both populations, offering valuable insights for future research and clinical trial design.

Intelligent noninvasive meningioma grading with a fully automatic segmentation using interpretable multiparametric deep learning

OBJECTIVES

To establish a robust interpretable multiparametric deep learning (DL) model for automatic noninvasive grading of meningiomas along with segmentation.

METHODS

In total, 257 patients with pathologically confirmed meningiomas (162 low-grade, 95 high-grade) who underwent a preoperative brain MRI, including T2-weighted (T2) and contrast-enhanced T1-weighted images (T1C), were included in the institutional training set. A two-stage DL grading model was constructed for segmentation and classification based on multiparametric three-dimensional U-net and ResNet. The models were validated in the external validation set consisting of 61 patients with meningiomas (46 low-grade, 15 high-grade). Relevance-weighted Class Activation Mapping (RCAM) method was used to interpret the DL features contributing to the prediction of the DL grading model.

RESULTS

On external validation, the combined T1C and T2 model showed a Dice coefficient of 0.910 in segmentation and the highest performance for meningioma grading compared to the T2 or T1C only models, with an area under the curve (AUC) of 0.770 (95% confidence interval: 0.644-0.895) and accuracy, sensitivity, and specificity of 72.1%, 73.3%, and 71.7%, respectively. The AUC and accuracy of the combined DL grading model were higher than those of the human readers (AUCs of 0.675-0.690 and accuracies of 65.6-68.9%, respectively). The RCAM of the DL grading model showed activated maps at the surface regions of meningiomas indicating that the model recognized the features at the tumor margin for grading.

CONCLUSIONS

An interpretable multiparametric DL model combining T1C and T2 can enable fully automatic grading of meningiomas along with segmentation.

KEY POINTS

• The multiparametric DL model showed robustness in grading and segmentation on external validation. • The diagnostic performance of the combined DL grading model was higher than that of the human readers. • The RCAM interpreted that DL grading model recognized the meaningful features at the tumor margin for grading.