Development and Validation of Clinical Diagnostic Model for Girls with Central Precocious Puberty: Machine-learning Approaches

Abstract

Background A brief gonadotropin-releasing hormone analogues (GnRHa) stimulation test which solely focused on LH 30-minute post-stimulation was considered to identify girls with central precocious puberty (CPP). However, it was tested using traditional statistical methods. With advanced computer science, we aimed to develop a machine learning-based diagnostic model that processed baseline CPP-related variables and a brief GnRHa stimulation test for CPP diagnosis. Methods We recruited girls suspected of precocious puberty and underwent a GnRHa stimulation test at Children Hospital 2, Vietnam, and Cathay General Hospital, Taiwan. Clinical data, bone age measurement, and 30-min post-stimulation blood test were used to build up the predictive model. The candidate model was developed by different machine learning algorithms that were mainly evaluated by sensitivity, specificity, the area under the receiver operator characteristic curve (AUC), and F1-score in internal and external validation data to classify girls as CPP and non-CPP at different time-points (0-min, 30-min, 60-min, and 120-min post-stimulation). Results Among the 614 girls diagnosed with PP, 524 (85.3%) had CPP. The random forest algorithm yielded the highest value of F1-score (0.976), specificity (0.893), positive predicted value (0.987), and relatively high value of AUC (0.972) that contributed to high probability to identify CPP. The performance metrics of the 30-min post-stimulation diagnostic model including sensitivity and specificity surpassed those of the 0-minute model (0-min) and were equivalent to those of the model obtained 60-min and 120-min post-stimulation. Hence, our machine learning-based model helps shorten the stimulation test to 30 minutes after GnRHa injection, in general, it requires 120 minutes for a completed GnRHa stimulation test. Conclusions We developed a diagnostic model based on clinical features and a single sample 30-minute post-stimulation to identify CPP in girls that can reduce distress for children caused by multiple blood samplings.

Development of a disease diagnostic model to predict the occurrence of central precocious puberty of female

“…To date, there is an increasing emphasis on basal sex hormone levels for diagnosing CPP because of the insurmountable drawbacks of the nonphysiological GnRH stimulation test, which does not truly re ect the level of gonadal development and requires multiple blood samples. Huynh [7] simpli ed the procedure of the GnRH stimulation test, reducing the duration and number of blood collections, however, it still used the method of stimulation test. Given that the GnRH stimulation test is cumbersome to perform and that overdiagnosis of CPP due to false positives after the GnRH stimulation test has been documented in the infant population[8], the possibility of replacing this test with a simpli ed evaluation panel including basal laboratory hormonal values, such as LH, and pelvic ultrasonography, which is noninvasive and relatively easy to perform, has been continuously reviewed over the years [9][10][11][12][13][14][15][16][17][18][19][20][21][22][23].…”

Development and Validation of a Combined MRI Radiomics, Imaging and Clinical Parameter‐Based Machine Learning Model for Identifying Idiopathic Central Precocious Puberty in Girls

“…Several studies have investigated whether only basic clinical information and various baseline sex hormones can be used to discriminate between ICPP and PPP, but none have been able to propose a reliable model that can be used for diagnosis. 18,20,21 An exploratory study has investigated the use of pelvic ultrasound to identify PP and premature thelarche (PT) and found that increased uterine and ovarian measurements may be an early and sensitive sign of PP. 21 One study developed a clinical risk model by ultrasound to assess the risk of CPP in girls with a model consistency index (C-index) of 0.85 in the training group.…”

“…It has been suggested that baseline LH is sufficient as a negative diagnostic power for ICPP, but a stable positive diagnostic power is lacking 31 . Moreover, a recent study used a random forest algorithm to build a classification model which included baseline clinical data, baseline sex hormones, and LH values at 30 minutes of a GnRH stimulation test which resulted in an excellent AUC (0.972) 18 . However, that study still adopted the reference standard GnRH stimulation test, and while the results were better than those in our study, patients still required multiple blood sampling, which was inconsistent with the goals of our study.…”

Imaging of Metabolic Bone Diseases: The Spine View, Part II

“…This can either lead to arrested and delayed growth or accelerated fusion of the physis and premature skeletal development, thereby depicting precocious puberty and referring patients to specialists early on in their disease. [48][49][50] Reduced variability, bone age accuracy, and rootmean-squared error improves after implementing AI-enabled diagnosis of radiologists when compared with AI alone, a radiologist alone, or a pooled cohort of experts. 50,51 AI applications demonstrate a promising role in detecting and monitoring metabolic bone diseases.…”