External validation of the paediatric risk of malignancy index

Objective This study aimed to validate the paediatric risk of malignancy index (PRMI), as previously published. Design External validation study. Setting Academic hospital: Radboud University Medical Center. Population Female paediatric patients under the age of 18 years diagnosed with, or treated f...

Full description

Saved in:
Bibliographic Details
Published in:BJOG : an international journal of obstetrics and gynaecology 2016-02, Vol.123 (3), p.448-452
Main Authors: Hermans, AJ, Kluivers, KB, Massuger, LF, Coppus, SF
Format: Article
Language:English
Subjects:
Adnexa Uteri
Adolescent
Adolescents
Cancer
Child
Child, Preschool
children
external validation
Female
Genital Neoplasms, Female - epidemiology
Gynecology
Humans
Infant
Infant, Newborn
Ovarian cancer
ovarian mass
ovary
Pediatrics
Retrospective Studies
Risk Assessment
risk prediction model
Tumors
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Objective This study aimed to validate the paediatric risk of malignancy index (PRMI), as previously published. Design External validation study. Setting Academic hospital: Radboud University Medical Center. Population Female paediatric patients under the age of 18 years diagnosed with, or treated for, an adnexal mass between January 1999 and October 2013. Methods Information was collected on diagnosis, presenting symptoms, and signs and imaging characteristics. The PRMI was calculated for each patient. Sensitivity, specificity, and positive and negative predictive values were calculated, and the results were visualised using a receiver operating characteristic curve (ROC curve). Main outcome measures Histological diagnosis, discriminative performance using the area under the curve (AUC) of the ROC curve and sensitivity and specificity. Results Seventy‐eight patients were included, with a median age of 12 years. A malignant mass was found in 17 patients (21.8%). The PRMI with a cut‐off value of 7 resulted in a sensitivity of 70.1% (95% CI 44.1–89.6%) and a specificity of 85.3% (95% CI 73.8–93.0%). The area under the ROC curve was 0.868 (95% CI 0.756–0.980). Conclusions The PRMI showed less discriminative capacity than originally published, but its performance was still good; however, further prospective validation studies are needed to define whether the model is useful in daily clinical practice.
ISSN:1470-0328
1471-0528
DOI:10.1111/1471-0528.13253