HealthDay News - New procedure has good sensitivity, but high number of false positives leads to low specificity
HealthDay News - Optical coherence tomography (OCT), a new imaging technology that allows noninvasive cross-sectional imaging, has high sensitivity for diagnosing cervical cancer, but low specificity, according to a study published in the March issue of Lasers in Surgery and Medicine.
Julia K.S. Gallwas, MD, of Ludwig Maximilian University Munich in Germany, and colleagues investigated the accuracy and reproducibility of OCT in characterizing and grading cervical intraepithelial neoplasia (CIN). OCT images were taken from suspicious and unsuspicious areas in 120 women undergoing colposcopy for suspected CIN. Each woman also underwent biopsy. The OCT images were evaluated separately by two blinded investigators and then compared with the actual histology based on standard grading: normal, inflammation, CIN1, CIN2, CIN3, squamous carcinoma. Sensitivity and specificity of OCT in detecting CIN, as well as interobserver agreement, were measured.
The researchers found that, depending on the chosen threshold, the sensitivity of OCT compared well with the expert colposcopy results. With the threshold at CIN1, the sensitivity was 98 and 96 percent for both investigators, respectively, and the specificity was 39 and 41 percent, respectively, due to a large number of false positive results. With the threshold at CIN2, sensitivity was 86 and 94 percent and specificity increased to 64 and 60 percent, respectively. Interobserver agreement measures were high.
"OCT is highly sensitive in identifying pre-invasive and invasive cancer of the uterine cervix," the authors write. "At present, OCT is not capable to replace colposcopy but it can be used as an adjunct either to guide biopsies or to define the extent of the lesions."
One of the study authors disclosed a financial relationship with Imalux, which provided equipment for the study.
AbstractFull Text (subscription or payment may be required)
Copyright © 2011 HealthDay. All rights reserved.
Can AI Enhance PET/MRI Assessment for Extraprostatic Tumor Extension in Patients with PCa?
December 17th 2024The use of an adjunctive machine learning model led to 17 and 21 percent improvements in the AUC and sensitivity rate, respectively, for PET/MRI in diagnosing extraprostatic tumor extension in patients with primary prostate cancer.
Can Radiomics Bolster Low-Dose CT Prognostic Assessment for High-Risk Lung Adenocarcinoma?
December 16th 2024A CT-based radiomic model offered over 10 percent higher specificity and positive predictive value for high-risk lung adenocarcinoma in comparison to a radiographic model, according to external validation testing in a recent study.