Diagnostic efficacy of computer extracted image features in optical coherence tomography of the precancerous cervix.

PURPOSE: To determine the diagnostic efficacy of optical coherence tomography (OCT) to identify cervical intraepithelial neoplasia (CIN) grade 2 or higher by computer-aided diagnosis (CADx). METHODS: OCT has been investigated as a screening/diagnostic tool in the management of preinvasive and early invasive cancers of the uterine cervix. In this study, an automated algorithm was developed to extract OCT image features and identify CIN 2 or higher. First, the cervical epithelium was detected by a combined watershed and active contour method. Second, four features were calculated: The thickness of the epithelium and its standard deviation and the contrast between the epithelium and the stroma and its standard deviation. Finally, linear discriminant analysis was applied to classify images into two categories: Normal/inflammation/CIN 1 and CIN 2/CIN 3. The algorithm was applied to 152 images (74 patients) obtained from an international study. RESULTS: The numbers of normal/inflammatory/CIN 1/CIN 2/CIN 3 images are 74, 29, 14, 24, and 11, respectively. Tenfold cross-validation predicted the algorithm achieved a sensitivity of 51% (95% CI: 36%-67%) and a specificity of 92% (95% CI: 86%-96%) with an empirical two-category prior probability estimated from the data set. Receiver operating characteristic analysis yielded an area under the curve of 0.86. CONCLUSIONS: The diagnostic efficacy of CADx in OCT imaging to differentiate high-grade CIN from normal/low grade CIN is demonstrated. The high specificity of OCT with CADx suggests further investigation as an effective secondary screening tool when combined with a highly sensitive primary screening tool.

Study of the diagnostic efficacy of real-time optical coherence tomography as an adjunct to unaided visual inspection with acetic acid for the diagnosis of preinvasive and invasive neoplasia of the uterine cervix.

OBJECTIVES: To determine the sensitivity and specificity of optical coherence tomography (OCT) as an adjunct to unaided visual inspection using acetic acid (VIA) in the detection of cervical intraepithelial neoplasia 2 (CIN 2) in a real-time clinical evaluation. BACKGROUND: This clinical study was a prospective cross-sectional comparative trial that screened 1000 patients (aged 30-50 years) in a low-resource setting. Women with abnormal cervical cytology or positive human papillomavirus (HPV) tests were referred for further evaluation including VIA, OCT imaging, colposcopy, and cervical biopsies. METHODS: The VIA diagnoses were coded by quadrant. The OCT was then performed in all VIA-positive areas and at the squamocolumnar junction in all 4 quadrants. All patients were colposcoped; assessed by quadrant with biopsies at 2, 4, 8, and 10 o'clock; all abnormal areas were biopsied; and endocervical curettage was performed. Data were analyzed using generalized estimating equations and logistic regression. RESULTS: Of the 1000 patients, 175 (17.5%) were HPV positive, 93 (9.3%) had abnormal cervical cytology greater than or equal to atypical squamous cells of undetermined significance, and 211 (21.1%) were either HPV positive or had abnormal cytology. The VIA, OCT, colposcopy, and biopsies were completed on 183 (86.7%) of 211 women. For VIA alone, the sensitivity and specificity in detecting lesions greater than or equal to CIN 2 was 43% and 96%. With the addition of OCT, the sensitivity increases to 62% with a specificity of 80%. CONCLUSIONS: With the addition of OCT, the sensitivity of VIA increased in all analyses for the detection of greater than or equal to CIN II, with a loss in specificity. We hope that the potential of this technology will be realized when a computer algorithm is generated to aid in image interpretation.

Diagnostic efficacy of real-time optical coherence tomography in the management of preinvasive and invasive neoplasia of the uterine cervix.

OBJECTIVE: Determine the sensitivity and specificity of optical coherence tomography (OCT) as an adjunct to colposcopy in the detection of cervical intraepithelial neoplasia (CIN) grade 2 or higher in a real-time clinical evaluation. BACKGROUND: Optical coherence tomography (OCT) uses infrared light similar to ultrasound pulse-echo imaging. Image resolution is optimal in the 1-to-3-mm range. This study is the third in our series of OCT investigations and our first real-time clinical trial. The study was conducted at the Peking University Shenzhen Hospital, Shenzhen, China. METHODS: Nonpregnant women 18 years or older with abnormal cervical cytologic findings or a positive high-risk human papillomavirus test result were recruited. Women were assessed; and diagnoses, recorded by cervical quadrant first with colposcopy, followed by colposcopic directed OCT. A biopsy of the abnormal areas was performed. In normal quadrants, biopsy specimens were obtained at the 2-, 4-, 8-, and 10-o'clock positions at the squamocolumnar junction depending on the quadrant. An endocervical curettage was also done. Individual OCT diagnoses were paired with colposcopic impressions and biopsy specimens to assess its role as a paired secondary screen. Data were analyzed using generalized estimating equations to control for correlation within a woman. RESULTS: One thousand two hundred thirty-seven paired diagnoses from 299 women were analyzed. Median age was 36 years. Ninety-six women (8%) had a diagnosis of CIN 2 or higher. Evaluation by quadrant showed that the sensitivity for CIN 2 or higher decreased by adding OCT to colposcopy, but the specificity increased from 83% to 93%. CONCLUSIONS: We continue to try to improve sensitivity by improving the near-infrared light source, decreasing the scan time to 8 frames per second, and using a larger diameter (5 mm) fiberoptic probe with a newly designed application specific probe sheath.

Optical coherence tomography as an adjunct to white light cystoscopy for intravesical real-time imaging and staging of bladder cancer.

OBJECTIVES: Optical coherence tomography (OCT) is a novel, real-time endoscopic imaging modality that permits delineation of microarchitectural features of bladder lesions. It may provide an extension of conventional cystoscopy by allowing noninvasive examination of bladder tissue at microscopic resolution (10 to 20 microm). The purpose of this study was to examine the application of OCT in augmenting the diagnosis and staging of bladder lesions. METHODS: We conducted a retrospective institutional review board-approved, single-institution, single-user review on the use of OCT as an adjunct to conventional cystoscopy in 32 patients with a history of bladder cancer (24), primary tumor (6), prostate cancer (1), or hematuria (1). We obtained OCT images of suspicious areas before biopsy or resection, interpreted them in real time, and subsequently compared them with pathology results. RESULTS: We obtained 94 images in 32 patients undergoing bladder biopsy or transurethral resection of bladder tumor. Age of the patients ranged from 49 to 84 years (mean, 59 years), with 25 men (78%) and 7 women (22%). We correlated 38 suspicious areas with biopsy findings. OCT imaging correctly identified tumors confined to the mucosa with a sensitivity and specificity of 90% and 89%, respectively. Muscle-invasive tumors were detected in 7 of 7 lesions with 100% sensitivity, 90% specificity, and 92% accuracy. CONCLUSIONS: Optical coherence tomography is a rapid, easy-to-use tool that can help differentiate Ta and T1 tumors and identify muscle-invasive bladder tumors. It provides real-time microarchitectural information that can aid in the evaluation of bladder tumors and adjacent and remote urothelium.