Radiomic features are being increasingly studied for clinical applications. We aimed to assess the agreement among radiomic features when computed by several groups by using different software packages under very tightly controlled conditions, which included standardized feature definitions and common image data sets. Ten sites (9 from the NCI's Quantitative Imaging Network] positron emission tomography–computed tomography working group plus one site from outside that group) participated in this project. Nine common quantitative imaging features were selected for comparison including features that describe morphology, intensity, shape, and texture. The common image data sets were: three 3D digital reference objects (DROs) and 10 patient image scans from the Lung Image Database Consortium data set using a specific lesion in each scan. Each object (DRO or lesion) was accompanied by an already-defined volume of interest, from which the features were calculated. Feature values for each object (DRO or lesion) were reported. The coefficient of variation (CV), expressed as a percentage, was calculated across software packages for each feature on each object. Thirteen sets of results were obtained for the DROs and patient data sets. Five of the 9 features showed excellent agreement with CV < 1%; 1 feature had moderate agreement (CV < 10%), and 3 features had larger variations (CV ≥ 10%) even after attempts at harmonization of feature calculations. This work highlights the value of feature definition standardization as well as the need to further clarify definitions for some features.

PurposeTo assess the interdevice agreement between swept-source Fourier-domain and time-domain anterior segment optical coherence tomography (AS-OCT).MethodsFifty-three eyes from 41 subjects underwent CASIA2 and Visante OCT imaging. One hundred eighty–degree axis images were measured with the built-in two-dimensional analysis software for the swept-source Fourier-domain AS-OCT (CASIA2) and a customized program for the time-domain AS-OCT (Visante OCT). In both devices, we examined the angle opening distance (AOD), trabecular iris space area (TISA), angle recess area (ARA), anterior chamber depth (ACD), anterior chamber width (ACW), and lens vault (LV). Bland-Altman plots and intraclass correlation (ICC) were performed. Orthogonal linear regression assessed any proportional bias.ResultsICC showed strong correlation for LV (0.925) and ACD (0.992) and moderate agreement for ACW (0.801). ICC suggested good agreement for all angle parameters (0.771–0.878) except temporal AOD500 (0.743) and ARA750 (nasal 0.481; temporal 0.481). There was a proportional bias in nasal ARA750 (slope 2.44, 95% confidence interval [CI]: 1.95–3.18), temporal ARA750 (slope 2.57, 95% CI: 2.04–3.40), and nasal TISA500 (slope 1.30, 95% CI: 1.12–1.54). Bland-Altman plots demonstrated in all measured parameters a minimal mean difference between the two devices (−0.089 to 0.063); however, evidence of constant bias was found in nasal AOD250, nasal AOD500, nasal AOD750, nasal ARA750, temporal AOD500, temporal AOD750, temporal ARA750, and ACD. Among the parameters with constant biases, CASIA2 tends to give the larger numbers.ConclusionsBoth devices had generally good agreement. However, there were proportional and constant biases in most angle parameters. Thus, it is not recommended that values be used interchangeably.

Objective: To compare the central corneal thickness (CCT) between different ethnicities and particularly Asian subethnic groups that may contribute to the different glaucoma diagnoses using the optical low-coherence reflectometry technique. Methods: A retrospective study of 6 years including 1512 eyes of 929 patients of the Beckman vision center, University of California, San Francisco from 2011 to 2017 had their biometric parameters, including CCT, measured with the Lenstar. Patients were categorized into African Americans, Caucasians, Hispanics, Pacific Islanders, and Asians. Asians were further subcategorized into Chinese, Vietnamese, Koreans, Filipinos, and Japanese. Results: Among 1356 patients who had their CCT measured by Lenstar from 2011 to 2017, 1512 eyes of 929 patients were included. The study population included 462 Caucasians (52.96%), 60 African Americans (6.46%), 92 Hispanics (9.9%), 32 Pacific Islanders (3.44%), 130 Chinese (13.99%), 52 Filipinos (5.6%), 37 Vietnamese (3.98%), 34 Koreans (3.66%), and 30 Japanese (3.23%). African Americans had the thinnest CCT with a mean of 518.62±40.3 followed by Asians with a mean of 539.29±34.1. Among the Asian study sample, the Chinese had the thinnest CCT with a mean of 537.66±32.5. CCT was adjusted for age, sex, glaucoma diagnosis, diabetes status, and prostaglandin analogs use for >12 months. Conclusions: Optical low-coherence reflectometry is a widely used technology, which can measure CCT. Our study confirms that African Americans have the thinnest corneas followed by Asians. In the latter group, relatively thin CCT may partly explain their high rates of normal-tension glaucoma.