• Korean Journal of Radiology
• /
• v.24 no.2
• /
• pp.155-165
• /
• 2023

Objective: Little is known about the effects of using different expert-determined reference standards when evaluating the performance of deep learning-based automatic detection (DLAD) models and their added value to radiologists. We assessed the concordance of expert-determined standards with a clinical gold standard (herein, pathological confirmation) and the effects of different expert-determined reference standards on the estimates of radiologists' diagnostic performance to detect malignant pulmonary nodules on chest radiographs with and without the assistance of a DLAD model. Materials and Methods: This study included chest radiographs from 50 patients with pathologically proven lung cancer and 50 controls. Five expert-determined standards were constructed using the interpretations of 10 experts: individual judgment by the most experienced expert, majority vote, consensus judgments of two and three experts, and a latent class analysis (LCA) model. In separate reader tests, additional 10 radiologists independently interpreted the radiographs and then assisted with the DLAD model. Their diagnostic performance was estimated using the clinical gold standard and various expert-determined standards as the reference standard, and the results were compared using the t test with Bonferroni correction. Results: The LCA model (sensitivity, 72.6%; specificity, 100%) was most similar to the clinical gold standard. When expert-determined standards were used, the sensitivities of radiologists and DLAD model alone were overestimated, and their specificities were underestimated (all p-values < 0.05). DLAD assistance diminished the overestimation of sensitivity but exaggerated the underestimation of specificity (all p-values < 0.001). The DLAD model improved sensitivity and specificity to a greater extent when using the clinical gold standard than when using the expert-determined standards (all p-values < 0.001), except for sensitivity with the LCA model (p = 0.094). Conclusion: The LCA model was most similar to the clinical gold standard for malignant pulmonary nodule detection on chest radiographs. Expert-determined standards caused bias in measuring the diagnostic performance of the artificial intelligence model.

• Applied Microscopy
• /
• v.41 no.2
• /
• pp.123-128
• /
• 2011

The micropyles on the prefertilized and artificial fertilized eggs of Epinephelus fasciatus were investigated using scanning electron microscope (SEM). The micropyles ($6.6{\pm}0.41\;{\mu}m$) of E. fasciatus eggs were found in the animal pole and theirs shape were observed as a flat crateriform of cylindrical shape. The micropylar vestibule arranged by 6~7 thickened spiral annuli on the ridge and contributed to differentiate and form fertilization cone for blocking to polyspermy by presenting swollen vestibule structure. As E. fasciatus eggs was pelagic, so chorionic surface was an uneven structures such as circular and fillar form nodules. Especially, various pores (0.15~0.55 ${\mu}m$, 230~270 pores) distributed at the only around micropyles, those pores radiately exhibited regular projection structures showing gill filament-shape. These ultrastructural characters of E. fasciatus eggs can be utilized in a taxonomical cue of grouper species.

• Journal of the Korean Society of Radiology
• /
• v.12 no.1
• /
• pp.23-30
• /
• 2018

To evaluate the effect of patient size on effective dose and image quality for Digital Chest Tomosynthesis(DTS) using additional 0.3 mm copper filtration. Eighty artificial nodules were placed in the thorax phantom("Lungman," Kyoto Kagaku, Japan), and Digital Chest Tomosynthesis(DTS) images of the phantom were acquired both with and without added 0.3 mm Cu filtration. To simulate patients of three sizes: small, average size and oversize, one or two 20-mm-thick layer of PMMA(polymethyl methacrylatek) blocks were placed on the phantom. The Effective dose was calculated using Monte Carlo simulations. Two evaluations of image quality methods have been employed. Three readers counted the number of nodules detected in the lung, and the measured contrast-to-noise ratios(CNRs) were used. Data were analyzed statistically. The ED reduced $26{\mu}Sv$ in a phantom, $33{\mu}Sv$ in one 20-mm-thick layer of PMMA block placed on the phantom, and $48{\mu}Sv$ in two 20-mm-thick layer of PMMA blocks placed on the phantom. The Effective dose(ED) differences between DTS with and without filtration were significant(p<0.05). In particular, when we used two 20-mm-thick layer of PMMA blocks placed on the phantom, the ED was significantly reduced by 36% compared with those without additional filtration. Nodule detection sensitivities were not different between with and without added filtration. Differences of CNRs were statistically insignificant(p>0.05). Use of additional filtration allows a considerable dose reduction during Digital Chest Tomosynthesis(DTS) without loss of image quality. In particular, additional filtration showed outstanding result for effective dose reduction on two 20-mm-thick layer of PMMA blocks placed on the phantom. It applies to overweight patients.

• KOREAN JOURNAL OF CROP SCIENCE
• /
• v.38 no.4
• /
• pp.336-342
• /
• 1993

To obtain the basic information necessary for the development of soybean varieties well adaptable to upland-paddy field rotational croppings, the difference of growth characteristics between upland and paddy-field including yield potentials of current recommending soybean varieties were evaluated. The growth characteristics, both above and under-ground, which were measured at flowering stage were generally greater in paddy-field but the number of root nodules was much greater in upland, thus the artificial inoculation was practically recommended for soybean growing in paddy-fields. Mean seed yield was generally higher in paddy-fields than in upland. All soybean varieties showed higher seed yield in the early planting date, April 20, were somewhat susceptible to soybean mosaic virus(SMV), thus they could be escaped from the disasterous endemic necrotic soybean mosaic virus(SMV-N). Soybean varieties showed over 4.0 tons/ha seed yield in the paddy-field were Williams 79, Union, SS77053, and Namhaekong. At the same time, Jangyeobkong and Danyeobkong were the most stable soybean varieties among the tested soybean varieties with less than 10% of coefficient variation values in all planting dates in paddy fields. Compared with Hwangkeumkong which is most widely being cultivated on farmer's fields, soybean varieties showed high yields in paddy-field were higher in plant height, less in the number of branches, and more in the number of nodes on main stem. At the same time, they had medium seed size which would bring the good germination and stands. Disease resistance especially for necrotic soybean mosaic virus was also one of the most decisive factors in seed yields for the early planted soybeans.