2005 Fiscal Year Final Research Report Summary
category classifications for breast cancer mass screening based on image perception
| Project/Area Number |
16591200
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Radiation science
|
|---|
| Research Institution | Nagoya University |
|---|
Principal Investigator |
SHIMAMOTO Kazuhiro Nagoya University, School of Medicine, professor, 医学部, 教授 (70178961)
|
|---|
| Project Period (FY) |
2004 – 2005
|
| Keywords | Breast cancer / Category classification / Mammography / Ultrasonography, mammary / Liquid crystal display / Monitor diagnosis / Kappa statistics / Reading environment |
|---|
| Research Abstract |
Various kinds of category classifications are used at clinical practice. The assessment categories for breast cancer screening, especially in the management of Category 3, are confusing, and it should be specified what should be done for both clinicians and patients : negative, there is nothing to comment on ; initial short interval follow-up is suggested, additional examinations will be required ; biopsy should be considered ; and appropriate management should be taken. Based on the category classification for breast mass-image forming lesions described by Japan Society of Breast and Thyroid Sonology (JABTS), observer variance in category assessment was evaluated using 100 cases of breast mass lesions. Each observer had their own scale for judgment, and
The influence of ambient room lighting conditions on soft-copy breast phantom image interpretation was evaluated, comparing cathode ray tube (CRT) monitors with liquid crystal displays (LCDs). Nine observers were asked to use a three-point scale to rate the visibility of various phantom objects (masses, specks and fibers) which were displayed on a 21-inch CRT (2,560 x 2,048) and a 21-inch LCD (2,560 x 2,048) under three different levels of ambient lighting conditions (20, 100 and 420 lux at the display center). Each phantom image was interpreted twice, and reproducibility of judgment and interobserver agreement was evaluated using kappa statistics. No significant differences were found among the three lighting conditions. Furthermore, intra- and interobserver agreement in judgments showed no effects of room illumination. Although the breast phantom objects were better visualized on LCDs than on CRT monitors, room illumination did not affect the performance score of soft-copy reading.
|
