PURPOSE: This study was performed to evaluate the usefulness of compression of radiologic images by using personal computer(PC) and the clinically acceptable compression ratio for personal archiving. MATERIALS AND METHODS: The 300 images of total 60 cases including 10 cases each of plain radiograph, angiography, barium study, US, CT, and MRI were obtained from video camera recorder and were digitalized at image board of PC with 628 x430 matrix and archived to 8 bit gray scale for BMP, TIFF, PCX, GIF, 16 bit color for TGA, and 8 bit gray scale and 24 bit color for JPEG. We compared the image quality on each of the 6 image file formats and various compression ratios by using a ROC(receiver operating characteristic) analysis. RESULTS: The file size order was TGA, BMP, PCX, TIFF, GIF, and the smallest JPEG. The excution time for compression and decompression of images ranged from 10 to 20 seconds and from 4 to 7 seconds, respectively. The compression ratio for TGA and TIFF, PCX, and GIF was less than 1.5:1 whereas it was more than 3:1 for JPEG. The higher the compression ratio of JPEG, the more degradation of image quality occurred although compression ratio below 20:1 made no considerable difference. In the remainder, the image quality was same before and after the compression. The digital images of US, CT, and MR showed lower degradation of image quality than plain radiograph, angiography, and barium study. CONCLUSION: The clinically acceptable limiting compression ratio of JPEG was 20: 1 which would be adequate for managing the large volume of image database on a PC-based digital imaging system.