[Diagnostic efficiency on digital snapshots of the standard radiology imaging]

The authors had experienced the telediagnosis on digital snapshots of standard radiology imaging (chest, abdomen, and bones), sent by e-mailing, to support the medical doctors working in remote areas of developing countries. In order to validate the overall procedure, the authors have set up a simulating model and estimated some parameters of accuracy of the diagnosis on digital snapshots against the golden standard of the diagnosis by direct look. The study concerned the standard X-ray tests of one hundred randomly-selected patients out of a hospital archive. Four years later the diagnosis by direct look, the team of radiologists carried out the blind cross check on the digital snapshots of the radiograms and stated their second diagnosis. Sensibility, specificity, predictive value of positives, predictive value of negatives and efficiency of the whole series have been 83.0, 95.1, 96.1, 79.6 and 88.0%. By breaking up the series by apparatus, the skeleton test shows similar data of the whole series. The chest test shows a specificity and predictive value of positives of 100.0%. Although the number of cases is low, the abdomen test apparently shows a sensibility and predictive value of negatives as high as 100%, but a lower specificity and predictive value of negatives (85.7 and 87.5%). Though this data is supportive to the validation of the procedure, even better results are supposedly achieved by increasing the quality of the snapshots and by improving the skills of using the software.