Main purpose
This research activity started from a specific need of a group of expert radiologists: whether and how to test the radiologists (with and without experience) in the field of mammography screening. We came up with an idea that became the main goal of this teamwork: the implementation of a self-assessment software to test the ability of radiologists in diagnostic mammography screening.
Input of the software:
- set of mammographic images to be diagnosed;
- set of same mammography previously analyzed by an expert radiologist ? reference area with the lesion (if present).
Output of the software:
- score in percentage (0-100 %) ? to assess the ability of the inexperienced radiologist as well as to test also experienced operators.
Method
The software implements three fundamental steps:
- Visualization (Fig. 1)
- Performed for all the patients included in the test
- Opening/closure of CC and MLO projection
- Zooming and contrast enhancement adjustments allowed
- Medical reporting (Fig. 2)
- Selection of the suspected lesions or …
- Inclusion of the patient among the negative cases.
Fig. 1: Example of visualization, MLO projection both left and right side of the same patient. On the left bottom corner are depicted the slider to adjust the contrast at desire.
Fig. 2: Example of medical reporting. In case of a recall of the patient, an area of variable size is placed over the suspected lesion.
- Evaluation/Comparison (Figure 3)
- Compare the reference (by experts) to the highlighted area (by the operator performing the test)
- Compute a numerical rating for each patient and draw a final conclusion
Fig. 3: Example of evaluation. The goodness of reporting will be evaluated by comparing the area of reference and the area located by the candidate
Main collaborations:
- Dott. Pietro Panizza, Istituto Nazonale dei Tumori, Milano
- Dott. Giuseppe Di Giulio, IRCCS Policlinico San Matteo di Pavia
- Dott. Beniamino Brancato, ISPO, Firenze