How MammoCoach works


Since MammoCoach is a HTML based software, the performance might fluctuate with different internet browsers. The best performance can be achieved with Chrome or Firefox. We recommend strongly to use one of the two.

You set up your account, and signify the yearly workload group that you belong to A: <500 B: 500 < 1,000 C: 1,000 < 2,000 D: 2,000 < 5,000 E: > 5,000 screening mammograms per year..

It is important to maximize the browser for your monitor and use a large screen resolution (minimum of 1280 by 800). Under "My Profile", you have the choice to set your viewer window to maximum size automatically or to set a fixed viewer window size. You decide which set up works best for you.

Due to constraints of the web, the images will initially be displayed at lower resolution, which allows the program to display the images in a reasonable speed. The magnification glass will allows you to see the magnified part of the image in full resolution. That is the mode to look for calcifications.

Our data sets can be read in comfortable batches of 100, with feedback available after every 100 cases.

Following completion of the total 1000 cases, you will be given a personalized summary with data about all your benchmarks, including cancer detection rate, recall rate, positive predictive values and your false positive recalls (quantitative analysis and feedback).

However, you will not only have the opportunity to mark ABNORMAL findings which will eventually determine your quantitative results and benchmarks but also you can mark NORMAL findings, based on following three criteria: a.) stability since prior study b.) normal morphology c.) finding seen on one view only. This will give you access to a qualitative analysis and feedback where you can see in the review mode where other users have decided not to call patient backs. Different colors will be used for different levels of expertise. Level of expertise is determined by amount of screening mammograms read per year.


You will be able to compare your results based on your NORMAL and ABNORMAL markers placed on the screening cases with the 'Gold standard'. Your assessment of the diagnostic cases will be also compared to the 'Gold standard'.

The 'Gold standard' is based on the assessment of the authors who contributed the cases. The validity of the author's assessment is determined by the fact that authors included only screening cases which the authors read as normal (date 3) and which have a normal previous study from 2 years ago (date 2) and in addition a normal previous study from 4 years ago (date 1). The study from date 2 was chosen as current study and the study from date 1 was chosen as previous study.

This guarantees maximum likelihood to exclude false negative cases. Otherwise all recalled, finally judged as BI-RADS 1/2 are being also followed by at least 2 years of normal screening cases as known to the author. Whereas all recall cases which ended up with biopsy will be correlated to the actual histology as known to the authors.

As soon as you complete your first batch of 100 screening cases, you will receive a message to review your recalled cases (which the authors also decided to recall). You will review the diagnostic images of your recalled cases and make a final decision between BI-RADS1/2 and BI-RADS ¾.

After this step your will see the statistics of your screening and diagnostic cases in form of a 2 by 2 table and a list of all the cases with associated markers (TP,TN,FP,FN).

In the Review Mode you will also be able to look through your cases, based on the provided 2 by 2 table indicating your False Negatives, False Positives, True Negatives and True Positives and compare your judgment with the judgment of peers, of different levels of expertise, who decided NOT to call back findings.

That approach also allows you to review the cases you did not call back but the authors called back, especially if they ended up getting a biopsy and showed a malignancy (FN).

Through that comprehensive approach MammoCoach will not only provide a source to increase your screening numbers to a more meaningful level but also will provide a unique learning experience mimicking the double read as practiced in European countries, by looking over the shoulders of peers of different expertise.

This will allow a continuous training to sharpen the screening skills and avoid unnecessary workups. At least 5 'super readers' reading more than 5000 screening mammograms will be included. These 5 'super readers' will be named in the 'About us' section of the webpage.

Retraining with Case collection and comparing performance:

Based on your False Negative and False Positive cases which are highlighted by MammoCoach a 'Learn' button will be displayed which brings up a large collection of teaching cases which correlates with the category (mass – calcifications – architectural distortion – asymmetry) of your miss. This is intended to retrain your eyes to recognize certain patterns on screening mammograms.

After having finished your individualized retraining, you will have the chance to go through 500 randomly picked cases out of the 1000 cases, including 10 % recalls and, about 2 cancers to see if your performance has improved. After that you will be provided again with the quantitative and qualitative feedback and statistics.