Loading...

quantusMM

ANALYSIS AND CLASSIFICATION OF MAMMOGRAPHS TO DETERMINE THE RISK OF BREAST PATHOLOGIES

Its technology is based on the quantitative analysis of the texture of the breast image. quantusMM classifies the images (craniocaudal perspective (CC) and oblique mid-lateral perspective (OML) of both breasts) and gives them a category which determines the malignancy status of the analysed breast.

Classification system applied in quantusMM

  • Class 1: Normal breast
  • Class 2: Probably benign findings (Probability less than or equal to 50%>
  • Class 3: Suspicion of malignancy (Probability between 50% and 95%)
  • Class 4: Highly suggestive of malignancy (Probability greater than or equal to 95%)
  • Class 5: Malignant breast

The system determines the probability of a sample being positive, depending on the false positives detected in the training sample. This is done on the basis of several thresholds, each with the sensitivity, specificity and positive and negative predictive values shown below.

For example, an image classified as "Class 5" is considered a safe positive since, at the chosen threshold, the occurrence of false positives (negative samples classified as positive) is 0% (specificity 100%), so that the positives predicted by the system will be true positives. In this case, the number of false negatives (positive specimens that are classified as negative) is higher than in the other lower malignancy classes (low sensitivity).

If an image does not show "Class 5" candidates, the system checks whether it belongs to a lower malignancy class, which would be "Class 4".

A specimen classified as "Class 4" is considered to have a high probability of malignancy since, with the chosen threshold, the occurrence of false positives (negative specimens that are classified as positive) is less than 5% (specificity > 95%), so most of the positives predicted by the system will be true positives. In this case, the number of false negatives (positive specimens that are classified as negative) is higher than in other classes of lower malignancy, but lower than in "Class 5", which is of higher malignancy (sensitivity is not as low).

If the image is also not classified as "Class 4", the process is repeated for each of the lower malignant classes, until "Class 1" is reached, which is considered benign.

"Class 1" is considered benign by the same argument explained for "Class 5", but using the values of sensitivity (at 100%) and negative predictive value (lower specificity).

Thus, the system determines the highest malignancy class to which each of the images to be analyzed belongs, and it is indicated in the corresponding report for each of the patient's mammographic images.

Comparative table of the reliability of quantusMM for each classification for Nodules

  Sensitivity Specificity PPV NPV
Class 1
Class 2 99.0% 23.0% 56.0% 96.0%
Class 3 95.0% 51.0% 65.0% 91.0%
Class 4 65.0% 95.0% 92.0% 73.0%
Class 5 37.0% 100.0% 99.0% 62.0%
* Preliminary data awaiting review by the Osakidetza clinical team.

Comparative table of the reliability of quantusMM for each classification for microcalcifications

  Sensitivity Specificity PPV NPV
Class 1
Class 2 99.0% 39.0% 54.0% 98.0%
Class 3 95.0% 75.0% 73.0% 95.0%
Class 4 75.0% 95.0% 91.0% 84.0%
Class 5 36.0% 100.0% 99.0% 69.0%
* Preliminary data awaiting review by the Osakidetza clinical team.

WHY DOES QUANTUSMM WORK?

quantusMM is presented as a novel Artificial Intelligence method based on state-of-the-art Deep Learning, with 2 different classification algorithms: one for nodules and one for microcalcifications.

Its technology is based on a quantitative analysis of the texture of the image. This analysis makes it possible to identify patterns associated with malignant nodules or microcalcifications, thus helping to identify the risk of breast cancer.

The original image is subjected to contrast enhancement filters and a division into patches so as not to lose image quality when the lesions to be detected are very small.

WHEN TO USE QUANTUSMM?

quantusMM has been designed with a clear focus on the general population and is intended to be a tool to assist clinicians, especially in mass screening of patients with risk factors, early detection of breast disease and prioritisation of waiting lists.

Therefore, any woman can take the test at any time.

quantusMM:



Non Invasive


Reliable


Fast