Mammography Image Comparison Analysis – Improving Early Detection Of Breast Cancer.
Mica is Automated Comparison alert platform of new and changed breast tissue. A decision support
information Hub and Medical Record platform for the Breast Cancer prevention arena.
Provides the user with a Personalized Journey app and access for his own mammograms
The idea behind Mika is to create artificial intelligence that can provide support and identify growths at early stages and in real-time, helping radiologists reduce screening time and reduce the number of radiologists required to verify screenings – from three radiologists to one.
The challenge was to create a UI that could include a lot of information on one screen, highlighting the following:
Design sprint schedule by days
First, to layout the project, we mapped out all relevant components such as: the patient, the radiologist, the second radiologist who verifies
the diagnosis, the back-office personnel, the senior physician, etc.
During the second step, we created an experience path for each user including the stops each user goes through to complete the process.
Pick a Target
Picking a target is a design sprint method that helps the team narrow down on a focus for the sprint. This exercise is all about finding the point in the customer journey that is most critical to get right. It’s where the highest risk of failure and biggest opportunity for success lies. The rest of the sprint flows from this decision
User Interface Components
User Interface Mocks
Home screen, I have compiled all the important patient information so that the radiologist has a clear snapshot of each patient before treatment. In addition, I added a filter that filters patients by was diagnosed / wasn’t Diagnosed.
The patient’s diagnostic screen. The challenge was to create an interface that would include a lot of information while utilizing minimal space. Information like diagnosis history, medical file, personal questionnaire, diagnosis from previous years, and the current diagnosis.
To make it easier on the radiologist, and to speed up the diagnosis process, I simplified the process so that after marking off the suspicious lump, a window which contain possible diagnosis pops up. All the radiologist needs to do is to check off the type of diagnosis and move on to the next step.
On the top right you can see two controls, “Radiologist” and “Mika Helper” are the diagnosis of the mammogram. In order to the differentiate between the two, the radiologist’s spot is in white circle and Mika’s is in a green square.
At the last stage, the radiologist explains the diagnosis considering the data he collected from the mammogram.
To streamline the system and the ongoing work of the radiologist, I put the “Diagnostics List” control on the top right. Clicking on it will open a list of all patient diagnosis.
The radiologist will always know where the diagnostic information originates from, as soon as they hover over the text, they will see a zoom in on the proper area, while the rest of the images remains at a normal size.
At any stage the radiologist can see the patient’s medical record without leaving the diagnosis screen. Allowing for faster diagnosis flow.
The radiologist work screen while working with the same breast but from a different angle – for this purpose I created a tool called: “Connect to” with which the radiologist can compare different sides of the same breast.
Comparison screen – comparing previous years images of the same breast with the current year image located on the left. On this screen I created a red sign that indicates of new lumps that were not in previous screenings. Another sign I created on this screen indicates lump growth compared to previous years, including growth percentages below the diagnosis (in white).
I created a hierarchically organized report which includes all the necessary the information that a third party (for example: a radiologist from another country) would need to understand the diagnosis clearly. First, I have the exam details and patient details, then I have the radiologist conclusions and lastly all the patient’s mammograms with the diagnosis.