It exploits similarities between unknown and known pictures to produce automatic insights about the unknown cases.
Starting from a flexible structure able to cover a wide range of domains, AIppo can be tailored to solve any specific problem.
AIppo CXR is the AI powered application for Chest X-Radiograph (CXR).
- provides risk indicators for a set of diseases
- prioritizes queue based on the need of assistance
- propose a list of similar cases from the known ones
The AI core of AIppo CXR consists of Deep Neural Network able to extract meaningful information from CXRs already elaborated.
AIppo CXR use this information to compare a new unknown case with known cases and compute the related risk indicators.
The doctor could always interpret the outcome by looking at the list of similar cases.
By collecting the final feedback for each new analysis, AIppo CXR continuously improve itself.
AIppo CXR comes as a Software as a Service (SaaS) product fully integrated with every health care organization.
It can be installed within medical network avoiding unnecessary data transfer or as a cloud service accessible from everywhere.
Simple APIs enable any IT infrastructure to receive and use AIppo CXR insights.
In most clinical settings, patient cases are analysed by radiologists in chronological order: first arrivals are served first.
This method can create conditions in which potentially infectious patients or patients in need for urgent care must remain for a long time in the waiting room. In this way, infection risk for those around them increases and care effectiveness decreases.
AIppo CXR can be integrated with Radiology Information System and provide a risk indicator for each patient in the queue.
By ordering the cases to be examined according to the risk indicator calculated by AIppo CXR, patients are prioritized based on their need of assistance. Those who can spread a contagious disease to other people are moved to the top of the reporting list. In this way, they can be identified and managed quickly and precisely.
For each case examined, AIppo CXR provides a risk indicator for each diseases of interest. The higher the indicator value, the more probable it is considered by AIppo CXR that the patient is suffering from the indicated disease.
Values can support the radiologist during the diagnosis, provide a second opinion on the case and report suspected cases quickly.
Moreover, AIppo CXR provides a most similar cases list related to the current patient, among those ones already diagnosed in the past. In this way, the doctor can compare the current unknown case with known cases, making the diagnosis process faster and more effective.
Thanks to the integration with Pictures Archiving and Communication System, for each similar case the doctor can access the Chest X-Radiograph and related information.
The similar cases list also allows the doctor to understand how AIppo CRX has determined the risk indicator.
It depends on the similarity of the unknown case to others known cases affected by the disease.
This allows the doctor to critically evaluate the system output and confirm or possibly correct it.
In fact, for each case, it is possible to communicate the final diagnosis to AIppo CRX. It will use this information to increase its knowledge base, improve the accuracy of its analyses and provide increasingly accurate risk indicators.
In a study with Università di Torino AIppo CXR has been tested for Covid-19 using a set of 536 Chest X-Radiographs collected and analysed by AOU San Luigi Gonzaga.
Preliminary results show significant advantages:
- Risk indicators can detect 89% of positive cases
- Reduction from 57% to 25% of negative cases examined before a positive one
- Analysis ready in less than 1 minute
AIppo CXR is under validation in clinical trial at AOU San Luigi Gonzaga.