Periodic Reporting for period 2 - d3pm (Sonio: Deep-Learning for Detection and Diagnostic of Prenatal Malformations)
Berichtszeitraum: 2024-01-01 bis 2024-12-31
Sonio’s mission is to contribute to access to high quality care for women and children worldwide thanks to artificial and collective intelligence.
On the mother's side, too many women still die due to pregnancy complications. In France, it has been estimated that 58% of the maternal deaths were probably or possibly avoidable.
On the foetus side, according to EUROCAT network, 2,7% of the foetus have at least one congenital malformations. Most of the studies conclude that 50% of the congenital malformations are not detected antenatally and that these screening rates vary a lot depending on the geographic location (i.e physicians expertise). An undiagnosed congenital malformation constitutes a greater risk of mortality for the newborn, especially, abnormalities requiring specific and urgent care at birth (e.g. transposition of the great arteries). Ultrasound is both relatively inexpensive and allows access to a large part of the foetal anatomy and the detection of obstetric complications. The limit is therefore rather access to clinical expertise rather than the technical improvement of ultrasound.
A very high number of potential malformations is accessible in all foetal anatomical regions (more than 400), knowing that these malformations can be isolated or part of a poly-malformative syndrome (9000 syndromes referenced in Orphanet, including 800 accessible on foetal ultrasound). In this context, learned societies have produced increasingly comprehensive and demanding protocols putting more pressure on practitioners. The latest recommendations from ISUOG 2022 include 20 mandatory ultrasound plans showing 42 anatomical structures. As a result, doctors may forget the mandatory views, leading to underdiagnosis. Even when the examination is exhaustive as to which views to acquire, a congenital anomaly can go unnoticed. In this context, artificial intelligence is a tool of choice for strengthening the skills of health practitioners. It has been shown that using a checklist can improve exam pass rates from 49% to 97%. However, checklists are little used in practice because their use is cumbersome and time-consuming. Sonio can democratise and generalise the use of a checklist. AI also has the potential to provide clinical and diagnostic information by detecting abnormal ultrasound signs on an image that may be related to a congenital malformation, recommending the next abnormalities to check for, and providing an interpretation on the phenotype/genotype correlation.
On the mother's side, too many women still die due to pregnancy complications. In France, it has been estimated that 58% of the maternal deaths were probably or possibly avoidable.
On the foetus side, according to EUROCAT network, 2,7% of the foetus have at least one congenital malformations. Most of the studies conclude that 50% of the congenital malformations are not detected antenatally and that these screening rates vary a lot depending on the geographic location (i.e physicians expertise). An undiagnosed congenital malformation constitutes a greater risk of mortality for the newborn, especially, abnormalities requiring specific and urgent care at birth (e.g. transposition of the great arteries). Ultrasound is both relatively inexpensive and allows access to a large part of the foetal anatomy and the detection of obstetric complications. The limit is therefore rather access to clinical expertise rather than the technical improvement of ultrasound.
A very high number of potential malformations is accessible in all foetal anatomical regions (more than 400), knowing that these malformations can be isolated or part of a poly-malformative syndrome (9000 syndromes referenced in Orphanet, including 800 accessible on foetal ultrasound). In this context, learned societies have produced increasingly comprehensive and demanding protocols putting more pressure on practitioners. The latest recommendations from ISUOG 2022 include 20 mandatory ultrasound plans showing 42 anatomical structures. As a result, doctors may forget the mandatory views, leading to underdiagnosis. Even when the examination is exhaustive as to which views to acquire, a congenital anomaly can go unnoticed. In this context, artificial intelligence is a tool of choice for strengthening the skills of health practitioners. It has been shown that using a checklist can improve exam pass rates from 49% to 97%. However, checklists are little used in practice because their use is cumbersome and time-consuming. Sonio can democratise and generalise the use of a checklist. AI also has the potential to provide clinical and diagnostic information by detecting abnormal ultrasound signs on an image that may be related to a congenital malformation, recommending the next abnormalities to check for, and providing an interpretation on the phenotype/genotype correlation.
Our scientific efforts financed by EIC have gone in three different main directions:
- improve the completeness of prenatal ultrasound examinations by building an AI tool able to recognize the different US planes and structures (Sonio Detect)
- improve the rate of abnormal ultrasound signs discovery at foetal ultrasound by building an image-based AI having high specificity and sensitivity on most important abnormal ultrasound signs (Sonio Suspect)
- improve the rate of prenatal rare disease diagnosis by providing an AI tool able to recommend next images to take or next anomaly to look for to avoid incorrect conclusion of an isolated malformation when it is in fact a rare disease (Sonio Expert).
We managed to obtain a FDA clearance in June 2023 for Sonio Detect V1 and a FDA clearance in April 2024 for a V2 with a wider perimeter. We went from a V1 with 13 foetal US views and 15 anatomical structures to a V2 with 26 Views and 53 structures.
Sonio Detect has demonstrated high performance in the set objectives of detecting standard views and anatomical structures for quality assessment purposes. But our investigations have also shown the limits of Sonio Detect, mainly trained on normal images (congenital malformations only represent 3% of births), when it is tested on pathological images. This is a clear and completely new demonstration of the interest of the Sonio Suspect that we are covering with funding from the EIC. Sonio Detect is already proving effective on certain pathologies such as “Agenesis of the corpus callosum” where a standard structure is either surnumerary or absent. Yet, Sonio Detect could also be prone to hallucinations and falsely detect absent anatomical structures in the context of other pathologies like renal agenesis where Sonio Detect may clear some images because the deviation that makes them pathological is too subtle to be detected. Our latest achievements have been the collection of large pathological databases, the annotation of these images and clips with their ultrasound semiology, as well as the training of Deep-Learning models on pathological images and normal images to better control these hallucinations. We have also built specific models to do automatic segmentation and measurement of anatomical structure to screen biometrics anomalies. These models have been the object of a 510(k) FDA clearance, Suspect V1, covering most of the congenital cardiopathies but also a brain and a digestive malformation. These models have obtained a strong performance during the FDA clinical validation and have been demonstrated to significatively increase the detection performance of physicians, we observed +34% sensitivity and +7% specificity during the reader study between the phase of the physicians was reading the images alone and the phase where he/she was assisted by the AI system.
Finally, we made some notable contributions to the effort of building an AI tool able to assist the physicians when trying to do prenatal diagnosis of a rare disease. The study we have performed in 2022 for the CE mark of Sonio Expert has been published in 2023 in the journal UOG. This first software which is commercialized for several years will eventually be integrated into the global platform.
- improve the completeness of prenatal ultrasound examinations by building an AI tool able to recognize the different US planes and structures (Sonio Detect)
- improve the rate of abnormal ultrasound signs discovery at foetal ultrasound by building an image-based AI having high specificity and sensitivity on most important abnormal ultrasound signs (Sonio Suspect)
- improve the rate of prenatal rare disease diagnosis by providing an AI tool able to recommend next images to take or next anomaly to look for to avoid incorrect conclusion of an isolated malformation when it is in fact a rare disease (Sonio Expert).
We managed to obtain a FDA clearance in June 2023 for Sonio Detect V1 and a FDA clearance in April 2024 for a V2 with a wider perimeter. We went from a V1 with 13 foetal US views and 15 anatomical structures to a V2 with 26 Views and 53 structures.
Sonio Detect has demonstrated high performance in the set objectives of detecting standard views and anatomical structures for quality assessment purposes. But our investigations have also shown the limits of Sonio Detect, mainly trained on normal images (congenital malformations only represent 3% of births), when it is tested on pathological images. This is a clear and completely new demonstration of the interest of the Sonio Suspect that we are covering with funding from the EIC. Sonio Detect is already proving effective on certain pathologies such as “Agenesis of the corpus callosum” where a standard structure is either surnumerary or absent. Yet, Sonio Detect could also be prone to hallucinations and falsely detect absent anatomical structures in the context of other pathologies like renal agenesis where Sonio Detect may clear some images because the deviation that makes them pathological is too subtle to be detected. Our latest achievements have been the collection of large pathological databases, the annotation of these images and clips with their ultrasound semiology, as well as the training of Deep-Learning models on pathological images and normal images to better control these hallucinations. We have also built specific models to do automatic segmentation and measurement of anatomical structure to screen biometrics anomalies. These models have been the object of a 510(k) FDA clearance, Suspect V1, covering most of the congenital cardiopathies but also a brain and a digestive malformation. These models have obtained a strong performance during the FDA clinical validation and have been demonstrated to significatively increase the detection performance of physicians, we observed +34% sensitivity and +7% specificity during the reader study between the phase of the physicians was reading the images alone and the phase where he/she was assisted by the AI system.
Finally, we made some notable contributions to the effort of building an AI tool able to assist the physicians when trying to do prenatal diagnosis of a rare disease. The study we have performed in 2022 for the CE mark of Sonio Expert has been published in 2023 in the journal UOG. This first software which is commercialized for several years will eventually be integrated into the global platform.
We obtained a FDA clearance in June 2023 for Sonio Detect V1 and a FDA clearance in April 2024 for a V2 with a wider perimeter. This AI system is embedded in a workflow solution, a reporting solution, that is commercialized in the US and deployed in more than 50 centers representing around 9 000 examinations performed monthly with the solution. CE marking was obtained in July 2024 and commercial launch is beginning in Europe.
This quality control tool should help to reduce obstetric complications and congenital malformations missed antenatally, allowing to decrease the risk of maternal and neonatal death or disability. The new version of Sonio Detect and Sonio Suspect, which presents perimeter and workflow features never achieved in this field, has the potential to meet these challenges.
This quality control tool should help to reduce obstetric complications and congenital malformations missed antenatally, allowing to decrease the risk of maternal and neonatal death or disability. The new version of Sonio Detect and Sonio Suspect, which presents perimeter and workflow features never achieved in this field, has the potential to meet these challenges.