Politecnico di Torino - Corso Duca degli Abruzzi, 24 - 10129 Torino, ITALY

+39 011 090 6100 info@tech-share.it

SynDiag – Automatic Pre-birth Syndromes Diagnosis

automatic medical analysisDecision supportdiagnosis of diseasesQuality of lifequantitative EchographyTechShareDay-d2w2

Introduction

Modern 3D ultrasound scans (US) are worldwide the most performed medical diagnosis exams. Unfortunately, they only provide an image of the fetus. SynDiag transforms automatically this image into a quantitative mathematical object, suitable for further elaboration, like distances and volumes analysis, and uses it to diagnose pathologies like cleft-lip and feto-alcohol syndrome.

Technical features

SynDiag can be installed as a plugin onto 3D US scanners already in use. It imports the registered US scans stack from the machine in DICOM format, the standard for medical images. Further steps are performed: data denoising, filtering, volumetric analysis alongside x, y and z directions, and data fusion to obtain the 3D mathematical description of the surface of the fetus face. Human intervention is never requested during the analysis and the physician doesn’t need any special training. The 3D fetus face is in PLY format (triangular mesh). Each point of the face is processed from a geometrical standpoint, in order to identify face’s landmarks like nose tip and lip extremes. Such data are then used to feed a clustering algorithm able to discriminate individuals manifesting a pathology. 80% success rate in quantitative model generation. 100% success rate in cleft-lip diagnosis, discriminating the side of the manifestations as well.

Possible Applications

  • Developed for gynecology, exportable to cardiology (80% of addressable pathologies covered);
  • Applicable to all pathologies manifesting malformations;
  • Generation of quantitative large database and normotype (average individual) definition.

Advantages

  • Quantitative analysis of the subject;
  • Fully automatic process;
  • Easy to use without special training;
  • Embeddable on existing technology;
  • Allows any mathematical elaboration;
  • Real time support to the physician;
  • Printable 3D model.