- Main research unit: Instituto de Telecomunicações (IT)
- Ricardo João Cruz Correia
- Pedro Manuel Vieira Marques
- Sandra Filipa Canário de Almeida
- Inês de Castro Dutra
- Fabio de Lima Hedayioglu
- Ye Can
- Sandra da Silva Mattos
- Start date: 2010 | End date: 2013
- Financing: € 120,000.00 (FCT)
Project description: In the ears of an experienced physician, a stethoscope yields important clinical information which can help an initial assessment of a patient's clinical condition and guide the subsequent need for more specialized exams. This is particularly true in chest Medicine, i.e. Cardiology and Pneumology, which is the reason why the stethoscope still maintains a key position in Medicine in the modern era. Auscultation, however, is a hard skill to master. The heart sounds are of low frequency and the intervals between events are in the order of milliseconds, requiring significant practice for a human ear to distinguish the subtle changes between a normal and a pathological heart sound. The use of a digitally enhanced stethoscope, adequate for training physicians to improve their basic skills in diagnosing and treating heart conditions, or as a stronger tool for world-wide screening of specific heart pathologies are but some examples of how state of the art technology can be used horizontally to benefit people at different economic, political or geographical levels. This motivates the key objective of the DigiScope project: develop the prototype of a digitally enhanced stethoscope, capable of automatically extracting clinical features from the collected data, as well as providing a clinical second opinion on specific heart pathologies.
A market inspection shows some commercial versions of electronic stethoscopes with different filtering options, but none provides computer assisted clinical data extraction. In addition, we found several isolated computer auscultation analysis systems referred in the literature, as described in one of our published articles [HedayO9]. Our review showed very promising work in audio feature extraction using signal processing methodologies (e.g. S1 and S2 segmentation [Liang97]; A2 and P2 segmentation [Xu00]), and much poorer results for heart pathology detection. Furthermore, almost all research did not enforce either rigorous and extensive clinical data collection or a vital clinical validation step, both essential for understanding the viability of digitally enhanced stethoscopes in routine clinical practice.
As this proposal will detail, building a DigiScope involves not only signal processing (clinical feature extraction) and datamining (pathology detection using audio features and patient record information) but successfully accomplishing this objective requires two other key tasks: our ability to not only collect heart audio signals, but to extract patient record information by communicating directly and efficiently with typically heterogeneous and complex hospital information systems; build a framework of supporting hospitals and Cardiologists both for providing access to the large amounts of data required by the signal processing and data mining tasks, and for a rigorous clinical validation step of the developed technology. Our research team includes experts in all these four tasks such as Dr. Ricardo Correia (8 years of experience in medical information systems), Dr. Miguel Coimbra (9 years of experience in signal and image processing, 5 of which in medical signal research), Dr. Inês Dutra (8 years of experience in data mining), Dr. Almeida (5 years of experience in Cardiology) and Dr. Sandra Mattos (22 years of experience in Cardiology). Our ‘silver bullet' is thus an experienced multidisciplinary team, which includes experts in all fields identified as vital for the success of DigiScope, most of which have had previous experience in multi-disciplinary (medical and engineering) research projects.
Our strategy lies in accomplishing a set of core objectives for the described four tasks that will guarantee the development of a hospital integrated, clinically validated digitally enhanced stethoscope. Although this is by itself ground-breaking, we will then have a set of ambitious high-risk, advanced objectives that will also guarantee the scientific novelty of this project. As such, we do not only expect to produce a final prototype but also a reasonable number of publications in high-impact journals and conferences, as well as what will probably be the largest clinically annotated database on audio signal processing for cardiology.