MICENEA project
X-ray mammography is the most widely used diagnostic technique in mass screening, because it allows for excellent spatial resolution and very good contrast. Despite being the 'golden standard' for breast cancer detection, mammography has drawbacks: the use of ionizing radiations, which limits frequent use; significant breast compression, which is often painful for the patient; X-ray images are not tridimensional; sensitivity decreases rapidly with breast density increasing, which makes it less efficient for younger patients.
Other methods like Ultrasound or Magnetic Resonance (MRI) have advantages over mammography and shortcomings, too. Both do not make use of ionizing radiations. Ultrasound has though lower resolution than mammography, is considered too operator dependent and characterized by low specificity. Contrast-enhanced breast MRI has a high sensitivity and if used in conjunction with mammography can improve diagnosis over standard MRI. However, MRI generates high magnetic fields, it is very expensive and time consuming. These concerns motivate the quest for other techniques to be used in combination with mammography - and indeed it is quite common in clinical practice to combine different imaging methods to get a more specific diagnosis - or as a standalone method for those cases in which mammography is not recommended.
This project, which gathers together researchers from the second University of Naples, the Politecnico di Torino and the University of Bologna in association with other Italian and international partners, has the following primary objectives.
- To design a microwave prototype system based on a novel approach that leverages new imaging algorithms, electromagnetic modeling methods, computation acceleration and ad-hoc developed antennas, RF circuits and digital hardware.
- To develop new algorithms for the automatic lesion detection which, according to the combined diagnostic philosophy outlined above, receives inputs from mammography, DOBI and the developed microwave prototype.
- To test both the microwave imaging system and the software for lesion detection in vivo, in a real clinical scenario.
A secondary objective of the project is the assessment of the new microwave imaging method as a technique particularly addressed to younger women than X-ray mammography currently targets. Frequency of test is key for early detection and is permitted by the absence of ionizing radiations. Younger patients are the part of the screening target where mammography is 'weaker', because of the higher breast density. The resolution achievable by a microwave imaging method is less than what mammography can get, yet it is expected to offer a good compromise between resolution, contrast sensitivity, penetration depth and cost. Also the insensitivity of the results to the execution of the operator and the ease of use both for the patient and the operator are welcome features.
With regard to the microwave system, new inverse and forward electromagnetic modeling algorithms will be developed which take advantage of new perspectives offered by parallel and cloud computing. Ad-hoc anatomical antennas working in the GHz frequency range will be designed and fabricated as well as the integrated circuits in CMOS 130 nm technology that drive the antennas, receive the scattered signals and digitize them. Finally, to minimize the execution time of the algorithms which elaborate the received signals, implementation with multi-threaded CPUs assisted by GPUs will be considered. To reach real time data processing, a board equipped with several FPGAs will be acquired and be part of a final system prototype.
Toward the goal of a combined detection system, novel properly designed lesion detection algorithms will be developed. In particular, a new image processing module will be designed to process images acquired by the different devices developed and used in this project. The processing aims to provide the software detection module with a representation of the image suitable for the algorithms that give hints about the presence of suspicious regions. The development of the combined automatic lesion detection system requires access to non-easily disposable Dynamic Optical Breast imaging (DOBI) equipments and images. This access is granted by one of the Italian medical research centers which expressed the will to join the consortium.
The aims of MICENEA project require a high degree of interdisciplinary and expertise, ranging from medicine and engineering to physics and computer science. To this end, the three research units which constitute the consortium include researchers whose competencies span all the needed expertise.