Affordable Medical X-ray Imaging for the Developing World: a Global Vision

Sorin Marcovici, Vlad Sukhovatkin and Oscar Cisek

XLV Diagnostics Inc, Thunder Bay, Ontario P7A 7T1, Canada

Abstract- Imaging technologies, already used routinely for diagnostics, are becoming an essential part of the work flow included in screening, treatment progress’ monitoring and outcomes’ evaluation. However, various imaging modalities considered standard practice for the medical institutions in the developed world remain today rather inaccessible for the populations at large in the developing countries. One of the limiting factors is the high cost of medical imaging equipment. To develop and produce low cost, medical imaging systems of good quality requires disruptive innovations and novel approaches to engineering. The sophistication and capabilities of ubiquitous consumer products available today provide a real opportunity to capitalize on their low cost components to drive down significantly the cost of medical imaging equipment. Following this approach, XLV Diagnostics, a Canadian start-up, is developing a completely new technology, X-ray Light Valve (XLV), for producing very affordable large area, flat panel, x-ray detectors based on which economical medical imaging systems can be built. The first XLV product will be a digital mammography machine for breast cancer screening that will be sold at a fraction of the prices asked for presently available machines. The paper will present the conceptual approach and innovative engineering to create from consumer type components a good quality digital imaging system and will show several of its parametric characteristics.

Keywords- Flat Panel Detectors, X-ray Light Valve, Liquid Crystals, Amorphous Selenium, Digital Mammography.

I. INTRODUCTION

X-ray imaging plays an increasingly important role in medical practice from diagnostics to treatment’s efficacy monitoring and outcomes’ evaluation. However, in many countries worldwide, the high cost of imaging equipments makes them inaccessible for large parts of the population. The developing world is faced with the challenge of using limited financial resources to procure and maintain large number of imaging systems needed for providing adequate medical treatments to their own, quite large population.

To date, most medical imaging systems were conceived and implemented to provide optimum performance during a variety of clinical circumstances and consequently, they incorporate many specialized features which drive up their price. On the other hand, many developing countries need inexpensive and good enough medical imaging systems to satisfy basic clinical needs. Among all imaging modalities, the x-ray imaging systems are used for very large number of procedures and reducing their costs represents a high impact priority. To develop and manufacture x-ray equipment and to achieve low cost yet effective medical imaging systems requires disruptive innovations and novel approaches to engineering.

II. X-RAY DIGITAL IMAGING

A. Present status quo At the core of most digital x-ray systems there is a large area Flat Panel Detector (FPD) whose basic function is to capture the x-ray quanta and translate the information they are carrying into electric signals which can be measured and reconstructed as images by the computing engine of the system. FPDs fall in two categories depending on the manner in which they implement quanta to electrical charges conversion: indirect and direct. The indirect FPDs convert the x-ray quanta in electrical charges by a two-step process: x-ray quanta to light and then light to electrical charges. Those FPDs incorporate some scintillating type material, either in columnar crystalline or amorphous form, for converting the x-ray quanta to light. In general, due to light spreading, their spacial resolution is limited to about 100 µm or more. The direct type FPDs convert in one step the x-ray quanta to electrical charges and they can achieve a spacial resolution of the order of 40 µm. Majority of direct type FPDs use amorphous Selenium to convert x-ray quanta to electrical charges and are the FPDs of choice for digital mammography due to their high spacial resolution. Both types of FPDs are built using two-dimensional arrays of either Thin-Film Transistors (TFT) or CMOS circuitry which are custom made and expensive. There were efforts to replace the TFT in indirect FPDs with various optical

subsystems but they introduced a secondary quantum sink that increases detector’s noise. In general, to minimize the detector noise, the front-end electronics must be custom made to provide extremely low electronic noise at high speed and wide dynamic range. This increases the cost of FPDs. For most digital mammography machines available today, the cost of the detector represents about half of the entire cost of the system. The designed-in topology of all FPDs, as implemented presently, does not allow for FPDs’ cost reduction below a fixed level determined by the inherent high costs of using several custom made components whose prices are volume insensitive. To develop affordable, medical x-ray imaging systems we need a new approach and way of thinking.

B. The way forward To produce good quality yet economic medical x-ray imaging machines, the cost of every major subsystem has to be brought down. This implies getting rid of expensive components that burden today the FPD cost and replacing them with inexpensive ones within creative, new designs. For digital mammography machines, for example, the first priority is to develop novel FPD technologies which by design guarantee a low cost. Many consumer products offer a plethora of inexpensive electronic components for sensing, computation, displays and networking. They also offer hints on how complex functionality can be obtained by using those components in a different manner than their originally intended use. The adaptation of consumer type components to construct digital detectors with equivalent or better performance than FPDs is the way to go forward. This paper will present the X-ray Light Valve (XLV), an innovative technology for x-ray detectors, and its specific implementation in a FPD for digital mammography.

C. XLV detectors for mammography The XLV large area, 25 cm x 30 cm, detector is a direct conversion type detector. The x-ray quanta are converted in electric charges by a 200 µm thick layer of amorphous Selenium, deposited on a thin sheet of simple glass. During exposure, a 2,000 V bias is applied across Selenium and creates a 10 V/µm field that moves the charges to opposing ends. At Selenium - glass interface, the two-dimensional, continuous, electrical charges’ distribution precisely maps the attenuation information without any additional noise. On the thin glass side opposite to Selenium layer a 10 µm Liquid Crystals (LC) layer twists its molecules in response to the corresponding local charges. In this way, the two-dimensional distribution of twists monotonically maps the

charges’ distribution without any significant, additional noise (molecular vibration noise contribution is extremely small). The LC materials and their assembly process are the same as the ones used for consumer products’ Liquid Crystals Displays (LCD). On the other side of LC layer, a linear optical scanner, quite similar to those used in fax machines, beams light through the twisted LC, reads back the reflected light modulated by the twists and digitizes the signals. XLV topology avoids the creation of a secondary quantum sink and provides for non-destructive, repetitive readings. Pixel size is selectable at digitization. The pixel size of this mammography detector is 42 µm. Due to its ability to read multiple times each charges distribution, the XLV detector could be programmed to digitize the same distribution with various pixel sizes and to output images with correspondingly various resolutions. Within limits, the light intensity of the linear optical scanner could also be increased to amplify the reflected light’s level. It is important to note that the high spatial resolution of amorphous Selenium detectors was successfully preserved within XLV topology. The expensive TFT matrix and the front-end electronics were replaced by low cost ubiquitous components like LC and linear optical scanner. Figure 1 presents XLV detector’s Modulation Transfer Function (MTF) in comparison with the MTF of several, commercially available FPDs for digital mammography. As observed from the curves, the XLV MTF performance compares favorably. Fig. 1 MTF comparison

The sharp lines’ edges definition and the good black-white contrast are clearly visible in Figure 2 and Figure 3 which show images of standard resolution targets.

Fig. 2 Linear resolution target

Fig. 3 Star resolution target

Figure 2 shows the unprocessed image of a linear resolution target and Figure 3 shows the unprocessed image of a star resolution target taken with the XLV detector. The Detective Quantum Efficiency (DQE) of the XLV detector is the same as the DQE of amorphous Selenium based FPDs. XLV detector very low noise and excellent low contrast resolution can be appreciated in the Figure 4. Fig.4 Scanned piranha

D. XLV based mammography machine XLV technologies are conducive to producing large area mammography detectors for a fraction of presently made FPDs costs. The next step necessary for bringing to market a good mammography machine at low cost is to reengineer the rest of the system by optimizing all electro-mechanical subsystems and eliminating unnecessary interfaces. Applying frugal engineering methodologies to make the subsystems modular and more compact and to minimize the number of cables and interconnections, translates into less expensive and more reliable machines. Clever use of off-the-shelf software for controlling machine operation and monitoring in real time its health contributes to further costs reduction. An additional benefit of compact engineering approach is the small foot print of the mammography machine, shown in Figure 5, which enables its use in limited spaces or in mobile vans.

III. FUTURE XLV DEVELOPMENTS

Following the commercial launch of the mammography machine equipped with the 25 cm x 30 cm XLV detector, a smaller XLV detector of 18 cm x 24 cm will be developed for screening smaller size breasts. The small detector will be interchangeable with the large detector within the same mammography machine. The XLV technology’s main attributes such as low cost, simplicity, reliability and easy scalability will also provide the basis on which various machines could be developed for medical x-ray imaging applications. Of great interest is a low cost, high throughput, digital x-ray chest system for tuberculosis screening in the developing countries.

Fig.5 XLV mammography machine

IV. CONCLUSIONS The industry must come up with innovative technologies and creative engineering approaches to answer the world-wide need for good quality yet economical medical x-ray imaging equipment. The XLV patented [1], [2], [3] technologies are showing practical ways for capitalizing today on the superior price-performance provided by consumer components, like LC and linear optical scanners, to build world class medical imaging equipment. They are the harbinger of a new wave of low cost medical devices and systems which will have a broad impact on all imaging modalities. XLV Diagnostics is actively sharing this global vision and looks forward to developing series of innovative x-ray medical imaging equipment. AKNOWLEDGEMENT The authors want to acknowledge with gratitude the seminal contributions of Prof. John Rowlands to the XLV scientific theory and its underlying basic technologies. REFERENCES

1. US 5,847,499 – Apparatus for generating multiple x-ray images of an object from a single x-ray exposure

2. US 6,052,432 – Method of generating multiple images of an object from a single x-ray exposure

3. US 7,687,792 – X-ray light valve based digital radiographic imaging system

Contact: Sorin Marcovici XLV Diagnostics Inc. 290 Munro Street

Thunder Bay, ON P7A 7T1 Canada

Email: sorin@xlvdiagnostics.com