The way we were (and how we got here): fifty years of technology changes in dental and maxillofacial radiology

Abstract

The history of the last 50 years (1970–2020) of technological changes and progresses for equipment and procedures in dental and maxillofacial radiology is related from the insider perspective of an industrial physicist and technologist who has been instrumental at innovating and developing medical equipment in different parts of the world. The onset and improvement of all major categories of dental and maxillofacial radiographic equipment is presented, from the standpoint of their practical acceptance and impact among common dentists and maxillofacial radiologists: X-ray sources and detectors for intraoral radiography, and panoramic systems, both film-based and digital (including photo-stimulated phosphor plates); and cone beam CT.

Introduction

“According to laws of aerodynamics, the bumble bee cannot fly; its body is too heavy for its wings and that’s the simple reason why. But the bumble bee doesn’t know this fact, and so it flies anyway for all to see.” (A.S. Waldrop) I have repeatedly encountered examples of this fake scientific aphorism (sometimes apocryphally attributed to Albert Einstein) during my professional life, and dental and maxillofacial radiology is a prominent area to demonstrate it.

More than half a century ago, in August 1968, the International Association of Dento-Maxillo-Facial Radiology was founded at the conclusion of its first congress in Santiago, Chile, hosted by Professor Gregorio Faivovich.

Even for those who were already adult at that time (and more so for those who were not yet born!), it is difficult now to realize or remember how different everyday life was back then, as far as requisite gadgets go. We did not have Internet and electronic mail, SMS, Whatsapp and socials, word processors, “smart devices,” tablets, cellular phones, personal computers, digital cameras, microprocessors, solid-state digital memories (RAM), electronic pocket calculators, fax machines, flat LCD color screens (we had only bulky monochromatic CRT terminals, without any graphic capability, and they were for big centralized industrial or scientific computers, not for personal use). Storage of digital data (for big computers) would occur through magnetic tape, perforated paper tape, or punched cards. Even photocopying was a still-immature awkward business, with ammonia-stinking paper that would fade away after some time; reports were handwritten and handed over to a secretary who would typewrite them in the required number of copies using carbon paper on tissue paper sheets.

Consequently, retrieving documents and data from the current XXI century, when a huge amount of information in standard digital formats is available through the Internet, or from various storage media, is immensely less challenging than for the XX century—especially if such information was not published on easily available public media. In writing this insider excursus about the technology changes of dental and maxillofacial radiology in the last 50 years, I had to rely largely on my own memory (with the help of a few other old-timers), and on a few scraps of scattered documents in my possession that have not been dispersed with the continuous changes and vicissitudes of the dental and radiological industry. My tale is not focused on theoretical scientific advancements as reported in professional papers and discussed in academia (which may precede mass adoption by many years), but rather about the practical conditions experienced by the ordinary dentist and maxillofacial radiologist. The year of introduction of commercially available equipment reported here are often approximate, because there may not be a precise and univocal date: prototypes or pre-production units are installed at pilot sites in advance of regular production, and frequently companies start commercialization of a new product only in the country of their domestic market, before making it available to the international market. The years indicated here refer to when the products are generally available in the international market.

For almost four decades, I have been a manager of R&D and innovation for several manufacturers of radiographic and electromedical equipment, and I have witnessed that the scenario of technologies and procedures for dental and maxillofacial radiology has also changed hugely (albeit perhaps not so much as for the paraphernalia of daily life).

X-ray sources for dental intraoral radiography

After World War II, the dental profession became accustomed to the “monoblock” (or “tubehead”) design of dental intra oral X-ray sources (i.e. with X-ray tube, high-voltage transformers, collimation, and other necessary components all enclosed in a sealed metallic housing: no more exposed high voltage wires, a significant cause of fatal accident in the prior decades), all suspended from the wall with a pantographic arm (Figure 1). The control was an electromechanical timer. Initially, the high-voltage was around 50 kV or even 45 kV, just adequate for proper dental imaging, especially with alternate current (AC) anodic power supply. In the 70’s and 80’s, induced by changes in regulations and standards under the pressure from professional opinion-leader radiologists (not from the dentists themselves, who were and are the overwhelming actual clientele), the manufacturers gradually increased the high voltage to 65 kV and eventually 70 kV in new products. Not all dentists were completely happy of the change, longing for the sharper image contrast associated with moderately low kV, as opposed to the “softer” images associated with higher kV and favored by trained (general) radiologists. The control became an electronic timer.

Figure 1.

The tubehead of Oralix 50 by Philips, a popular dental intraoral X-ray of the ‘70 s (50 kV AC, short pointed cone). AC, alternate current.

Various studies published in the scientific literature over several decades have claimed and proven that any high-voltage in the 50–90 kV range is adequate for dental intraoral imaging. There is a broad optimum of X-ray hardness (i.e. kV) at 70 kV for best image quality—depending, however, also on the detector, the diagnostic task, and the observer preferences and habits.

In the same timeframe, in new products the “short cone” formerly used (or “coning device” actually shaped as a pointed cone) was being replaced with a “long cone”—and concurrently, the older bisecting technique was generally supplanted by paralleling technique (note*). This was made possible by the greater X-ray beam flux (dose rate) attained with the higher kV, which is required for retaining adequate dosage (without increasing too much the exposure time) with the larger source-to-detector distance that paralleling technique implies. This also brought about the wide adoption of film-holder and positioner devices (inconvenient with pointed short cones and bisecting technique), especially in USA where full-mouth surveys are common.

[Note *: a cone (or coning device) combines in a single device a positioning indicating device (PID) and a collimator, also formally known as beam limiting device (BLD).]

It should be clarified that when I say “short” and “long” in this context, I mean typically about 10 cm (4”) and 20 cm (8”) respectively. Later, when all new collimators become suitable for paralleling technique, a commercial “long cone” would typically mean 30 cm (12”), a size that never become really popular. As to the 40 cm (16”) cones (obtained with special modifications of the standard ones) which are occasionally mentioned in scientific literature, essentially they have never existed outside some academic milieus. The 80’s also saw the introduction of the first rectangular-field dental collimators, which slowly begun to attain popularity initially in Sweden, then in North Europe in general.

There is a particular model of dental X-ray source that deserves a special mention for its unicity: General Electric GE 1000 (later by Gendex). This device (or its direct predecessor) was developed and launched in the 50’s and was produced virtually unchanged for 50 years, quite a record! Undoubtedly, it must have been state-of-the-art when it was first launched, but become a very mature and ordinary technology in the subsequent decades. However, it possessed a unique feature: like in the bigger X-ray sources for medical radiology, the operator had the capability of directly setting all three technique factors—anodic current (mA), AC anodic high-voltage (kVp), exposure time (s)—independently from each other and over a broad range. Of course, the control was achieved with rotary switches, potentiometers and analog dials, given the technology of the time. This made it precious in certain dental schools for teaching and investigational purposes—although not much so for the general dentist who rather seeks ease and immediacy of use. It was also quite sturdily built, as it weighted twice as much as the average dental X-ray source, with a cyclopean suspension arm that let it be slammed against the wall by careless students, without too much damage (to the arm, not to the wall!). It was produced and sold until the early 2000s in ever dwindling minuscule numbers, before being discontinued to the dismay of dental old-school instructors.

It is a mantra of all industry that it must be “market-oriented”, i.e. it must strive to satisfy the expectations and needs of the user, both expressed and unaware. But the practical application of this principle usually passes through asking the input of the salesman in the field. Hence, “market-oriented” often becomes the needs and expectations of the salesmen, who may be more skilled and motivated at immediately selling products rather than with technology and clinical applications. What almost all salesmen want above all is the same product as their competitors, but better performing, bigger, faster … and less expensive; and the better performance should be quantifiable with a scalar number. For a dental X-ray source that quantifiable scalar number initially was the kV, which was requested to increase above that offered by competitors, even if the optimum had already been attained, under the idea that “the bigger, the better.”

As an example of the above, in the 80’s the company for which I worked, that had already updated its offering for North America with a 70 kV model, was requested by its sales force to augment the product line with a 75 kV model (AC), launched in the second half of that decade. It was a memorable commercial failure, with only very few hundred units sold in North America over a span of several years. Scaling up with kV is not at all linear with size, weight, and cost, and in fact the resulting design was bulky and expensive, and images generally less contrasted (hence, less palatable) than with the prior models.

In the 80’s, the advent of DC dental intraoral X-ray sources, that is with direct current (DC) anodic high-voltage supply, changed this paradigm. DC X-ray sources offer many advantages over AC (especially in association with the fast digital detectors that would come into use in the 90’s, as we shall see, less so with chemically processed X-ray films): lighter construction, much more accurate exposure time and possibility to impart very short irradiations, stability of the technique factors; the radiation hardness with a DC X-ray source is approximately equivalent to that from an AC X-ray source with anodic voltage 5 kVp higher.

The first commercial DC dental intraoral X-ray sources was the American-made Intrex by SS White in 1980,¹ the second was MinRay DC by Soredex of Finland in 1984.²

There is no basic technical reason why a DC and an AC dental X-ray source of good quality construction should have a significantly different cost (although a cheap design can be achieved for AC systems, without compromising too much the reliability, by renouncing on features and performances). But at the beginning, DC systems were offered at prices generally higher than AC systems, on the ground of the additional cost for the electronic converter board, not present in the AC systems. So, initially salesmen needed an argument to push the new technology into the market. However, the argument chosen, the “scalar figure” usually flaunted, was an erroneous one: it was claimed that DC technology involved significantly less radiation dose for the patient than AC.

I believe that, to some extent, this questionable belief originated from a misunderstanding of the early studies conducted on the first DC dental intraoral X-ray sources, including studies conducted around 1985 by Erkki Tammisalo, Ebba Helmrot and Olof Eckerdal in Sweden, in which the radiation dose from Min-Ray DC was measured and compared with that from the Philips Oralix 65 AC dental X-ray source.^1–4 In these studies, it was found that the DC unit implies a slightly lower (a decrease of 20% or less) entrance dose (skin dose) to the (simulated) patient. Immediate comparison of entrance dose readout between X-ray sources of different kind is problematic, since a different set of technique factors may be optimal for the different cases. But the main problem was that many recipients of this study didn’t understand (or choose not to understand) the difference between entrance dose (or skin dose), and total absorbed dose and/or equivalent dose (E). Entrance dose alone is an easily measured practical index that can be associated with radiation-induced risk only in a general way, whereas small differences are substantially irrelevant (especially at the very-low-doses involved with dental radiology). A higher kV (=harder) X-ray beam deposits less energy to the skin than in the deeper tissues, but the situation is further complicated by the shielding action of the intraoral detector itself.

At the beginning, the adoption of DC dental intraoral X-ray sources was frequently fraught with reliability problems, and the failure of the power stage of the converter board was not infrequent. The technology for (moderate) power electronics was still relatively immature in the 80’s and early 90’s, and technical competence was not easy to find. We had to wait for the next millennium for this technology to fully mature and achieve complete reliability. I have observed that, oddly, systems designed in Europe tended to fail more frequently in America, and vice versa; I believe this might have been caused by the different nature of the electrical power mains in the two continents (not just voltage and frequency, but, e.g. also network impedance and protection devices).

The 70’s and 80’s saw much activity not only with the X-ray source, but also with the X-ray detector, i.e. at that time the X-ray film. The discussion resonated in the professional and scientific literature about whether and when the faster intraoral X-ray E-film would eventually supplant the prior D-film in the market (the even-faster F-film was also considered). Proponents of the E-class film contended and demonstrated, in many scientific articles, that the E-film—when properly processed—offers essentially the same image quality, contrast, sharpness, and noise properties, as the slower D-film, therefore it should be universally adopted because of the advantages in reduced radiation dose to patient (and also less cause of movement blurring). “When properly processed”: this was the key; many ordinary dentists were aware that their chemical processing of dental X-ray films frequently run short of being optimal (spent developer, wrong temperature and time, inadequate safelights, …). The perception was that the slower D-film was more tolerant to mishandling, whereas in such condition the E-film turned out grainier. This has slowed down the general replacement of the D-film by the E-film, especially in USA where the latter never fully took over the former—until the rise of digital radiology made the matter moot.

The United States was also rather dissimilar from the rest of the world in the much more widespread adoption of automatic film processors at (small) dental clinics, starting from the late 80’s. Undoubtedly, this was caused by the frequent habit of submitting patients to periodic full-mouth intraoral radiographic surveys, a practice not universally adopted elsewhere. This required processing many dental intraoral X-ray films in a short period of time, which was inconvenient and lengthy if done manually (or semi-manually). The industry offered various models of automatic processors for dental X-ray films, which for the most part were designed essentially as a scaling-down of the larger models used in medical radiology. One notable exception was the fairly compact British-made Velopex by Medivance (still being produced), which transports the X-ray films through the processing baths via a nylon mesh.

Panoramic radiography

Notwithstanding the importance of intraoral radiography, the device, technology, and application whose expansion and consolidation characterized the last quarter of the XX century was panoramic radiography. Panoramic radiography as we know today had been conceived and also practically developed back in the 50’s, with the pioneering work of Yrjo Veli Paatero in Finland, who had been anticipated—but with different and more embryonic technical approaches—e.g. by K. Heckmann in Germany and by Hisaji Numata in Japan and supplemented, e.g. by Sydney Blackman in UK, J. Duchamel in France, and D. Hudson and Henry Hollman in USA.^5–9

By the 70’s, many commercial models were established in the marked, notably the Orthopantomograph by the Finnish Palomex resulting from the implementation of the original works of Paatero by Timo Nieminen, and co-produced and marketed by Siemens (that provided the X-ray generation sub system); and, e.g. Panoral by Ritter X-ray; Panellipse by General Electrics; Panorex by SS White; various models by Japanese brands, e.g. Asahi Roentgen, Yoshida, Hida, and J Morita (Figure 2); the old Rotograph produced by the short-lived Watson in UK and the subsequent homonymous long-lived Rotograph (1975) by FIAD (later Villa Sistemi Medicali, Italy) (Figure 3); and Cranex by Soredex. In America, this technique was generally known as “panoramic,” in Europe and also in Japan as “pantomography” or “orthopantomography” (OPT, OPTG) (a term that was claimed by Siemens as its own trade name, but that had already been used priorly in public scientific literature since 1958, e.g. by Paatero himself). Finn, Japanese, American, Italian, German, and lately Korean brands have dominated the world market for this kind of devices ever since.

Figure 2.

Panex by J. Morita, a panoramic machine from 1971. (Picture courtesy of Yoshito Sugihara, J.Morita Mfg. Corporation)

Figure 3.

Rotograph (first series) by FIAD, a panoramic machine of the late 70’s. (Picture courtesy of Fabio Curti, Villa Sistemi Medicali)

In USA, initially the fortune of panoramic radiology was greatly facilitated by its procurement by the Armed Forces, who needed an effective method to mass screen military personnel during and after the Korean War and later the Vietnam War, also for the purpose (sadly) of post-mortem identification of corpses via the dentition.

The technology of the early systems focused on implementing the proper cinematic trajectories through purely mechanical means, such as cams and/or eccentric levers. All movements—that of the rotating gantry and that of the film cassette—were synchronously driven by a single electric motor. With this design, essentially only one kind of trajectory and projection was possible in a given machine. Compared to today, the early systems were fairy massive, weighting several hundred of kg and requiring shipment in separate parts and partial assemblage on site; sometimes, the walls had to be reinforced in order to bear the weight of the machine. However, these machines usually offered the advantage of being very durable.

The SS White Panorex displaced the position of the rotational centre for the right and left side for better projection geometry and to avoid casting the shadow of the spine onto the anterior teeth. Initially, this was achieved by shifting (by a predetermined distance) the patient himself (i.e., the chair onto which the patient was seated), in a later design by shifting the mechanical centre of rotation in the machine. The general consensus was that the image quality was quite good geometrywise, but the overall image was made of two contiguous pieces, separated by a white (radiolucent) band at the middle. This made counting the (partially overlapping) teeth in the dentition less immediate, which was deemed by many to be inconvenient. The Panellipse by General Electrics owed its name to the elliptical trajectory of the centre of rotation. The European machines stemming from Paatero work soon implemented projection geometries based upon the uninterrupted shift between three rotational centres (for the right, central, and left parts of the jaws).

Cranex by Soredex, Finland, was the first model featuring a DC power generator for the X-ray source, in 1977. The company was established as a start-up by a group of technical managers who defected from Palomex when the top management at Siemens stated that DC power generators would have no market in dental radiology.

It was in the 80’s and early 90’s that panoramic systems become a common staple of private dentists (at least in Europe, North America, and Japan). That period also saw substantial advances in the detailed comprehension of the theoretical basis of rotational panoramic radiography, through many studies by the late Ulf Welander (then at the University of Umeå, Sweden, and Editor-in-Chief of DentomaxilloFacial Radiology) and of his allied team at the UTHCSA led by William Doss McDavid.^10–13 Besides providing for a comprehensive mathematical framework for rotational panoramic radiography, they also pioneered (with a working prototype) the foundations of digital panoramic radiology.^14,15 They used a linear array of sensors (stripe arrays as customary today were not yet available). No “tomographic” effects are possible with a strictly linear array, and their practical results showed (surprisingly to many) that the tomographic effect is not essential to produce panoramic images similar to those we are accustomed to see. Among the very many other contributions to the theory of panoramic radiography, the work by E.H. Tammisalo, by Jan Van Aken, by Gerard Sanderink, by Masaru Shiojima, and by William Scarfe should not be forgotten.¹⁶

I run myself into an instance of the bumblebee myth in the mid-80’s, when I set forth to lead the development of a new panoramic system (Philips Orthoralix SD), where the complex movements required to implement the panoramic projections would be driven, robotwise, by a set of independently operated stepping motors under the control of a microprocessor (“software-driven,” rather than by mechanical cams, levers, and chains), one of the first of this kind to become commercially available.¹⁷ Some senior expert at the headquarters’ Research & Development facilities declared that this approach could not work, because the intrinsically discrete movement originated by stepping motors would result into intolerable striping artefacts. Like the bumblebee, I went forth on my way, and of course, no striping was even detected by anybody (the mechanical inertia of the system damps out completely the discreteness). The same technology soon was gradually adopted also by all other manufacturers; in 1986 also Planmeca had launched a panoramic machine (PM 2002 CC) with microprocessor-controlled operation, and—for the first time—patient-oriented sideway (instead of facing the wall) which facilitated positioning (Figure 4). This new technology made possible selecting different types of radiographic projections in the same system, attuned to the specific diagnostic task, such as, for instance, special projections for the TMJ and for the frontal sinuses. This was (and still is) a feature greatly advertised to demonstrate and promote the high class of a panoramic system (although I think that only a tiny fraction of all users actually take much advantage of it). One of the features most sought by several manufacturers, and most difficult to satisfactorily implement, was the transversal scannographic cuts (a.k.a. transtomography) of the jaw (a sequence of linear pseudo tomographies) for the purpose of implant planning, which was subsequently somewhat superseeded by the advent of CBCT. Soredex’s Scanora (1988) (Figure 5) and later Cranex Tome (1996) were panoramic systems that included a small-field spiral tomography function for the maxillofacial region, primarily for transverse tomographic cuts of the jaws for implant planning purpose.

Figure 4.

Planmeca PM 2002 CC (late 80’s), one of the first panoramic machines to be microprocessor-controlled (with SCARA robotic technology), and the first to feature patient standing at 90° from the wall, thus facilitating its accurate positioning by the operator. (Picture courtesy of Timo Muller, Planmeca)

Figure 5.

Scanora by Soredex, a panoramic system that included spiral-tomographic capabilities for cross-sectional views of the jaws and the TMJs (early 90’s).

The new-generation panoramic systems produced and marketed since the late 80’s were also superior in that all used DC X-ray generators, and generally had a leaner and lighter construction than their predecessors. From this time, almost all commercial panoramic machines had the option—or a version—of an additional arm with cephalostat (sometime nabbed a “teleradiograph”) for cephalometric radiography, for orthodontic planning purpose. The geometrical arrangement of such arm and cephalostat with respect to the X-ray source is supposed to implement the same projection geometry (with Source–Object Distance of 150 cm = 60 in.) according to the criteria in the fundamental Broadbent-Bolton “Roentgenographic Cephalometer” described by Wingate Todd in 1931, which is generally considered a basis for cephalometric tracing and orthodontic treatment.^18,19 I will not enter into the specific details of this, but I observe that a seemingly simple device as a cephalometric arm has been the object, over the years, of a surprisingly vast and complex assortment of insights and controversies, about humble items such as, for instance, the wedge filter and the ear plugs – much of which is now superseded by the advent of digital and three-dimensional (3D) technologies.

The 80’s also saw the demise of a technology that had previously enjoyed some popularity with a niche of users: intraoral panoramic radiography—offered by Siemens (Status-X), Philips (StatOralix), and Koch & Sterzel (Panoramix). These systems used a tiny “pencil-like” X-ray source to be positioned inside the oral cavity, with the detector (the film with intensifying screen) wrapped externally around the jaws. The resulting image was quite sharp, but of course the geometrical accuracy was completely lacking. Also, having a high-voltage X-ray tube placed inside the mouth was not well accepted by the patient, and the skin dose to the oral mucosa and the effective dose to the other radiosensitive organs in the region was dismayingly high by today’s standards, so with the growing awareness about radiation risk reduction, this technique was dismissed.

Other uncommon devices and technological applications saw their culmination in the 80’s and—for the most part—their demise in the subsequent years, in conjunction with the spread of digital imaging. For instance: Zonarc by Palomex/Siemens, a remarkable variation of rotational panoramic radiography for supine patients (instead of standing or seated) featuring different selectable modes of projection (Figure 6); and the Axial-Tom, Com-Cat by ISI, Quint Sectograph, dedicated linear or spiral tomographs for the head (especially for the TMJ).

Figure 6.

Zonarc by Siemens (early 80’s), a panoramic machine for supine patient with multiple choice of image layer shapes.

As in the case of intraoral imaging, the 70’s and 80’s witnessed significant progress also with the detectors for panoramic and cephalometric radiography, which then consisted of cassettes with film and intensifying screen combinations. The antiquated calcium tungstate (CaWO₄) intensifying screen, dating back to the early times of radiology, was replaced by new materials based upon rare-earths compounds, where the wavelength of the emitted fluorescent light was specifically attuned to the spectral sensitivity of the corresponding X-ray film. There was a small number of important manufacturers, notably Kodak, Fujifilm, 3M-Ferrania, DuPont, and Agfa/CaWo. The battle raged in the market and in the scientific and professional publications as to whether the “blue” or the “green” combination was superior, until the introduction of digital imaging rendered the issue moot. A topic of note was that the new rare-earth-based screens—with the possible exception of the DuPont make—generally provided more contrast (perhaps too much contrast for the panoramic application) with respect to the older CaWO₄ screens.

Digital intraoral radiography

The first approaches to electronic imaging pertinent to dental and maxillofacial radiology was xerography, pursued with various experimental studies during the 70’s and 80’s.^20–23 However, it never got much traction in industry and the clinical arena. Digital imaging, the great technological revolution in dentomaxillofacial radiology, begun its practical maturation in the 80’s.

When talking of digital imaging in radiology, it is customary to divide between computed radiography (CR) and digital radiography (DR). The former technology (CR) (which is also referred to as photo-stimulated phosphors plates or PSP, and also as storage phosphors) is regarded by some as an incomplete form of digital imaging (because of the analogy with the conduct in chemically processed film-based “analog” radiography), transitional toward the latter (DR) which would be the only ultimate and full form of digital imaging. I disagree with this perception, and see both CR and DR as two valid technologies that complement each other for different applications in oral radiology. I am comforted in this view by the observation that PSP has retained its diffusion and popularity in dental and maxillofacial radiology a quarter of a century after its appearance (although not drawing much attention anymore in scientific literature: perhaps all has been written already!).

But the first foray into digital or electronic imaging for dental radiology came with DR technologies, at the beginning of the 80’s. My own first exposure to such innovative technology took place in May 1984, at a business product meeting in Stresa, Italy, where I met a young Paul Van de Stelt, then freshly appointed professor at the recently established ACTA in Amsterdam, NL. He presented the prototype of a novel fluoroscopic intraoral system at which he had been working in co-operation with the Philips corporation. The imaging part of the system consisted of a hook-shaped fiber-optics, with a scintillator screen glued to the front face, that conveyed the image forming on the scintillator screen to a video camera at the opposite end, all of which was rigidly connected to the X-ray source supported by an articulated arm. Interestingly, the video chain was not perceived as the only important part of the system; the ergonomics and the geometrical arrangement of the X-ray source also was. A lot of thinking had gone into how to maneuver the source & detector gantry while retaining isocentric projection for the object being fluoroscoped. All this was evidently inspired by C-arm interventional fluoroscopy, which then constituted the kind of electronic imaging already popular in general radiology.

It is difficult to understand, from today’s perspective, that at that time the main, or only, scope conceived for electronic radiography in dentistry was to check in real time (i.e. live) the progression of a endodontic file down the root during a canal treatment, to prevent drilling beyond the apex into the bone. Hence, interventional fluoroscopic C-arms was the reference application.

In addition to Van de Stelt’s team in Amsterdam, also another team of Philips LEP in Paris was co-operating with Jean Pierre Camus of Reims, France, in a project very similar in scope and technical approach to the former (aside from the fiber-optics not being curved). A live demonstration with a prototype of canal treatment on a volunteer, in front of a public audience of hundreds of French dentists, was organized as a part of the 1988 meeting of SITAD (Salon des Industries et Techniques des Equipements pour l'Art Dentaire) in Paris. The outcome was an embarrassing fiasco. The equipment stopped to work after a fraction of a minute, midway through the treatment, during which only vague shadows of something that could have been constructed as an endodontic file and the outline of a tooth were discernable on the large screen above the surgical theatre.

This outcome should not be surprising. Handling microchips and other miniaturized and delicate parts necessary for the video chain is a risky and tricky business even for well-equipped laboratories and technically competent research team, and the reliability of early prototypes is finicky, until a robust manufacturing know-how is established, which may take years. Taking into account the modest X-ray flux from a dental radiographic source, the inherent inefficiencies and losses in the optical chain, and the approximately 20 ms time window for each frame, it was to be expected that the image signal-noise-ratio would be awfully poor, which was in fact the case. Furthermore, it become clear that the radiation cumulative load to the operator (the dentist, even more than to the patient) would be inacceptable. It became also evident that a real-time live fluoroscopy in endodontics would be very impractical, or impossible, because the dentist could not simultaneously keep the detector in position and drill the tooth; and unnecessary, because the progress of the file could be checked with a discrete sequence of images as well. At that time, I was in charge of supervising the business and technical feasibility of those projects, and after that event I immediately recommended to the company that they be stopped and the focus be moved away from fluoroscopic to single-shot techniques.

Small-area CCD image sensors were already well established for optical video cameras, in combination with optical components. But for intraoral oral radiology one needs a sensor with an active area sufficiently large for direct coverage of at least one tooth (which was considered to be about 20 × 30 mm), because X-rays cannot practically be refracted and focused like optical light can (except in several-metres-long orbiting space X-ray telescopes). The problem was that such large-area sensors didn’t exist yet; the feasibility of fabricating them with sufficient reliability, production yield, and not-too-astronomical cost was questioned and doubted by many; and the mere investment cost necessary for production even of one prototype was in the order of half a million dollars, money that companies were not willing to venture for a yet unproved business. The closest thing existing (i.e., the largest sensor commercially available) was the 7883 CCD from Thompson Semiconductor that measured 9 × 13 mm (and was meant for video cameras, not radiography).

Already in the early 80’s Francis Mouyen, a French dentist from Toulouse, had embarked in a path that would start the industrial and practical application of electronic imaging in dental radiology. Mouyen was not a trained engineer or a physicist, in those fields he was rather a passionate amateur, but he understood well the practical needs and expectations of his profession, and had vision and ebullient enthusiasm. Having filed for patents since 1983, he persuaded Trophy, then a small dental X-ray equipment company active predominantly in their French domestic market, to fund him with a very modest grant to build, demonstrate, and validate the prototype of a dental intraoral electronic radiography system. It worked. He had the brilliant idea to resize the visible-light image radiographically forming on a 17 × 26 mm intensifying screen down to the 9 × 13 mm of the CCD sensor, through a carefully fabricated short tapering fiber optics, shaped like a squat truncated pyramid. The whole detector, 14 mm thick, could thus be accommodated inside the oral cavity.

Francis Mouyen was the bumblebee that, notwithstanding the lack of sound technological background, the paucity of funds, the absence of a prior references, dared to fly high and went where no man had gone before. He should forever be remembered and commended for that accomplishment.

That led to production and commercialization of Trophy RVG 25000 in 1987 (RVG staying for Radio Visio Graphy) (Figure 7).^24–27 To today’s standards, the performances of these early RVG systems were rather questionable. Image quality was far below the current standards; the bulky detector was uncomfortable for the patient and difficult to position, notwithstanding an active area so modest to border usability; and the price was quite high compared to what we are accustomed now. In spite of these shortcomings it was an instant success, at the beginning in France then in the rest of Europe and elsewhere, so great was the need and demand for electronic imaging. The commercial success in France was also helped by the fact that the French pertinent authority later established a reimbursement rate for dental X-rays made with this new technology, at a level fivefold higher than with the chemical-processed X-ray film. This rang a loud bell for all other manufacturers of dental radiographic equipment, who then realized that investing in electronic imaging was not only possible and desirable, but even a necessity in order to retain prominence in the market.

Figure 7.

A young Francis Mouyen and a young patient being radiographically examined with Trophy RVG 25000 (late 80’s), the first dental electronic imaging system. (Picture courtesy of Bruno Ehrmann, CS Dental)

The shrewd reader may have noticed that I have carefully refrained from using the term “digital imaging” in the last few sections, and used “electronic imaging” instead. This is because the technology implemented in the early RVG systems, and their prototypal predecessors, although “electronic” was not “digital”. Images were not captured and exposed as digital files (in some standard format like, e.g. the now common JPEG or TIFF). Instead, they were transmitted to the analog CRT display as a continuous video signal. When another radiograph was taken (or if the equipment was switched off) the prior image was lost forever. The only way to retain a record of the radiograph was to print it with a (optional) thermal paper printer. Of course, this bears the imprint of the fluoroscopic interventional C-arms of that age, where the “digital memory” was an extra option. This was to change very soon, and already the second version, RVG32000 launched in 1990, had the option for producing TIFF digital image files, through a frame-grabber and the associate software program mini-Julie.^28,29

In partnership with the Trophy company, Francis Mouyen applied for a number of patents with broad claims for products that they would manufacture. In essence, they tried to secure the priority for any kind of electronic imaging application in dentistry. In fact, when the first direct competitors appeared, initially Trophy attempted to aggressively use his patent to shut the others off the market. At the 1992 exhibition of the ADF in Paris, a Trophy representative surveyed the booths of competitors accompanied by a uniformed policeman (a “flic”), threatening them of immediate coerced closure and harsh penal consequences for patent infringement. In fact, the French patent application that was ultimately granted by the patent authorities described a very specific invention encompassing a combination of various features that had all to be present. The extension to dentistry of technologies that were already established for general radiography was not considered to be a significant invention. So, the claims in his patents had to be drastically restricted to certain specific technical aspects of the product, resulting into Mouyen’s second patent related to the X-ray detector, filed in 1992 and published in 1995. Maybe, the legal strictures of the patenting process have hampered Francis Mouyen from taking full credit and full advantage from his innovative impetus.

For those who are not familiar on how the patenting process works, I need to clarify that granting a patent does not imply that the Patent Office validates and certifies that the invention actually works, or that it is based upon solid scientific grounds. The essential duties of a Patent Office are just to verify that: (i) the patent application and description is formally structured in the prescribed manner; (ii) the “prior art” is described and referenced; (iii) the invention and a “preferred embodiment” is fully described (it is supposed that a prototype has been fabricated); (iv) the “claims” are fully and clearly stated; (v) the same invention or parts thereof, leading to the claims, has not been previously publicly “disclosed” (in the country for which the patent is applied for), for instance by publishing in the press and media, or at conferences, or by any other openly accessible manner; and of course that (vi) all fees and taxes have duly been paid. That the invention really works the way it is alleged or makes scientific sense is not a concern of the Patent Office. I think that was also the case for some aspects of the patents that Mouyen got granted. In fiber optics, a small fraction of the fibers may be made of opaque extramural absorbent (EMA) fibers, interspersed among the mass of other optically transparent fibers, whose purpose is actually to suppress the light scattered out of the optical fibers, which would otherwise fog the image. One essential feature of the Mouyen patents was that the fiber optics should incorporate “metallic particles” translucent to visible light but blocking X-rays (sic !?) (presumably EMAs made of high-atomic-number material) to shield X-rays from directly hitting the underlying semi-conductor chip, and damaging it. A simple physical analysis reveals that this assumption is unjustified. 12 mm of glass of the fiber optics interposed between the entry face and the semi-conductor chip already absorb some 90% of the X-ray traversing the scintillator screen, and the presence a small percent of fully X-ray-absorbing EMA would not significantly change it.

I believe Mouyen may have been frustrated by the failures and malfunctioning that he must have encountered with the very early prototypes (for the reasons to which I have already alluded), and convinced himself that the cause was radiation damage. This led to an amusing episode that I will report later. It has also had one long-lasting repercussion: scintillator-coated fiber-optics faceplates (typically 3 mm thin) for intraoral detectors are still advertised as advisable in order to provide radiation protection for the underlaying semi-conductor sensor. But semi-conductor sensors specific for X-ray imaging are designed to withstand, without degradation, cumulative loads of radiation dose which are orders of magnitude greater that what they can reasonably undergo in a lifetime of service in dentistry. I am not aware of any commercial detector ever returned from the field with failures or symptoms of degradation that could be attributed to radiation damage; not even one.

In the immediately following years, i.e. the early 90’s, a real avalanche of innovations hit the dental radiology milieu, with several intraoral digital imaging systems becoming commercially available, that used a full-area sensor without the inconvenience of image-resizing fiber optics.

The first one was Visualix by Philips Dental Systems (soon to morph into Gendex Dental Systems) launched in October 1991 at the Expodental trade show in Milan, in conjunction with the annual FDI meeting. It was a truly digital system with a flat intraoral detector using a purposely developed (and radiation hardened) CCD sensor 18 × 24 mm, 288 × 386 isometric pixels 63 µm, 8-bit grayscale, directly coupled with a Gd₂O₂S (gadox) scintillator (after an initial stint with direct X-ray conversion in the silicon sensor). It operated as a peripheral of a standard personal computer (DOS-based, Windows wasn’t out yet), and saved images in standard file format (TIFF, JPEG, BMP) that were stored in the computer memory and could be exported (Figure 8).^30,31

Figure 8.

The author at the EADMFR congress in Turin, Italy, in July 1993, holding an early version of the Gendex Visualix dental intraoral X-ray digital detector, with holder/positioner), next to a Gendex Oralix DC, one of the first direct-voltage dental X-ray sources, with rectangular collimator/PID. (Picture courtesy of Yoshihiko Hayakawa, Kitami Technical College). PID, positioning indicating device

That was followed almost immediately by Sens-A-Ray of Regam Medical Systems, with very similar characteristics, developed in Sundsvall, Sweden, by Per Nelvig (a dental radiologist) with the help of his brother Lars Nelvig (an electronic engineer), in cooperation with EEV, a large English electronic components manufacturer, and with some (modest) financial support by the Swedish government.³² The Nelvig brothers had worked at the development of their invention for some years, and a functional prototype was built already in 1985, but the paucity of financial resources caused the project to progress at relatively slow pace and delayed industrialization.

Many other companies followed suit in Europe and elsewhere (Siemens, J Morita, …) and by the mid-90‘s the users could choose from a diverse range of products. The Flashdent by Villa Sistemi Medicali, released in 1991, used a design for the detector also based upon resizing (focusing) the scintillating image onto a smaller sensor via optical media; this approach was fraught with some inherent physical difficulties, and ultimately did not enjoy much success. Trophy also abandoned the design with the truncated-pyramid fiber optics, which must have been the most expensive component of their system and a was cause of positioning difficulties and discomfort for the patient, and adopted a full-area sensor with RVG-S in 1993.

The spread of general awareness and knowledge about dental digital radiography was greatly promoted by a series of five symposia—“Digital Imaging in Dental Radiology”—organized by Paul Van de Stelt and Gerard Sanderink of ACTA, between 1990 and 1998 in the Netherlands (this bi-annual meeting was later succeeded by the computerized maxillofacial imaging (CMI) congress founded and convened by Allan G. Farman). The second symposium held in Amsterdam in June 1992 was especially memorable. Various manufacturers introduced their newly developed digital dental radiographic systems. Among those, Ulf Welander presented Sens-A-Ray, on behalf of Regam Medical Systems, showing intraoral images obtained with a new full-area CCD sensor that directly converted X-rays into charges in the silicon substrate. At the conclusion of the presentation, Francis Mouyen, who was in the public, contended from the floor that those images could not possibly have been taken with a detector like the one described, because the unshielded direct exposure to X-radiation would promptly deteriorate the sensor and cause the appearance of severe permanent artifacts. Undeterred, Ulf Welander calmly turned around to watch at the X-ray images being displayed, and just commented: “Well, if these are artifacts, they are artifacts that look pretty much like teeth.” That settled the matter.

At the same meeting another momentous event took place: from Louisville, Kentucky, Allan G. Farman demonstrated the transmission in quasi-real-time of a digital dental radiograph freshly taken with a Trophy RVG, as an attachment to an e-mail. In the associated abstract, he concluded that “Pilot investigations using electronic mail networks … have proven successful for the transmission of 32 kByte images with intercontinental relay in as little as 20 s. However, the rate of transmission does vary with traffic in the system and image reconstruction on receipt can take several minutes. …”. The audience stared in amazement.

This far, digital dental radiographic systems had spread mostly in Europe, but by 1992 David Schick of New York filed for a US patent for a dental intraoral radiographic detector, and later begun production.³³ In 1997, he filed for another patent about using a so-called active pixel sensor (APS) CMOS; this is a technology that was already well-known and established at that time, but he could have the patent granted by restricting the scope to X-ray detectors for dental radiography, a limitation that the US patent office deemed sufficient for the requirement of “innovation” (whether or not that could have stood a rigorous legal challenge, is left to be seen).³⁴ Anyway, Schick had the merit of toppling the cliché that the (less-expensive) CMOS technology was not suitable for X-ray imaging, and inferior to CCD, and rapidly obtained a vast commercial success in the USA.

In the years to come, the debate resounded in the professional press about the presumed superiority of the image quality with CCD-based detectors over CMOS-based ones, or vice-versa, arguably incited by the commercial interests of companies relying on one or the other technology. Certain articles in the scientific press compared apples to oranges, i.e. compared and rated the performances of specific detectors as if being CCD or CMOS was the only determinant, neglecting a multitude of other unrelated factors. Time has demonstrated that the argument was irrelevant, and the image quality achievable with a CMOS sensor is at least as good as with CCD sensors, plus CMOS offers other benefits such as the greatly reduced power consumption (which in turn brings along various advantages) and lower manufacturing costs. In fact, almost all manufactures have gradually converted to CMOS, including the successors of Trophy (at the beginning a resolute supporter of CCD) with a product that still stands among the best available.

By the early 2000s, a real plethora of digital intraoral detectors from a multitude of vendors was commercially available.³⁵ In their case, the quantifiable scalar number touted to support their worth was, and still is, the “geometrical (or theoretical) maximum spatial resolution,” in line-pairs/mm, although, being a theorical maximum, that is just one of the several factors affecting the actual practical spatial resolving power or acuity, and the latter being just one of the factors (together with signal-to-noise ratio and grayscale latitude) that determine the overall imaging diagnostic value.

Alongside with the hardware (i.e. the detector and the associate electronics) also the related imaging software moved forth in great strides during the 90’s. There were (and still are) two distinct lines of conceptual architecture: those of imaging software programs basically developed by the equipment vendors, and those originated by software houses already active in the dental environment with practice management software, that is programs to organize the patients and materials workflow in the (private) dental practice and the related administration. The former focus on processing and analysis of the X-ray images themselves, the latter on database arrangement of images and patient information. Interestingly, in certain countries (notably France and USA) a lot of traction to the diffusion of intraoral digital imaging systems came from the practice management software, that had already introduced the required computers, and familiarity with the same, to a vast network of private dental practices. The software houses perceive the dental imaging system as an optional adjunct to their practice management software, and the interconnection between practice/patient management and imaging often is achieved through a “bridge” between two unrelated (and often poorly harmonized) pieces of software programs.

This has brought along the still-debated and not well-resolved issue of intraoperativity and accessibility of image and patient databases between equipment from different vendors. Starting from the late 90’s, a lot of efforts have been put into addressing this through the DICOM standard (Digital Imaging and Communication in Medicine), especially with the work of the DICOM Working Group for Dentistry, devotedly led and animated for several years by Allan G. Farman and composed of representatives from many of the major dental industries and from academia. DICOM has become, for many years now, an indispensable and undisputed requisite for medical diagnostic systems in hospital environment. However, the situation with private dental practices is quite different. Most dentists do not feel the urge to share their images and records with other users outside their practice (and may not even know what DICOM is). One might wonder why DICOM hasn’t anyway promptly become the standard universally adopted by the manufacturers. The problem is that: (i) at least initially, it did not comprise provisions specifically suited to dentistry; and (ii) it is (or is perceived as being) extremely complex, to the point that very special competences are required to properly implement it—competences that the large medical equipment corporations necessarily have, but the medium-to-small size dental company often don’t. Consequently, implementation of DICOM functionalities by dental companies was often achieved through standard software libraries licensed from specialized independent software houses, at not-negligible cost per individual license. Such cost is relatively minor in relation to that of large medical radiographic equipment, but not so for the less expensive dental radiographic equipment, hence it is usually offered as a separate pay-for option. Only a fraction of dentists has proved willing to pay for such option. Over the years, the DICOM standard and its cognizance has evolved for better, and its diffusion in dentistry has increased with the appearance of the more complex CBCT systems, but the situation is still in progress.

In 1997, the economist Clayton M. Christensen (recently passed away) authored a best seller book by the title “The innovator’s dilemma.”³⁶ The central thesis expounded in the book is that large, mature, market-dominating corporations are structurally incapable of disrupting innovation (a term that he coined), because new technologies would become profitable only in the medium-to-long terms, while in the short-term they would compete against those established products that make their immediate profit (that short term profit against which the management is gauged and rewarded or penalized by the company’s proprietors and/or the stock market analysts). The only way for such corporations to innovate would be to spin-off or acquire smaller autonomous entities. I think one of the best examples of this construct was Kodak in the 90’s. At the beginning of that decade, Kodak dominated the film market and the X-ray film market as well, with over 90% market share in America and similar leading positions in many other parts of the world. As the ascent of digital radiology in dentistry become apparent, in 1990–1991 Kodak Medical Systems reacted by setting up a scientific advisory panel, staffed with several among the most illustrious dental and maxillofacial radiologists in the world, with the purpose (more or less explicit) of … finding arguments to support the superiority of the X-ray film and slow down the adoption of dental digital radiography! The fear was that their highly profitable market of X-ray film (and chemicals, and screens) would be compromised (as in fact the case was in a few years). In retrospective that was like trying to stop a waterfall with bare hands. The paradox was that, in the same time span, the photographic film and camera division of Kodak was resolutely trying to migrate to digital photography in earnest. In a few years Kodak Medical Systems must have realized that they also needed to jump on the bandwagon right away, and caught the lost time by acquiring Trophy in France, Practiceworks (a major software house for practice management software in USA), and in 2005 Orex (an Israeli manufacturer of dental panoramic CR scanners).

The practical availability of X-ray digital detectors has made possible the development and application to dental intraoral imaging, since the very early 90’s, of tuned aperture CT (TACT), a kind of tomosynthesis that provides stereoscopic or pseudo-3D imaging from a small number of conventional radiographic projections taken with different geometry (i.e. from different angulations). Its application to dentistry was developed largely thanks to the researches of Richard Webber and coworkers at the University of North Carolina, and has seen practical implementation in commercial equipment, notably by Instrumentarium of Finland that enabled it in panoramic machines since 1997. TACT is said to provide superior diagnostic performances, respect to conventional two-dimensional radiography, e.g. in the detection of caries and cracks, and in the analysis of periodontal bone. However, it has not spread to general adoption and, so far, remains mostly a tool for a minority of specialists and aficionados.

Digital panoramic radiography

The 90’s saw the adoption of CR also for dental and maxillofacial applications. The production of photo-stimulated phosphor plates and of the related scanners for medical radiography, and the relevant patents, was (and still is) essentially monopolized by three great medical device manufacturers: Fuji Film in Japan, Agfa in Europe (via the controlled CaWo in Germany), Kodak in North America.

Already in 1985 Isamu Kashima from Kanagawa, Japan, had shown that photo-stimulated phosphor plates could be advantageously used for panoramic radiography.³⁷ A distinct practical advantage of PSP and CR in panoramic radiography was that the user already owning (or having access to) a film-based X-ray machine would not need to scrap and replace his (expensive) panoramic machine in order to “go digital.” He would just need to replace the intensifying-screen cassette with one with a PSP plate, and add a CR/PSP scanner; in a large radiological clinic environment, the latter might have already been available for general medical radiology.

But the early experimental field applications of this technology encountered a bizarre obstacle, whose explanation requires a short excursus about how electro-medical companies set the specifications of their products. Generally speaking, they are specified to conform at least with the minimum essential provisions of accepted international standards. Conversely, the standards are written to encompass and describe what the larger and leading manufacturers have produced and successfully distributed to the market. For X-ray films—and, by extension, intensifying screens, and, by further extension, PSP plates—the applicable standard is published by ISO (International Standards Organization), and substantially describes the sizes offered by Kodak over a long period of time. However, the cassette hosting film and screen is considered to be a part or accessory of a medical electrical equipment, and as such is of pertinence of another standardization body, IEC (International Electrotechnical Commission). One may think that the sets of sizes foreseen in both standards coincides (since standard cassettes are made to accommodate standard films and screens and plates) and in fact that is the case … with one exception: panoramic size (15 × 30 cm and the now obsolete 5 × 12 in.). Initially, most medical CR scanners already available in hospitals had not been designed to accept cassettes in the 15 × 30 cm panoramic size, although PSP plates of that size could be easily made (e.g. cutting down a larger format). Probably, most of the people from large and leading manufacturers that manned the working groups of IEC at that time were not involved with panoramic machines (and possibly barely knew about them), and were not aware of the need to include that size in the standard, and in the capability of the CR scanners.

In the early 90’s I had worked at projects to circumvent this obstacle, like e.g. a modified panoramic machine that would carry a 24 × 30 cm cassette with a PSP plate where only the lower part would be exposed to the X-ray beam and produce an image for the customary 15 cm height when scanned, and to other like contraptions. But it was clear that dental-specific CR scanners would be eventually needed for panoramic (and cephalometric) and for intraoral radiography as well, also because the ordinary dentist customer could not afford the large cost of a regular medical CR scanner.

However, the first commercial dental PSP scanner was for intraoral plates only, the Digora by Soredex launched in 1994. Basically, its design was a scaling down of the architecture used for large-area medical CR scanners, based upon a linear progressing laser scanning beam, and it would do one intraoral plate at the time. Since a cassette cannot be used intraorally, dental intraoral PSP plates must be complemented with disposable opaque plastic covers or pouches (or “barriers”), having the double purpose of preventing cross-infection among patients, and avoiding the ambient light to partially erase the latent image on the plate after the radiographic exposure and before the scanning.

That was followed in 1997 by DenOptix from Gendex, developed in California, the first dental PSP scanner to accommodate both panoramic/cephalometric size imaging plates and intraoral plates. At about the same time, also Orex, an Israeli start-up, also presented a scanner for panoramic and cephalometric PSP plates (but not intraoral); this was the only company to use Kodak PSP outside Kodak itself, in fact it was later acquired by Kodak. After a few years also Dϋrr Dental of Germany (Air Technique in USA) launched a dental panoramic and intra oral PSP scanner (Scan-X), and other companies followed later. All the above-mentioned systems adopted different kinds of design than the original Digora and the medical CR systems, e.g. based upon a rotating drum.

While the immediate yearning for digital panoramic had been quenched by PSP imaging plates and scanners (using existing panoramic machines), that for direct and immediate panoramic imaging, with a digital sensor, persisted.³⁸ Already in 1995 a kit (DXIS) for conversion of panoramic systems to “fully digital” was available from SIGNET of Catalin Stoichita, a start-up in Paris. The problem with such kind of add-on kits was that they required an invasive adaptation into the hosting panoramic machine, which—among else—would invalidate the warranty and technical support from the manufacturer of the same.

Digital panoramic radiography is usually achieved by the so-called time deferred integration (TDI), where the signal accumulating in the short (but not infinitesimal) width of the detector integrates in a manner similar to what physically happens with an analog detector (film or plate). Such mechanism is inherent with the operation of CCD stripe sensors, and is electronically emulated in case of CMOS stripe sensors.

By late 90’s, the time was ripe for fully integrated digital panoramic systems to be originally designed by various major manufacturers. In 1997 it was DigiPan by Trophy, which was distributed in OEM as an upgrade for panoramic machines of other manufactures like the OP100 by Instrumentarium, Panoura by Yoshida, Orthoralix DPI by Gendex, etcetera. In the same general timeframe also other vendors launched machines with their own digital detector: Dimax by Planmeca (Finland); Orthophos DS by Siemens (later Sirona–Germany); and by J Morita. Almost all other manufactures followed suite soon.

The need to complement direct digital panoramic also with (optional) direct digital cephalometrics was soon apparent. The latter was (and is) more costly than the former, given the need for a longer scanning detector to cover the full height of the head (typically 24 cm). Also, the cephalometric scanning time would be a few seconds, which is not well received by some orthodontists who wish instant imaging to preclude the (very hypothetical) problem of geometrical distortions caused by patient movements. The work is still in progress to speed-up the scanning or to achieve single-shot cephalometrics with a large area flat panel detector.

By the 2000’s, a new technology became available to enhance the diagnostic value of panoramic radiography: laminographic tomosynthesis reconstruction.^39,40 This technology was introduced and then patented by AJAT of Helsinki, Finland, a start-up founded in 2001 and led by Konstantinos Spartiotis and partially financed by the Finnish government.⁴¹ This line of development was also investigated in Japan at about the same time. AJAT implemented an add-on kit that combined a new type of high-speed, high-resolution and high-efficiency digital imaging detector made of a combination of single-crystal Cadmium Telluride (CdTe) detecting material with an underlying CMOS array sensor, and an image tomosynthetic reconstruction method consisting in the collection of a large number of raw images that are then processed by a reconstruction algorithm into a panoramic image after the acquisition. Actually, the CdTe-based detector and the reconstruction method are two separate and distinct aspects that complement each other advantageously but not necessarily. In fact, the laminographic tomosynthesis reconstruction has since been adopted in a number of commercial machines without being associated with a CdTe detector, e.g. by Instrumentarium, Vatech, and Sirona, offering the advantage that the position and shape of the position layer can be determined after the panoramic irradiation (conceptually, even automatically), thus relieving from the need of accurate patient positioning. In spite of early experiments, e.g. by Asahi Roentgen in Japan, the widespread adoption of CdTe-based detector has so far been slow, probably because of cost, production capacity, and patent infringement issues.

Cone beam CT

At the end of the 90’s, the other great revolution in maxillofacial radiology begun: volumetric radiography (3D) via cone beam CT (CBCT). The theoretical and mathematical basis of CBCT have been well-known among mathematicians for many years.²¹ In 1978, a sole CBCT system was built and installed at the Mayo Clinic in Rochester, Minnesota, for cardiac and pulmonary diagnostic applications.⁴² It was 5.6 × 6.3 m large, weighted about 17 tons, used 14 separate X-ray sources and image-intensifier X-ray detectors, and required two large shielded rooms to operate; it costed a fortune in yearly maintenance; and cardiology and pulmonary do not seem to be the best elective applications for CBCT, which requires a stationary radiographic object. Nonetheless, it was mightily popular and appreciated at Mayo Clinic, and was reluctantly decommissioned more than 20 years later, when the obsolescence of the main components made maintenance impractical or impossible.

Already since 1994 QR, a small start-up in Verona, Italy, then virtually unknown in the dental milieu, technically led by Pierluigi Mozzo and Attilio Tacconi, physicists, undertook to develop and patent a maxillofacial-specific CBCT, named NewTom 9000, using a 9” image intensifier detector (Figure 9). At the very beginning of 1997, the first three units were sold and permanently installed at the private radiology practices of Giovanni Polizzi in Verona, Italy, Paolo Sartorato in Noale, Venice, Italy, and of Silvio Diego Bianchi in Turin, Italy. The new machine was also presented at the 1998 ECR in Vienna, and a paper was published in the November 1998 issue of EJR, which is still one of the most highly cited articles in our specialty.^43,44 At the start, many experts expressed doubts that a system like that could work or have any diagnostic value, but the bumblebee kept flying.

Figure 9.

The prototype of NewTom 9000 by QR (1996), the first commercial dentomaxillofacial CBCT machine. (Picture courtesy of Attilio Tacconi). CBCT cone beam CT.

Nevertheless, at the very beginning the dental profession for the most part failed to notice this disrupting innovation, because the inventors and the company were not rooted in the dental milieu, and the journals and exhibitions where it was first presented were medical rather than dental and maxillofacial.

In July 1999, at the fifth CMI (computed maxillofacial imaging) section of the CARS (computer-assisted radiology and surgery) congress in Paris, the convenor Allan G. Farman announced that one of the planned featured lectures in the last day of the congress would be replaced by a different conference, to present an innovation of great importance. This was unheard off! I had never ever witnessed, at a major international congress prepared long in advance, that a planned lecture would be substituted at the last moment, except for force majeure. The next day Pierluigi Mozzo explained the operation of CBCT to the audience and showed clinical images, with an image quality that he himself admitted was just adequate, but—he said—could be expected to improve in the future. This time the audience of international maxillofacial radiologists definitively took notice (although his presentation never made into the proceedings of the congress, which had already been edited in advance), and the CBCT revolution took off in earnest (Figure 10). NewTom begun sales in North America in 2000, promoted by the renowned orthodontist Carl Guggino.

Figure 10.

An early NewTom 9000 installed at the radiology department of the Sundsvall Hospital, Sweden, with Per Nelvig and staff. (Picture courtesy of Claudiano Tagliareni)

Simultaneously, also in Japan practical work had independently begun to develop commercial CBCT machines, animated primarily by the indefatigable Yoshinori Araki, and by others.^45,46 J Morita introduced the first Accuitomo small-field CBCT machine after 2000. Shortly later, following the researches of Rika Baba and others, Hitachi also launched Mercuray, a sophisticated and powerful CBCT machine with a large field detector.^47,48 Baba and Ueda also investigated applications of CBCT in chest and orthopedic imaging, and the adoption of large-area flat panels detectors, which however did not result into commercial products.^49–52 But Hitachi was (and is) a purely medical company with no application experience and commercial footprint in dentistry, so their product was overdesigned (and overpriced) for dentomaxillofacial applications and achieved limited penetration outside Japan.

Initially, all available CBCT models utilized image intensifiers, bulky and heavy, as image detector, which constrained size, weight, and image quality. In the late 90’s, Predrag Sukovic, a post-graduate student of Biomedical Engineering at the University of Michigan at Ann Harbor, and his mentor Neal Clinthorne at the Department of Radiology of the Medical School, investigated at the application of the TFT flat panel as image detector for 3D radiography. It was already evident that the flat panel offers many advantages over the image intensifier, in image quality and else, and its cost was in the process of decreasing to a level compatible with industrial deployment. With the fundamental help by Sharon L Brooks at the Dental School (later Editor-in-Chief of DMFR), by 2002 they had prototyped a maxillofacial CBCT, based upon the mechanical structure of an existing panoramic machine (by Panoramic Corporation) and featuring a large-area TFT flat panel as image detector.^53,54 In 2003 Sukovic and Clinthorne started up a company called Xoran (which would later manufacture CBCT machines for ENT applications, initially with the name MiniCat), and licensed the design for dentomaxillofacial applications to ISI (Imaging Sciences International) of Hatfield, Pennsylvania, which re-adapted it on the mechanical structure of an old and discontinued panoramic machine of theirs, and since 2004 produced it with the name iCat (Figure 11). The first unit (“DentoCat”) was delivered and installed at the Dental School of the University of Michigan in February 2004. iCat enjoyed vast popularity in North America during the 2000s, thus fostering a veritable rush to implement this 3D technology by all other vendors in maxillofacial radiology (like it had been for digital radiography one decade earlier).

Figure 11.

A young Predrag Sukovic sitting in the first pre-production unit of DentoCat/iCat (that he was instrumental at developing), at the Dental School of the University of Michigan in 2003. (Picture courtesy of Neil Clinthorne, University of Michigan)

At the 2005 IDS in Cologne, Planmeca presented a prototype of small-field CBCT, with a 12 × 12 cm flat panel, that was built as a modification of their panoramic machine, hence more compact and leaner than the other system then available. The machine begun being commercialized 2 years later, in 2007. By the early 2010s, some 40 or 50 models from different vendors were commercially available.

The following decade, the 2010s, is recent history, and witnessed the continuous success, evolution, and expansion of CBCT. Among else, a vast variety of hybrid CBCT & panoramic (and cephalometric) machines have become available and popular (outnumbering the purely CBCT systems), with separate detectors for CBCT and Panoramic or, more recently, a single detector for both modalities. More and more machines have adopted the offset detector technique, that makes possible to increase—up to a factor 2—the radius of the reconstructed volume for a given width of the detector’s active area. This, however, requires that a scan over a complete 360° full rotation, or more, is performed, which may not be possible in machines whose original mechanical design was just for panoramic, where usually the rotation is only about 270°.

In the vast majority (or the totality) of maxillofacial CBCT machines, the 3D reconstruction is based upon the well-known and well-proven FDK (Feldkamp–Davis–Kress) filtered back projection algorithm (it is virtually impossible to have an accurate and comprehensive information on this, since all manufacturers jealously guard it as a secret). Various scientific studies have advocated the transition to other more advanced and accurate 3D reconstruction methods (ART, iterative, etcetera), which require more computing power and substantial software and mathematical development, but the challenge generally has not yet been taken up; evidently, in maxillofacial imaging the approximate FDK algorithm is good enough for the task (for now).

A technique that has its roots many years earlier with classic CT scanning has also received great impetus, that is the use of 3D radiographic data from CBCT scanning, and their integration and merging with visible-light surface data from facial photography and endoscopic intraoral optical scanner or camera, plus the associated 3D printing of surgical models, templates and masks, for the purpose of dental implant planning and computer-guided surgery.

Handheld dental X-ray sources

This historical excursus would not be complete without mentioning an innovative type of product that has not stirred very much attention in the scientific literature, but is proving a game-changer of its kind in the field: handheld dental X-ray sources (distinct from X-ray sources that are just portable). Probably, the first of this type of products was Nomad, designed and initially manufactured by Aribex of Orem, Utah, founded in 2003 and led by Clark Turner, a physicist. Nomad was initially intended for forensics and for special applications in the field, and was being presented at a forensic meeting in Bangkok, Thailand, when the terrible tsunami of 2004 hit the Far East coast. The few units that were present at the meeting, and others promptly shipped from the factory, were immediately conveyed to the disaster area to help in the rescue operations (identification of corpses, etcetera). This brought the product, and the very concept of handheld dental X-ray sources, into the spotlight. From that moment, its adoption spread speedily among dentists in the USA, where nowadays it is among the most popular dental X-ray sources. Almost simultaneously with Nomad, also several companies in South Korea entered the market with a variety of models. In Far East, the devices produced by Korean manufacturer (Vatech, Genoray, Dexcowin, Rexstar, …) have enjoyed a large, growing popularity. In continental Europe, the penetration is still marginal, mostly due to regulatory restrictions and radiation safety concerns, that many studies in USA have refuted as unfounded or excessive.^55–60

Conclusion

What else?

There is no conclusion.

The technological progress does not stop, does not come to an end or conclusion. We may expect for sure that the existing technologies and products will continue to enjoy evolutionary improvements in the years to come. There isn’t a crystal ball to predict what will be the next disruptive innovations, and when. But I dare to say that, at some future time, we may expect to see the ascent in practical, daily radiology, including dental and maxillofacial radiology, of:

• More effective X-ray image detectors, with dual energy and/or single-photon counting capabilities (and variation thereof), and new detecting materials.

• Phase-contrast X-ray imaging (as opposed to mass-attenuation radiography).

• Improved tomosynthesis.

• Artificial intelligence assisted diagnosis.

• Last but not least, a new generation of X-ray sources (for instance, tubes with cold cathode emission), after a substantial technological stagnation lasting for almost a century!