General: Clinical 3D Imaging- Has Its Time Finally Arrived?

Elliot K. Fishman, M.D. , Professor of Radiology and Oncology , Johns Hopkins University School of Medicine

Three-dimensional imaging (3D) for CT applications was introduced shortly after clinical CT scanning became a reality in the late 1970's. Whether through work done by Gabor Herman and associates at the University of Pennsylvania or Mike Vannier and associates at the Mallinckrodt Institute of Radiology, 3D imaging was viewed as a way of abstracting more information from a series of transaxial CT scan slices. Not suprisingly early applications involved bone especially in areas like the skull and craniofacial regions (regions of high CT attenuation and anatomic zones less affected by patient motion or breathing). Although most radiologists at the time were not enthusiastic about 3D reconstructions our referring physicians found them extremely helpful in patient management decisions in complex cases. Over the next 15 or so years 3D imaging continued to evolve with the introduction of faster computer processing times, lower priced workstations with better price/performance profiles and new rendering algorithms (i.e. volume rendering). Yet, despite these and many other advances 3D imaging continued to be a study performed in a select group of institutions for a limited set of applications.

It is debatable why the progress of 3D imaging in the radiologic environment was so slow but a number of reasons have been suggested including:

• high cost of workstations
• perceived notion that 3D had limited clinical applications
• 3D was felt to be of value only to the referring physician but not to the radiologist
• difficulty in using 3D workstations due to poor system design
• a "killer app" (application) had not been developed
• major vendors like Siemens Medical Systems and GE did not push
• 3D as a mainstream product
• poor reimbursement for 3D studies (especially the physician component)

What really began to change the equation was the development of spiral CT and the ability to obtain true volume data sets which were ideal for 3D or volume imaging. With the continued development of spiral CT scanning from a technology than could acquire 12 seconds of data to a technology that could acquire up to 100 seconds of data things began to change. New applications for CT began to develop based on these new technologies. The role of 3D imaging was becoming more of a core function of CT and inseparable especially with applications like CT angiography and virtual endoscopy. The introduction of multidetector CT and its advances for vascular imaging continued this development cycle which has been driving 3D imaging to become more of a standard exam rather than a unique procedure. In fact, every scanner manufacturer now recommends or ships a workstation capable of 3D imaging with their high-end scanners (multidetector CT scanners (MDCT)). Yet, there is still the feeling among some radiologists that 3D imaging is not yet suitable for their practice. This seeming contradiction may seem hard to explain but is based on in great part on the resistance of radiology and radiologists to change.

The biggest limitations to the use of 3D imaging (and other post-processing tools) in the clinical environment include:

• a lack of understanding of the advantages provided by these techniques both from a clinical and patient care perspective
• a lack of understanding of how to use these new techniques
• a lack of understanding how to merge new technologies into a busy clinical practice that already may be overwhelmed by the volume of work and/or a staffing shortage (both radiologists and technologists)
• resistance to change especially changes in work distribution and flow

These limitations can be translated into a need for:

• education
• training
• clarification of work flow issues
• staffing

I believe that education can come from any of several sources including:

• CME courses including courses with hands on sessions. For example the RSNA as part of its annual meeting has hands on sessions on the use of computers including medical workstations. Siemens Medical Systems has sponsored a hands on course using the 3DVirtuoso workstation the past two years in Orlando.
• reading the radiologic literature (and pertinent literature from other subspecialties) and noting the clinical role of 3D imaging especially as it applies to CT angiography.
• getting information from the vendor of your workstation including detailed hands on training on the use of the workstation and better system documentation.
• Web-based educational sites like www.CTisus.com where all of the 3D protocols are available.

Training reflects more on the ability to obtain technical expertise on a 3D imaging system. Although every workstation vendor provides some form of hands on training it usually is but 2 days and this may be unsatisfactory for either the radiology technologist or radiologist. It is not suprising that the most common complaint about a workstation and its use is lack of sufficient training. This problem can be solved by either the 3D vendors providing enhanced training (including through web based training) on or off site or for the interested parties to go to sites with similar equipment and learn in a more hands on method. Unless there is improvement in the training available the use of 3D imaging will continue to lag other technologies.

Workflow issues and staffing are both separate but closely intertwined problems. The decision as to who does the 3D imaging (radiologist vs. technologist) and where the workstation is located are decisions that are made by individual institutions. Although my experience is one where the radiologists do the actual 3D imaging (including creating the images and filming them) other sites have found a dedicated technologist (with radiologist oversite) to be an ideal strategy. The advantages of the radiologist only works in cases where the radiologist(s) is dedicated to committing the necessary time and effort to the enterprise. This is becoming more of an issue where most institutions are understaffed and trying to cope with the clinical load without adding new studies. However, this is short sighted as using a technique like CT angiography will decrease the staffing (both radiologist and technologist) needed for more invasive procedures like classic angiography. In addition our view that future of imaging revolves around direct 3D viewing replacing axial CT scanned based imaging which will require primary radiologist participation. One factor that will increase the radiologist's willingness to be the primary person for the 3D image analysis is the availability of true real time volume rendering. The Siemens 3DVirtuoso with the VolumePro upgrade will make this wish a reality.

Other sites have found that a dedicated technologist can perform most of the routine studies and the radiologist works in a more supervisory role as well as doing the more difficult/complex cases. Advantages of this work flow relate to less of a commitment of the radiologists time and may provide more continuity especially in those groups where a radiologist is not based at any one hospital or office. In this model selection of the technologist is critical as they should be an individual who is willing to learn what the purpose of each study is and is committed to continuing education. The person must be self-motivated and committed to the project. The technologist will also need people skills to deal with both the radiologist and the referring physician. This workflow issue is critical to the success of any 3D program and will need to be decided on a case to case basis.

Although it would be ideal to have multiple workstations connected over a high speed network capable of doing 3D imaging this is rarely the case today. The decision as to where to physically place the workstation is therefore critical. I have found that it is ideal to have the workstation away from the scanner suite in a separate room or office. This allows consultation with referring physicians without interrupting the primary function of the CT scanner which is to scan patients. This separate 3D suite or lab allows for the centralization of function especially when a number of different scanners and/or modalities are networked to a single workstation. For example at Hopkins our 3D lab is connected by a 100 megabyte backbone to scanners in the hospital, the adjacent outpatient center, the adjacent oncology center, the emergency room and a remote site 10 miles away. All images seamlessly reach the workstation for postprocessing.

Workflow issues are obviously a critical factor in the success of a 3D operation. The timely performance of a CT scan will be negated if there is a time lag till the 3D images are generated. Although most 3D studies do not require an immediate turnaround other applications are very time sensitive. These applications include acetabular fracture repair (in select cases), suspected mesenteric ischemia and suspected aortic dissection. Training of enough staff members to cover these off-hours cases is needed to provide the 24/7 coverage demanded today.

The one problem with this central location is that as 3D visualization becomes a primary interpretation tool it would need to be located in the scanner suite or in the area where films are interpreted. This would potentially require a number of workstations which would be cost effective if used to enhance the primary interpretation. Implementation of this paradigm

Multidetector CT is probably the final brick that will push 3D imaging into the mainstream. Although I will not discuss the specific clinical advantages of MDCT it is easy to conclude that any 3D application that could be done previously can be done better due to a combination of factors including narrower collimation, higher resolution imaging and faster scan times. MDCT also has resulted in many for CT are now clinical reality. These include areas like coronary artery angiography and peripheral CT angiography. However, even more than that is the practical reality of MDCT. While in the prespiral era a scan sequence of 35-50 images were the rule with single detector spiral CT equally impractical (user may become to tired from looking at too many images). 3D imaging or volume imaging may prove to be an alternative. Viewing the entire data set as a volume with an interactive 3D real time display may be ideal. The ability to interactively segment out organs or organ systems will help with more accurate detection of disease as well as quantification of disease volumes. The use of an interactive mode will also speed up the viewing process for both the radiologist and the referring physician.

Providing this capability is a challenge today. In order to view images interactively it is generally accepted that 15 frames per second would be needed. This means that for a (table 4).

Obviously once images are viewed in a volume format 3D will become truly mainstream.

Conclusion

The modification of an established workflow pattern is difficult if not impossible to change. This is especially true if the old system worked well and its members are satisfied with its performance. To paraphrase an old saying " everyone wants progress but no one wants change". It is only when the system becomes unworkable or unsatisfactory does the window for change open. The introduction of MDCT will provide the impetus by creating an environment where a new paradigm will be needed. We look forward to these changes and the potential innovative solutions that will be its result.

TABLE 1

The factors, which are driving 3D imaging into the realm of a commonly used and accepted clinical study (by the radiologic community) include:

• a better understanding of the clinical value added of 3D imaging
• the growth of CT angiography and the demand for clinical studies by the referring physicians
• better reimbursements for 3D
• wealth of supporting data in the radiologic literature
• easier to use 3D workstations

TABLE 2

• a dedicated 3D lab
• anywhere there is space to put a workstation
• in the CT reading area
• in the referring physicians office, clinic and/or the O.R.
• near the CT scanner

TABLE 3

The biggest limitations to the use of 3D imaging and other post-processing tools are:

• a lack of understanding of the advantages
• provided by these techniques
• a lack of understanding of how to use these
• new techniques
• a lack of understanding how to merge new technologies into a busy clinical practice that
• already may be overwhelmed by the volume
• of work and/or a staffing shortage (both radiologists and technologists)

References

1. www.CTisus.com contains all the CT protocols for single and multidetector CT as well as complete references for all of the clinical applications