From CT imaging to progress made in image-guided radiation therapy, effective image management is crucial to oncology imaging. Storage, retrieval and tools for comparing images must all come together in a cohesive, seamless process for effective and efficient cancer diagnosis, staging, re-staging and treatment. New offerings are making image management itself easier to manage.
“The amount of images we’re dealing with in oncology is a hundred times more than just a few years ago,” says Vasudha Lingareddy, MD, medical director of the Edward Cancer Center department of radiation oncology in Naperville, Ill. The 40,000-square-foot facility opened in February 2005, and has seen double-digit volume growth each year since. As a result, plans are underway to open a second outpatient cancer center in another Chicago suburb in late 2008 or early 2009.
Filmless for efficiency
From day one, the facility was designed to be filmless. “We worked toward everything being electronic,” says John Fan, PhD, senior medical physicist. The center started out with a Clinac linear accelerator with On-board Imaging and obtained a Trilogy linear accelerator for IMRT within a year of opening, both from Varian Medical Systems. The equipment helps with image management because “it is easier than film,” says Fan. “It’s more efficient and the images are better. There is less wait time for imaging prior to treatment” which lets the clinicians treat patients more efficiently.
The center uses CT cone beam imaging, fiducial markers and image fusion for its cancer patients. Radiation oncology clinicians routinely use PET, MR as well as simulation CT images for treatment planning. These images must be fused together to provide correct tumor and anatomical locations. All of these techniques require a lot of storage. The hospital provides redundancy for the facility through its PACS. “Some people set up a little server in their department and they’re limited,” says Fan. “The whole thing has to fit into the general scheme of data storage.”
Fan has built upon the Varian Aria framework for image management. The team also uses Arian Offline Review software on the physician end to review cine and port films. “That gives us a task list for which films need to be checked. That software is very helpful in managing those images,” says Fan.
The hospital’s radiology department stores more data, but the cancer center is catching up. Since the facility has dramatically exceeded planned growth, more hard drives are added as necessary. “Nowadays, it’s pretty cheap,” says Fan. That’s a very good thing since “4D imaging and cone beam scanning take up a lot of storage space.” A 4D CT study set with 1,000 to 2,000 images can take 0.5 to 1 GB of storage. A cone beam study set occupies 30 MB. Daily cone beam CT scans for 45 days results in 1.4 GB of cone beam CT images for one patient.
An integrated solution
Fan and Lingareddy say it was an easy decision to go with single technology vendor, Varian. “The accelerator, treatment planning system and image management system all have to talk to each other,” says Lingareddy. “Integration is key. A piecemeal solution is nothing but a headache.”
The team used another treatment planning system in the past but “as the technology gets more complicated it has to seamlessly talk to each other,” says Lingareddy. “We’re moving toward dynamic adaptive therapy so the information has to flow back and forth instantaneously.”
That makes communication between different vendors’ systems impractical. Although data and images can be pushed through different systems with the DICOM protocol, any time you use the DICOM transfer you need a quality assurance program in place to check the integrity of data. Also, DICOM RT protocol does not include all the treatment parameters in radiation oncology. Some parameters will be lost during transfer/conversion. “Thinking about that amount of work is unbelievable,” Lingareddy says. As a result, “almost from day one, we decided to use an integrated solution. With an integrated solution, data and images from treatment planning to treatment delivery are stored in one database. There is no need for data/image transfer and conversion. Down the road, our field is heading towards dynamic adaptive radiation therapy which requires daily imaging and daily modification of treatment plan while the patient is waiting on the treatment table. Instantaneous information flow is a must.”
Storage is top of mind
For John DeMarco, MD, associate clinical professor, chief of clinical physics at the University of California-San Diego department of radiation oncology, storage is a big concern when considering new oncology imaging technologies. DeMarco has two linear accelerators from Siemens Medical Solutions, both with electronic portal imaging devices to take digital images to ensure patient alignment. While that replaced a lot of the hard-copy film the department used in the past, “with these devices, we have to figure out how to store the images along with the corresponding images we’re comparing against.”
When the UCSD department became a Siemens customer, they installed the company’s MB3000 archiving system. “The system provides very good workflow. You can archive from any machine, any CT scanner, to this archiving system.” However, the ongoing technological changes may mean that the system isn’t the ideal candidate for the current generation of equipment. “Because Siemens is selling so many systems because they work so well and people are excited about the technology, they’re now moving toward the next generation of archiving to better meet large dataset storage requirements.”
In general, “image acquisition and image guidance is taking off, but storage and archiving is lagging behind,” DeMarco says. A regular radiology department uses PACS with large databases for storing and archiving. Radiation therapy is moving toward a mini-PACS, he says, not a full-blown PACS. At the conclusion of a patient’s treatment, all of their digital reconstructed radiographs and matching portal images need to be stored and kept forever.
With more and more image-guided treatment—virtually all of radiation therapy is done with some 3D imaging—a separate storage process is required. “We can either store on the existing UCSD PACS or archive the images locally in the department in case the patient comes back for retreatment,” he says. “We need to remember what we did and what the anatomy looked like 5 or 10 years ago.”
On top of retention requirements, electronic portal imaging devices have led to treatment that is “a paradigm shift in how we do our image guidance for patient setup and positioning,” says DeMarco. “That comes at a cost. Instead of taking a single planar image once a week, now we could be taking a CT of a patient every day over the course of treatment.” While that provides an unprecedented way to evaluate whether the patient is in the correct position for treatment, it’s a huge challenge in image storage and image movement throughout the department.
UCSD recently installed Siemens’ INVISION Patient Care Solutions patient information system. That plays a role in reworking how these large datasets are archived. But with the massive amount of data, “now we have to keep better track of which patients are being imaged by which modality and of storage space on the local imaging device at the linear accelerator,” says DeMarco.
Data growth is a constant IT issue but budgetary concerns haven’t been a problem for DeMarco—yet. “If we keep track of the storage problem and we add storage on an incremental basis, we have no problems getting the expenditure approved. If we wanted to make a large paradigm shift, like spend a lot of money on a new PACS for radiation therapy, that would be difficult.” Eventually, he says, radiation therapy will turn into a separate radiology department with large-scale storage issues. “Radiation therapy will need its own sophisticated PACS.” There are scheduling and other logistical issues to consider, but it might be cheaper and easier for a hospital to invest in a mini PACS and manage it on a local level. “I like the local solution myself, but is it cost effective? That’s hard to answer at this point.”
Renate Muller-Runkel, MD, medical physicist at St. Margaret Mercy Health Center in Hammond, Ind., uses a Kodak 2000RT CR Plus system from Carestream Health which offers an all-in-one CR solution for therapy, simulation and dosimetry imaging. “Image management here is very important because it is of the utmost importance that the beam is in the right position,” she says.
The team inputs all new patients’ digital reconstructed radiographs into the system. Then, once the patient is on the treatment table, the therapist takes the cine film, read in the Kodak system, and compares it side-by-side with the intended target. The therapist can see how precise the beam positioning is—whether the target moved a half centimeter posteriorly or to the left or right. This process is done on a routine basis at least once a week, Muller-Runkel says, “so that we know our treatment is aimed exactly at the tumor.”
Once a physician has requested a change in position, another image is taken the next day before turning on the beam. Patients typically get one portal image done per week. Many are undergoing IMRT treatments, but some undergo conventional 3D therapy. Muller-Runkel’s team of two physicians and five therapists treats about 25 patients a day.
The hospital uses an electronic medical system so all data are stored redundantly and Muller-Runkel reports that the entire process has put an end to lost films.
She says she prefers the Kodak system to the new linear accelerators that usually come with a portal imager. Her system offers a larger field size that can be captured on a flat panel imager and increased accuracy. She would like to get a second machine for another location but the budget may not allow for that just now.
More images, more storage
The image volume required in oncology means more daily images and more work for physicians and physicists, says Jerome Landry, MD, a radiation oncologist at the Emory Winship Cancer Institute in Atlanta. The facility uses IMRT and IGRT equipment from Varian Medical Systems.
When patients receive IMRT and IGRT therapy, images are checked daily or every other day. “At Emory, we track how well the physicians are doing with weekly checking. We typically run about 90 to 100 percent compliance.” They do have to spend extra time to be compliant, however. “From my point of view, the technology is so advanced that the 30 minutes added to my day to get online and look at these images is not that big of a deal when I consider that I’m hitting the tumor more accurately.”
To help improve efficiency with extracranial stereotactic radiosurgery, Landry says the group has written a program called Aerial that allows for more accurate quality assurance of gantry angles and set-up precision. “When we do treatment, we can bring that data in quickly. That gives us more comfort with our treatment and helps us save time.”
Aerial allows for quick treatment in critical situations at any site. “One can remotely look at images from multiple sites,” Landry says. “The patient doesn’t necessarily have to come back to Emory if something needs to be done immediately. I can visualize images online in real time.”
The last word
Data storage requirements in oncology imaging are only going to increase. Each step taken to further imaging technology means larger datasets and more imaging. Vendors are aware of this and are working on more storage solutions to help providers keep up with the rapid growth.