Accurate diagnosis and staging of cancer of the prostate continue to present problems for radiologists and urologists. Prostate cancer volume is a significant indicator of several tumor prognostic factors, including extracapsular spread. However, the use of sextant biopsy, transrectal ultrasound, or MRI to estimate cancer volume preoperatively has been disappointing. The challenge of achieving sufficient contrast and spatial resolution has, until now, been unanswered.

Researchers at Beth Israel Deaconess Medical Center in Boston have reported initial promising data using 3T MRI with combined pelvic phased-array and endorectal coils to noninvasively assess tumor volume using parametrically analyzed dynamic contrast enhanced MRI (DCE-MRI). The studies were conducted by a collaborative team of pathologists, urologists, radiation oncologists, medical oncologists, and radiologists.

While high-resolution MRI of the prostate at 1.5T has gained some acceptance for pretherapeutic staging of prostate cancer—most important, whether it is an intracapsular tumor or whether there is extracapsular spread of disease—researchers at Beth Israel and Harvard Medical School, Boston, wanted to evaluate the potential clinical application of combined pelvic phased-array and endorectal coils at 3T.

Axial images of the prostate (mid third). Note in the left peripheral zone an area with suspicious kinetics, colored red. The corresponding T2-W image shows little signal changes in that area. The tumor (biopsy-proven cancer on the left side, midprostate) cannot be delineated clearly. Note the high spatial resolution (prostate capsule and periprostatic vessels) as well detailed intraprostatic anatomy achievable at 3T. Images courtesy of B. Nicolas Bloch, MD.

"The reason we went there was signal. The currency of MR is signal," says Neil M. Rofsky, MD, director of MRI at Beth Israel and associate professor of radiology at Harvard. "With higher field strength, you achieve thinner sections, while preserving image quality. If your slices are too thin, the noise creeps in, and you won't be able to define the true details of the image that we desire."

In the combined coils study, six patients were examined using 1.5T and 3T scanners with pelvic phased-array surface coil and a disposable endorectal prostate coil. Beth Israel helped to modify the indwelling probe so that it would work well at the higher field strength.

According to the study's findings, published in the August 2004 issue of the journal Academic Radiology, the researchers were able to "acquire T2-W fast spin echo images with 1.5 mm slices, field of view 12, matrix 320 x 192 (voxel size 0.35 mm³), with excellent anatomic detail and good T2 contrast. A 1.5 mm axial slice thickness permitted high-quality multiplanar reconstructions with clear visualization of small patho-anatomic structures. Dynamic contrast-enhanced gradient echo images showed excellent spatial resolution (voxel size, 0.38mm³) and temporal resolution."

In their findings, the researchers stated that their initial results support the clinical utility of endorectal 3T for noninvasive evaluation of the prostate with image features and quality not achievable at 1.5T. Using modified T2-fast spin echo and dynamic contrast-enhanced gradient echo sequences, the researchers obtained whole gland coverage with 35-38 µm³ resolution, without interfering artifacts.

Subsequently, the improved image quality has facilitated the characterization of disease processes. Anatomic information in the dynamic images has enabled the researchers to clearly distinguish between intracapsular and extracapsular contrast enhancement. "If you were to compare a phased-array coil 3T prostate exam to an endorectal 3T prostate exam, the fidelity of the image with the latter is much greater, particularly when you start imaging with higher combinations of temporal and spatial resolution," Rofsky says.

SHIFT TO 3T FOR PROSTATE

The research has already impacted practice patterns at Beth Israel. Almost all of the work on the prostate is now being done at 3T. "3T offers the higher spatial resolution and gives us the opportunity to really push spectroscopy to the next level," Rofsky says.

While the question of whether 3T's greater resolution translates into greater diagnostic certainty is still being answered, Rofsky says it is "certainly going in that direction."

"It's translating into greater diagnostic performance, but we and other sites are still generating the data," Rofsky says. "I'm confident that 3T will add tremendous value diagnostically."

Researchers from Beth Israel and Harvard, as well as General Hospital Vienna and Medical University Vienna in Austria, and The Weizmann Institute of Science, Rehovot, Israel, recently began to investigate a second area of concern: patients who have a rising prostate serum antigen (PSA) and repeated episodes of negative biopsy results.

Currently, the MRI diagnosis of prostate cancer relies predominantly on T2-weighted imaging, a noncontrast approach that relies on a signal difference of the innate tissue between cancer and the background tissue. "That approach is not always able to detect cancers. It's a very tough study to look at," Rofsky says.

Spectroscopy, which looks at metabolic features of the gland and the tumors, is another option in detecting prostate cancer. "The pitfall of spectroscopy is that the spatial resolution is not particularly refined, and therefore, issues like volume [and] staging haven't worked out that well," Rofsky cautions.

"But it has been able to offer more specificity in the diagnosis—that is, some nodule that may appear could be due to inflammatory changes, and by adding the spectroscopic information, we can actually get information about the metabolites—and that will help give us a handle on whether the nodule is likely to be benign or malignant," Rofsky says.

NEW SOFTWARE APPROACH

Through the combined utilization of MRI and software known as three-time point (3TP), the researchers have been studying how to help practitioners better designate the location of a tumor. Rofsky says dynamic contrast-enhanced 3TP imaging, along with the advent of potential new agents, presents new opportunities down the line to monitor tumors noninvasively.

The 3TP software—developed in the 1990s by Hadassa Degani, PhD, a professor at the Weizmann Institute—uses a dynamic contrast-enhanced approach, which is then mapped parametrically as a color-coding of the likelihood of cancer, based on the kinetic properties of the tumor—with red indicating a high likelihood of cancer, green indicating an intermediate likelihood, and blue an unlikely instance of cancer. "The 3TP technology simplifies our evaluation and gives us a very refined analysis of the entire gland," Rofsky says. "If we see these red nodules, we can provide a biopsy target for the urologist."

Before 3TP, physicians would have to find an area they thought was suspicious, drop an electronic cursor over it, obtain signal intensity measurements over time, and try to identify the enhancement patterns.

Researchers at Beth Israel have studied the use of prostate 3TP since October 2003 to evaluate the accuracy of DCE-MRI in determining tumor volume in the prostate gland.

According to the initial reporting of their study, reported at the 2005 Society of Computed Body Tomography and Magnetic Resonance meeting and the 2005 International Society for Magnetic Resonance in Medicine meeting, MRI of the prostate was performed on a 1.5T unit with combined surface and endorectal coils in 10 patients prior to prostatectomy. For purposes of the study, the researchers acquired high spatial resolution dynamic Gd-DTPA enhanced gradient echo (8.1/4; 01 min 35 s; FOV 160 mm; 2 pre-, 5 post-contrast acquisitions) sequences. The DCE-MRI images were analyzed pixel by pixel with the 3TP pharmacokinetic model. Tumor areas were marked on whole mount histology and MRI slices by two readers, and the volumes of each tumor, total tumor volume, total volume of the whole gland, and the ratio of the total tumor volume/total gland volume were assessed with customized software developed in house. The MRI-derived volumes and the histopathologically determined volumes were compared.

Initial results of the study found that the parametrically analyzed MRI (via 3TP) detected 29 out of 34 tumors correctly, including all tumor foci >143 mm.³ There were five false positives (55-150 mm³), and five false negatives (20-143 mm³).The researchers found an excellent linear correlation (R= 0.89) between the tumor volumes (histology) and the tumor volumes determined by DCE-MRI.

"Our preliminary data is suggesting that we now have the ability to calculate the volume of tumor load, which is a new parameter that's never been accessible before this technology," Rofsky says. "That may provide very important information about the biology of an individual's tumor, rather than relying on traditional tables that summate likelihoods of cancer based on the digital rectal exam, the Gleason score [which uses a numerical grade to describe the cancer features from biopsy specimens], and the PSA."

Rofsky is hopeful that the ability to identify tumors and more accurately target biopsies will lead to earlier diagnoses of prostate cancer.

"We're not yet biopsying tumors under MRI guidance, but when our urologists come down to look at these images, they have a mental note of where they need to go, and they will biopsy with an informed map in their mind," he says. "We've been able to identify tumors, and the targeted biopsy results have been positive. So rather than repeating the 'blind' biopsy time and time again with repeatedly negative results, we can use the MR data to get to the diagnosis earlier and optimize the treatment options."

Danielle Cohen is associate editor of Decisions in Imaging Economics.