PET/CT guides management of cancer patients
Imaging centers learn radiotherapy planning techniques by actually doing them
By: Karen Sandrick
A year ago, nuclear medicine physicians and radiation oncologists at the University of Wisconsin's Comprehensive Cancer Center weren't sure how best to acquire data from fusion PET/CT scanning for cancer treatment planning, particularly targeted radiotherapy.
When they first obtained their PET/CT machine, nuclear medicine physicians usually positioned patients to gather anatomic and functional information for disease staging rather than for possible radiotherapy. Patients often had to be rescanned in different positions to simulate the conditions in which they would be treated, adding costs that were not reimbursed.
The physicians also followed a standard PET/CT imaging protocol, consisting of an ultrafast low-dose CT followed by PET. This protocol did not provide enough precision for radiotherapy planning, said Dr. Minesh Mehta, director of radiation oncology at Wisconsin.
Now that the center has been using the combined modality routinely for patients with cancer of the lung, head and neck, esophagus, cervix, lymphatic system, and occasionally the breast, PET/CT imaging follows a distinct radiotherapy protocol. After patients are placed in a body mold, face mask, or other immobilization device, they receive the FDG isotope. At the time of scanning, patients undergo a low-dose CT followed by PET. As soon as PET imaging is completed, the CT couch is indexed back to front, a power injector is hooked up, and patients receive full-dose intravenous CT contrast.
In transmitting full-dose CT imaging data along with low-dose CT and PET findings to the computerized radiotherapy treatment planning system, the protocol fuses functional information with anatomic detail without artificially increasing the uptake of FDG associated with the transformation of CT scans into 511-keV transmission maps. Iodinated contrast media look like water at 511 keV, and correction on the basis of Hounsfield units produces an artifact by making blood vessels appear darker than surrounding tissues.
The Comprehensive Cancer Center is tackling other PET/CT methodological issues as well. Mehta and his colleagues are exploring software solutions to correct motion differences between CT and PET in imaging tumors in the chest and abdomen. There is always some degree of tumor displacement between the CT and PET data sets in the thorax, he said, because CT images are obtained during one phase of the respiratory cycle while PET scans are averaged over several breathing cycles during the 20-minute acquisition time.
The Wisconsin physicians are also trying to determine the appropriate standardized uptake value threshold that differentiates malignancies from benign masses in the radiation treatment planning system. At this point, no one knows whether the SUV should be 4, 7, or 10, Mehta said.
His colleagues at the center are reviewing their experience and testing other tracers, including fluorodeoxy-thymidine (FLT), to overcome image interpretation errors such as PET cold spots that turn out to be cancers and hot spots that are benign.
"Sensitivity and specificity for PET/CT are still emerging, and differences in findings may depend on tumor heterogeneity," Mehta said. "Tumors that do not grow very fast or do not have a high metabolic rate may not pick up the PET isotope and therefore appear to be cold. Other conditions, like inflammation, may be superimposed on top of malignancies and produce false-positive results."
Yet even in the face of procedural trial and error, Mehta is documenting phenomenal changes in radiotherapy treatment as a result of PET/CT. In direct comparisons of radiotherapy treatment plans based on PET/CT and on CT information alone, he is identifying when the tumor volume differs by 20% or more, the radiation dose fluctuates by 10% or more, or the overall treatment plan changes altogether, from curative to palliative intent.
Mehta calculates an index of conformality, which assigns a value of zero when tumor volume contours determined by CT and PET/CT imaging are identical and a value of one when the contours are completely different. The mean conformality index is 0.43 in early studies of patients with esophageal cancer and 0.49 in patients with lung cancer. These figures indicate that with PET/CT the average tumor volume treated by radiotherapy varies by 40% to 50%, he said.
SIMILAR RESULTS
Other cancer centers are seeing similar effects on treatment. Dr. Arnold C. Paulino, an associate professor of radiation oncology at Emory University, estimates that PET/CT has altered the management of 35 patients with intact tumors by about 20%. He has found that he can provide better radiation coverage of tumors in patients with non-small cell lung cancer by differentiating malignancies from atelectasis on PET/CT.
Paulino has also widened the scope of treatment to encompass otherwise occult cancers. He expanded the treatment plan for a 60-year-old man with oropharyngeal cancer when PET/CT revealed a nodule in the lung that turned out to be a second primary cancer.
Radiation oncologists at the University of Zurich in Switzerland are more confident about reducing the volume of tumor to target for treatment when they are presented with PET/CT findings rather than CT data alone. The standard deviation error between two oncologists declined from 30 mL with CT to about 10 mL with PET/CT in one study, said Dr. Gustav von Schulthess, director of nuclear medicine.
It is far too early to determine whether changes in cancer treatment based on PET/CT data will positively affect disease recurrence rates or mortality, but some radiation oncologists are reporting improved quality of life among cancer patients. Based on imaging data from PET/CT, the tumor treatment volume has dropped 25% to 40%, and the radiation dose has increased 20% to 25% at Holy Name Hospital in Teaneck, NJ. As a result, fewer lung cancer patients have difficulty swallowing after radiotherapy, and patients with stomach cancer have less nausea and loss of appetite, said Dr. Charles Vialotti, chief of radiation oncology.
Even when PET alone is sufficient for visualizing cancer in lymph nodes in some patients, PET/CT is streamlining the workup. According to Dr. Perry Grigsby, PET has revolutionized the treatment of women with cervical cancer. Until PET, radiation oncologists could tell only that lymph nodes were enlarged in the pelvis, para-aortic area, and supraclavicular regions on MR or CT. But PET flags disease involvement whether lymph nodes are enlarged or not, said Grigsby, a professor of radiation oncology at the Mallinckrodt Institute of Radiology.
In a study of 101 women, Grigsby found that while CT identified only 77 cervical cancers, PET confirmed malignancies in 100 women. PET detected three times more abnormal lymph nodes in the pelvis than CT (67 versus 20) and abdomen (21 versus seven). And although PET found eight abnormal supraclavicular lymph nodes, CT found none. Nevertheless, it is not PET/CT's ability to provide a precise location of cancer activity that is helpful to Grigsby because he doesn't confine radiation to a few millimeters; he treats involved lymph nodes and several centimeters of tissue around them. The fusion technique is valuable, he said, because it avoids another imaging test such as a CT or an intravenous pyelogram to learn whether a patient has hydronephrosis.
NEW TERRITORY
In addition to cancer treatment planning, PET/CT is beginning to carve out new territory in cancer treatment follow-up. Dr. Jacqueline Brunetti, associate medical director of radiology at Holy Name Hospital, points out that physicians are in the very early stages of evaluating the effectiveness of cancer treatment with PET/CT. But because the fusion technique reveals reductions in the biological activity of tumors, which occur far more quickly than decreases in anatomical tumor structure, it may be preferable to CT in the immediate post-treatment period. PET/CT might be used as soon as the first month after chemotherapy or within three months of radiotherapy, she said.
It may also be suitable for imaging patients between courses of therapy, just after radiotherapy for microscopic disease and before a boost dose to the primary tumor. And it may add accuracy in the follow-up of patients treated with chemotherapy. The more explicit localization of tiny abnormalities with PET/CT than with conventional PET may reduce the risk of missing a persistent or recurrent cancer, Brunetti said.
By identifying sites of persistent or increased abnormal FDG uptake or new hot spots, fusion PET/CT may detect early recurrences of cancer, said Dr. Carlos Perez, chair of radiation oncology at the Mallinckrodt Institute and Barnes Jewish Hospital in St. Louis.
LEARNING CURVE
Because PET/CT is so new, investigators are still learning not only about what it can add to cancer treatment and follow-up but how it should be done to enhance image quality. One question, von Schulthess said, is how much CT is needed. Can imagers get away with a low-dose CT? Do they need to add a high-dose unenhanced CT? Or do they need IV contrast-enhanced CT on top of high-dose unenhanced scans?
As a rule, von Schulthess obtains a low-dose CT for transmission correction, and when requested, a contrast-enhanced CT, which takes another 30 seconds. Other centers in Germany, such as the University of Essen, administer a low-dose continuous IV infusion of contrast material.
An area that has not been fully explored is whether specific CT protocols should be designed for given diseases. In one of the first studies to examine this issue, von Schulthess found that standard PET/CT was sufficient in lymphoma patients; a contrast-enhanced CT did not provide further diagnostic information.
Respiratory gating is another object of debate. Gating sharpens the appearance of lesions on PET, which affects chemotherapy and radiotherapy planning but not the surgical approach. Surgeons use PET/CT images to orient themselves to the patient, the extent of disease, and the path to malignancy, according to von Schulthess.
But respiratory gating to achieve perfect coregistration of PET and CT is not critical for diagnosis, he said. Even if the PET lesion is not directly on top of the CT lesion in the chest, it can be used as a pointer so physicians know they will find a CT lesion within 1 cm or a maximum of 2 cm. If a lymph node next to the liver remains hidden, an unenhanced CT can be followed by a contrast-enhanced study. Lesions in the chest are usually not difficult to find because of the soft-tissue air contrast on CT. And motion inferior to the diaphragm and along the uppermost part of the abdomen can often be corrected by breath-holding.
Holy Name Hospital and Memorial Sloan-Kettering Cancer Center, among others, find that respiratory gating makes sense, however. These centers provide gated radiotherapy to patients with tumors in the chest or abdomen and to candidates for external-beam limited partial breast radiotherapy. The latter is delivered through an implantable balloon to the localized lumpectomy bed.
The two hospitals are participating in a multicenter trial of respiratory gating that will accrue between 300 and 400 patients over the next six to nine months. An infrared camera will be used to record a patient's natural breathing pattern by tracking the movement of sensors on the chest during inhalation and exhalation. A fluoroscope will continually track changes in the position of target tumors with respiration until it finds a portion of the respiratory cycle in which the tumor is essentially immobile. The PET/CT scanner then will be restricted to acquiring images at three points in that time period.
Because everyone's breathing is different, some patients' tumors remain still when they are at full inspiration, some move less when they are at full expiration, and others are somewhere in between, Vialotti said.
"We are determining the portion of the respiratory cycle when the tumor moves within a certain predefined range-2 or 3 or 1 mm-and programming the PET/CT scanner to image the radioactive glucose concentration or the anatomy at that same phase of the breathing cycle," he said.
Newer, more tumor-specific radioactive tracers are also in development. FDG-PET, the workhorse tracer for PET/CT, is primitive in many ways, according to Mehta, and it is not active in some malignancies such as prostate and brain cancer.
Radioactive choline has been more useful in PET/CT imaging in a small number of men with prostate cancer, von Schulthess said. Tyrosine will probably be a more effective agent in illuminating malignant hot spots in the brain if PET/CT proves to have a role in brain tumor imaging.
(FLT) has shown promise in assessing women after chemotherapy for breast cancer. A study reported at the 2003 meeting of the Society of Nuclear Medicine showed that FLT-PET more accurately predicted a response to the breast cancer antigen CA27.29, as well as the overall tumor response to chemotherapy, than FDG-PET in 11 newly diagnosed women with breast cancer.
Tracers have been slow to emerge primarily because the PET market has been small. Only when a solid base of PET scanners is installed worldwide will interest in formulating new radiopharmaceuticals grow, according to von Schulthess.
But neither the dearth of radiotracers and the limited number of PET/CT scanners, nor the continued tinkering with imaging practicalities will erode the role of PET/CT in oncology treatment planning and follow-up. PET/CT has already replaced CT for some types of cancer, such as bronchogenic carcinoma. And the technology is becoming indispensable in others as it gains acceptance by clinicians.
"Physicians have developed so much confidence in PET/CT, the cases are usually clear," von Schulthess said. "And it tells everything about the patient. It's a one-stop-shop oncology exam."
MS. SANDRICK is a freelance writer in Chicago.
|
