Multislice CT emerges as gold standard for chest imaging

Boosted by the introduction of multislice machines, CT has been steadily overtaking other imaging modalities in visualizing the chest. CT is far more specific than chest radiography and faster and more global than nuclear medicine scanning. Although MR imaging is more sensitive to differences in fat and soft-tissue contrast and provides more precise demarcations of tissue planes, it doesn't return much signal from pulmonary parenchyma.

Patients often prefer MSCT to MRI because they can be on and off the table in five to 10 minutes instead of being confined in a magnet for 45 minutes or an hour. CT also has become capable of producing sophisticated image reconstructions. Coronal reformations that depict the thorax from the posterior/anterior to anterior perspective can be performed more or less automatically, and the paddlewheel reconstruction can home in on the pulmonary arteries and elucidate the winding routes of their branches.

Because multislice technology improves image quality and quickens acquisition time, radiologists can reduce the dose of radiation. Unlike the heart or the mediastinum and blood vessels, which require higher doses of radiation, the lung with its air-filled pockets is low in attenuation and naturally appears black, while abnormalities stand out in white. As a result, Princess Margaret Hospital in Toronto has cut radiation doses for CT in half over the last 18 months, based on phantom work.

Dr. Narinder Paul, head of thoracic radiology for the University Health Network in Toronto, expects to reduce radiation even further when examining patients strictly for pathology in the lung. Solid, variable-contrast, and unknown nodules could be detected at one-fourth the current clinical dose of 200 to 300 mGy/cm, according to findings from a paper presented at the 2004 RSNA meeting. Solid nodules could be found with a radiation dose of 50 mGy/cm or less.

Except for special cases-PET/CT for staging lung cancer, chest radiography for confirming the presence of severe acute respiratory syndrome, and MRI for evaluating regional lung involvement from asthma-MSCT is the new standard of reference for imaging the chest.

PULMONARY EMBOLISM

Pulmonary angiography was the frontline imaging tool for finding emboli in the lungs until the mid- to late 1990s, and Doppler ultrasound was the most common way to look for the origins of emboli in the lower extremities as recently as 2000. Both imaging methods have been replaced by a single spiral CT scan of the chest and lower limbs in most institutions in Europe.

In Lausanne, Switzerland, 1000-bed University Hospital performs 600 to 700 examinations for pulmonary embolism a year, and most are CT studies. Only about 150 are pulmonary angiograms, said Dr. Pierre Schnyder, chair of radiology.

Time-consuming Doppler ultrasound, which replaced venography in the lower limbs about 10 years ago, has in turn been replaced by rapid CT scanning to look for the origin of emboli in the legs, he said.

Ventilation/perfusion (V/Q) scanning, another mainstay of investigations for pulmonary embolism, is used much less frequently in academic medical centers and also in many community hospitals in the U.S., said Dr. Charles S. White, a professor of radiology at the University of Maryland, Baltimore.

The switch to CT that began in the early 1990s with single-array detector machines accelerated with the release of the first four-row detector scanners, which produced higher quality images in a shorter acquisition time. The move to CT has surged with the addition of even faster eight- and 16-row scanners, Schnyder said.

"Single-row machines were fantastic at the time, but compared with what we get now, they were rather slow. We could acquire 12 to 15 cm in 30 seconds, no more. Now, with 16-row machines, the same region of interest can be examined in five seconds, which is fine for 98% of patients. Almost any patient can hold their breath for five seconds, even if they have some degree of respiratory decompensation or impairment," he said.

Although single-row CT scanning has been highly effective at ruling out pulmonary embolism, with a negative predictive value of 99% to 100%, MSCT detects pulmonary emboli at a finer stage, said Dr. Geoffrey Rubin, chief of cardiovascular imaging and an associate professor of radiology at Stanford University School of Medicine. A study of more than 1200 pulmonary embolism examinations presented at the 2004 RSNA meeting found 41% of the emboli in segmental arteries, 27% in lobal regions, and 17% at the subsegmental level distally, which never would have been found with a single-row machine.

LUNG CANCER

Unless radiologists need to answer a specific question in the mediastinum, such as whether lung cancer has invaded the trachea or esophagus, MRI is not a major player in diagnosing lung cancer. Despite a number of improvements in faster scans, different pulse sequences, and various suppression techniques, MRI doesn't visualize small nodules, and the signal characteristics are not descriptive enough to be of much value in terms of differentiating benign lesions from malignancies, said Dr. David Yankelevitz, a professor of radiology at Columbia University.

PET scanning is limited because of its inability to see small nodules. Because it has difficulty drilling down to lesions less than a centimeter in size, PET's main value is in staging lung cancer, Yankelevitz said. The use of PET to check for spread of lung cancer is on the rise in the U.S. and Europe.

PET/CT is hampered by unacceptable rates of false positives and false negatives, according to Rubin. Small cancers may be negative on a PET scan, and large noncancerous lesions that are actively inflammatory can be positive.

Despite strong advocates such as Yankelevitz, many groups consider the data too weak to support routine CT screening for lung cancer. It is still unclear whether CT is finding early adenocarcinomas that eventually will grow and become malignant or precursor lung cancer lesions that will not ultimately kill a patient.

"Frequently, abnormalities are detected in the lung that are consistent with a diagnosis of lung cancer, but the majority will not be cancer. These are basically localized, parenchymal opacities that are less than 3 cm in diameter. What we've learned with the introduction of multirow CT is that the finer we cut the lungs with our sections, the more nodules we see," Rubin said.

As computer-aided detection systems increase even further the number of nodules spotted on CT screening exams, the output will be hundreds of lesions that have to be followed up even though they most likely won't be cancerous, he said.

Software that allows volumetric assessment of lung nodules may enable radiologists to nail down true changes in disease progression by adding a third dimension to the measurement of nodules over time, White said. Volumetric or flat-panel CT, which is in the prototype stage, may increase the precision of lung cancer detection by acquiring sections as thin as 300 microns.

"One of the problems with 2-mm nodules is that the diameter of those lesions approaches the thickness of the CT section. So our confidence in characterizing those lesions or even saying that they exist among the noise is limited. If we have a 300-micron section, we might be more confident that there really is a 2-mm lesion," Rubin said.

INTERSTITIAL LUNG DISEASE

In the past, radiologists would perform high-resolution CT and typically skip 90% of the lung. They essentially obtained 1-mm scans every 10 mm. With MSCT, radiologists can scan the entire lung in a single breath-hold and capture 1-mm sections throughout the parenchyma. The result is an axial evaluation of lung tissue as well as a direct coronal reconstruction or simply a coronal or sagittal view of an axial section.

MSCT can therefore discern gradients of interstitial disease from the apex to the base of the lung or appreciate conditions such as bronchiectasis that are seen more clearly on a thin, high-resolution, coronal image than on an axial scan, White said.

SARS

Imaging is performed primarily not to diagnose SARS but to exclude other causes of the symptoms, such as straightforward pneumonia or tuberculosis. Chest x-ray, the principal imaging modality, is inexpensive, quick, and portable. An x-ray unit can be taken directly into a ventilated isolation room to scan a patient, and it can be easily washed down afterward to reduce the chance of transmitting infection.

Yet chest x-rays are normal for 20% to 25% of patients at high risk for SARS because they have been in an environment where they could have come in contact with infected individuals. These patients often are sent for a low-dose CT examination of the chest to confirm or rule out the presence of atypical consolidation.

When needed, cross-sectional imaging has been carefully controlled to avoid removing patients suspected of having SARS from isolation, reduce the risk of contamination of other individuals while SARS patients are being transported to the imaging department, and eliminate the downtime required to clean and ventilate the suite following CT scanning, Paul said.

EMPHYSEMA AND SMALL AIRWAYS DISEASE

Although V/Q scanning has been used to assess ventilation, and quantitative software systems for measuring the degree of emphysema have existed for a decade, imaging tools now allow far more powerful and precise quantification of not only the progression of emphysema but the pattern of disease involvement in the lung apex or base that may be amenable to surgery.

Commercial software packages for CT determine how much of the lung is emphysematous by counting the pixels that fall below a specific HU threshold. This measurement can track the change in pulmonary dynamics following lung reduction surgery or the degree of stabilization or worsening of disease after medical therapy.

Volumetric reconstructions that spread out the trachea as if it had been excised and mounted for pathological analysis can be done throughout the respiratory cycle to assess the degree of collapse of the structure or to look for the presence of tumors and other abnormalities in patients with tracheal malasia.

MR ventilation that follows the path of hyperpolarized helium as it passes to the alveolar spaces following inhalation identifies areas of poor ventilation that may indicate emphysema or other ventilatory problems.

ASTHMA

For decades, the evaluation of asthma has relied on pulmonary function tests, but blowing into a spirometer provides only a global assessment of lung function. It cannot show whether the apical segment of the right upper lobe or the base of the left lower lobe is most affected by inflammation.

Pulmonary function testing also cannot determine how aerosols distribute within the lung, whether the drugs accumulate in their inflammatory targets, whether aerosols differ in their ability to reach the outer part of the lung, or whether the inflammatory response changes after treatment.

Investigators are beginning to consider the use of imaging modalities to analyze asthma more definitively. High-resolution CT could be used in a semifunctional way to explore, for example, the thickening of the bronchial wall as a result of chronic or recurrent inflammation during inhalation and expiration by examining the air traffic in affected parts of the lung, Paul said.

This would, however, require repeated exposure to radiation by asthmatic patients, most of whom are young. Therefore, researchers are leaning toward MRI, which is poor in delineating lung tissue but excellent for analyzing lung perfusion and ventilation by revealing the distribution of gadolinium or hyperpolarized helium.

"It takes quite a lot of infrastructure in terms of research science support to get hyperpolarized helium, which is not cheap. But I think it's the way forward," Paul said. "Now that we have seen CT technology take a huge leap with multislice scanners, we inevitably will see MR technology take a leap in order to compete. And when we get quicker acquisition, better software and hardware, we will see more MR applications in the lung, including asthma, which is on the horizon."

Ms. Sandrick is a contributing editor for Diagnostic Imaging.