The oldest standard study in radiology is the plain film of the chest. This classic study still has value, but today, other methods are replacing it for many indications. This article reviews current procedures for three of the most common pulmonary problems: emboli, nodules of an unknown nature, and lung cancer. It also looks briefly at current thinking about screening for lung cancer.
Pulmonary Emboli: Choices
Pulmonary emboli are common: by one estimate, there is one embolus per 1,000 persons per year in the United States. Approximately 10% of these patients die within the first hour; without treatment, another 30% probably will die. With treatment-heparin with or without a thrombolytic agent-the mortality rate is 8% or less. However, some of those deaths will be caused by complications of the treatment, so treatment without a firm diagnosis is unwise. How does one make the diagnosis quickly and accurately?
The standard method has been the ventilation/perfusion (V/Q) scan, a technically demanding study that is not easy to interpret. At present, the revised PIOPED (Prospective Investigation of Pulmonary Embolism Diagnosis) criteria1 are used to assign a probability that the patient has a pulmonary embolus.
Efforts began early in the 1990s to improve the accuracy of V/Q scan interpretation. A recent report from Massachusetts General Hospital in Boston showed that a radiologist plus a computer can be more accurate than either alone.2 Leading the team was James A. Scott, MD, associate professor of radiology, Harvard Medical School.
“Computers are only slightly less accurate than a human radiologist in interpreting V/Q scans,” Scott reports. “We also found that computers and humans have slightly different strengths and weaknesses. So when the computer and the human work together, the deficiencies of one are compensated for by the strengths of the other. For example, the radiologist might believe a scan shows a pulmonary embolism, whereas the computer says it is not. The radiologist would then say that the scan is indeterminate. The same conclusion would be reached if the computer says the scan is positive while the radiologist says it is negative. Using the computer as a consultant would create more uncertainty, but it would decrease the number of errors.”
An increasingly popular method of evaluating patients for pulmonary emboli is helical or electron-beam CT. So great has been the enthusiasm at the Mallinckrodt Institute of Radiology at Washington University, St Louis, that two more scanners have been purchased to handle this and other on-call uses.3 Various studies of CT have shown a sensitivity as high as 95% and a specificity as great as 97% for central emboli, although the accuracy may be only about 65% for peripheral emboli.4,5 Note, however, that many of these data were obtained with older scanners and may underestimate the accuracy of present technology. A team of researchers at the University of Munich recently reported using a combination of angiography and dynamic imaging on an electron-beam CT scanner for a comprehensive assessment of patients with possible pulmonary embolus.6 At present, the utility of CT as a routine method of diagnosing pulmonary embolism is unclear, and clinical trials with contemporary scanners continue.7
The greatest problem in diagnosing pulmonary emboli, however, may not be what imaging study to order but rather when to use imaging of any kind. At Thomas Jefferson University Hospital in Philadelphia, Charles M. Intenzo, MD, director of nuclear medicine, had three technologists resign because of the large number of off-hour requests for V/Q scans that produced normal findings. He described his experience at the 86th Scientific Assembly and Annual Meeting of the Radiological Society of North America (RSNA) in Chicago last year.8
“Because of medicolegal fears, the emergency room physicians were ordering too many V/Q scans,” he notes. “I can understand their point of view. My problem, both as a physician and as an administrator, is that if these tests are ordered at night, I have to call in ancillary personnel.”
Intenzo and his colleagues found that the number of V/Q scans ordered in the emergency department almost doubled-from 152 to 292-between 1995 and 1999. During this same period, there was a dramatic rise in the number of scans that proved to be normal: from 33% to 72%.
“The thresholds for ordering the study were simply too low,” he stresses. “I have been showing the emergency department staff how many of the studies that are ordered are normal, but you cannot change practices overnight.
“There is a shortage of radiologic technologists, so we cannot afford to lose them because they are tired of being called in at night to do scans that turn out to be normal. My point to the emergency department community is that if you keep overutilizing the studies, you won’t be able to do any studies because there will not be any technologists.”
Changing from V/Q scanning to CT would not solve the overutilization problem.
“As I said in Chicago, if we had been using helical CT to evaluate patients for pulmonary embolism, the head of our CT division would have presented this study!” Intenzo exclaims.
The Perplexing Nodule
Pulmonary nodules are common, and most of them are not cancers. For example, in a series of 1,520 self-referred patients who underwent low-dose helical CT scanning at the Mayo Clinic as part of a lung cancer screening trial, 1,358 pulmonary nodules were detected at baseline, of which only 23 were cancers. Is a lung lesion benign? Is it lung cancer? Is it a metastasis from a cancer elsewhere in the body?
One of the least expensive ways of distinguishing benign and malignant lung lesions is repeated scanning with calculation of changes in volume to determine the doubling time. David F. Yankelevitz, MD, associate professor of clinical radiology at Weill Medical College-Cornell Medical Center in New York City, reported at the RSNA meeting that this method can be highly sensitive, calculating lesion volume within 2%. “If you wrap adhesive tape around a cardboard tube, it usually requires three windings to make a sufficient difference to be seen by the human eye,” he explains. “Contemporary CT scanners can see a single layer of tape, approximately 0.1 mm.” In a small-scale trial, Yankelevitz and his associates found that it was possible to detect the growth of a malignant pulmonary nodule by examining two CT scans obtained 30 days apart. All 15 nodules could be classified correctly as malignant or benign.9
Another of the simpler methods of characterizing pulmonary nodules is determination of the extent of enhancement of the lesion after injection of contrast medium. In an international trial involving patients with 356 pulmonary lesions, 171 of which were proven malignant, the cancers enhanced by a median of 38.1 Hounsfield units (HU) on scans obtained 1, 2, 3, and 4 minutes after contrast administration, whereas the benign lesions showed a median increase of 10.0 HU. If an increase of no more than 15 HU was used as the cut-off for benign lesions, this method had a sensitivity of 98%, a specificity of 58%, and an accuracy of 77% for identifying cancers.10
Other ways of identifying malignancy in pulmonary nodules are gaining popularity. One is injection of a 99mTc-labeled polypeptide that binds to somatostatin receptors (expressed on the surface of many cancer cells). Localization of the label is detected with single-photon emission CT. In one trial, this polypeptide correctly identified or excluded cancer in 27 of 30 patients with solitary pulmonary nodules.11
An increasingly popular method of characterizing pulmonary nodules is positron emission tomography with fluorodeoxyglucose (FDG-PET). Because cancers in general have higher metabolic rates than non-malignant tissue, the labeled glucose concentrates preferentially in tumors. A recent meta-analysis of trials examining 1,474 pulmonary lesions with FDG-PET found a sensitivity of 96.8% and a specificity of 77.8% for cancer.12 Significantly, the size of the lesions did not affect accuracy, although the authors of the study noted that only a few nodules <1 cm were studied.
In April 2001, HCFA expanded the list of indications for PET for which reimbursement can be obtained. Diagnosis, staging, and restaging of non-small-cell lung cancer is now covered provided the study is done on a dedicated PET scanner. The agency is continuing to examine the issue of PET performed with coincidence gamma cameras.
Nodule is Malignant: What Next?
Once a nodule has been confirmed to be malignant, the next step is to determine the stage of the disease. There still is some debate about what imaging methods to use. Atul Mehta, MD, head of bronchology,? Cleveland Clinic, reviewed almost 1,500 studies of lung cancer staging, including nearly 1,100 on the use of CT. Even with the fourth-generation scanners, the accuracy is only about 84%. “Can we live with that accuracy?” Mehta asked in a presentation at Chest 2000 in San Francisco last October. “Is CT acceptable as a sole staging tool? The answer to that question is ‘no.'” Nevertheless, he believes that all patients suspected of having lung cancer should have a CT scan. Mehta also pointed out that although a lymph node larger than 10 mm is routinely considered abnormal, that figure is arbitrary. Smaller nodes may contain tumor.
There is growing interest in the role of PET with fluorine 18-labeled deoxyglucose (FDG) as a staging tool for lung cancer, and this use is covered by Medicare for some histologic types of disease. Two papers published last year indicate the potential. Pieterman and associates from Groningen University Hospital in The Netherlands found the sensitivity and specificity for mediastinal metastases in non-small-cell lung cancer to be 91% and 86%, respectively.13 For CT, the corresponding values were 75% and 66%. Particularly noteworthy was the discovery of distant metastases in 11 of the 102 patients that had not been identified by standard examinations. Farrell and colleagues from Duke University demonstrated the utility of FDG-PET in patients with stage I non-small-cell lung cancer.14 A comparison with the histopathologic findings showed FDG-PET to have staged 86% of the patients accurately. The positive and negative predictive values for PET were 47% and 97%, respectively. “A negative PET scan in these patients suggests that mediastinoscopy is unnecessary and that these patients can proceed directly to thoracotomy,” they wrote. In his presentation at Chest 2000, Mehta offered one caution: falsely positive results can be obtained if there is inflammation in the mediastinum, such as in patients with tuberculosis or histoplasmosis.
Screening for Lung Cancer?
An estimated 156,900 Americans will die of lung cancer in 2000. Another 56,3000 will die of colorectal cancer, 41,200 of breast cancer, and 31,900 of prostate cancer.
Screening for colorectal, breast, and prostate cancer is common, with millions of Americans participating. So why is screening for lung cancer not routine?
The largest study of lung cancer screening to date was the Mayo Lung Project, in which persons at risk were screened regularly by chest radiographs and sputum studies.15 Only 29% of the lung cancers that were discovered were resectable, and the mortality rate was the same in the screened and unscreened populations. A similarly designed study at Memorial Sloan-Kettering Cancer Center in New York produced the same findings.16 The American Cancer Society, the World Health Organization, and the American College of Radiology have stated that lung cancer screening is not warranted with present methods.
Perhaps newer technology should cause us to take another look at the issue. Commenting on the Mayo Lung Project during the RSNA session, Claudia I. Henschke, MD, PhD, professor of radiology at New York Weill Cornell Medical Center, remarked that the “Mayo Lung Project was a dismal failure and has kept us from screening for more than 20 years.” The fact is, she said, that the results are irrelevant today, when the far more sensitive low-dose helical CT is available.
She and her colleagues in the Early Lung Cancer Action Project (ELCAP) recently described the findings in 1,000 asymptomatic persons aged 60 or older who had a significant smoking history.17 All underwent low-dose helical CT scans and chest radiographs. The CT study found 27 cancers, 26 of which were resectable and 23 of which were in stage I. In contrast, chest radiography identified only seven of these cancers, including four of those in stage I.? In a series of 1,369 Japanese adults at high risk of lung cancer, helical CT scans revealed 15 lung cancers, only four of which were seen on the chest radiographs. Nearly all of the lesions (93%) found by CT were in stage I.18 During the second RSNA session, Stephen J. Swenson, MD, department of diagnostic radiology at the Mayo Clinic, discussed the early findings of the new Mayo Lung Project, which has enrolled 1,520 asymptomatic persons. During the baseline scan, 23 lung cancers were found: 21 by CT and two by sputum examination. Eighteen cancers were resected in the hope of cure, with 11 being in stage IA and 14 being <2 cm.
Definition of the value of conventional chest films in reducing lung cancer mortality may come from the 16-year study recently organized by the NCI’s Early Detection Branch. This study has randomized 74,000 patients to be screened or not for lung, colorectal, and prostate or ovarian cancer, with the aim of determining whether mortality from these cancers can be reduced. The value of low-dose helical CT will not be known until we have many years of follow-up of the trials just described.
1. Freitas FE, Sarosi MG, Nagle CC, Yeomans ME, Freitas AE, Juni JE. The use of modified PIOPED criteria in clinical practice. J Nucl Med. 1995;36:1573-1578.
2. Scott JA, Palmer EL, Fischman AJ. How well can radiologists using neural network software diagnose pulmonary embolism? AJR Am J Roentgenol. 2000;175:399-405.
3. Woodard PK. Response [letter]. Radiology. 2001;218:913.
4. Teigen CL, Maus TP, Sheedy PF, et al. Pulmonary embolism: diagnosis with contrast-enhanced electron-beam CT and comparison with pulmonary angiography. Radiology. 1995;194:313-319.
5. Goodman LR, Curtin JJ, Mewissen MW, et al. Detection of pulmonary embolism in patients with unresolved clinical and scintigraphic diagnosis: helical CT versus angiography. AJR Am J Roentgenol. 1995;164:1369-1374.
6. Schoepf UJ, Bruening R, Konschitzky H, et al. Pulmonary embolism: comprehensive diagnosis by using electron-beam CT for detection of emboli and assessment
of pulmonary blood flow. Radiology. 2000;217:693-700.
7. Rathbun SW, Raskob GE, Whitsett TL. Sensitivity and specificity of helical computed tomography in the diagnosis of pulmonary embolism: a systematic review. Ann Intern Med. 2000;132:227-232.
8. Intenzo CM, Kim SM, Alexander M, Yanovskaya L. Use and abuse of lung scintigraphy by emergency room physicians: a five year trend [abstract 504]. Radiology. 2000;217(P):323.
9. Yankelevitz DF, Gupta R, Zhao B, Henschke CI. Small pulmonary nodules: evaluation with repeat CT-preliminary experience. Radiology. 1999;212:561-566.
10. Swenson SJ, Viggiano RW, Midthun DE, et al. Lung nodule enhancement at CT: multicenter study. Radiology. 2000;214:73-80.
11. Blum JE, Handmaker H, Rinne NA. The utility of a somatostatin-type receptor binding peptide radiopharmaceutical (P829) in the evaluation of solitary pulmonary nodules. Chest. 1999;115:224-232.
12. Gould MK, Maclean CC, Kuschner WG, Rydzak CE, Owens DK. Accuracy of positron emission tomography for diagnosis of pulmonary nodules and mass lesions: a meta-analysis [comment appears in JAMA. 2001;285:936-937]. JAMA. 2001;285:914-924.
13. Pieterman RM, van Putten JW, Meuzelaar JJ, et al. Preoperative staging of non-small-cell lung cancer with positron-emission tomography [comment appears in N Engl J Med. 2000;343:290-292]. N Engl J Med. 2000;343:254-261.
?14. Farrell MA, McAdams HP, Herndon JE, Patz EF Jr. Non-small cell lung cancer: FDG PET for nodal staging in patients with Stage I disease. Radiology. 2000;215:886-890.
15. Fontana RS, Sanderson DR, Woolner LB, Taylor WF, Miller WE, Muhm JR. Lung cancer screening: the Mayo program. J Occup Health. 1986;28:746-750.
16. Flehinger BJ, Melamed MR, Zaman MB, Heelan RT, Perchick WB, Martini N. Early lung cancer detection: results of the initial (prevalence) radiologic and cytologic screening I the Memorial Sloan-Kettering study. Am Rev Respir Dis. 1984;130:555-560.
17. Henschke CI, McCauley DI, Yankelevitz DF, et al. Early Lung Cancer Action Project: overall design and findings from baseline screening. Lancet. 1999;354:99-105.
18. Kaneko M, Eguchi K, Ohmatsu H, et al. Peripheral lung cancer: screening and detection with low-dose spiral CT versus radiography. Radiology. 1996;201:798-802.
Judith Gunn Bronson, MS, is a contributing writer for Decisions in Axis Imaging News.