The chest has long been a popular target for imaging, but there have been great changes recently in how it is done and why.

Today, a CT scan is more likely to be ordered than a plain radiograph. During the decade ending in December 2001, Massachusetts General Hospital noted both a large increase in the number of chest studies obtained per patient and a significant rise in the number of chest CT scans, not only for inpatients and emergency admissions, but also for outpatients. 1 A team at the GE Global Research Center endorsed this observation, commenting that “no other single modality is more often used in chest imaging today [than CT].” 2

How is CT being applied in the chest? One of the most exciting stories is the progress in screening for lung cancer. As Claudia I. Henschke, MD, and her colleagues of the Department of Radiology at Weill Medical College of Cornell University recently pointed out, this cancer is an ideal target for screening. 3 First, it is common, accounting for 12% of new cancers and 28% of cancer deaths in the United States each year. 4 Second, the population at greatest risk is known. Third, although the traditional view is that the disease is systemic almost from the beginning and thus that cure is rare, this may not be true. Certainly, if a lung cancer is not diagnosed until it produces symptoms or signs, it is “essentially incurable,” 3 but localized lung cancer has an almost 50% 5-year survival rate, 5 and according to some of the recent International Early Lung Cancer Action Project (I-ELCAP) data, even better than that. Thus, screening might pay the same high dividends in lung cancer that it has in prostate and breast cancer.

The first screening efforts, which relied on chest radiography and, in some trials, sputum examination, were disappointing. However, low-dose CT may succeed. That was a take-home lesson from the 2004 Annual Meeting of the Radiological Society of North America.


Figure. A: Benign nodule. Screen-detected nodule with round shape and smooth, sharply defined borders. B: Malignant nodule. Screen detected nodule with oval shape and lobulated, irregular border with a few small spiculations. Images courtesy of David Yankelevitz, MD, Weill Medical College, Cornell University.

Henschke and her colleagues were among three groups presenting data on screening at RSNA. Their study, I-ELCAP, started in 1993 and involves 33 institutions enrolling almost 30,000 subjects, who are receiving regular low-dose multidetector-array CT (MDCT) scans (see sample protocol, page 38). Of the 376 lung cancers detected by the initial screen, 82% were in Stage I and were resected, yielding an 8-year lung-cancer case-fatality rate of only 4%. 6 In a subgroup of subjects, 123 (1%) had mediastinal lesion, but only four were cancers, so mediastinal masses identified by screening should be approached conservatively. 7 Elsewhere, these researchers have found that uncalcified nodules smaller than 0.5 mm do not warrant immediate work-up but can safely be monitored, with further studies undertaken only if the masses grow. 8

Investigators at the H. Lee Moffitt Cancer Center and Research Institute at the University of South Florida are using low-dose spiral CT and sputum examination in 1,151 subjects at least 45 years of age with a mean of 57.9 pack-years of smoking. 9 At baseline, 35% of the scans were abnormal, leading to diagnosis of 28 prevalence neoplasms. In the follow-up incidence screenings, 14% of the scans were abnormal, with 18 cancers being detected. Altogether, 79% of these lesions were radiologic Stage I, but only 60% were pathologic Stage I, leading the investigators to caution that “[s]mall cancer size does not always correlate with early stage disease.” These investigators pursue masses detected at screening using positron emission tomography with 18 fluorine deoxyglucose (FDG-PET) in hopes of distinguishing cancer and benign lesions noninvasively and have found a sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for malignancy of 84%, 86%, 96%, and 55%, respectively. 10 The false-negative PET scans usually marked lesions with ground-glass opacities, which appear to be less aggressive. 11

The third screening trial described is being carried out in Korea and has enrolled 6,406 asymptomatic subjects, of whom only 48% are smokers. 12 At the initial scan, 35% of the subjects had at least one suspect nodule, and 19 subjects had lung cancer. Curiously, only 47% of the cancers occurred in smokers, although follow-up scans found that the cancer doubling time was much faster in smokers than in nonsmokers (5 versus 15 months).

Of course, screening scans are of no value if lesions that actually are present are not found. Can computer-assisted detection (CAD) simplify the review of the enormous numbers of images screening programs generate?

At Stanford University School of Medicine, Rubin and colleagues examined 20 outpatients for pulmonary nodules by MDCT using 1.25-mm section thickness and a 0.6-mm interval. A total of 195 noncalcified nodules with a diameter of at least 3 mm were detected. Reading by two radiologists had a mean sensitivity of 63%; addition of CAD increased this value to 76%. When the CAD program was adjusted to produce no more than three false-positive readings per scan, the “sensitivity was substantially higher than with conventional double reading” even by these highly experienced radiologists. 13

Another trial of CAD was described at the RSNA by Das and colleagues of Aachen, Germany, who are studying 178 former power plant workers with significant asbestos exposure. 14 Those investigators are using a low-dose protocol with a 16-slice MDCT scanner. Images are reconstructed in 1-mm slices with an increment of 0.5 mm. Suspect lesions were found in eight patients at baseline, and CT-guided needle biopsy demonstrated cancer in five, with the others still being worked up at the time of the report. Asbestos-induced pulmonary changes were visible in 80 patients. For differentiation of benign and malignant nodules, repeat CT scans at 3, 6, and 12 months are planned.

*Slightly different protocols are used for other four-slice and for single-slice scanners; for details and a discussion of the full requirements for a lung cancer-screening program, see: Henschke CI, Yankelevitz DF, McCauley DI, Libby DM, Pasmantier MW, Smith JP. Guidelines for the use of spiral computed tomography inscreening for lung cancer. Eur Respir J. 2003; Suppl 39:45s–51s. Full text is available through Pub Med.

Other presenters at RSNA also testified to the value of CAD, both for identifying nodules on screening or other lung CT scans and for helping determine which nodules are malignant. According to a team from Paris, CAD improves sensitivity by 5% to 10% compared with double reading. 15 In another series, CAD found 114 nodules in 37 patients that were overlooked by two human readers. 16 Notably, 18 of these nodules were larger than 4 mm, and the CAD discoveries had an impact on patient management in 10 cases. The I-ELCAP investigators likewise found that CAD improved sensitivity. 18 In retrospective analyses, computer analysis helped determine which nodules were malignant 19 and increased reader confidence in a diagnosis of malignancy. 20

Where does screening stand now?

“If you can find a cancer early that, in the absence of treatment, would be fatal and then provide effective treatment, you reduce deaths from lung cancer. That’s logical. The I-ELCAP has shown that CT can find lung cancer early,” says David F. Yankelevitz, MD, a member of the I-ELCAP team.

This does not mean that everyone in a high-risk group should be screened.

“The decision about whether to be screened needs to be individualized,” Yankelevitz stresses. “For example, if a person’s life expectancy is not at least 10 years, screening does not make sense. We want to be able to tell people what their benefit might be from undergoing screening: what is the likelihood a cancer will be found, what is the likelihood that cancer will be curable, how likely is it they will not die soon from some other cause?”


Once a cancer is found, by screening or otherwise, how should it be staged? Imaging advances and minimally invasive procedures have not only improved staging accuracy but changed the work-up algorithms. 21 The American College of Chest Physicians has developed an algorithm that includes CT of the chest and upper abdomen and, if available, FDG-PET.22 Magnetic resonance imaging is advised if the tumor is in the superior sulcus and in patients who have neurological signs or symptoms. Screening for extrathoracic disease is said not to be necessary in asymptomatic patients whose disease is in Stage I or II.

Brink and associates from Freiburg, Germany, recently demonstrated the value of FDG-PET for primary staging of small-cell lung cancers. 23 Complete agreement between PET and other studies was documented in 75 of 120 consecutive patients, and PET proved to be correct for 47 of the 65 sites where it and other modalities disagreed. The only staging error resulted from the inability of PET to detect a brain metastasis. Notably, PET was significantly more sensitive than CT in detecting extrathoracic lymph-node and other metastases except those to the brain, and the investigators suspect PET will reduce the number of other imaging studies and invasive procedures needed.

The newly introduced combination of PET and CT may be even better. The investigators in the Korean screening trial studied PET/CT as a means of preoperative staging in 106 patients. As judged by the findings at surgery, PET/CT was significantly more accurate than CT alone. 24 Similar results were reported by investigators from Essen, Germany, 25 in whose experience PET/CT provided an accurate T stage in all 12 patients and an accurate N stage in 27 of 28.

A staging approach that avoids radiation is ventilatory-gated short inversion time-inversion recovery (STIR) turbo spin-echo (TSE) MRI, which was studied in 110 patients with confirmed non-small-cell lung cancer (NSCLC). When a lungsaline ratio of 0.6 was the threshold for a finding of mediastinal and hilar lymph-node metastases, the sensitivity was 93% and the specificity 87%. 26

Another possible role of imaging in the work-up for lung cancer is determining whether a patient will be able to tolerate pulmonary resection. Radiologists at Kobe University Graduate School of Medicine in Japan explored the ability of dynamic perfusion MRI to predict the quality of the postoperative lung function and concluded that the study was a feasible alternative to perfusion scintigraphy for this purpose. 27


Imaging may contribute to lung-cancer management in other ways. Physicians from the Department of Chest Medicine at Show Chwan Memorial Hospital in Changhua, Taiwan, have demonstrated that the extent of 99mTc-tetrofosmin uptake reflects tumor-cell expression of p-glycoprotein and multidrug resistance-related protein-1. 28 In a series of 30 patients, half of whom had a good response to chemotherapy and half of whom did not, the tumor-to-background ratios in the tetrofosmin studies were predictive of tumor response. In another study of 105 metastatic pulmonary lesions, a prototype software program applied to low-dose MDCT scans proved capable of detecting clinically significant changes in lesion volume during chemotherapy, 29 perhaps enabling early change to another regimen.

Another role for CT is guiding percutaneous ablation of unresectable primary or metastatic lung cancers. In a multicenter trial, radiofrequency ablation (RFA) was performed in 14 patients with biopsy-proved 1- to 3-cm NSCLC who were not candidates for surgery or radiotherapy. 30 Nine of the 11 patients with follow-up for at least 6 months showed progressive lesion shrinkage, and complete ablation was confirmed in seven patients. The 1-year survival rate was 81%, with a disease-specific survival rate of 100%. Another team, which focused on tumors that had eroded a rib, found necrosis exceeding 90% in 26 of 30 lesions, with all of the patients who had been treated for pain obtaining at least some relief. 31 In a larger series, 126 patients with inoperable lung lesions were treated by RFA for either palliation or local disease control. More than half of the latter were alive at the time of the report. 32 Radiologists from Italy also reported good results with CT-guided RFA. 33 An alternative is laser ablation with CT fluoroscopy guidance, which is being studied by a team in Germany. 34 Complete regression was documented for 42% of metastases and partial regression for 35%. Loss of contrast enhancement on follow-up CT was early evidence of success. The best response was obtained with lesions of 2.5 to 4.0 cm.

Not all of the results have been so favorable. A team from Los Angeles applied CT-guided RFA in 16 patients with 22 unresectable lung cancers. Two of the patients relapsed at the treated site and six elsewhere, and these investigators suggested that the curative potential of RFA is “questionable.” 35 However, it is noteworthy that half of these patients already had high-stage disease or disease resistant to previous treatment.


Determining whether a patient has a pulmonary embolus has long been difficult. The traditional study was the radionuclide ventilationperfusion (V/Q) scan, but CT and CT angiography (CTA) have largely displaced it, not only for diagnosis1 but for follow-up of thrombolysis or surgical embolectomy. 36 In the experience of one emergency department, the availability of MDCT led to a 234% increase in CTA examinations for pulmonary emboli, although it also increased the likelihood of a negative study. 37 The limitation of the ability of CTA to reveal isolated peripheral emboli does not appear to have an adverse effect on patient outcome. 38

Will CTA be displaced, in turn, by MRI or MR angiography (MRA)? Radiologists at Kobe University Graduate School of Medicine compared chest radiography, contrast-enhanced MDCT, MRA, V/Q scans, and angiography (the reference standard) in 48 patients. 39 The sensitivity and specificity of MRA were 92% and 84%, respectively, significantly better than the values for the traditional examination. At the Kerckhoff Heart Center in Nauheim, Germany, real-time true fast MRI with steady-state precession (TrueFISP) on a 1.5T scanner had 100% specificity and sensitivity of 83% to 97%. 40 The study took less than 3 minutes and did not require contrast, and the investigators concluded that real-time MRI can be reliable for the emergency diagnosis of pulmonary emboli.

We have come a long way from plain chest film.

Judith Gunn Bronson, MS, is a contributing writer for Decisions in Axis Imaging News.


  1. Wittram C, Meehan MJ, Halpern EF, Shepard JA, McLoud TC, Thrall JH. Trends in thoracic radiology over a decade at a large academic medical center. J Thorac Imaging. 2004;19:164-170.
  2. Walter D, De Man B, Iatrou M, Edic PM. Future generation CT imaging. Thorac Surg Clin. 2004;14:135-149.
  3. Henschke CI, Yankelevitz DF, McCauley DI, Libby DM, Pasmantier MW, Smith JP. Guidelines for the use of spiral computed tomography in screening for lung cancer. Eur Respir J. 2003;Suppl 39:45s-51s.
  4. Jemal A, Tiwari, RC, Murray T, et al. Cancer statistics, 2004. CA Cancer J Clin. 2004;54:8-29.
  5. Jemal A, Clegg LX, Ward E, et al. Annual report to the nation on the status of cancer, 1975-2001, with a special feature regarding survival. Cancer. 2004;101:3-27.
  6. Henschke CI, Sone S, Markowitz S, et al. International Early Lung Cancer Action Project (I-ELCAP): evaluation of low-dose CT screening [abstract]. Radiology. 2004;233:263. Abstract SSA04-01.
  7. Henschke CI, Lee I, Rifkin M, et al. CT screening for lung cancer: mediastinal masses in the Early Lung Cancer Project and the New York Early Lung Cancer Action Project [abstract]. Radiology. 2004;233:264. Abstract SSA04-06.
  8. Henschke CI, Yankelevitz DF, Naidich DP, et al. CT screening for lung cancer: suspiciousness of nodules according to size on baseline scans. Radiology. 2004;231:164-168.
  9. Clark RA, Hazelton TR, Coppage L, Wadhwa NR, Tockman M. Lung cancer screening with computed tomography: initial results of a cohort screening trial [abstract]. Radiology. 2004;233:263. Abstract SSA04-02.
  10. Clark RA, Berman CG, Eikman EA, Hazelton TR, Coppage L, Tockman M. PET for the diagnosis of screening-detected lung lesions [abstract]. Radiology. 2004;233:382. Abstract SSE04-01.
  11. Lee KS, Jeong YJ, Han J, Kim BT, Kim H, Kwon OJ. T1 non-small cell lung cancer: imaging and histopathologic findings and their prognostic implications. Radiographics. 2004;24:1617-1636.
  12. Chung MJ, Lee KS, Chong S, et al. Lung cancer screening with low-dose helical CT in Korea: experiences at Samsung Medical Center for recent five years [abstract]. Radiology. 2004;233:264. Abstract SSA04-03.
  13. Rubin GD, Lyo JK, Paik DS, et al. Pulmonary nodules on multi-detector row CT scans: performance comparison of radiologists and computer-aided detection. Radiology. 2005;234:274-283.
  14. Das M, Felten MK, Flohr TG, Kraus T, Günther RW, Wildberger JE. Low dose multi-detector row CT lung cancer assessment in an asbestos exposed high-risk cohort: initial results from screening 178 patients [abstract]. Radiology. 2004;233:264. Abstract SSA04-04.
  15. Brochu B, Beigelman-Aubry C, Raffy P, Wood SA, Grenier PA. Computer-aided detection (CAD) of lung nodules on high-resolution MDCT screening exams [abstract]. Radiology. 2004;233:422. Abstract SSG06-01.
  16. Gebhard HH, Lauffer L, Costello P, Schoepf UJ. Prospective evaluation of an algorithm for the automated detection of lung nodules at routine chest CT in a clinical environment [abstract]. Radiology. 2004;233:423. Abstract SSG06-04.
  17. Li F, Li Q, Aoyama M, Sone S, MacMahon H, Doi K. Evaluation of a computerized scheme for differentiating benign from malignant lung nodules on high-resolution CT [abstract]. Radiology. 2004;233:422-423. Abstract SSG06-02.
  18. McCulloch CC, Henschke CI, Kazerooni EA, Leader JK, Turner W, Avila R. CAD-assisted radiologist sensitivity improvement in the detection of solid and subsolid nodules in low-dose CT lung cancer screening exams [abstract]. Radiology. 2004;233:608-609. Abstract SSQ05-05.
  19. Awai K, Murao K, Ozawa A, et al. Estimation of malignancy of pulmonary nodules at thin-slice helical CT: the performance of radiologists is improved by computer-aided diagnosis [abstract]. Radiology. 2004;233:608. Abstract SSQ05-02.
  20. Shah SK, Rogers S, De Zoysa K, et al. Solitary pulmonary nodule diagnosis on CT: results of a large-scale observer study [abstract]. Radiology. 2004;233:423. Abstract SSG06-03.
  21. Sihoe AD, Yim AP. Lung cancer staging. J Surg Res. 2004;117:92-106.
  22. Kent MS, Port JL, Altorki NK. Current state of imaging for lung cancer staging. Thorac Surg Clin. 2004;14:1-13.
  23. Brink I, Schumacher T, Mix M, et al. Impact of [18F]FDG-PET on the primary staging of small-cell lung cancer. Eur J Nucl Med Mol Imaging. 2004;31:1614-1620.
  24. Shim SS, Lee KS, Kim B, Lee EJ, Choi JY, Chung MJ. Accuracy of integrated PET/CT using fluorodeoxyglucose in preoperative staging on non-small cell lung cancer: a prospective comparison with standalone CT [abstract]. Radiology. 2004;233:383. Abstract SSE04-02.
  25. Veit P, Vogt FM, Freudenberg LS, Barkhausen J, Ruehm SG, Antoch G. Comparison of whole body PET/CT and whole body MRI in patients with newly diagnosed non-small cell lung cancer: initial results [abstract]. Radiology. 2004;233:383. Abstract SSE04-04.
  26. Ohno Y, Hatabu H, Takenaka D, et al. Metastases in mediastinal and hilar lymph nodes in patients with non-small cell lung cancer: quantitative and qualitative assessment with STIR turbo spin-echo MR imaging. Radiology. 2004;231:872-879.
  27. Ohno Y, Hatabu H, Higashino T, et al. Dynamic perfusion MRI versus perfusion scintigraphy: prediction of postoperative lung function in patients with lung cancer. AJR Am J Roentgenol. 2004;182:73-78.
  28. Kuo TH, Liu FY, Chuang CY, Wu HS, Wang JJ, Kao A. To predict response chemotherapy using technetium-99m tetrofosmin chest images in patients with untreated small cell lung cancer and compare with p-glycoprotein, multidrug resistance related protein-1, and lung resistance-related protein expression. Nucl Med Biol. 2003;30:627-632.
  29. Kuhnigk J, Bornemann L, Dicken V, Wormanns D, Krass S, Peitgen H. In vivo validation of automated CT-based volumetric assessment of lung metastases [abstract SSG06-06]. Radiology. 2004;233:423-424.
  30. Lencioni RA, Crocetti L, Cioni R, Suh RD, Wallace A, Bartolozzi C. Percutaneous radiofrequency ablation of stage IA non-small cell lung cancer: a prospective multicenter clinical trial [abstract]. Radiology. 2004;233:512. Abstract SSK05-03.
  31. vanSonnenberg E, Shankar S, Morrison PR, et al. Radiofrequency ablation of thoracic lesions II: initial clinical experience-technical and multidisciplinary considerations in 30 patients [abstract. Radiology. 2004;233:512. Abstract SSK05-05.
  32. Simon CJ, Dupuy DE, Mayo-Smith WW, DiPetrillo TA, Ready NE. Complications of image-guided percutaneous lung radiofrequency ablation: experience with 126 patients (total of 163 lesions treated) in 6 years [abstract]. Radiology. 2004;233:513. Abstract SSK05-07.
  33. Belfiore G, Tedeschi E, Moggio G, Coffi R, Cincotti FC, Rossi R. CT-guided radiofrequency palliative ablation of unresectable lung cancer: clinical and imaging findings at one-year follow-up [abstract]. Radiology. 2004;233:513. Abstract SSK05-06.
  34. Weigel C, Schuchmann S, Kirsch M, Hosten N. Percutaneous laser ablation of pulmonary metastases performed during CT fluoroscopy under conscious sedation: technical success, complications, and technique effectiveness during a 4-year period [abstract]. Radiology. 2004;233:513. Abstract SSK05-08.
  35. Suh RD, Wallace A, Goldin JG. CT-guided radiofrequency ablation for unresectable NSCLC: short and long-term results at one year [abstract]. Radiology. 2004;233:512. Abstract SSK05-04.
  36. Quiroz R, Zou KH, Kipfmueller F, Costello P, Goldhaber SZ, Schoepf UJ. Negative spiral CT safely rules out acute PE: a systematic analysis of patient outcome studies [abstract]. Radiology. 2004;233:605. Abstract SSQ04-02.
  37. Donohoo JH, Mayo-Smith WW, Pezzullo JA, Dupuy DE. Multidetector CT pulmonary angiography: utilization and yield in a busy emergency department [abstract]. Radiology. 2004;233:606. Abstract SSQ04-05.
  38. Kipfmueller F, Quiroz R, Kucher N, Costello P, Goldhaber SZ, Schoepf UJ. Recovery from right ventricular pressure overload after thrombolysis or embolectomy for acute, severe pulmonary embolism (PE): assessment at spiral CT [abstract]. Radiology. 2004;233:606. Abstract SSQ04-07.
  39. Ohno Y, Higashino T, Takenaka D, et al. MR angiography with sensitivity encoding (SENSE) for suspected pulmonary embolism: comparison with MDCT and ventilation-perfusion scintigraphy. AJR Am J Roentgenol. 2004;183:91-98.
  40. Kluge A, Muller C, Hansel J, Gerriets T, Bachmann G. Real-time MR with TrueFISP for the detection of acute pulmonary embolism: initial clinical experience. Eur Radiol. 2004;14:709-718.