| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Perspective |
1 Department of Radiology and Center for Imaging Science, Samsung Medical
Center, Sungkyunkwan University School of Medicine, 50, Ilwon-dong,
Kangnam-gu, Seoul 135-710, South Korea.
2 Present address: Department of Diagnostic Radiology, Pusan National University
Hospital, Pusan National University School of Medicine and Medical Research
Institute, Pusan, Korea.
Received December 13, 2005; accepted after revision May 25, 2006.
Address correspondence to K. S. Lee
(kyungs.lee{at}samsung.com).
OBJECTIVE. The purpose of our study is to improve radiologists' understanding of the clinical issues involved in making a diagnosis and to guide further diagnostic workup and treatment of solitary pulmonary nodules (SPNs).
CONCLUSION. Information on the morphologic and hemodynamic characteristics of SPNs provided by dynamic helical CT, with high specificity and reasonably high accuracy, can be used for initial assessment. PET/CT is more sensitive at detecting malignancy than dynamic helical CT, and all malignant nodules may be potentially diagnosed as malignant by both techniques. Therefore, PET/CT may be selectively performed to characterize SPNs that show indeterminate results at dynamic helical CT.
Keywords: chest imaging • dynamic CT • lung • lung neoplasms • PET/CT • screening • solitary pulmonary nodules
CT screening has increased the detection rate of small pulmonary nodules (SPNs). The tissue characterization of subcentimeter nodules, still a challenge to radiologists, can be performed using serial volume measurements from CT. Video-assisted thoracoscopic surgery removal after nodule localization may be performed for the diagnosis and treatment of a subcentimeter nodule. In this perspective, we propose how to improve our understanding of the clinical issues involved in making a diagnosis and to guide further diagnostic workup and treatment of SPNs.
SPNs are defined as focal, round, or oval areas of increased opacity in the
lung with diameters of 
3 cm
[1]. The characterization of
SPNs is a major concern not only to radiologists but also to clinicians
because malignant lesions account for only 60-80% of resected pulmonary
nodules
[2-4].
The goal of the radiologic evaluation of SPNs is to noninvasively
differentiate benign from malignant lesions as accurately as possible.
Morphologic evaluations can help differentiate benign and malignant nodules
when they have typical benign or malignant features, but there is considerable
overlap between benign and malignant nodules in terms of their morphologic
presentations [5].
Various strategies other than morphologic evaluations have been applied to the differentiation of malignant and benign nodules, which include growth rate assessment [6], Bayesian analysis [7], and hemodynamic characteristics on dynamic helical CT [8-10]. In addition, the assessment of nodular metabolic characteristics on 18F-FDG PET [11] and pathologic evaluations using transthoracic needle aspiration, transthoracic needle biopsy, or video-assisted thoracoscopic surgery have also been used for characterizing SPNs. However, no single diagnostic algorithm can be applied to all cases.
The aim of this perspective is to improve our understanding of the clinical issues involved in making a diagnosis and to guide further diagnostic workup and treatment of SPNs.
Detection
Although new diagnostic techniques have been introduced, the detection of
lung nodules on imaging is difficult. CT screening has increased the detection
rate of small nodular attenuations, including those of early peripheral lung
cancer
[12-14].
Despite the higher spatial and contrast resolutions of CT, nodular lesions are
missed on chest CT. Missed lesions occur because of failures of detection. In
a study by Ko et al. [15], a
small nodule size of 5 mm in diameter (nodule detection sensitivity:
5 mm vs > 5 mm, 74% vs 82%), ground-glass opacity nodules (nodule detection
sensitivity: ground-glass opacity vs solid, 65% vs 83%), and lesion location
(nodule detection sensitivity: central vs peripheral, 61% vs 80%) were shown
to be major factors that contribute to the difficulty in detecting nodules.
Nodule detection can be improved by advances in computer-aided detection (CAD)
systems that are being developed and evaluated to provide a second perspective
for nodule detection on CT. The use of CAD can help improve radiologist
performance for the detection of unidentified lung cancers during lung cancer
screening with CT [16,
17]. However, Lee et al.
[18] showed that the
sensitivity of a CAD system (81%) did not significantly differ from that of
radiologists (85%). Radiologists were more sensitive at detecting nodules
attached to other structures (Fig.
1), whereas CAD was better at detecting isolated nodules and those
that were 5 mm in diameter
[18].
|
|
|
|
Evaluation of the specific morphologic features of SPNs can help differentiate benign from malignant nodules. Small, smooth nodules with well-defined margins are suggestive of, but not diagnostic for, benignity; and a lobulated contour or an irregular or spiculated margin with distortion of adjacent vessels is typically associated with malignancy [5]. Diffuse, laminated, central nodular, and popcornlike calcifications within nodules suggest benignity. On the other hand, eccentric or stippled calcifications have been described in malignant nodules. Fat or calcification may be observed in up to 50% of pulmonary hamartoma [19]. In a study by Jeong et al. [10], multivariate analysis was used to identify criteria independently associated with a diagnosis of a malignant nodule that had a higher odds ratio for malignancy than other radiologic finding criteria; the criteria were a lobulated margin, a spiculated margin, and the absence of a satellite nodule. Considerable overlap exists between the internal characteristics (air bronchogram, cavitation, wall thickness, attenuation, and so forth) of benign and malignant nodules. Initial morphologic evaluations of SPNs often result in nonspecific findings and further evaluations to exclude malignancy.
When interpreting screening CT images, radiologists must search for lung nodules, differentiate malignant lesions from benign nodules, and finally, recommend follow-up actions for these detected lesions. The results of the Early Lung Cancer Action Project (ELCAP) [13] suggested that nodules with pure (nonsolid) or mixed (partially solid) ground-glass opacity on thin-section CT are more likely to be malignant than are those with solid opacity. Li et al. [20] more specifically evaluated the characteristic thin-section CT findings of malignant and benign nodules detected on screening CT. Among nodules with pure ground-glass opacity, a round shape was found more frequently for malignant lesions (65%) than for benign lesions (17%); mixed ground-glass opacity, a subtype that has ground-glass opacity in the periphery and a high-attenuation zone in the center, was seen much more often in malignant lesions (41%) than in benign lesions (7%). Among solid nodules, a polygonal shape or a smooth margin was present less frequently in malignant than in benign lesions.
Growth Rate Assessment
Determination of growth rate by comparing sizes on current and prior images is an important and cost-effective step in the evaluation of SPNs. The absence of detectable growth over a 2-year period of observation is a reliable criterion for establishing that a pulmonary nodule is benign [21]. Despite widespread acceptance, the value of 2-year stability for the diagnosis of a benign nodule may not be sufficient for ground-glass opacities and suspected bronchioloalveolar carcinoma [22]. In addition, it is difficult to reliably detect growth in small (< 1 cm) nodules. To overcome this limitation, it has been proposed that the growth rate of small nodules be assessed using serial volume measurements rather than diameter [23]. Computer software programs that automatically enable a nodule volume calculation have become widely available. Moreover, computer-aided 3D quantitative volume measurement methods have been developed and applied clinically [23-25] (Figs. 2A, 2B, and 2C).
However, all these volumetric methods are focused on solid pulmonary nodules. Few reports have dealt with nodule volumetric methods for ground-glass opacity nodules [26]. Volumetry for ground-glass opacity or semisolid (a solid nodule containing a ground-glass opacity component) nodules is more difficult than that of purely solid nodules, because these nodules have a lower nodule-to-lung parenchyma contrast ratio than solid nodules. Precise nodule volumetry methods for ground-glass opacity or semisolid nodules are under development.
Other factors that should be considered in performing in vivo nodule volumetry are motion artifacts, the presence of an adjacent normal structure, and small nodule size. Nodule volume is crucially dependent on the underlying cardiac phase. In addition, cardiovascular motion is not conveyed proportionally to various pulmonary segments. The motion is prominent in areas adjacent to the vascular structures and the heart. Nodule segmentation from adjacent normal structures is important for accurate volume measurement [26].
Bayesian Analysis
Clinical correlations continue to play a critical role in the assessment of SPNs. In an attempt to more accurately define various known risk factors, both clinical and radiologic, an increasing number of investigators have used Bayesian analysis [7]. The odds likelihood ratio form of the Bayes theorem allows the probability of malignancy to be calculated by estimating likelihood ratios for various individual radiographic and clinical characteristics derived from previous literature. Likelihood ratios are measures of the probability of a positive test result or finding in a patient with disease divided by the probability of a positive test result or finding in a patient without disease.
Using these calculations, it is possible to combine individual probabilities into an overall estimate of the odds favoring malignancy in SPNs. The hierarchy of radiologic and clinical likelihood ratios for malignancy include, in decreasing rank, a cavity of 3 16 mm in thickness, irregular or spiculated margin on CT scans, patient complaints of hemoptysis, a patient history of malignancy, patient age > 70 years, nodule size of 21-30 mm in diameter, nodule growth rate of 7-465 days, an ill-defined nodule on chest radiographs, patient a current smoker, and nodules with indeterminate calcification on CT scans [7]. To date, investigators have reported accuracies ranging between 53% and 96% for predicting the likelihood of malignancy [7]. Unfortunately, traditional Bayesian analysis does not include considerations of more advanced imaging techniques such as dynamic helical CT and PET, which are useful for the preoperative stratification of benign and malignant nodules.
In contrast to manual work in the Bayesian method, an artificial neural network is a computerized processing device the design of which is inspired by the design and functioning of animal brains and their components. Most neural networks have some sort of "training" rule whereby the weights of connections are adjusted on the basis of presented patterns. In differentiating malignant from benign nodules, as many as useful parameters are needed as input data for a better performance with the artificial neural network. By extracting objective features from chest radiographs and connecting them to a computerized method, Nakamura et al. [27] proved that an artificial neural network has the potential to improve the diagnostic accuracy of radiologists in the distinction of malignant and benign SPNs. If CT morphologic features are obtained and provided, a more sophisticated artificial neural network system will be provided.
Hemodynamic Characteristics on Dynamic Helical CT
The evaluation of tumor vascularity with dynamic helical CT has proved to be useful in the differentiation of malignant and benign nodules [8, 28-33]. Various threshold attenuation values have been reported to be useful for distinguishing malignant from benign nodules on dynamic helical CT [8-10, 28-33] (Table 1). The threshold attenuation values refer to the cutoff Hounsfield units of increased attenuation after contrast injection for differentiating malignant from benign nodules. Yamashita et al. [31] reported that a maximum attenuation of 20-60 H appears to be a good predictor of malignancy, and a report by Swensen et al. [8] in 2000 is also noteworthy, in that the authors reported a threshold value of 15 H produced a sensitivity of 98%, a specificity of 58%, and an accuracy of 77% for malignant nodules. Since 2000, the cutoff values for the differentiation between benign and malignant nodules have been set at 15 H.
|
|
|
|
With the advent of MDCT, we have the advantage of shorter acquisition times, greater coverage, and superior image resolution along the z-axis [34]. Image clusters obtained at a given time throughout a nodule can be acquired sequentially by using a helical technique at short time intervals after the IV injection of contrast medium, thus allowing the same or very similar scans to be obtained through a nodule at various times to compare extent of enhancement. In a dynamic study with MDCT [9], higher peak enhancement was obtained than in previous studies performed using conventional or single-detector helical CT [8, 28-33], and thus higher attenuation cutoff values can be used for differentiation. Actually, with a cutoff value of 30 H of net enhancement, overall diagnostic accuracy (sensitivity of 99%, specificity of 54%, positive predictive value of 71%, negative predictive value of 97%, and an accuracy of 78%) was similar to that in previous studies [8, 9, 28-31] performed using single-detector helical CT [9].
However, in early studies that focused on the early phase of dynamic CT, some overlap was found between malignant and benign nodules—for example, active granulomas and benign vascular tumors. Although the results of previous dynamic studies showed high sensitivity for the diagnosis of malignant nodules, specificities were too low. In addition, approximately 50% of indeterminate lung nodules that had an eventual benign surgical diagnosis required patient hospitalization for their surgical removal, which is expensive and involves morbidities and mortality [35, 36]. Thus, a need was created for noninvasive imaging techniques for the specific diagnosis of indeterminate lung nodules.
Evaluation of SPNs by analyzing combined wash-in and washout
characteristics on dynamic helical CT allows more precise evaluations of
nodule hemodynamics. In addition, the efficacy of tissue characterization has
improved, and now sensitivities and specificities of more than 90% are
achieved using evaluations of washout patterns in the delayed dynamic phase
[10]. In a study by Jeong et
al. [10], malignant nodules
were characterized using a wash-in of 3 25 H and a washout of 5-31
H (Figs. 3A,
3B,
3C,
4A,
4B, and
4C). Benign nodules can be
characterized using a wash-in of < 25 H, a wash-in of 3 25 H in
combination with a washout of > 31 H, or a wash-in of 3 25 H and
persistent enhancement without washout
[10]. When diagnostic criteria
of both wash-in of 3 25 H and washout of 5-31 H were applied for
malignancy, sensitivity, specificity, and accuracy for malignancy were 81-94%,
90-93%, and 85-92%, respectively
[10,
37]. According to multivariate
analysis and after controlling for the effect of other diagnostic factors,
Jeong et al. found that the following diagnostic criteria indicated a
malignant nodule: 
25 H wash-in and 5-31 H washout (p = 0.001;
odds ratio, 25.7), lobulated margin (p = 0.011; odds ratio, 41.7),
spiculated margin (p = 0.006; odds ratio, 35.3), and absence of a
satellite nodule (p = 0.004; odds ratio, 13.8).
|
|
|
Metabolic Characteristics on 18F-FDG PET
In an effort to improve the diagnostic accuracy of imaging pulmonary lesions, PET with 18F-FDG has been used. Malignant cells have upregulated metabolisms and proliferate rapidly. Comparable enhancements of glucose and 18F-FDG uptake in malignant cells have permitted malignancy to be detected on PET, which is considered an accurate, noninvasive diagnostic test, with a sensitivity of 88-96% and a specificity of 70-90% for malignant nodules [11, 38-42] (Table 2). Integrated PET/CT provides more anatomic detail and improved staging accuracy of non-small cell lung cancer versus PET alone or CT alone [43]. In a recent comparative study [37] of dynamic helical CT and integrated PET/CT for SPNs, the sensitivity, specificity, and accuracy for predicting malignant nodule on dynamic helical CT and integrated PET/CT were 81%, 93%, 85% and 96%, 88%, 93%, respectively. In that study, all malignant nodules were interpreted correctly by at least one of these two techniques, dynamic helical CT or PET/CT [37] (Figs. 3A, 3B, 3C, 4A, 4B, and 4C).
|
SPNs with increased 18F-FDG uptake should be considered malignant, although false-positive results can be obtained in patients with infectious and inflammatory processes such as active tuberculosis, histoplasmosis, and rheumatoid nodules [6, 44]. The high specificity of 18F-FDG PET for the diagnosis of benign lesions has important clinical usefulness. Lesions with low 18F-FDG uptake may be considered benign. However, false-negative results may be seen in primary pulmonary malignancies such as carcinoids, bronchioloalveolar carcinomas, adenocarcinomas with a predominantly bronchioloalveolar carcinoma component (Figs. 5A, 5B, and 5C), and malignant SPNs of < 10 mm in diameter [11, 45, 46] (Figs. 6A, 6B, and 6C). Fluorine-18 FDG PET yields false-negative results in about 5% of all stage T1 lung cancers, but in only 3% of stage T1 lung cancers greater than 5 mm in diameter [47]. The long-term survival of patients with a negative PET scan for lung cancer suggests that these tumors behave indolently.
|
|
|
|
|
|
Characterization of Subcentimeter Nodules
Since the introduction of helical and MDCT, the detection of small pulmonary nodules of < 10 mm has become routine. However, characterization of nodules of < 10 mm in diameter is a challenge to radiologists. Although nodules < 10 mm in diameter have a low chance of being malignant, the reported percentage of malignancy varies according to the series [48-51]. In one report [48], nodules of < 10 mm in diameter in nonprimary lobes in lung cancer patients have only a 4% chance of being malignant. In another report, 18% of nodules < 10 mm in diameter in patients with extrathoracic malignancy were malignant [49]. On the other hand, Munden et al. [50] reported that 58% of all nodules < 10 mm in diameter were malignant and 41% of nodules < 10 mm in diameter in patients without previous malignancy represented malignancy.
Currently, serial volume measurements of nodules are regarded as the most
reliable technique for the characterization of small nodules
[23]. However, no report has
dealt with a large number of patients. Dynamic helical CT is another reliable
technique for nodule characterization. However, thin-section ( 1.0-mm
section thickness) images with large coverage along the z-axis are
required, especially for nodule characterization located in the lower lung
zone, where the respiratory excursion of the lungs is great. PET or PET/CT is
suboptimal for the characterization of subcentimeter nodules because the
spatial resolution (currently 7 mm at most) of those techniques is
insufficient for detecting malignant subcentimeter nodules
[37] (Figs.
6A,
6B, and
6C). Therefore, it is
necessary to provide practical guidelines for the follow-up and management of
indeterminate small pulmonary nodules
[52].
Very small (3-5 mm) well-defined nodules (sometimes called "ditzels") have been encountered more often with the advent of MDCT. Treatment decisions regarding these nodules vary depending on the patient's age, the risk of malignancy, and the risk of developing granulomatous disease. Regarding radiologists' interpretation of and management decision for these nodules [53], most radiologists recommend short-term follow-up, with less aggressive recommendations in patients who have a lower likelihood of malignancy and more aggressive recommendations in patients with a higher likelihood of malignancy.
Transthoracic Needle Biopsy or Aspiration
The use of transthoracic needle aspiration or transthoracic needle biopsy for assessing solitary nodules is well established [54-57]. The most common indication for transthoracic needle aspiration or transthoracic needle biopsy is an indeterminate SPN requiring a preoperative diagnosis, especially in patients who are unfit for surgery and who need histologic diagnosis for nonsurgical treatment planning [58]. Recent advances in transthoracic needle aspiration or biopsy, including the use of CT guidance, immediate cytopathology, the introduction of core biopsy techniques, and postbiopsy positional restrictions (to limit the incidence of pneumothoraces) have made this option still more attractive. However, the accuracy of transthoracic needle aspiration for the definitive diagnosis of benign disease has proven to be limited, typically less than 50% [56]. The effect of small lesion size (< 15 mm) is somewhat more controversial [54, 55]. Li et al. [55] reported a significant difference in the diagnostic accuracies of transthoracic needle biopsy for small and large lesions (74% vs 96%, respectively). However, more recently, Wescott et al. [54], in a study of 74 biopsies of 64 small (< 15 mm) lesions, reported a sensitivity of 93%, a specificity of 100%, and an accuracy of 95%. These discrepancies in the diagnostic accuracy of transthoracic needle biopsy for small lesions likely reflect a combination of level of experience and the number of attempted biopsies.
Although transthoracic needle aspiration or transthoracic needle biopsy is sensitive for the diagnosis of intrathoracic malignancy, the ability to differentiate between the various cell types of bronchogenic carcinoma is less well established [58]. In addition, transthoracic needle aspiration or biopsy may be technically limited in SPNs that are in a high apical location or too close to the hemidiaphragm.
|
|
|
In patients with indeterminate SPNs, video-assisted thoracoscopic surgery is indicated for lesions that are inaccessible to transthoracic needle biopsy, or when transthoracic needle biopsy is unlikely to provide a definitive benign or malignant diagnosis because of either lesion characteristics or biopsy limitations [59]. Video-assisted thorascopic surgery is performed, usually with the patient under general anesthesia, using a double-lumen endobronchial tube to allow ventilation of the contralateral lung. After collapse of the ipsilateral lung, three incisions are typically made and a telescope coupled to a video camera is introduced. Video-assisted thorascopic surgery nodule removal after nodule localization using a pulmonary nodule marking system (hookwire or tattooing with dye) is useful for the diagnosis and treatment of small SPNs. The dedicated nodule marking system is efficient in terms of localization and stable fixation of small (4-15 mm in diameter) subpleural (usually within 2 cm from the pleura or the fissure) pulmonary nodules. A small amount of pneumothorax occurs as a complication in up to 20% of patients during the placement of the nodule marking system. Once located, most nodules are removed for diagnosis or treatment. In about 5% of patients, wire dislocation may hamper nodule removal [60-63].
Prognosis of Malignant SPNs
The prevalence of nodal metastasis in patients with peripheral lung cancer—that is, stage T1 lung cancer—manifested as an SPN is considered to be low. However, several studies have reported a relatively high prevalence of mediastinal lymph node metastases [64, 65]. Aoki et al. [22] reported that adenocarcinoma appearing as localized ground-glass opacity shows slow growth, and Kim et al. [66] reported that the extent of ground-glass opacity is significantly greater in patients without recurrence, nodal metastases, or distant metastases. In addition, Jung et al. [67] suggested that the prevalence of extrathoracic metastasis is significantly less in small peripheral lung cancer with ground-glass opacity than in similar lesions without ground-glass opacity. Aoki et al. [68] evaluated the prognostic importance of preoperative thin-section CT findings in peripheral lung adenocarcinoma. All adenocarcinomas smaller than 2 cm in which ground-glass opacity accounted for more than 50% of tumor volume were free of lymph node metastasis and vessel invasion, and all these patients remain alive and without recurrence 10 years later. Coarse spiculation and thickening of bronchovascular bundles around tumors were observed more frequently in tumors with lymph node metastasis or vessel invasion than in those without these findings.
|
|
|
A growing malignant nodule needs its own blood supply from adjacent tissues; the blood supply is essentially required for tumor growth and metastatic spread. The blood supply process may be caused by the increased release of angiogenic factors, such as vascular endothelial growth factor, from a malignant nodule and the subsequent increase in the extent of microvessel density [9, 73]. An increased microvessel density leads to increased capillary perfusion and permeability and is frequently associated with strong enhancement of a malignant nodule on CT [9, 74]. Therefore, the extent of enhancement can be interpreted as a reflection of tumor vascularity (Figs. 7A, 7B, and 7C). Histologic evaluations of tumor microvessel densities and of expressions of vascular endothelial growth factor are important prognostic factors in non-small cell lung cancers [75, 76]. In other words, the likelihood of metastatic disease increases as the number of intratumoral microvessels increases in lung cancers.
Recently, Guo et al. [77]
investigated the correlation between microvessel density—which is
determined using various endothelial antibodies (immunostaining methods for
the presence of CD31, CD105, CD34, or 
-smooth muscle actin)—and
18F-FDG uptake; they compared the prognostic impact of those
factors in lung adenocarcinomas. They concluded that CD105 microvessel density
reflects active angiogenesis, and that it is a better indicator of prognosis
in patients with lung adenocarcinoma than other antibodies are (microvessel
density extent determined using CD105 was negatively correlated with patient
prognosis). Moreover, 18F-FDG uptake at PET was found to be
significantly correlated with active angiogenesis as determined by CD105
microvessel density.
Prediction of Hilar or Mediastinal Nodal Metastasis in Malignant SPNs
Efforts have been made to identify CT findings that indicate a propensity
to metastasize based on analyses of morphologic characteristics, such as size
or the marginal characteristics of primary tumors
[69,
70,
78,
79]. However, results have
been unsatisfactory and controversial, especially in solid malignant nodules.
Shim et al. [80] suggested
that the extent of nodule enhancement is related to a propensity toward
mediastinal or hilar nodal metastasis; tumor size, marginal characteristics,
and the presence of necrosis or bronchovascular thickening showed no
correlation with mediastinal or hilar nodal metastasis. Those authors reported
that stage T1 lung cancers that show high peak enhancement or net enhancement
on dynamic CT have a high likelihood of mediastinal nodal metastasis, with a
sensitivity of 62%, specificity of 76%, and accuracy of 74%, after applying
cutoff values of 110 H of peak enhancement, and a sensitivity of 77%,
specificity of 70%, and accuracy of 71% when applying a cutoff value of 60 H
of net enhancement.
|
The evaluation of extent of nodule enhancement on dynamic helical CT of malignant SPNs allows better (although not statistically significant) sensitivity for the prediction of mediastinal nodal metastasis on a patient basis than PET/CT, whereas PET/CT was found to have perfect specificity and positive predictive values [80]. Therefore, mediastinoscopy may be recommended in patients with nodules that show high enhancement on dynamic helical CT even though PET/CT does not suggest the presence of mediastinal nodal metastasis (Figs. 8A, 8B, and 8C).
Current and Future Strategies
It is apparent that despite the considerable technical advances of imaging
techniques, the diagnostic workup and management of patients with SPNs still
relies on clinical perspectives, and that no single clinically based
diagnostic algorithm can be applied to all cases. According to a study by
Jeong et al. [10], each of the
diagnostic criteria for malignancy—that is, 25 H wash-in and 5-31 H
washout (p = 0.001; odds ratio, 25.7), a lobulated margin (p
= 0.011; odds ratio, 41.7), a spiculated margin (p = 0.006; odds
ratio, 35.3), and the absence of a satellite nodule (p = 0.004; odds
ratio, 13.8)—is useful for malignant nodule diagnosis by multivariate
analysis after controlling for the effects of other diagnostic factors. In
light of previous results [10,
37], SPNs may be initially
evaluated using dynamic helical CT, which is readily available in most
institutions and which provides morphologic and hemodynamic information about
nodules. PET/CT can then be used as the next step to evaluate patients with a
possible malignancy.
In terms of devising diagnostic strategies for SPNs, the following questions should be raised: Should both examinations be performed despite the possible radiation hazards and the higher cost? If so, which one is technically superior, and which is the more cost-effective? On dynamic helical CT, nodules with 3 25 H wash-in and 5-31 H washout can be diagnosed as malignant with high specificity [10, 37]. Moreover, nodules with 3 25 H wash-in that have persistent enhancement without washout or with > 31 H washout can be considered benign with a 71-95% negative predictive value, although they still have the potential of being malignant. PET/CT of these patients increases diagnostic sensitivity [37]. Nodules with < 25 H wash-in have only a low possibility of being malignant (range, 0-5%) [10, 37].
However, in cases of nodules with < 25 H wash-in and nodules with 3 25 H wash-in and persistent enhancement without washout or with > 31 H washout, when they show malignant morphologic features of a lobulated and spiculated margin but no satellite nodules, evaluation will be improved by subsequent PET/CT. If nodules have malignant morphologic features of a lobulated or spiculated margin and no satellite lesions, PET/CT may be recommended even though a hemodynamic study suggests benignancy, especially for nodules having 3 25 H wash-in.
The accuracies of transthoracic needle aspiration and biopsy require elaboration. Previously described diagnostic yields of transthoracic needle aspiration or biopsy for assessing SPNs [55, 57] are lower than those obtained on PET/CT alone or on dynamic helical CT and PET/CT in combination [37], which suggests that transthoracic needle aspiration or biopsy cannot totally replace PET/CT or dynamic helical CT for the characterization of SPNs.
MDCT data may be beneficial for assessing 3D nodule perfusion and may be more accurate than the 2D analysis currently used [83]. In the future, 3D nodule enhancement and dynamic imaging may further improve the assessment of SPNs.
Conclusion
CT screening, with or without the help of a CAD system, has increased the detection rate of small nodular lesions, including that of early peripheral lung cancer. Dynamic helical CT, providing information about morphologic and hemodynamic characteristics with high specificity and reasonably high accuracy, can be used for the initial assessment of SPNs. PET/CT is more sensitive for detecting malignancy than dynamic helical CT, and all malignant nodules may be potentially diagnosed as malignant by these two techniques. PET/CT may be selectively performed to characterize SPNs when dynamic helical CT shows inconsistent results between morphologic and hemodynamic characteristics (Fig. 9). Serial volume measurements, and thus calculation of volume doubling time, are currently the most reliable methods for the tissue characterization of subcentimeter nodules. When an SPN is highly likely to be malignant, removal of the nodule by video-assisted thorascopic surgery, after nodule localization using the pulmonary nodule marking system, may be performed for diagnosis and treatment.
References
