16 PET/CT of the Thorax H. C. Steinert

(1)

16 PET/CT of the Thorax

H. C. Steinert and G. K. von Schulthess

H. C. Steinert, MD

Nuclear Medicine, Department of Medical Radiology, University Hospital, 8091 Zurich, Switzerland G. K. von Schulthess, MD

Professor and Director, Nuclear Medicine, Department of Medical Radiology, University Hospital, 8091 Zurich, Switzerland

CONTENTS

16.1 Basic Aspects of PET/CT Imaging 225 16.1.1 Introduction 225

16.1.2 PET Imaging 225 16.1.3 Radiopharmaceuticals 226 16.1.4 PET/CT 227

16.1.5 Clinical Protocols in PET/CT 227 16.1.6 Critical Appraisal of Clinical PET/CT 228 16.2 Clinical Applications of PET/CT in the Chest 230 16.2.1 Solitary Lung Nodule 230

16.2.2 T Staging 230 16.2.3 N Staging 231 16.2.4 M Staging 231

16.2.5 Recurrent Lung Cancer 233 16.2.6 Pitfalls 233

16.2.7 Small Cell Lung Cancer 233 16.2.8 Malignant Pleural Mesothelioma 233

References 234

16.1 Basic Aspects of PET/CT Imaging

16.1.1 Introduction

Positron emission tomography (PET)/CT is a new imaging modality. This is true despite the fact that traditionally PET belongs to nuclear medicine and that CT belongs to radiology. As a consequence, both fraternities derive from the respective imaging modality, that they should “own” PET/CT. Looking at PET/CT from this political perspective is not con- ducive to optimizing the information obtained. Our

experience with PET/CT suggests that the CT portion enhances PET’s unmatched ability to identify tumor foci; thus, the clinician can relate better to a PET scan when he knows where the foci are located. The PET scans contain little anatomic information, and know- ing exactly where a lesion is located is sometimes critical, so the role of CT is not only to make PET more palatable to clinicians. A metastasis of a colo- rectal carcinoma in the liver hilus is completely dif- ferent from one in the adjacent liver tissue. In order to clarify the notion that PET/CT is a new modality, observe the following: ﬂ uorodeoxyglucose (FDG), the radiopharmaceutical used predominantly in clinical practice is such a good “contrast agent,” that many of the routines established in standard CT imaging may be obsolete when designing imaging protocols for PET/CT. Having a small lymph node take up FDG will obviate the necessity to resort to subtle analyses of lesion shape, contrast media enhancement, and size on CT; thus, PET/CT may require less “CT-ology”

than the CT specialists are used to, and maybe more PET-driven CT image analysis. As an example, early results suggest that using a low-dose CT without intravenous contrast agent in conjunction with PET may be adequate in most settings of tumor staging (Hany et al. 2002a); thus, the future of PET/CT will rely on imaging specialists who are both competent readers of PET and CT images. To state this drasti- cally, no radiologist has ever had the idea that one specialist should look at T1-weighted MR scans, the other at T2-weighted scans, and the diagnosis is then made in conference! But this is precisely what some departments introducing PET/CT are currently practicing.

Thus, the approach to PET/CT in this chapter tries

to avoid political issues and assumes that the read-

ers of PET/CT images are competent in PET imaging

and CT imaging. The future will show how this will

be achieved politically. We are fortunate enough in

the Department of Nuclear Medicine at the Zurich

University Hospital to have several board-certiﬁ ed

radiologists on the staff, in addition to board-certi-

ﬁ ed nuclear physicians.

(2)

16.1.2 PET Imaging

The PET imaging is a nuclear medicine imaging modality, using radiopharmaceuticals such as standard nuclear medicine. The principal difference is that PET uses positron emitters rather than radionuclides, which emit gamma rays or X-rays. The ensuing two sections are a summary. More in-depth information may be obtained from von Schulthess (2003). The positron emitted by a nucleus decays according to a positron decay scheme, i.e., the nuclei of positron emitters, which have a neutron deﬁ cit, and the proton decays to a neutron with concomitant emission of a positron.

The positron, being a particle of antimatter, is not stable in matter. After emission, it is slowed down and at low kinetic energies it will annihilate with one of the elec- trons present plentifully in matter. This slow-down pro- cess occurs typically within 1–3 mm from the site where the nuclide emitted the positron, and is dependent on the emission energy of the positron. As this emission occurs randomly into any spatial direction, this is one of the main determinants of the lowest spatial resolution achievable in PET imaging. The annihilation reaction results in two gamma rays of 511 keV, which part from the site of annihilation at an angle of almost exactly 180°.

This angle is dictated by the physical laws of conserva- tion of energy and momentum: 511 keV is the equiva- lent energy of the electron (and positron) mass (two such particles annihilate to give rise to two gamma rays), and the angle is a result of momentum conversation. In fact, the angle is not precisely 180° but is determined by the net momentum of the electron and the positron at the time of annihilation and may vary by one or two degrees. This represents a second limit to the theoretical spatial resolution achievable in a PET scanner, which is around 2–3 mm, providing optimal scanner geometry and using F-18 as radiolabel.

In a high-end PET scanner the two gamma rays emitted are subsequently detected by a ring scintillation detector consisting of thousands of detector elements, which are connected by coincidence circuitry. If a detector detects one gamma photon, for a few nanosec- onds all other relevant detector elements are ready to detect the second photon from the annihilation event.

If indeed a second event is recorded, the radioactive decay of the positron emitter has had to take place on a (almost) straight line connecting the two detector ele- ments responding. The PET is thus a map of lines con- necting two detectors. It is intuitively clear that regions where many of these lines intersect correspond to hot spots observed in PET. Not all coincidences are true coincidences, but there are also random coincidences

and coincidences of gamma photons where one or both have been Compton scattered. The PET scanners have mechanisms to correct for these “false” events, but a dis- cussion would be beyond the scope of this chapter and the reader is referred to von Schulthess (2003).

The PET, such as single photon emission computed tomography (SPECT) is an emission scanning method.

This means that gamma rays from annihilation events buried deep inside the body are less likely to exit the patients’ body than those coming from annihilations close to the patients’ body surface. In order to obtain homogeneous images and to be able to quantify PET images, a so-called attenuation correction of the PET data, which are called emission data, has to be done.

Attenuation in PET is measured by using photons from a radioactive source, which rotates around the patient very much like a CT tube, and uses the gamma detectors of the PET scanner; however, the photon ﬂ ux of these sources is very low and attenuation correction scans are thus lengthy procedures. The attenuation data to correct the PET emission data are introduced into the image reconstruction process of the emission images and result in attenuation-corrected PET images. This is currently considered state-of-the-art PET imaging. A typical rule of thumb is that the measurement time for attenuation correction in PET systems is approximately half of that used to obtain emission data; thus, a 30-min PET scan is divided into 20-min emission image acqui- sitions and 10-min transmission image acquisition.

16.1.3 Radiopharmaceuticals

As stated, PET imaging requires radiopharmaceuticals

such as conventional nuclear medicine imaging. The

standard positron emitters available from cyclotrons

of 10–30 MeV are carbon-11, nitrogen-13, oxygen-15,

and ﬂ uorine-18. C-11, N-13, and O-15 have half-lives

of less than 30-min and are impractical for clinical

use beyond centers disposing of a cyclotron. Com-

pounds using F-18 as a label have physical half-lives

of 110-min; thus, central production and limited-area

shipment of F-18 based compounds is possible. By far,

the most widely used compound is F-18 FDG, which is

a glucose analog. It is taken up into cells and 6-phos-

phorylated like glucose, but no further metabolism is

notable; thus, FDG labels cellular glucose uptake and

metabolism, the latter because hexokinase responsible

for 6-phosphorylation is the rate limiting enzyme in

glycolysis. FDG happens to be an “unbelievably” good

molecule to label many malignant tumors and active

inﬂ ammatory sites, and thus it is FDG-PET, which is

(3)

currently in predominant use. Glucose is obligatorily used in the brain and excreted by the kidneys, but not substantially accumulated in other organs except the heart. As a result, the lesion-to-background ratio is excellent in most situations except for brain lesions.

There is muscular uptake in muscles exercised and uptake into nuchal fat in some patients as recently identiﬁ ed, but the pathophysiological signiﬁ cance of this fat uptake is not known at present (Hany et al.

2002b). The clinical data discussed in this chapter are based on FDG-PET only.

Many other F-18 labeled compounds have been synthesized and are under clinical evaluation. Of interest are a vast array of receptor ligands such as F-18 DOPA, amino acid metabolism tracers, such as F-18 ethyl-tyrosine (FET), F-18-based cell membrane proliferation markers, which are choline derivatives, and DNA synthesis markers such as F-18 thymidine (FLT). The clinical experience with these markers is currently too limited to decide what their future clinical role may be. FET may be a better marker of brain tumors than FDG, because it does not accumu- late much in the normal brain, and choline deriva- tives may have similar properties but in addition may be good markers in prostate carcinoma, where FDG mostly fails to identify lesions.

In units where a cyclotron is available, additional compounds can be used clinically. The most notable are H2O-15 water and N-13H3 ammonia for the quantitative assessment of brain and heart perfusion, respectively. In addition, many C-11 compounds can be used in such units, but the short half-life of 20-min of C-11 makes synthesis of relevant compounds dif- ﬁ cult and costly.

16.1.4 PET/CT

There are several good reasons for combining PET and CT into a single in-line imaging system, where the patient is moved from the CT to the PET gantry by simple table motion (Kinahan et al. 1998; von Schulthess 1999 ).

Such a system provides PET and CT images per- fectly co-registered by hardware arrangement pro- vided that the patient does not move. This is of major importance because PET scans of the body provide little anatomic information, and software image co- registration has been shown to be cumbersome and inaccurate by many authors.

At least as important and maybe the main incen- tive for manufacturers to promote and produce PET/

CT scanners is the fact that attenuation correction in a PET/CT scanner can be done as well with the CT data obtained as anatomic reference. This obviates the lengthy standard attenuation procedure used in PET and improves patient throughput by approxi- mately 30%, when a state-of-the-art CT scanner is incorporated into a PET/CT scanner. A CT scan is in fact an attenuation map, albeit measured with 70–140 keV polychromatic X-rays, whereas the PET attenuation scan is obtained with relatively mono- chromatic photons at 511 keV. Measurements have demonstrated, however, that the Hounsﬁ eld CT maps can be transformed into PET attenuation maps with relatively simple transformations requiring minimal computational efforts (Burger et al. 2002). The qualitative and quantitative properties of PET are maintained when CT is used for attenuation correc- tion (Kamel et al. 2002a).

The higher patient throughput of a PET/CT scan- ner compared with a PET-only scanner may offset the additional equipment cost at least in environments where the number of patients to be scanned per PET scanner is high. A net cost advantage results from the fact that FDG decays rapidly with a half-life of 110-min; thus, faster scanning results in more efﬁ - cient use of FDG, which may be delivered as an initial dose for several patients.

While a combination of PET and MR is also con- ceivable, using MR is difﬁ cult, because the MR images are not attenuation maps like the CT images. In the brain, where MR is superior to CT for anatomic imag- ing, software image co-registration is much simpler than with body data, and thus a PET/CT scanner is mainly an instrument which is useful in imaging of the human body from the head down to the feet.

It is likely that these synergistic effects obtained when adding CT to PET are responsible for the almost explosive growth of PET/CT imaging worldwide. One and a half years after the commercial announcement probably close to 100 PET/CT systems are installed and another 100 have been ordered.

16.1.5 Clinical Protocols in PET/CT

As PET/CT is a novel technology, the clinical proto-

cols in PET/CT are subject to fast evolution. At our

institution we routinely give diluted bowel contrast

agent 1 h prior to scanning (Dizendorf et al. 2002)

followed by supine FDG injection and patient rest

of at least 30 min. After bladder voiding just prior

to scanning, we ﬁ rst perform a low-dose CT scan

(4)

at 40 mAs (Hany et al. 2002a). This covers 100 cm of axial ﬁ eld of view in less than 30 s with a slice thickness of 4.5 mm, which is matched to the PET slice thickness. The CT scanner is a state-of-the-art 4-slice MDCT with a gantry rotation of 0.5 s mini- mum. Extensive evaluation of the CT breath-hold breathing pattern best matching the free-breathing pattern of PET data acquisition has led us to con- clude that an unforced end-expiratory state is best, during the period where the CT images the regions adjacent to the diaphragm (Goerres et al. 2002);

thus, patients are instructed to exhale and hold their breath when the CT scanner scans this body region.

As stated, CT scanning is accomplished in less than 30 s. Subsequently, PET scanning is started from the pelvis up. Starting PET scanning in the pelvis rather than the head obviates a potential major PET/CT mismatch in the bladder region. This is due to bladder ﬁ lling between CT data acquisition and the relatively lengthy PET scan, if started at the head. The PET scan is performed using six to seven table positions. As the PET scanner covers an axial ﬁ eld of view of around 15 cm with 32 slices per table position, approximately 90–105 cm of the patient are covered, which include the anatomic regions from the brain to the upper thighs in almost all patients.

With table position imaging times of 3–4 min, typi- cal scan length of a PET/CT partial-body scan cov- ering the patient from the head to the mid-upper thighs is 20–30 min.

After this “baseline” PET/CT data acquisition, additional standard CT protocols can be run depend- ing upon the clinical requirements. As is discussed in the second part dealing with lung tumor imaging, it is desirable in some settings to also perform a CT scan enhanced with i.v. contrast, which can better delin- eate the lesions in relation to vascular structures. The overall protocol design for such all-encompassing PET/CT examinations is not deﬁ ned and the next years will have to indicate where and when additional contrast-enhanced CT scanning is useful.

Other groups have advocated the use of i.v. con- trast-enhanced CT scans from the beginning and as only CT scan. It is our opinion that this is less than optimal for two reasons: ﬁ rstly, in many settings, CT contrast is not needed, as FDG is mostly a much better “contrast agent” than the contrast agents used in CT. Secondly, vascular contrast, which for vessel delineation is dense and transient, causes the CT data not to be ideal attenuation maps for the subse- quent PET scan. It has to be emphasized that iodin- ated contrast behaves in a way that it can be difﬁ cult to use a scan enhanced with intravenous contrast

agent for attenuation correction (Dizendorf et al.

2003); iodine looks like bone at 80–140 keV but like soft tissue at 511 keV; and thus the Hounsﬁ eld- to 511 keV transformation used for attenuation correc- tion may not work consistently without segmenting the vessels out of the CT and replacing the vessels with soft tissue density on the CT data). The addi- tional radiation burden to a patient with a low-dose CT at 40 mAs is in the range of 2–3 mSv, that of PET around 10 mSv, thus making a PET/CT examination with the above protocol one with a lower radiation dose or comparable to standard CT.

16.1.6 Critical Appraisal of Clinical PET/CT

Clinical experience with PET/CT is currently limited;

however, after performing approximately 3000 clini- cal scans thus far and having been the ﬁ rst users of a clinical system (an initial prototype combining a low-end PET and CT has been operational at the uni- versity of Pittsburgh since 1998), we can conﬁ dently appraise at least some aspects of clinical PET/CT.

PET/CT is easy to use and in the overwhelming number of patients, image co-registration is excellent.

This is probably due to the fact that oncology patients, who represent the major patient group imaged, are generally cooperative. PET/CT is more specifi c than PET. To our surprise, routine availability of an ana- tomic reference frame has led to the identifi cation of various pitfalls not fully appreciated in PET prior to the introduction of PET/CT. The superimposition of an FDG avid focus onto an anatomically identifi able structure makes the image interpreter more confi - dent of what he sees and reduces interobserver vari- ability. This also applies to the use of PET/CT data in radiation planning. The software infrastructure has now been developed which permits transfer of PET and CT data into the planning environment of radia- tion oncologists by DICOM standard formats. The PET/CT is more sensitive than PET because some lesions, particularly in the lung, which are too small to be noted on PET, are recognized on CT.

Viewing PET and CT images next to each other

is not adequate for lesion localization once lesions

are below approximately 2 cm, and it is particularly

in such lesions where PET excels and where stan-

dard morphological criteria of malignancy in CT

no longer are very useful; thus, co-registration and

fusion or linked cursor viewing are mandatory. All

currently available data suggest that the software

approach is at least logistically difﬁ cult and obviously

(5)

Fig. 16.1a–g. A 55-year-old woman with a large cell carcinoma in the left lung.

PET/CT was performed for preoperative staging. a Maximum intensity projection (MIP) PET scan, b, e axial unenhanced low-dose CT scans, c, f axial CT-attenua- tion corrected PET scans, and d, g axial co-registered PET/CT scans. Diagnosis of PET/CT imaging: lung cancer in the left hilar region. Lymph node metastases in the aorto-pulmonary window, subcarinally, and in the contralateral mediastinum, making the patient inoperable

g a

c

e b

d

f

(6)

does not provide easily accessible data for attenua- tion correction.

Any new modality has also new artifacts and pitfalls. The most relevant artifact identiﬁ ed so far are misregistration artifacts around the diaphragm (Fig. 16.2d). They lead to abdominal lesions appar- ently located in the supradiaphragmatic lung zones and a photon deﬁ cit of variable importance overlying the diaphragmatic domes and looking like “bananas”

on the coronal CT attenuation corrected PET scans.

A second artifact, which may be more prominent in CT-corrected PET scans than PET scans corrected with the standard transmission scan methods, is that arising around metallic implants. Slight misregistra- tions can lead to overcorrections, which then result in focal or linear regions of apparently increased FDG uptake. The last artifact comes from major misreg- istration due to patient motion between the acquisi- tion of the two scans.

16.2 Clinical Applications of PET/CT

in the Chest

Lung cancer staging is based on the TNM system and requires accurate characterization of the pri- mary tumor (T stage), regional lymph nodes (N stage), and extrathoracic metastases (M stage).

Accurate tumor staging is essential for choosing the appropriate treatment strategy. The majority of PET imaging work has been done in non-small cell lung cancer (NSCLC). It has been shown that FDG PET is highly accurate in classifying lung nodules as benign or malignant. Whole-body PET improves the rate of detection of mediastinal lymph node metastases as well as extrathoracic metastases when compared with conventional imaging methods such as CT, MR, ultrasound, or bone scan. Since commercial PET scanners provide nominal spatial resolution of 4.5–6 mm in the center of the axial ﬁ eld of view, even lesions less than 1 cm with an increased FDG uptake can be detected. This represents a critical advantage of PET over CT and MR. Recently, integrated PET/CT scanners have been introduced. Integrated PET/CT enables the exact matching of focal abnormalities on PET to anatomic structures on CT. First clinical results of integrated PET/CT imaging show that PET/

CT is superior in diagnostic accuracy in comparison with PET alone, CT alone, and visual correlation of PET and CT (Hany et al. 2002a; Lardinois et al.

2003 ).

16.2.1 Solitary Lung Nodule

The ability of PET to separate between benign and malignant lesions is high, but not perfect. For benign lesions, a high specifi city for FDG PET has been demonstrated. It has been shown that FDG PET is highly accurate in differentiating malignant from benign solitary pulmonary nodules (0.6–3 cm) when radiographic fi ndings were indeterminate (Gupta et al. 1996). In a series of 61 patients, PET had a sensitivity of 93% and a specifi city of 88% for detecting malignancy; however, FDG PET may show negative results for pulmonary carcinoid tumors and bronchiolo-alveolar lung carcinoma. Lesions with increased FDG uptake should be considered malignant, although false-positive results have been reported in cases of infl ammatory and infectious pro- cesses, such as histoplasmosis, aspergillosis, or active tuberculosis. The PET is clinically useful in patients with a solitary pulmonary nodule less than 3 cm in diameter, especially where biopsy may be risky or where the nodule carries a low risk for malignancy based on patients’ history or radiographic fi ndings.

With integrated PET/CT an additional certainty to the presence or absence of FDG uptake in the pul- monary nodule can be achieved.

16.2.2 T Staging

Without image fusion, the use of PET in T staging

lung cancer is limited. Recently, it has been shown that

integrated PET/CT is superior to CT alone, PET alone,

and visual correlation of PET and CT in T staging of

patients with NSCLC (Lardinois et al. 2003). Due to

the exact anatomic correlation of the extent of FDG

uptake, the delineation of the primary tumor can be

deﬁ ned precisely; therefore, the diagnosis of chest

wall inﬁ ltration and the mediastinal invasion by the

tumor is improved. Lesions with chest wall inﬁ ltration

are classiﬁ ed as stage T3 and are potentially resect-

able. Integrated PET/CT provides important informa-

tion on mediastinal inﬁ ltration, too; however, PET/CT

imaging is unable to distinguish contiguity of tumor

with the mediastinum from the direct invasion of the

walls of mediastinal structures. It has been shown

that FDG PET is a useful tool for the differentiation

between tumor and peritumoral atelectasis. This is

particularly important for the planning of radio-

therapy in patients with lung cancer associated with

atelectasis. The information provided by FDG PET

(7)

Fig. 16.2a–g. A 69-year-old woman. Computed tomography demonstrated a lung cancer in the upper lobe of the right lung. PET/CT was performed for preoperative staging. a MIP PET scan, b, e axial unenhanced low-dose CT scans, c, f axial CT- attenuation corrected PET scans, and d, g axial coregistered PET/CT scans. Diagno- sis of PET/CT imaging: Lung cancer in the right lung with central necrosis. Small ipsilateral mediastinal lymph-node metastasis. Liver metastasis in segment IV. Note slight respiration induced mismatch of FDG uptake and anatomic tumor extent in d. This is due to the fact that PET is acquired during free breathing, whereas CT is acquired during respiratory arrest (end expiration is preferable)

g a

c

e b

d

f

(8)

results in a change in the radiation ﬁ eld in approxi- mately 30–40% of patients (Nestle et al. 1999).

16.2.3 N Staging

The PET has proven to be a very effective staging modality for mediastinal nodal staging (Steinert et al. 1997; Vansteenkiste et al. 1998; Dwamena et al. 1999; Pieterman et al. 2000; Hellwig et al.

2001). CT and MR imaging are limited in depicting small mediastinal lymph node metastases. Several studies have demonstrated that FDG PET is signiﬁ - cantly more accurate than CT in determination of nodal status (Gupta et al. 1996; Nestle et al. 1999;

Steinert et al. 1997). In our own study of 47 patients, PET assigned the correct N stage in 96% of cases and CT was correct in 79% of cases (Steinert et al. 1997).

Dwamena et al. (1999) performed a meta-analytic comparison of PET and CT in mediastinal staging of NSCLC. The mean sensitivity and speciﬁ city (±95%

CI) were 0.79±0.03 and 0.91±0.02, respectively, for PET and 0.60±0.02 and 0.77±0.02, respectively, for CT (Dwamena et al. 1999). These results were con- ﬁ rmed in another meta-analysis with a total of more than 1000 patients (Hellwig et al. 2001).

Even if mediastinoscopy remains the gold standard for mediastinal staging, not all mediastinal lymph nodes are routinely accessed by use of mediastinos- copy, particularly in the para-aortic region and in the aorto-pulmonary window. The limited view through the scope and the single direction in which biopsies can be carried out prevents 100% accuracy. The accuracy of mediastinoscopy is approximately 90% and is surgeon dependent (Patterson et al. 1987). Recently, it has been demonstrated that PET is useful to assist medi- astinoscopy. Due to the knowledge of PET information, mediastinoscopy revealed additional mediastinal dis- ease in 6% of patients (Kernstine et al. 2002).

Exact allocation of focal abnormalities on PET to specifi c lymph nodes is diffi cult or even impossible due to the poor anatomic information provided by PET alone. The presence and site of lymph node metastases should be recorded according to the revised American Thoracic Society lymph node station-mapping system (Mountain and Dresler 1997). In patients with bulky mediastinal disease or multilevel nodal involvement the assessment of N stage is easy; however, the exact localization of lymph node metastases in the hilus is diffi cult. Lymph nodes distal to the mediastinal pleural refl ection and within the visceral pleura are classifi ed as N1 nodes. Lymph

nodes within the mediastinal pleural envelope are classifi ed as N2 nodes. Because the pleura is visible neither in CT nor in PET, the exact classifi cation of a hilar node as N1 node or N2 node remains diffi - cult. The diffi culty of PET is the localization of small single nodes, particularly in patients with a mediasti- nal shift due to an atelectasis or anatomical variants.

In our experience, integrated PET/CT imaging will become the new standard of mediastinal staging.

The high reliability of integrated PET/CT in the exact localization of extrathoracic vs intrathoracic and mediastinal vs hilar lymph nodes might have very important therapeutic implications.

Until now, our group used non-enhanced CT scans for integrated PET/CT imaging. We could not ethically justify the use of vascular contrast material, because all patients had a conventional contrast- enhanced CT for staging before. In non-enhanced CT scans delineation of vessels was considerably poorer than with contrast enhancement, or impos- sible; however, in our patient series non-enhanced PET/CT scans are suffi cient for planning surgery in approximately 80% of patients. Further evaluation is necessary to defi ne conditions in which the applica- tion of intravascular contrast material might have an additional diagnostic impact in integrated PET/CT imaging; however, regarding infi ltration of hilar and mediastinal vessels, a relatively low sensitivity, speci- fi city, and accuracy (68, 72, and 70%, respectively) of conventional CT scan with contrast enhancement has been observed (Rendina et al. 1987).

Microscopic foci of metastases within very small lymph nodes cannot be detected with any imag- ing modality. If there is no increased FDG uptake in PET, integrated PET/CT will not provide further information based on FDG accumulation. Recently, it has been reported that FDG PET after induction therapy is less accurate in mediastinal staging than in staging of untreated NSCLC (Akhurst et al. 2002).

The PET overstaged nodal status in 33% of patients, understaged nodal status in 15%, and was correct in 52%. Future studies are required to correlate FDG PET prior to and after treatment.

16.2.4 M Staging

Whole-body FDG PET is an excellent method to screen for extrathoracic metastases (Weder et al.

1998). In a meta-analysis of 581 patients, sensitivity,

speciﬁ city, and accuracy of FDG PET were 94, 97,

and 96%, respectively (Hellwig et al. 2001). Cur-

(9)

rent imaging methods are inadequate for accurate M staging of patients. The PET detects unexpected extrathoracic metastases in 10–20% of patients and changes therapeutic management in approximately 20% of patients. The FDG PET is more accurate than CT in the evaluation of adrenal metastases (Erasmus et al. 1997). Marom et al. (1999) compared the accu- racy of FDG PET to conventional imaging in 100 patients with newly diagnosed NSCLC. Comparing bone scintigraphy and FDG PET in detection bone metastases, the accuracy was 87 and 98%, respec- tively. All hepatic metastases were correctly identi- ﬁ ed with PET and CT. With CT, however, benign liver lesions were overstaged as metastases; thus, accuracy of PET was superior to CT in the diagnosis of liver metastases.

The clinical signiﬁ cance of a single focal abnor- mality on PET remains unclear, especially when no morphological alterations occur on CT images. The advantage of integrated PET/CT imaging is the exact localization of a focal abnormality on PET. This was the case in 20% of all patients with extrathoracic metastases in our study on the value of integrated PET/CT (Lardinois et al. 2003).

16.2.5 Recurrent Lung Cancer

A very high accuracy of FDG PET in distinguishing recurrent disease from benign treatment effects has been shown. If PET images demonstrate areas of tumor viability, they can direct biopsy for patho- logic conﬁ rmation. Patients should be evaluated a minimum of 2 months after completion of therapy.

Otherwise post-therapeutic healing processes or radiation pneumonitis may result in false-positive PET fi ndings. These abnormal fi ndings return to normal at variable times without further interven- tion. In the experience of Inoue et al. (1995), a cur- vilinear contour of increased FDG accumulation was seen mostly in infl ammatory lesions, whereas focal nodular uptake was seen mostly in recurrent tumors.

Their data suggest that FDG PET can be clinically used for selecting biopsy sites because of its high sensitivity in detecting recurrent lung cancer.

16.2.6 Pitfalls

False-negative FDG PET results have been reported in pulmonary carcinoid tumors and in bronchiolo-

alveolar carcinomas. Some active infectious or infl ammatory lesions may have an increased FDG uptake. Tuberculosis, eosinophilic lung disease, histoplasmosis, aspergillosis, and other infections may have a signifi cant uptake of FDG (Strauss 1996). Furthermore, sarcoid shows a typical bilateral relatively symmetric hilar uptake pattern. Therefore, lesions with an increased FDG accumulation should be histologically confi rmed; however, most chronic infl ammatory processes do not signifi cantly take up FDG.

It is well known that active muscles accumulate FDG. In some patients with lung cancer an intense focal FDG accumulation is seen in the lower anterior neck just lateral to the midline. Co-registered PET/

CT images revealed that the focal FDG uptake was localized in the internal laryngeal muscles (Kamel et al. 2002b). This ﬁ nding is a result of compensatory laryngeal muscle activation caused by contralateral recurrent laryngeal nerve palsy due to direct nerve invasion by lung cancer of the left mediastinum or lung apices. The knowledge of this ﬁ nding is impor- tant to avoid false-positive PET results.

16.2.7 Small Cell Lung Cancer

The staging procedures for small cell lung cancer (SCLC) do not differ from those for NSCLC. The pri- mary role for imaging is to separate accurately limited disease (LD) from extended disease (ED). Based on the widespread dissemination of SCLC, a battery of imaging tests is performed such as CT of the chest and abdomen, CT, or MRI of the brain and a bone scan.

Recently, it has been shown that whole-body FDG PET is a useful tool for staging SCLC (Schumacher et al. 2001). The FDG PET is superior to conventional staging in the detection of all involved sites, and particularly in the assessment of mediastinal lymph node metastases. Our ﬁ rst experience demonstrate that integrated PET/CT imaging in SCLC is a highly valuable tool for planning radiation treatment. It is useful for accurate target deﬁ nition by reducing the probability of overlooking involved areas.

16.2.8 Malignant Pleural Mesothelioma

Similarly to lung cancer, excellent FDG uptake in

malignant pleural mesothelioma (MPM) has been

previously described (Marom et al. 2002).

(10)

Schneider et al. (2000) demonstrated that PET is particularly valuable for distinguishing between benign and malignant pleural processes. The FDG is not taken up in pleural ﬁ brosis; thus, differential diagnosis of the pleural lesions is possible.

PET imaging is useful in localizing the areas involved with MPM; however, PET and CT are unable to differentiate MPM from pleural adenocarcinoma, so that histology is needed for conﬁ rmation.

The role of PET is to document the extent of pleural disease, to establish mediastinal lymph node involve- ment, to evaluate tumor invasion, and to diagnose recurrence. Our experience demonstrates that inte- grated PET/CT imaging is an excellent method for staging patients with MPM. With the co-registration of anatomic and metabolic information, the extent of the tumor can be precisely deﬁ ned. Small mediastinal lymph node metastases can be detected and precisely localized. Integrated PET/CT imaging is helpful in identifying the optimal biopsy site thereby increasing diagnostic accuracy of the histological examination.

References

Akhurst T, Downey RJ, Ginsberg MS et al. (2002) An initial experience with FDG-PET in the imaging of residual disease after induction therapy for lung cancer. Ann Thorac Surg 73:259–266

Burger G, Goerres S, Schoenes A, Buck A, Lonn HR, Schulthess GK von (2002) PET attenuation coefﬁ cients from CT images: experimental evaluation of the transformation of CT into PET 511-keV attenuation coefﬁ cients. Eur J Nucl Med 29:922–927

Dizendorf E, Hany TF, Buck A, Schulthess GK von, Burger C (in press) Cause and magnitude of the error induced by oral CT contrast agent in CT-based attenuation correction of PET emission studies. J Nucl Med

Dizendorf EV, Treyer V, Schulthess GK von, Hany TF (2002) Application of oral contrast media in coregistrated positron emission tomography–CT. Am J Roentgenol 179:477–481 Dwamena BA, Sonnad SS, Angobaldo JO, Wahl RL (1999)

Metastases from non-small cell lung cancer: mediastinal staging in the 1990s – meta-analytic comparison of PET and CT. Radiology 213:530–536

Erasmus JJ, Patz EF, McAdams HP et al. (1997) Evaluation of adrenal masses in patients with bronchogenic carcinoma using 18F-ﬂ uorodeoxyglucose positron emission tomography. AJR 168:1357–1362

Goerres GW, Kamel E, Heidelberg TN, Schwitter MR, Burger C, Schulthess GK von (2002) PET–CT image co-registration in the thorax: inﬂ uence of respiration. Eur J Nucl Med 29:

351–360

Gupta NC, Maloof J, Gunel E (1996) Probability of malignancy in solitary pulmonary nodules using ﬂ urine-18-FDF and PET. J Nucl Med 37:943–948

Hany TF, Steinert HC, Goerres GW, Buck A, Schulthess GK

von (2002a) Improvement of diagnostic accuracy of PET imaging using a high performance in-line PET–CT system:

preliminary results. Radiology 225:575–581

Hany TF, Gharehpapagh E, Kamel EM, Buck A, Himms-Hagen J, Schulthess GK von (2002b) Brown adipose tissue: a factor to consider in symmetrical tracer uptake in the neck and upper chest region. Eur J Nucl Med 29:1393–1398

Hellwig D, Ukena D, Paulsen F, Bamberg M, Kirsch CM (2001) Metaanalyse zum Stellenwert der Positronen-Emissions- Tomographie mit F-18-Fluorodesoxyglukose (FDG-PET) bei Lungentumoren (in German). Pneumologie 55:

367–377

Inoue T, Kim E, Komaki R et al. (1995) Detecting recurrent or residual lung cancer with FDG-PET. J Nucl Med 36:

788–793

Kamel E, Hany TF, Burger C et al. (2002a) CT vs 68Ge attenuation correction in a combined PET/CT system: evaluation of the effect of lowering the CT tube current. Eur J Med 29:312–318

Kamel E, Goerres GW, Burger C, Schulthess GK von, Steinert HC (2002b) Detection of recurrent laryngeal nerve palsy in patients with lung cancer using PET–CT image fusion.

Radiology 224:153–156

Kernstine KH, McLaughlin KA, Menda Y et al. (2002) Can FDG PET reduce the need for mediastinoscopy in potentially resectable nonsmall cell lung cancer? Ann Thorac Surg 73:394–402

Kinahan PE, Townsend DW, Beyer T, Sashin D (1998) Attenu- ation correction for a combined 3D PET/CT scanner. Med Phys 25:2046–2053

Lardinois D, Weder W, Hany TF, Kamel EM, Korom S, Seifert B, von Schulthess GK, Steinert HC (2003) Integrated PET/CT imaging improves staging on non-small-cell lung cancer. N Engl J Med. 348(25):2500–2507

Marom EM, McAdams HP, Erasmus JJ (1999) Staging non- small cell lung cancer with whole-body PET. Radiology 212:803–809

Marom EM, Erasmus JJ, Pass HI, Patz EF Jr (2002) The role of imaging in malignant pleural MPM. Semin Oncol 29:

26–35

Mountain CF, Dresler CM (1997) Regional lymph node clas- siﬁ cation for lung cancer staging. Chest 11:1718–1723 Nestle U, Walter K, Schmidt S, Licht N et al. (1999) 18F-deoxyglu-

cose positron emission tomography (FDG-PET) for the planning of radiotherapy in lung cancer: high impact in patients with atelectasis. Int J Radiat Oncol Biol Phys 44:593–597 Patterson GA, Ginsberg RJ, Poon PY et al. (1987) A prospec-

tive evaluation of magnetic resonance imaging, computed tomography, and mediastinoscopy in the preoperative assessment of mediastinal node status in bronchogenic carcinoma. J Thorac Cardiovasc Surg 94:679–684

Pieterman RM, van Putten JWG, Meuzelaar JJ et al. (2000) Preoperative staging of non-small-cell lung cancer with positron-emission tommography. N Engl J Med 343:

254–261

Rendina EA, Bognolo DA, Mineo TC et al. (1987) Computed tomography for the evaluation of intrathoracic invasion by lung cancer. J Thorac Cardiovasc Surg 94:57–63

Schneider DB, Clary-Macy C, Challa S et al. (2000) Positron emission tomography with F18-ﬂ uorodeoxyglucose in the staging and preoperative evaluation of malignant pleural MPM. J Thorac Cardiovasc Surg 120:128–133

Schumacher T, Brink I, Mix M et al. (2001) FDG-PET imaging

(11)

for the staging and follow-up of small cell lung cancer. Eur J Nucl Med 28:483–488

Steinert HC, Hauser M, Allemann F et al. (1997) Non-small cell lung cancer: nodal staging with FDG PET versus CT with correlative lymph node mapping and sampling. Radiology 202:441–446

Strauss LG (1996) Fluorine-18 deoxyglucose and false-positive results: a major problem in the diagnosis of oncological patients. Eur J Nucl Med 23:1409–1414

Vansteenkiste JF, Stroobants SG, de Leyn PR et al. (1998) Lymph node staging in non-small cell lung cancer with

FDG PET scan: a prospective study on 690 lymph node stations from 68 patients. J Clin Oncol 16:2142–2149 Schulthess GK von (ed) (2003) Clinical molecular anatomic

imaging: PET, PET/CT and SPECT/CT. Lippincott Williams and Willkins, New York

Schulthess GK von (200) Cost considerations regarding an integrated CT–PET system. Eur Radiol 10:S377–S380 Weder W, Schmid R, Bruchhaus H, Hillinger S, Schulthess GK

von, Steinert HC (1998) Detection of extrathoracic metastases by positron emission tomography in lung cancer. Ann Thorac Surg 66:886–893

(12)