John O Prior
Nuclear Medicine Department
CHUV University Hospital Lausanne
Lung cancer is commonly affecting around three million new patients each year and is the leading cause of cancer- related deaths in many countries. Nonsmall-cell lung carcinoma (NSCLC) accounts for 80% of all lung cancers and remains a dreadful disease, with a five-year overall survival rate of 14%. At the time of diagnosis, lung cancer has generally spread to adjacent hilar or mediastinal lymph nodes in about 25% of patients, and 35–45% of patients have extrathoracic metastases. A systematic approach to diagnosis and staging is necessary to optimise therapy for each individual patient. Positron emission tomography (PET) with (18)F-fluorodeoxyglucose (FDG) has emerged as an important tool for characterising pulmonary nodules and lung cancer staging.(1,2)
The uptake of (18)F-FDG in untreated NSCLC is directly proportional to the expression of glucose transporter-1 and to the grade of malignancy within a given tumour type. Tumours that are highly aggressive demonstrate a markedly increased glycolytic activity. Conversely, slow-growing, more indolent tumours such as bronchoalveolar carcinomas or primary pulmonary carcinoids with lower glucose utilisation rate may have reduced detection sensitivity. Other known false-negatives are the imaging of small lesions <0.7cm due to partial volume effect, finite spatial resolution of PET scanners and respiratory movements. Additionally, (18)F-FDG uptake is not specific to neoplasia, and many inflammatory processes can lead to false-positive results (pneumonia, tuberculosis, aspergillosis, sarcoidosis, histoplasmosis, rheumatoid nodules and other granulomatous diseases). PET has been found to have >90% accuracy in separating a single pulmonary nodule into benign or malignant disease.(1) Thus, it is reasonable to adopt a watchful waiting attitude in PET-negative nodules, while PET-positive nodules warrant histopathological confirmation. PET may also serve as “metabolic” biopsy in the context of severe emphysema or bleeding disorders. Several studies have demonstrated that PET is a useful adjunct in the diagnosis and staging of NSCLC for routine clinical use.(3) It has higher accuracy for diagnosing and staging and can reduce the number of unnecessary thoracotomies in patients with potentially resectable disease.(2) Current guidelines recommend preoperative PET in patients without distant metastatic disease on computed tomography (CT) and mediastinal sampling of PET-positive abnormal lymph nodes before primary tumour surgery (both grade of evidence B).(4) Following surgical resection, about 40% of patients with clinical stage I disease and about 60% of patients with clinical stage II disease ultimately experience recurrent disease. This is presumably due to the presence of micrometastatic disease at the time of presentation. It is known that neither PET nor other imaging modalities can currently detect such micrometastases. In fact, a recent in-vitro study has estimated a threshold of 106 malignant cells for a tumour to be detected in the clinical setting.(5) This corresponds to a size of approximately 1mm, which is below the intrinsic PET resolution (4–6mm) but nevertheless was detected by PET in phantoms.(5)
Staging of NSCLC with PET
Once the diagnosis of NSCLC has been made, prognosis and treatment can be determined according to the TNM staging system.(6) Until recently, the preoperative management of NSCLC consisted of identifying N2, N3 or M1 disease, which are in general not amenable to surgery. However, with the recent advent of neoadjuvant radiochemotherapy before resection showing improved survival,(7) the accurate preoperative determination of T2, T3 and N1 status has regained new interest.
Extent of primary tumour (T)
CT imaging is intrinsically better than PET alone at defining the extent of primary tumour and pleural chest involvement because of its higher resolution.
Extent of regional lymph node involvement (N)
PET has shown consistently higher accuracy (sensitivity 85%, specificity 90%) than CT (sensitivity 61%, specificity 79%) in noninvasive staging of mediastinum in many studies,(2) particularly for assessing N1 and N2 disease. PET is also superior to CT regarding negative predictive value (87–100%). The opinion that negative PET findings in the mediastinum obviate the need for invasive mediastinoscopy currently prevails, but a few studies advocate the general use of lymph node biopsies for staging the mediastinum. However, risks are inherent to mediastinoscopy, with 0.2% mortality and 0.2–2.5% morbidity rates. A recent meta-analysis has found that negative PET in lymph nodes >16mm leads to a post-test probability for N2 disease of 21%.(8) This indicates that mediastinoscopy prior to thoracotomy in such large-sized PET-negative lymph nodes may prevent unnecessary thoracotomies. When positive, PET is considered nonspecific, and mediastinoscopy is generally recommended.(4)
Distant metastases (M)
Their presence at the time of diagnosis generally precludes surgery. PET has been shown to detect distant involvement undetected by conventional staging procedure in about 15% of patients and can negate false-positive or equivocal findings in conventional imaging.(9) Adrenal masses are found in up to 20% of patients with NSCLC but represent metastases in only 60% of the cases. PET can help characterise such lesions as benign or malignant with a high negative predictive value. Bone metastases from NSCLC are frequently encountered. PET was found to be more sensitive for osteolytic bone metastases than conventional scintigraphy, with fewer false-positive results due to degenerative or ancient posttraumatic bone disease. For liver metastases in NSCLC, studies aiming to show the superiority of one technique over the other brought mixed results. Yet, PET can sometimes find liver metastases in indeterminate contrast-enhanced CT abnormalities. PET may be helpful in determining the malignant or benign nature of pleural effusion diagnosed by CT. For brain imaging, optimal detection can be achieved with contrast-enhanced CT or MRI, and a dedicated PET brain acquisition is rarely performed in routine clinical practice. Nevertheless, guidelines advise against routine neuroimaging unless symptoms or signs of central nervous system involvement are noted.
Assigning disease status
Stages 0, I, II and IIIA are considered as potentially resectable from the perspective of residual lung function and overall surgical risk. Stage IIIB (extensive, unresectable invasion of local structures and/or contralateral mediastinal or any supraclavicular or scalene lymph node involvement) and stage IV (presence of metastases) have traditionally been treated with radiochemotherapy. There is currently interest in combining surgery and neoadjuvant radiochemotherapy for patients with stage IIIA disease or in postoperative chemotherapy for stage I patients.
Since its clinical introduction in 2001, integrated PET/CT has become the fastest-growing imaging modality on the market. It allows the fusion of morphological CT and functional PET images acquired closely in time from each other, translating into better localisation of FDG-avid lesions (see Figure 1). This technology has rapidly gained widespread acceptance, and practically every new commercial installation is an integrated PET/CT device.(10) Results of a number of nonrandomised studies have found that integrated PET/CT outperforms CT or PET alone to accurately stage NSCLC (see Table 1).(11–16) In addition, these integrated systems offer improved image quality, shorter imaging times (by about 30%) and increased patient convenience. The advantages of PET/CT over PET for T-staging are essentially derived from the CT component, which allows a more precise delineation of the tumour. In case of infiltration of hilar or mediastinal vessels, the PET/CT may be performed with a contrast agent for increased accuracy. Of interest, in one study, PET/CT was also shown to be more sensitive and accurate to characterise single pulmonary nodule than helical CT, even when using state-of-the-art dynamic CT acquisition protocols.(17)
With regard to the nodal staging, PET/CT had superior accuracy over PET alone in three out of five studies performed to date.(11,13,16) In the other two studies, PET/CT had only a slightly but not significantly better accuracy than PET alone, but these studies may have been under-powered.(12,15) This improvement is brought by a superior ability to localise metastatic lymph nodes. For instance, the distinction of a solitary lymph node metastasis in the hilum is notoriously difficult, and the information brought by PET/CT may be crucial to differentiate nonresectable N2 disease from resectable N1 disease.(10) The identification of an occult supraclavicular lymph node (N3 disease) also has important therapeutic implications and is known to occur in about 15% of patients, which would otherwise be qualified as N2 disease on the basis of mediastinoscopy alone.
For M-staging, PET/CT can pinpoint the exact localisation of a focal FDG-avid abnormality, which may be present in 20% of NSCLC patients with extrathoracic disease.(11) PET/CT may decrease the number of false-positive findings found by PET alone (eg, with CT helping to distinguish degenerative disease from bone metastases or to detect osteosclerotic metastases that often have normal metabolic activity on PET). In addition, PET/CT seems to provide excellent positive predictive value for adrenal metastases.(18) PET/CT clearly affects staging of NSCLC, as shown in a prospective series of 350 patients with solitary pulmonary nodule, in which about 10% had findings related to other malignancies or benign lesions that would have been overlooked with conventional staging.(19)
Regarding the overall TNM staging, PET/CT is significantly more accurate than PET, with increased observer confidence resulting in fewer equivocal findings.(12,15,16) For centres where PET/C is not available, PET fusion imaging with CT may be the next best choice, although it is subject to many practical limitations related to differences in patients and internal organs positioning between the PET and the CT studies, often performed on separate days.(15,16)
Change in patient management due to PET/CT
With more accurate staging comes more appropriate treatment and, thus, fewer invasive unnecessary interventions. It has been found that, in about 20–40% of patients, the use of PET/CT impacted patient management, with the exclusion of surgery in about 15% of the patients due to unexpected detection of distant metastases. (11–13) In a recent multisite Dutch study in 465 patients, Herder et al found that including PET alone as first-line imaging in patient with (suspected) lung cancer did not reduce the overall number of diagnostic tests and procedures, except the number of invasive tests requiring general anaesthesia. This strategy reduced the number of mediastinoscopies without significantly changing the overall costs.(20)
Future role for PET/CT in lung cancer
The prognosis of NSCLC is primarily determined by disease stage, but other independent factors may be important, such as tumour aggressiveness and metabolic activity. Preliminary data indicate that patients with higher (18)F-FDG uptake in the tumour(21) or diffuse uptake in the bone marrow(22) may have a shorter survival time. Another emerging area of PET/CT is the assessment of the metabolic response to chemo-therapy early in the course of treatment (as early as after the first cycle), with decrease in tumour uptake being correlated with the final response. This could allow switching from ineffective first-line chemotherapy to a more efficacious second-line treatment. An important area for PET/CT seems to be the restaging after preoperative neoadjuvant radiochemotherapy in patients with locally advanced lung disease, such as demonstrated recently by Pöttgen et al, with significant correlations between tumour metabolic activity, response and outcome.(23) Indeed, patients who do not respond to neoadjuvant chemotherapy rarely benefit from surgery. After therapy, PET has been shown to be helpful to determine recurrent lung cancer in the presence of new intrathoracic masses, and PET/CT may further help in this setting, although a broader use as a surveillance tool remains controversial and is to be determined in well-designed, prospective, multisite trials.
Radiation therapy planning is another field benefiting from PET/CT. It has been found that the addition of PET changes gross tumour volume in 20–100% of the patients. However, outcome studies are needed to determine the impact on tumour response and patient survival. Important technical problems remain to be solved, such as criteria to define active tumour extent and how to manage respiratory motion (eg, with respiratory-gated PET/CT).
Radiopharmaceuticals other than (18)F-FDG are of great interest. Clinical studies are being performed in lung cancer patients aiming at targeting cell membrane synthesis ((11)C-methionine and (11)C-choline), lipid ((11)C-acetate) and DNA synthesis ((18)F-fluorothymidine), which may present a higher specificity, at least theoretically. Preliminary human studies seem to confirm this higher specificity but so far show a poor sensitivity. More studies are needed to assess their role in monitoring therapy. Hypoxia imaging with (60)Cu-ATSM or (18)F- fluoromisonidazole is also of great interest for radiation oncologists and also to oncologists with the newest chemotherapeutic agents targeting hypoxic cells (eg, tirapazamine).(24)
PET/CT in small-cell lung cancer and mesothelioma
Accounting for about 20% of lung cancers, small-cell lung carcinoma (SCLC) is an aggressive disease, often disseminated at the time of diagnosis. Here, a simpler TNM staging system is applied. Disease limited to one hemithorax (“limited disease”) is treated by radiochemotherapy, while more extensive disease is conventionally treated by chemotherapy alone. Fewer studies have examined the use of PET (none for PET/CT) in SCLC. Initial staging is more accurate with PET, which usually shows more extensive disease than conventional imaging, except in the brain, as demonstrated by Brink et al in a prospective study in 120 patients.(25) Similarly to NSCLC, tumour uptake seems to provide prognostic information independent of disease stage, with higher uptake yielding to worse prognosis. In mesothelioma (<2% of lung cancers), limited data show PET/CT to be more promising over CT alone to determine patients potentially responding to curative aggressive surgery, but it mainly leads to upstaging of patients and a decreased rate of unneeded surgery.(26)
Although the overall survival rate of lung cancer remains dismal, accurate assessment of the stage of the disease is crucial and constitutes the foundation for promising multidisciplinary combined treatments involving surgical resection, radiation therapy and/or chemotherapy. While PET alone has already brought major advances in patients with lung cancer, the future of PET/CT looks even more promising and extends well beyond lung cancer staging.
- Gould MK et al. JAMA 2001;285:914-24.
- Gould MK et al.Ann Intern Med 2003;139:879-92.
- Vansteenkiste J, Fischer BM, Dooms C, Mortensen J. Lancet Oncol 2004;5:531-40.
- Silvestri GA, et al. Chest 2003;123 Suppl:147S-56S.
- Fischer BM, et al. Eur J Nucl Med Mol Imaging 2006;DOI: 10.1007/s00259-005-0038-6.
- Mountain CF. Chest 1997;111:1710-7.
- Depierre A, et al. J Clin Oncol 2002;20:247-53.
- de Langen AJ, et al.. Eur J Cardiothorac Surg 2006;29:26-9.
- Pieterman RM,et al. N Engl J Med 2000;343:254-61.
- von Schulthess GK, Steinert HC, Hany TF. Integrated PET/CT: current applications and future directions. Radiology 2006;238:405-22.
- Lardinois D, et al. N Engl J Med 2003;348:2500-7.
- Antoch G, et al. Radiology 2003;229:526-33.
- Cerfolio RJ, et al. Ann Thorac Surg 2004;78:1017-23.
- Shim SS, et al.Radiology 2005;236:1011-9.
- Halpern BS,et al. Chest 2005;128:2289-97.
- De Wever W, et al. Eur Radiol 2006;DOI:10.1007/s00330-006-0284-4.
- Yi CA, et al. PET/CT. J Nucl Med 2006;47:443-50.
- Blake MA, et al. Radiology 2006;238:970-7.
- Lardinois D, et al. J Clin Oncol 2005;23:6846-
- Herder GJ, et al. J Clin Oncol 2006;24:1800-6.
- Cerfolio RJ, Bryant AS, Ohja B, Bartolucci AA. J Thorac Cardiovasc Surg 2005;130:151-9.
- Prevost S, Boucher L, Larivee P, et al. J Nucl Med 2006;47:559-65.
- Pottgen C, Levegrun S, Theegarten D, et al. Clin Cancer Res 2006;12:97-106.
- Hicks RJ, Rischin D, Fisher R, et al. Eur J Nucl Med Mol Imaging 2005;32:1384-91.
- Brink I, Schumacher T, Mix M, et al. Eur J Nucl Med Mol Imaging 2004;31:1614-20.
- Steinert HC, Santos Dellea MM, Burger C, Stahel R. Lung Cancer 2005;49 Suppl 1:S33-5.