Tumor Biol. DOI 10.1007/s13277-014-2102-y
REVIEW
PET/CT assessment of neuroendocrine tumors of the lung with special emphasis on bronchial carcinoids Filippo Lococo & Alfredo Cesario & Massimiliano Paci & Angelina Filice & Annibale Versari & Cristian Rapicetta & Tommaso Ricchetti & Giorgio Sgarbi & Marco Alifano & Alberto Cavazza & Giorgio Treglia
Received: 4 April 2014 / Accepted: 13 May 2014 # International Society of Oncology and BioMarkers (ISOBM) 2014
Abstract Pulmonary neuroendocrine tumors (pNETs) arise from bronchial mucosal cells known as enterochromaffin cells which are part of the diffuse neuroendocrine system. The pathological spectrum of pNETs ranges from low-/intermediate-grade neoplasms such as bronchial carcinoids (BCs), also known as typical or atypical carcinoids, to high-grade neoplasms as large-cell neuroendocrine carcinoma and small-cell lung cancer. The tumor biology of pNETs still represents a matter of open debate. The distinct features among the different pNETs include not only their pathologic characteristics but also their clinical behavior, epidemiology, treatment, and prognosis. In this sense, a correct pathological identification in the preoperative setting is a key element for planning the best strategy of care in pNETs and especially in BCs. Controversial results have been reported on the diagnostic accuracy of fluorine-18-fluorodeoxyglucose positron emission tomography or positron emission tomography/computed F. Lococo (*) : M. Paci : C. Rapicetta : T. Ricchetti : G. Sgarbi Unit of Thoracic Surgery, IRCCS-Arcispedale Santa Maria Nuova, Reggio Emilia, Italy e-mail: [email protected] A. Cesario IRCCS-San Raffaele della Pisana, Rome, Italy A. Filice : A. Versari Department of Nuclear Medicine, Reggio Emilia, Italy M. Alifano Department of Thoracic Surgery, Descartes University Hospital Cochin, Paris, France A. Cavazza Unit of Pathology, IRCCS-Arcispedale Santa Maria Nuova, Reggio Emilia, Italy G. Treglia Department of Nuclear Medicine, Oncology Institute of Southern Switzerland, Bellinzona, Switzerland
tomography (F-18-FDG PET or PET/CT) in BCs. On the other hand, there is increasing evidence supporting the use of PET with somatostatin analogues (DOTA-TOC, DOTANOC, or DOTA-TATE) labeled with gallium-68 (Ga-68) in pNETs. Herein, we review the pertinent literature aiming to better define the current state of art of PET/CT in the detection and histological differentiation of pNETs with special emphasis on BCs. Keywords Neuroendocrine tumors . Pulmonary tumors . PET/CT . Functional imaging
Introduction Pulmonary neuroendocrine tumors (pNETs) represent a spectrum of neoplastic entities distinguished not only on the basis of pathologic characteristics but also regarding their clinical behavior, epidemiology, treatment, and prognosis. About the bronchial carcinoids (BCs), typical carcinoids (TCs) are indolent neoplasms with a good prognosis, whereas atypical carcinoids (ACs) have a less indolent behavior with a certain propensity for metastatic spread. Both these pNETs are optimally treated with complete surgical excision. Conversely, more aggressive pNETs, such as large-cell neuroendocrine carcinoma and small-cell lung cancer (LCNEC and SCLC, respectively), often present with local invasion, regional lymph nodal metastases, and distant spread; as a result, they show a poor prognosis and usually may not be candidates for surgical resection [1, 2]. Recent evidences [2, 3] suggest that even when surgery is indicated in well- or intermediate-differentiated pNETs, the extension of both pulmonary resection and lymph nodal dissection are determined by the cyto/histology characteristics of BCs. Unfortunately, TCs and ACs share structural
Tumor Biol.
radiological findings and a clear differentiation between these pNETs is not possible through radiological findings only [4, 5]. In this context, the functional imaging evaluation by using nuclear medicine techniques has improved in the last two decades with the aim of helping the physicians in the challenging management of pNETs [6]. In particular, positron emission tomography (PET) by using different tracers is particularly useful in the work-up of pNETs being able to detect functional abnormalities which often precede the onset of morphological lesions on conventional radiological imaging, such as computed tomography (CT) [7]. Tracer uptake at PET imaging may be evaluated visually or by using semiquantitative measures such as the maximal standardized uptake value (SUVmax). Finally, the development of hybrid PET/CT tomographs has greatly contributed to a more accurate delineation of areas of increased tracer uptake, overcoming the limits of patients repositioning when the two images were acquired independently and fused afterwards [7]. In the last years, several PET tracers have been proposed for the assessment of disease extent, restaging, and therapy response in pNETs [6]. However, in this review, we focus our attention and deeply discuss on the potential role of PET or PET/CT using fluorine-18-fluorodeoxyglucose (F-18-FDG) and somatostatin analogues labeled with gallium-68 (Ga-68DOTA-peptides) in pNETs with a particular attention on clinical and surgical implications in BCs.
Histopathological classification of pulmonary NETs It is well known that pNETs arise from Kulchitzky cells that are normally present in the bronchial mucosa and share the common morphologic features of neuroendocrine tumors including organoid nesting, palisading, rosettes, or a trabecular growth pattern. Travis and colleagues have proposed in 1991 [8] a histological classification of pNETs (revised in 2004 [9]) that includes TCs as low-grade tumors, ACs as intermediategrade malignancies, and LCNEC and SCLC as high-grade malignancies (Table 1). Moreover, WHO recognized as a different entity the so-called diffuse idiopathic pulmonary neuroendocrine cell hyperplasia (DIPNECH), a preinvasive precursor of carcinoid tumors [10].
The pathological distinction of pNETs is based on histological features, particularly mitotic rate and presence of necrosis. BCs, both TCs and ACs, consist of small nests or interconnecting trabeculae of uniform cells separated by a prominent vascular stroma and numerous thin-walled blood vessels. Mitoses and necrosis are the histopathologic features that distinguish TCs from ACs, in particular ACs have small foci of necrosis and 2–10 mitoses per 10 high-power fields (Table 1 and Fig. 1). A correct pathological identification of pNETs during the preoperative setting is a key element for planning the best strategy of care, considering the different biological behavior of the various histology subtypes. Nevertheless, a correct preoperative pathological differentiation between TCs and ACs (often based on cytology only) is generally hard to obtain. Similarly, the histological characterization between the WHO histotypes during frozen-section intraoperative assessment is something extremely difficult. In this context, the preoperative PET evaluation by using different tracers could represent a sort of “noninvasive biopsy” trying to correlate as accurately as possible the uptake pattern of pNETs with the histological subtypes.
Role of F-18-FDG PET (PET/CT) scan Since F-18-FDG is a glucose analogue, this tracer may be very useful in detecting malignant lesions which usually present high glucose metabolism and consequently increased F-18FDG uptake [7, 11]. Slow-growing pulmonary tumors usually exhibit a lower F-18-FDG uptake when compared to aggressive pulmonary malignancies [12]. As a result of that, LCNECs and SCLCs show higher F-18-FDG uptake compared to BCs [13, 14], while, to date, a significant increased F-18-FDG uptake in DIPNECH cases has not been reported [10]. Whereas F-18-FDG PET and PET/CT are widely used in evaluating aggressive tumors such as high-grade pNETs, in particular for staging, restaging, or treatment response assessment [7], many physicians consider F-18-FDG PET or PET/ CT as tools of limited value for the evaluation of BCs, due to the commonly slow growth of these tumors. Several articles in the literature evaluated the detection rate of F-18-FDG PET or PET/CT in BCs reporting conflicting
Table 1 Grading and diagnostic criteria of pulmonary NETs based on the 2004 WHO classification
Grade Mitosesa Necrosis Morphology a
Typical carcinoid
Atypical carcinoid
Large-cell neuroendocrine carcinoma
Small-cell lung cancer
Low 10 Often (diffuse) Poorly differentiated (large cells)
High >10 Frequent (diffuse) Poorly differentiated (small cells)
×10 high-power fields
Tumor Biol.
Fig. 1 a Typical carcinoid (Hematoxylin-Eosin, ×100), consisting in well-vascularized nests composed of bland epithelioid cells with large cytoplasm. No necrosis was present, and mitotic activity was negligible. b
Atypical carcinoid (Hematoxylin-Eosin, ×100), with an area of comedolike necrosis. This tumor had three mitoses per 10 high-power fields
results in this setting (detection rate ranged from 14 % to values >90 %) [15–25]. The main findings of these articles are described in Table 2. A recent prospective study on 32 patients with clinical suspicion of BC demonstrated that F-18-FDG PET/CT had a sensitivity, specificity, and accuracy in detecting BCs of 78, 11, and 59 %, respectively. Nevertheless, F-18-FDG PET/CT was positive in all cases of AC and false-negative in eight
cases of TC (sensitivity was 62 and 100 % for TCs and ACs, respectively) [22]. Interestingly, the method of analysis of F-18-FDG uptake in the pulmonary nodule is pivotal for an adequate evaluation of the accuracy of PET or PET/CT in BCs. In fact, as recently remarked by Stefani et al. [23], the detection rate of F-18-FDG PET or PET/CT for BCs was generally investigated on the basis of a visual assessment (usually considering as positive
Table 2 F-18-FDG PET or PET/CT findings in bronchial carcinoids: data from the literature (case reports excluded) First author (year)
No. of Histology pulmonary of BCs NETs (BCs)
Detection rate (%) of BCs [positivity criterion]
Detection rate (%) of TCs versus ACs Mean SUVmax Mean SUVmax [positivity criterion] in TCsa in ACsa
Erasmus 7 (7) (1998) Kruger (2006) 13 (13) Daniels 16 (16) (2007)
6 TCs, 1 AC
1/7 (14) [uptake>mediastinum] N.A.
N.A.
12 TCs, 1 AC 11 TCs, 5 ATs
7/13 (54) [SUVmax >2.5] 12/16 (75) [uptake> mediastinum]
8.5 N.A.
Chong (2007) 37 (7)
2 TCs, 5 ACs
3/7 (43) [uptake>mediastinum] 3.3±0.1
Kayani (2009) 18 (13)
11 TCs, 2 ACs 9/13 (69) [uptake> mediastinum] 13 TCs, 7 ACs 14/20 (70) [uptake> mediastinum] 6 TCs, 4 ACs 8/10b (80) [SUVmax >2.5]
2.9±0.8
0/2 (0) vs. 3/5 (60) [uptake> mediastinum] 13.1±2.0 7/11 (64) vs. 2/2 (100) [uptake> mediastinum] 5.3±2.0 7/13 (54) vs. 7/7 (100) [uptake> mediastinum] 7.9±5.4 4/6 (67) vs. 4/4 (100) [SUVmax >2.5]
24 TCs, 1 AC
3.6±2.8
5.0
5.28
5.08
2.7±1.6 2.88
8.1±4.1 4.37
Jindal (2011)
20 (20)
Kuyumcu 10 (10)b (2012) Stefani (2013) 25 (25) Alpay (2013)
27 (27)
Moore (2013) 29 (29) Venkitaraman 26 (26) (2014)
13/25 (52) [SUVmax >2.5] 24/25 (96) [SUVmax ≥1.5] 17 TCs, 8 ACs, 23/25 (92) [SUVmax >2.5] 2 OCs 23 TCs, 6 ACs N.A. 21 TCs, 5 ACs 18/26 (69) [uptake> mediastinum]
3.0±1.5 N.A.
5.5±4.0 2.1±0.9
1/6 (17) vs. 0/1 (0) [uptake> mediastinum] 6/12 (50) vs. 1/1 (100) [SUVmax>2.5] 8/11 (73) vs. 4/5 (80) [uptake> mediastinum]
4.3±2.0
12/24 (50) vs. 1/1 (100) [SUVmax >2.5] 23/24 (96) vs. 1/1 (100) [SUVmax ≥1.5] 16/17 (94) vs. 7/8 (88) [SUVmax >2.5] N.A. 13/21 (62) vs. 5/5 (100) [uptake> mediastinum]
BCs bronchial carcinoids, ACs atypical carcinoids, TCs typical carcinoids, OCs oncocytic carcinoids, N.A. not available a
Data extrapolated by the tables of the pertinent articles
b
Patients assessed before surgery
Tumor Biol.
an uptake superior to the mediastinal background) or a SUVmax cutoff (usually considering a cutoff of 2.5 as applied in the first studies on the use of F-18-FDG PET in lung carcinoma). However, BCs, especially TCs, usually show a lower metabolic activity compared to lung carcinoma and an increased detection rate of F-18-FDG PET in BCs could be obtained by using SUVmax cutoff values lower than 1.5 and considering the normal lung, and not the mediastinum, as background region for the visual assessment [23]. F-18-FDG uptake in pNETs was found to be related to some biologic features. In particular, Kaira et al. clearly demonstrated that F-18-FDG uptake is determined by the presence of glucose metabolism, hypoxia, and angiogenesis in pNETs [13]. A correlation between F-18-FDG uptake and the expression of glucose transporters (Glut-1) in pNETs has been also reported [14]. About the hystological type, the detection rate of F-18-FDG PET or PET/CT seems to be overall higher in ACs compared to TCs due to the more aggressive behavior and high proliferation rate of ACs (Table 2). In this regard, a recent retrospective study [24] reported that ACs usually show a higher F-18-FDG uptake compared to TCs and a SUVmax≥6 had a predictive value >95 % for distinguishing ACs from TCs. The significant higher SUVmax in ACs compared to TCs at F-18FDG PET was also evident in other studies [16, 19–21]. Conversely, a recent article did not demonstrate a significant difference in SUVmax between TCs and ACs [25]. Whereas F-18-FDG uptake is usually related to the tumor size for lung carcinomas [26], controversial results about the possible correlation between F-18-FDG uptake, visually or semi-quantitatively assessed, and tumor size in BCs were reported [15, 16, 18, 21, 23, 24]. Furthermore, a correlation between functional imaging staging by F-18-FDG PET or PET/CT and surgical staging could not be definitively assessed due to the limited number of metastatic BCs included in the pertinent articles [15–25]. Finally, SUVmax has been reported to be related to the patients’ survival in poorly differentiated pNETs (LCNECs and SCLCs) [17, 26–28]. On the other hand, the prognostic value of F-18-FDG PET or PET/CT in BCs is far to be defined and studies on a large cohort of patients are warranted. In conclusion, F-18-FDG PET or PET/CT seems to be suboptimal in detecting BCs, whereas higher detection rate and SUVmax in ACs vs. TCs are often reported. Nevertheless, when a BC is strongly suspected by clinical or radiological findings, even a low F-18-FDG uptake should not be considered as a negative result and, in such cases, a surgical resection or a biopsy could be performed.
Role of Ga-68-DOTA-peptides PET (PET/CT) scan Neuroendocrine tumors including pNETs usually overexpress somatostatin receptors (SSTRs) on their cell surface. SSTRs density is related to the degree of the tumor differentiation. Therefore, well-differentiated pNETs usually express a higher density of SSTRs compared to poorly differentiated pNETs [29]. In the last years, the recent development of novel tracers, in particular somatostatin analogues labeled with gallium-68 (Ga-68-DOTA-peptides), has allowed an accurate imaging of SSTRs by PET. In detail, Ga-68-DOTA-peptides bind to SSTRs overexpressed on neuroendocrine tumor cells. The structure of these radiopharmaceuticals includes the somatostatin analogue binding to SSTR (TOC, NOC, TATE), a chelant (DOTA) and a positron-emitting isotope (Ga-68) [6, 30]. The most relevant difference among these compounds relies in a variable affinity to SSTR subtypes: all can bind to subtypes 2, but only DOTA-NOC presents a good affinity for subtypes 2, 3, and 5. Despite the observed differences in receptor binding affinity between the different somatostatin analogues used for PET, they seems to provide similar diagnostic accuracy and clinical impact in pNETs [6]. The uptake of these tracers is not dependent on the cellular metabolism (as compared ,for example, to F-18-DOPA or F-18-FDG) and noninvasively provides information on SSTR expression with direct therapeutical implications. For example, metastatic pNETs with high SSTR expression could be suitable for peptide radioreceptor therapy [6, 30]. Finally, beyond the easier labeling and synthesis process [31], Ga-68-DOTA-peptides PET offer several advantages over somatostatin receptor scintigraphy with Indium-111pentetreotide for diagnosis and therapy planning of pNETs, including the higher affinity for SSTRs type 2, the superior resolution derived by the use of PET technique, and the more rapid examination time [6, 32, 33]. Several articles in the literature have demonstrated the usefulness of Ga-68-DOTA-peptides PET or PET/CT in patients with neuroendocrine tumors [34], but a limited number of articles was focused on pNETs only [35] (Table 3). In particular, the rarity of BCs strongly limits the acquisition of large and robust evidences on the use of new diagnostic tools in this setting. Moreover, the employment of Ga-68-DOTApeptides, despite increasing, is still limited to specialized centers [6]. Ambrosini et al. [36] evaluated 11 patients with BCs by using Ga-68-DOTA-NOC PET/CT. This method detected a higher number of lesions as compared to CT scan only (37 versus 21, respectively) providing additional information in 82 % (9 of 11) of patients, contributing to a better evaluation of the extent of the disease and changing the management in about 30 % of cases.
Tumor Biol. Table 3 Ga-68-DOTA-peptides PET or PET/CT findings in selected articles from the literature focused on pulmonary carcinoids (case reports excluded) First author (year)
Tracer
Ambrosini (1998) Kayani (2009) [19]
DOTA-NOC 11 DOTA-TATE 18
Jindal (2010, 2011) DOTA-TOC Venkitaraman (2014) DOTA-TOC
No. of Histology patients
20 32
11 BCs 11 TCs, 2 ACs, 5 other pNETs 13 TCs, 7 ACs 21 TCs, 5 ACs, 6 other tumors
Detection Mean SUVmax Mean SUVmax Detection rate (%) of in ACsa TCs versus ACs rate (%) of BCs in TCsa 9/9 (100) 13/13 (100)
N.A. 40±30.7
N.A. 4.9±0.28
N.A. 11/11 (100) vs. 2/2 (100)
19/20 (95) 25/26 (96)
36.7±18.1 21.5
9.6±5.9 15.4
13/13 (100) vs. 6/7 (86) 21/21 (100) vs. 4/5 (80)
pNETs pulmonary neuroendocrine tumors, BCs bronchial carcinoids, ACa atypical carcinoids, TCa typical carcinoids, N.A. not available a
Data extrapolated by tables of the pertinent articles
Kayani et al. [19] performed Ga-68-DOTA-TATE PET/CT in 18 patients with pNETs. All TCs were positive at Ga-68DOTA-TATE PET/CT. ACs and high-grade pNETs had less Ga-68-DOTA-TATE avidity compared to TCs. Two cases of DIPNECH did not show significant uptake of Ga-68-DOTATATE. Jindal et al. [20, 37] evaluated 20 patients with BCs (13 TCs and 7 ACs) by using Ga-68-DOTA-TOC PET/CT. The Fig. 2 Increased uptake of radiolabeled somatostatin analogues (a, b; arrows) and low/ absent uptake of F-18-FDG (c, d) in a case of typical carcinoid of the right lung (images from IRCCS, Arcispedale Santa Maria Nuova, Reggio Emilia, Italy)
overall detection rate of this method for BCs was 95 % (19 of 20 cases were positive). All TCs showed significant uptake of Ga-68-DOTA-TOC. One AC lacked any significant tracer uptake. TCs also showed significantly higher uptake of Ga68-DOTA-TOC compared to ACs. In one patient, Ga-68DOTA-TOC PET/CT facilitated the detection of additional lesions compared to conventional imaging. These findings were confirmed by a further prospective study of the same
Tumor Biol.
group [22] which reported an overall sensitivity, specificity, and accuracy of Ga-68-DOTA-TOC PET/CT in the diagnosis of BCs of 96, 100, and 97 %, respectively, confirming a significant higher SUVmax in TCs compared to ACs. About high-grade pNETs, a recent study of Sollini et al. [38] on 24 patients with extensive disease SCLC reported that Ga-68-DOTA-peptides PET/CT was positive in 83 % of patients, demonstrating enhanced SSTR expression in 50 % of cases. Based on these preliminary results, the authors suggest a potential role of somatostatin receptor PET/CT in selecting SCLC patients who are suitable for peptide radioreceptor therapy. Overall, the detection rate of Ga-68-DOTA-peptides PET or PET/CT in BCs is very high and these techniques may provide additional information compared to conventional imaging techniques. Since TCs have been reported to express the SSTRs more abundantly if compared with higher grade pNETs [29], it seems logical that TCs show higher uptake of Ga-68-DOTA-peptides [19]. Some false-negative findings of Ga-68-DOTA-peptides PET or PET/CT in pNETs may be due to tumors with low expression of SSTRs (mainly cases of intermediate and high-grade pNETs). Although inflammatory diseases can yield false-positive results at somatostatin receptor imaging since SSTRs are overexpressed by the inflammatory cells [39–41], false-positive findings of Ga-68-DOTApeptides PET in pNETs are rarely described [42]. Whereas the role of somatostatin receptor PET or PET/CT seems to be well Fig. 3 Moderate uptake of radiolabeled somatostatin analogues (a, b; arrows) and increased F-18-FDG uptake (c, d; arrows) in a case of atypical carcinoid of the left lung (images from IRCCS, Arcispedale Santa Maria Nuova, Reggio Emilia, Italy)
defined in low-grade pNETs (TCs), further studies should be carried out to assess the value of this functional imaging method in DIPNECH, and intermediate- and high-grade pNETs.
Dual tracer PET evaluation using Ga-68-DOTA-peptides and F-18-FDG The evaluation of solitary pulmonary nodule (SPN) is a serious diagnostic challenge for which clinicians may use various imaging modalities and algorithms to detect and predict the likelihood of malignancy. SPNs (especially the round-shape nodules) are difficult to accurately diagnose based on limited sensitivity of noninvasive imaging, technical limitations of biopsy (especially in small-size lesions), and the high frequency of benign lesions detected during radiological evaluation. In this setting, functional imaging techniques could potentially help the physicians in such challenging process, carrying out useful information for the diagnosis and histological differentiation of the SPNs. F-18-FDG PET or PET/CT may be very useful in distinguishing benign lesions from malignant ones but when a BC is suspected, the low/absent metabolic activity of the lesion does not completely exclude such diagnosis. On the other hand, in case of increased F-18-FDG uptake, the suspicious of BC enters in the differential diagnosis process with other entities (including other lung tumors and
Tumor Biol.
inflammatory diseases). Differently, Ga-68-DOTA-peptides are more specific tracers in the detection of pNETs. In particular, a positivity at Ga-68-DOTA-peptides PET is highly predictive for a definitive diagnosis of pNET (excluding rare false-positive findings). Nevertheless, the absence of Ga68-DOTA-peptides uptake does not exclude the diagnosis of pNET, especially in intermediate- and high-grade subtypes. At the light of these considerations, the scientific community has started to investigate the opportunity of using a combined Ga-68-DOTA-peptides and F-18-FDG PET/CT evaluation for suspected pNETs [43, 44]. Actually, only few reports [18, 19, 21, 45–47] on small series in the literature assessed the use of Ga-68DOTA-peptides in comparison to F-18-FDG for the evaluation of histological subtypes of pNETs. Kumar et al. [46] evaluated the combination of F-18-FDG PET/CT scan and Ga-68-DOTA-peptides PET/CT scan in differentiating seven patients with bronchial masses including three BCs (two TCs and one AC). The TCs had mild F-18FDG uptake and high Ga-68-DOTA-TOC uptake. The AC had moderate uptake of F-18-FDG and high 68Ga DOTATOC uptake. Kayani et al. [19] compared Ga-68-DOTA-peptides and F-18-FDG PET/CT findings in 18 patients with pNETs, including 13 BCs (11 TCs and 2 ACs). TCs showed significantly higher uptake of Ga-68-DOTA-peptides and significantly lower uptake of F-18-FDG than did pNETs of higher grade. Moreover, no false-positive uptake of Ga-68-DOTA-TATE was observed in this study population, but there were three sites of false-positive uptake of F-18-FDG due to inflammation. Jindal et al. [20] studied 20 patients with BCs (13 TCs and 7 ACs) by using F-18-FDG and Ga-68-DOTATOC. TCs showed a higher uptake of Ga-68-DOTATOC than ACs, while the latter showed increased F-18-FDG uptake. The ratio of SUVmax of Ga-68DOTA-TOC uptake to that of F-18-FDG was significantly higher in TCs than in ACs (p
REVIEW
PET/CT assessment of neuroendocrine tumors of the lung with special emphasis on bronchial carcinoids Filippo Lococo & Alfredo Cesario & Massimiliano Paci & Angelina Filice & Annibale Versari & Cristian Rapicetta & Tommaso Ricchetti & Giorgio Sgarbi & Marco Alifano & Alberto Cavazza & Giorgio Treglia
Received: 4 April 2014 / Accepted: 13 May 2014 # International Society of Oncology and BioMarkers (ISOBM) 2014
Abstract Pulmonary neuroendocrine tumors (pNETs) arise from bronchial mucosal cells known as enterochromaffin cells which are part of the diffuse neuroendocrine system. The pathological spectrum of pNETs ranges from low-/intermediate-grade neoplasms such as bronchial carcinoids (BCs), also known as typical or atypical carcinoids, to high-grade neoplasms as large-cell neuroendocrine carcinoma and small-cell lung cancer. The tumor biology of pNETs still represents a matter of open debate. The distinct features among the different pNETs include not only their pathologic characteristics but also their clinical behavior, epidemiology, treatment, and prognosis. In this sense, a correct pathological identification in the preoperative setting is a key element for planning the best strategy of care in pNETs and especially in BCs. Controversial results have been reported on the diagnostic accuracy of fluorine-18-fluorodeoxyglucose positron emission tomography or positron emission tomography/computed F. Lococo (*) : M. Paci : C. Rapicetta : T. Ricchetti : G. Sgarbi Unit of Thoracic Surgery, IRCCS-Arcispedale Santa Maria Nuova, Reggio Emilia, Italy e-mail: [email protected] A. Cesario IRCCS-San Raffaele della Pisana, Rome, Italy A. Filice : A. Versari Department of Nuclear Medicine, Reggio Emilia, Italy M. Alifano Department of Thoracic Surgery, Descartes University Hospital Cochin, Paris, France A. Cavazza Unit of Pathology, IRCCS-Arcispedale Santa Maria Nuova, Reggio Emilia, Italy G. Treglia Department of Nuclear Medicine, Oncology Institute of Southern Switzerland, Bellinzona, Switzerland
tomography (F-18-FDG PET or PET/CT) in BCs. On the other hand, there is increasing evidence supporting the use of PET with somatostatin analogues (DOTA-TOC, DOTANOC, or DOTA-TATE) labeled with gallium-68 (Ga-68) in pNETs. Herein, we review the pertinent literature aiming to better define the current state of art of PET/CT in the detection and histological differentiation of pNETs with special emphasis on BCs. Keywords Neuroendocrine tumors . Pulmonary tumors . PET/CT . Functional imaging
Introduction Pulmonary neuroendocrine tumors (pNETs) represent a spectrum of neoplastic entities distinguished not only on the basis of pathologic characteristics but also regarding their clinical behavior, epidemiology, treatment, and prognosis. About the bronchial carcinoids (BCs), typical carcinoids (TCs) are indolent neoplasms with a good prognosis, whereas atypical carcinoids (ACs) have a less indolent behavior with a certain propensity for metastatic spread. Both these pNETs are optimally treated with complete surgical excision. Conversely, more aggressive pNETs, such as large-cell neuroendocrine carcinoma and small-cell lung cancer (LCNEC and SCLC, respectively), often present with local invasion, regional lymph nodal metastases, and distant spread; as a result, they show a poor prognosis and usually may not be candidates for surgical resection [1, 2]. Recent evidences [2, 3] suggest that even when surgery is indicated in well- or intermediate-differentiated pNETs, the extension of both pulmonary resection and lymph nodal dissection are determined by the cyto/histology characteristics of BCs. Unfortunately, TCs and ACs share structural
Tumor Biol.
radiological findings and a clear differentiation between these pNETs is not possible through radiological findings only [4, 5]. In this context, the functional imaging evaluation by using nuclear medicine techniques has improved in the last two decades with the aim of helping the physicians in the challenging management of pNETs [6]. In particular, positron emission tomography (PET) by using different tracers is particularly useful in the work-up of pNETs being able to detect functional abnormalities which often precede the onset of morphological lesions on conventional radiological imaging, such as computed tomography (CT) [7]. Tracer uptake at PET imaging may be evaluated visually or by using semiquantitative measures such as the maximal standardized uptake value (SUVmax). Finally, the development of hybrid PET/CT tomographs has greatly contributed to a more accurate delineation of areas of increased tracer uptake, overcoming the limits of patients repositioning when the two images were acquired independently and fused afterwards [7]. In the last years, several PET tracers have been proposed for the assessment of disease extent, restaging, and therapy response in pNETs [6]. However, in this review, we focus our attention and deeply discuss on the potential role of PET or PET/CT using fluorine-18-fluorodeoxyglucose (F-18-FDG) and somatostatin analogues labeled with gallium-68 (Ga-68DOTA-peptides) in pNETs with a particular attention on clinical and surgical implications in BCs.
Histopathological classification of pulmonary NETs It is well known that pNETs arise from Kulchitzky cells that are normally present in the bronchial mucosa and share the common morphologic features of neuroendocrine tumors including organoid nesting, palisading, rosettes, or a trabecular growth pattern. Travis and colleagues have proposed in 1991 [8] a histological classification of pNETs (revised in 2004 [9]) that includes TCs as low-grade tumors, ACs as intermediategrade malignancies, and LCNEC and SCLC as high-grade malignancies (Table 1). Moreover, WHO recognized as a different entity the so-called diffuse idiopathic pulmonary neuroendocrine cell hyperplasia (DIPNECH), a preinvasive precursor of carcinoid tumors [10].
The pathological distinction of pNETs is based on histological features, particularly mitotic rate and presence of necrosis. BCs, both TCs and ACs, consist of small nests or interconnecting trabeculae of uniform cells separated by a prominent vascular stroma and numerous thin-walled blood vessels. Mitoses and necrosis are the histopathologic features that distinguish TCs from ACs, in particular ACs have small foci of necrosis and 2–10 mitoses per 10 high-power fields (Table 1 and Fig. 1). A correct pathological identification of pNETs during the preoperative setting is a key element for planning the best strategy of care, considering the different biological behavior of the various histology subtypes. Nevertheless, a correct preoperative pathological differentiation between TCs and ACs (often based on cytology only) is generally hard to obtain. Similarly, the histological characterization between the WHO histotypes during frozen-section intraoperative assessment is something extremely difficult. In this context, the preoperative PET evaluation by using different tracers could represent a sort of “noninvasive biopsy” trying to correlate as accurately as possible the uptake pattern of pNETs with the histological subtypes.
Role of F-18-FDG PET (PET/CT) scan Since F-18-FDG is a glucose analogue, this tracer may be very useful in detecting malignant lesions which usually present high glucose metabolism and consequently increased F-18FDG uptake [7, 11]. Slow-growing pulmonary tumors usually exhibit a lower F-18-FDG uptake when compared to aggressive pulmonary malignancies [12]. As a result of that, LCNECs and SCLCs show higher F-18-FDG uptake compared to BCs [13, 14], while, to date, a significant increased F-18-FDG uptake in DIPNECH cases has not been reported [10]. Whereas F-18-FDG PET and PET/CT are widely used in evaluating aggressive tumors such as high-grade pNETs, in particular for staging, restaging, or treatment response assessment [7], many physicians consider F-18-FDG PET or PET/ CT as tools of limited value for the evaluation of BCs, due to the commonly slow growth of these tumors. Several articles in the literature evaluated the detection rate of F-18-FDG PET or PET/CT in BCs reporting conflicting
Table 1 Grading and diagnostic criteria of pulmonary NETs based on the 2004 WHO classification
Grade Mitosesa Necrosis Morphology a
Typical carcinoid
Atypical carcinoid
Large-cell neuroendocrine carcinoma
Small-cell lung cancer
Low 10 Often (diffuse) Poorly differentiated (large cells)
High >10 Frequent (diffuse) Poorly differentiated (small cells)
×10 high-power fields
Tumor Biol.
Fig. 1 a Typical carcinoid (Hematoxylin-Eosin, ×100), consisting in well-vascularized nests composed of bland epithelioid cells with large cytoplasm. No necrosis was present, and mitotic activity was negligible. b
Atypical carcinoid (Hematoxylin-Eosin, ×100), with an area of comedolike necrosis. This tumor had three mitoses per 10 high-power fields
results in this setting (detection rate ranged from 14 % to values >90 %) [15–25]. The main findings of these articles are described in Table 2. A recent prospective study on 32 patients with clinical suspicion of BC demonstrated that F-18-FDG PET/CT had a sensitivity, specificity, and accuracy in detecting BCs of 78, 11, and 59 %, respectively. Nevertheless, F-18-FDG PET/CT was positive in all cases of AC and false-negative in eight
cases of TC (sensitivity was 62 and 100 % for TCs and ACs, respectively) [22]. Interestingly, the method of analysis of F-18-FDG uptake in the pulmonary nodule is pivotal for an adequate evaluation of the accuracy of PET or PET/CT in BCs. In fact, as recently remarked by Stefani et al. [23], the detection rate of F-18-FDG PET or PET/CT for BCs was generally investigated on the basis of a visual assessment (usually considering as positive
Table 2 F-18-FDG PET or PET/CT findings in bronchial carcinoids: data from the literature (case reports excluded) First author (year)
No. of Histology pulmonary of BCs NETs (BCs)
Detection rate (%) of BCs [positivity criterion]
Detection rate (%) of TCs versus ACs Mean SUVmax Mean SUVmax [positivity criterion] in TCsa in ACsa
Erasmus 7 (7) (1998) Kruger (2006) 13 (13) Daniels 16 (16) (2007)
6 TCs, 1 AC
1/7 (14) [uptake>mediastinum] N.A.
N.A.
12 TCs, 1 AC 11 TCs, 5 ATs
7/13 (54) [SUVmax >2.5] 12/16 (75) [uptake> mediastinum]
8.5 N.A.
Chong (2007) 37 (7)
2 TCs, 5 ACs
3/7 (43) [uptake>mediastinum] 3.3±0.1
Kayani (2009) 18 (13)
11 TCs, 2 ACs 9/13 (69) [uptake> mediastinum] 13 TCs, 7 ACs 14/20 (70) [uptake> mediastinum] 6 TCs, 4 ACs 8/10b (80) [SUVmax >2.5]
2.9±0.8
0/2 (0) vs. 3/5 (60) [uptake> mediastinum] 13.1±2.0 7/11 (64) vs. 2/2 (100) [uptake> mediastinum] 5.3±2.0 7/13 (54) vs. 7/7 (100) [uptake> mediastinum] 7.9±5.4 4/6 (67) vs. 4/4 (100) [SUVmax >2.5]
24 TCs, 1 AC
3.6±2.8
5.0
5.28
5.08
2.7±1.6 2.88
8.1±4.1 4.37
Jindal (2011)
20 (20)
Kuyumcu 10 (10)b (2012) Stefani (2013) 25 (25) Alpay (2013)
27 (27)
Moore (2013) 29 (29) Venkitaraman 26 (26) (2014)
13/25 (52) [SUVmax >2.5] 24/25 (96) [SUVmax ≥1.5] 17 TCs, 8 ACs, 23/25 (92) [SUVmax >2.5] 2 OCs 23 TCs, 6 ACs N.A. 21 TCs, 5 ACs 18/26 (69) [uptake> mediastinum]
3.0±1.5 N.A.
5.5±4.0 2.1±0.9
1/6 (17) vs. 0/1 (0) [uptake> mediastinum] 6/12 (50) vs. 1/1 (100) [SUVmax>2.5] 8/11 (73) vs. 4/5 (80) [uptake> mediastinum]
4.3±2.0
12/24 (50) vs. 1/1 (100) [SUVmax >2.5] 23/24 (96) vs. 1/1 (100) [SUVmax ≥1.5] 16/17 (94) vs. 7/8 (88) [SUVmax >2.5] N.A. 13/21 (62) vs. 5/5 (100) [uptake> mediastinum]
BCs bronchial carcinoids, ACs atypical carcinoids, TCs typical carcinoids, OCs oncocytic carcinoids, N.A. not available a
Data extrapolated by the tables of the pertinent articles
b
Patients assessed before surgery
Tumor Biol.
an uptake superior to the mediastinal background) or a SUVmax cutoff (usually considering a cutoff of 2.5 as applied in the first studies on the use of F-18-FDG PET in lung carcinoma). However, BCs, especially TCs, usually show a lower metabolic activity compared to lung carcinoma and an increased detection rate of F-18-FDG PET in BCs could be obtained by using SUVmax cutoff values lower than 1.5 and considering the normal lung, and not the mediastinum, as background region for the visual assessment [23]. F-18-FDG uptake in pNETs was found to be related to some biologic features. In particular, Kaira et al. clearly demonstrated that F-18-FDG uptake is determined by the presence of glucose metabolism, hypoxia, and angiogenesis in pNETs [13]. A correlation between F-18-FDG uptake and the expression of glucose transporters (Glut-1) in pNETs has been also reported [14]. About the hystological type, the detection rate of F-18-FDG PET or PET/CT seems to be overall higher in ACs compared to TCs due to the more aggressive behavior and high proliferation rate of ACs (Table 2). In this regard, a recent retrospective study [24] reported that ACs usually show a higher F-18-FDG uptake compared to TCs and a SUVmax≥6 had a predictive value >95 % for distinguishing ACs from TCs. The significant higher SUVmax in ACs compared to TCs at F-18FDG PET was also evident in other studies [16, 19–21]. Conversely, a recent article did not demonstrate a significant difference in SUVmax between TCs and ACs [25]. Whereas F-18-FDG uptake is usually related to the tumor size for lung carcinomas [26], controversial results about the possible correlation between F-18-FDG uptake, visually or semi-quantitatively assessed, and tumor size in BCs were reported [15, 16, 18, 21, 23, 24]. Furthermore, a correlation between functional imaging staging by F-18-FDG PET or PET/CT and surgical staging could not be definitively assessed due to the limited number of metastatic BCs included in the pertinent articles [15–25]. Finally, SUVmax has been reported to be related to the patients’ survival in poorly differentiated pNETs (LCNECs and SCLCs) [17, 26–28]. On the other hand, the prognostic value of F-18-FDG PET or PET/CT in BCs is far to be defined and studies on a large cohort of patients are warranted. In conclusion, F-18-FDG PET or PET/CT seems to be suboptimal in detecting BCs, whereas higher detection rate and SUVmax in ACs vs. TCs are often reported. Nevertheless, when a BC is strongly suspected by clinical or radiological findings, even a low F-18-FDG uptake should not be considered as a negative result and, in such cases, a surgical resection or a biopsy could be performed.
Role of Ga-68-DOTA-peptides PET (PET/CT) scan Neuroendocrine tumors including pNETs usually overexpress somatostatin receptors (SSTRs) on their cell surface. SSTRs density is related to the degree of the tumor differentiation. Therefore, well-differentiated pNETs usually express a higher density of SSTRs compared to poorly differentiated pNETs [29]. In the last years, the recent development of novel tracers, in particular somatostatin analogues labeled with gallium-68 (Ga-68-DOTA-peptides), has allowed an accurate imaging of SSTRs by PET. In detail, Ga-68-DOTA-peptides bind to SSTRs overexpressed on neuroendocrine tumor cells. The structure of these radiopharmaceuticals includes the somatostatin analogue binding to SSTR (TOC, NOC, TATE), a chelant (DOTA) and a positron-emitting isotope (Ga-68) [6, 30]. The most relevant difference among these compounds relies in a variable affinity to SSTR subtypes: all can bind to subtypes 2, but only DOTA-NOC presents a good affinity for subtypes 2, 3, and 5. Despite the observed differences in receptor binding affinity between the different somatostatin analogues used for PET, they seems to provide similar diagnostic accuracy and clinical impact in pNETs [6]. The uptake of these tracers is not dependent on the cellular metabolism (as compared ,for example, to F-18-DOPA or F-18-FDG) and noninvasively provides information on SSTR expression with direct therapeutical implications. For example, metastatic pNETs with high SSTR expression could be suitable for peptide radioreceptor therapy [6, 30]. Finally, beyond the easier labeling and synthesis process [31], Ga-68-DOTA-peptides PET offer several advantages over somatostatin receptor scintigraphy with Indium-111pentetreotide for diagnosis and therapy planning of pNETs, including the higher affinity for SSTRs type 2, the superior resolution derived by the use of PET technique, and the more rapid examination time [6, 32, 33]. Several articles in the literature have demonstrated the usefulness of Ga-68-DOTA-peptides PET or PET/CT in patients with neuroendocrine tumors [34], but a limited number of articles was focused on pNETs only [35] (Table 3). In particular, the rarity of BCs strongly limits the acquisition of large and robust evidences on the use of new diagnostic tools in this setting. Moreover, the employment of Ga-68-DOTApeptides, despite increasing, is still limited to specialized centers [6]. Ambrosini et al. [36] evaluated 11 patients with BCs by using Ga-68-DOTA-NOC PET/CT. This method detected a higher number of lesions as compared to CT scan only (37 versus 21, respectively) providing additional information in 82 % (9 of 11) of patients, contributing to a better evaluation of the extent of the disease and changing the management in about 30 % of cases.
Tumor Biol. Table 3 Ga-68-DOTA-peptides PET or PET/CT findings in selected articles from the literature focused on pulmonary carcinoids (case reports excluded) First author (year)
Tracer
Ambrosini (1998) Kayani (2009) [19]
DOTA-NOC 11 DOTA-TATE 18
Jindal (2010, 2011) DOTA-TOC Venkitaraman (2014) DOTA-TOC
No. of Histology patients
20 32
11 BCs 11 TCs, 2 ACs, 5 other pNETs 13 TCs, 7 ACs 21 TCs, 5 ACs, 6 other tumors
Detection Mean SUVmax Mean SUVmax Detection rate (%) of in ACsa TCs versus ACs rate (%) of BCs in TCsa 9/9 (100) 13/13 (100)
N.A. 40±30.7
N.A. 4.9±0.28
N.A. 11/11 (100) vs. 2/2 (100)
19/20 (95) 25/26 (96)
36.7±18.1 21.5
9.6±5.9 15.4
13/13 (100) vs. 6/7 (86) 21/21 (100) vs. 4/5 (80)
pNETs pulmonary neuroendocrine tumors, BCs bronchial carcinoids, ACa atypical carcinoids, TCa typical carcinoids, N.A. not available a
Data extrapolated by tables of the pertinent articles
Kayani et al. [19] performed Ga-68-DOTA-TATE PET/CT in 18 patients with pNETs. All TCs were positive at Ga-68DOTA-TATE PET/CT. ACs and high-grade pNETs had less Ga-68-DOTA-TATE avidity compared to TCs. Two cases of DIPNECH did not show significant uptake of Ga-68-DOTATATE. Jindal et al. [20, 37] evaluated 20 patients with BCs (13 TCs and 7 ACs) by using Ga-68-DOTA-TOC PET/CT. The Fig. 2 Increased uptake of radiolabeled somatostatin analogues (a, b; arrows) and low/ absent uptake of F-18-FDG (c, d) in a case of typical carcinoid of the right lung (images from IRCCS, Arcispedale Santa Maria Nuova, Reggio Emilia, Italy)
overall detection rate of this method for BCs was 95 % (19 of 20 cases were positive). All TCs showed significant uptake of Ga-68-DOTA-TOC. One AC lacked any significant tracer uptake. TCs also showed significantly higher uptake of Ga68-DOTA-TOC compared to ACs. In one patient, Ga-68DOTA-TOC PET/CT facilitated the detection of additional lesions compared to conventional imaging. These findings were confirmed by a further prospective study of the same
Tumor Biol.
group [22] which reported an overall sensitivity, specificity, and accuracy of Ga-68-DOTA-TOC PET/CT in the diagnosis of BCs of 96, 100, and 97 %, respectively, confirming a significant higher SUVmax in TCs compared to ACs. About high-grade pNETs, a recent study of Sollini et al. [38] on 24 patients with extensive disease SCLC reported that Ga-68-DOTA-peptides PET/CT was positive in 83 % of patients, demonstrating enhanced SSTR expression in 50 % of cases. Based on these preliminary results, the authors suggest a potential role of somatostatin receptor PET/CT in selecting SCLC patients who are suitable for peptide radioreceptor therapy. Overall, the detection rate of Ga-68-DOTA-peptides PET or PET/CT in BCs is very high and these techniques may provide additional information compared to conventional imaging techniques. Since TCs have been reported to express the SSTRs more abundantly if compared with higher grade pNETs [29], it seems logical that TCs show higher uptake of Ga-68-DOTA-peptides [19]. Some false-negative findings of Ga-68-DOTA-peptides PET or PET/CT in pNETs may be due to tumors with low expression of SSTRs (mainly cases of intermediate and high-grade pNETs). Although inflammatory diseases can yield false-positive results at somatostatin receptor imaging since SSTRs are overexpressed by the inflammatory cells [39–41], false-positive findings of Ga-68-DOTApeptides PET in pNETs are rarely described [42]. Whereas the role of somatostatin receptor PET or PET/CT seems to be well Fig. 3 Moderate uptake of radiolabeled somatostatin analogues (a, b; arrows) and increased F-18-FDG uptake (c, d; arrows) in a case of atypical carcinoid of the left lung (images from IRCCS, Arcispedale Santa Maria Nuova, Reggio Emilia, Italy)
defined in low-grade pNETs (TCs), further studies should be carried out to assess the value of this functional imaging method in DIPNECH, and intermediate- and high-grade pNETs.
Dual tracer PET evaluation using Ga-68-DOTA-peptides and F-18-FDG The evaluation of solitary pulmonary nodule (SPN) is a serious diagnostic challenge for which clinicians may use various imaging modalities and algorithms to detect and predict the likelihood of malignancy. SPNs (especially the round-shape nodules) are difficult to accurately diagnose based on limited sensitivity of noninvasive imaging, technical limitations of biopsy (especially in small-size lesions), and the high frequency of benign lesions detected during radiological evaluation. In this setting, functional imaging techniques could potentially help the physicians in such challenging process, carrying out useful information for the diagnosis and histological differentiation of the SPNs. F-18-FDG PET or PET/CT may be very useful in distinguishing benign lesions from malignant ones but when a BC is suspected, the low/absent metabolic activity of the lesion does not completely exclude such diagnosis. On the other hand, in case of increased F-18-FDG uptake, the suspicious of BC enters in the differential diagnosis process with other entities (including other lung tumors and
Tumor Biol.
inflammatory diseases). Differently, Ga-68-DOTA-peptides are more specific tracers in the detection of pNETs. In particular, a positivity at Ga-68-DOTA-peptides PET is highly predictive for a definitive diagnosis of pNET (excluding rare false-positive findings). Nevertheless, the absence of Ga68-DOTA-peptides uptake does not exclude the diagnosis of pNET, especially in intermediate- and high-grade subtypes. At the light of these considerations, the scientific community has started to investigate the opportunity of using a combined Ga-68-DOTA-peptides and F-18-FDG PET/CT evaluation for suspected pNETs [43, 44]. Actually, only few reports [18, 19, 21, 45–47] on small series in the literature assessed the use of Ga-68DOTA-peptides in comparison to F-18-FDG for the evaluation of histological subtypes of pNETs. Kumar et al. [46] evaluated the combination of F-18-FDG PET/CT scan and Ga-68-DOTA-peptides PET/CT scan in differentiating seven patients with bronchial masses including three BCs (two TCs and one AC). The TCs had mild F-18FDG uptake and high Ga-68-DOTA-TOC uptake. The AC had moderate uptake of F-18-FDG and high 68Ga DOTATOC uptake. Kayani et al. [19] compared Ga-68-DOTA-peptides and F-18-FDG PET/CT findings in 18 patients with pNETs, including 13 BCs (11 TCs and 2 ACs). TCs showed significantly higher uptake of Ga-68-DOTA-peptides and significantly lower uptake of F-18-FDG than did pNETs of higher grade. Moreover, no false-positive uptake of Ga-68-DOTA-TATE was observed in this study population, but there were three sites of false-positive uptake of F-18-FDG due to inflammation. Jindal et al. [20] studied 20 patients with BCs (13 TCs and 7 ACs) by using F-18-FDG and Ga-68-DOTATOC. TCs showed a higher uptake of Ga-68-DOTATOC than ACs, while the latter showed increased F-18-FDG uptake. The ratio of SUVmax of Ga-68DOTA-TOC uptake to that of F-18-FDG was significantly higher in TCs than in ACs (p
