BEST EVIDENCE TOPIC – THORACIC
Interactive CardioVascular and Thoracic Surgery 19 (2014) 129–134 doi:10.1093/icvts/ivu068 Advance Access publication 19 March 2014
Is a positron emission tomography–computed tomography scan useful in the staging of thymic epithelial neoplasms? Andrea Vitia,*, Alberto Terzib, Andrea Bianchic and Luca Bertolaccinia a b c
Thoracic Surgery Unit, S. Croce e Carle Hospital, Cuneo, Italy Thoracic Surgery Unit, Sacred Heart Hospital, Negrar, Verona, Italy Nuclear Medicine Service, S. Croce e Carle Hospital, Cuneo, Italy
* Corresponding author. Division of Thoracic Surgery, S. Croce e Carle Hospital, Via Michele Coppino 26, 12100 Cuneo, Italy. Tel: +39-0171-641450; fax: +39-0171-642491; e-mail: [email protected] (A. Viti). Received 11 November 2013; received in revised form 12 February 2014; accepted 25 February 2014
Abstract A best evidence topic in thoracic surgery was written according to a structured protocol. The question addressed concerned the employment of 18-fluorine fluorodeoxyglucose positron emission tomography/computed tomography (PET–CT) in the preoperative evaluation of thymic epithelial neoplasms (TENs). We reviewed its role as a predictor of Masaoka stage and histology (according to the WHO). In clinical practice, stage is considered the most important determinant in the therapeutic approach. A total of 265 papers were found as a result of the reported search, of which 14 represented the best evidence to answer the clinical question. The authors, journal, date and country of publication, patient group studied, study type, relevant outcomes and results of these papers are tabulated. All the studies are retrospective analyses. Patient numbers varied from 13 to 58. Our research showed that PET–CT could clearly add information about the histology of the tumour. Thymic carcinoma constantly showed a higher standard uptake value (SUVmax) than thymomas. Furthermore, in one retrospective study of 36 patients, when using a derived PET indicator, the T/M ratio (ratio between SUVmax of the tumour and SUVmax of mediastinum, as conventionally measured at the level of the aortic arc), PET–CT could also differentiate between low- and high-risk thymomas (low risk vs high risk P = 0.01). In another study, a cut-off value of T/M ratio of 2.75 was identified between low- and high-risk TENs. The role of PET–CT in prediction of stage is harder to recognize. In one study, there was a statistically significant correlation between SUVmax, T/M ratio and Masaoka stage (P = 0.781 and 0.718, respectively). When analysing the data from the three larger series on this topic (58, 51 and 47 patients, the latter group selected in a multicentre study), only one study identified a correlation between SUVmax and Masaoka stage (Spearman correlation coefficient 0.30, P = 0.0436), while the other two failed to identify a relationship between SUVmax and stage. In three retrospective studies, PET–CT identified otherwise undetected distant metastases, thereby modifying the clinical approach. Actually, when evaluating the correlation between stage and PET–CT indicators (in particular SUVmax), results are controversial and would need further speculation.
In patients with [thymic epithelial neoplasms], does [CT scanning alone] vs the addition of a [PET–CT scan] significantly alter information on [staging or histology]?
to the pericardium. The patient also had a positron emission tomography (PET-CT) scan and the lesion was found to have a standard uptake value (SUVmax) of 8.3. The patient underwent minimally invasive procedure. However, due to invasion of the pericardium, conversion to sternotomy was required. Pathological evaluation confirmed a B3 thymoma with infiltration into the pericardium (Masaoka-Koga Stage III). Afterwards, you wondered whether PET–CT would be useful in the preoperative evaluation of all patients with mediastinal masses and, in particular, its role in predicting histology and stage of thymic epithelial neoplasms (TENs). You resolve to check the literature yourself.
CLINICAL SCENARIO
SEARCH STRATEGY
A 57-year old patient presented with an incidental finding of an isolated thymic mass 3 cm in diameter after a computed tomography (CT) scan of the thorax. It looked like it might be adherent
A Medline search from 1990 to December 2013 was performed using the PubMed interface with the following terms: [Thymoma OR Thymic Neoplasms] AND [Positron Emission Tomography OR
INTRODUCTION A best evidence topic was constructed according to a structured protocol. This is fully described in the ICVTS [1].
THREE-PART QUESTION
© The Author 2014. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.
BEST EVIDENCE TOPIC
Keywords: Review • Thymoma, positron emission tomography • 18-Fluorine fluorodeoxyglucose positron emission tomography/ computed tomography, staging
A. Viti et al. / Interactive CardioVascular and Thoracic Surgery
130
Table 1: Best evidence papers Author, date, journal and country Study type (level of evidence)
Patient group
Outcomes
Key results
Comments
Sung et al. (2006), J Nucl Med, Korea [2]
33 patients (16 thymic carcinomas included)
Parameters evaluated: SUVmax and intralesional pattern of uptake correlation with histology
SUVmax significantly higher in carcinomas than in thymomas (P < 0.001)
Integrated PET–CT can be useful for differentiating between thymoma and carcinoma
Retrospective analysis (level 3)
Possible role in staging
Carcinomas showed a more homogeneous uptake of 18-fluorine fluorodeoxyglucose (18F-FDG) than thymomas (P < 0.001)
PET–CT useful for detecting lymph node metastases, undetected with stand-alone computed tomography
In 2 cases, otherwise undetected metastases were evidenced Kumar et al. (2009), Ann Nucl Med, India [3]
23 patients (including 4 carcinomas and 5 hyperplasias)
Parameter evaluated: SUVmax to characterize different thymic lesions
Retrospective analysis (level 3) Igai et al. (2010), Eur J Cardiothorac Surg, Japan [4]
No statistically significant difference was observed between low-risk and high risk TENs 13 patients (including 5 thymic carcinomas)
Retrospective analysis (level 3) Fukumoto et al. (2012), Eur J Cardiothorac Surg, Japan [5]
Parameter evaluated: SUVmax and histological characterization
58 patients (11 carcinomas and 3 carcinoids)
Parameter evaluated: SUVmax and histological characterization Correlation between SUVmax and staging evaluated
40 patients (Type A = 1, AB = 12, B1 = 11, B2 = 7, B3 = 6 and C = 3)
Retrospective analysis (level 3)
Parameter evaluated: SUVmax, 18-FDG and 11-carbon acetate (AC) for histological characterization Correlation between SUVmax and staging evaluated Radiolabelled 11-carbon acetate is a marker of plasma membrane synthesis
Korst et al. (2014), J Thorac Cardiovasc Surg, USA [7]
No relationship between dimension and SUVmax small number of patients for each group Pretreatment SUVmax may be useful for differentiating thymoma from thymic carcinoma
SUVmax significantly higher in carcinomas (P < 0.001). No significant differences between lowand high-risk groups. SUVmax in advanced stages (III and IV) was not significantly higher (P = 0.06)
FDG SUVmax: Type A/AB: 3.2 ± 0.7; Type B1: 4.8 ± 2.0; Type B2: 3.7 ± 1.2; Type B3: 5.0 ± 1.4 and Type C: 9.2 ± 2.4. The uptake value was higher in carcinomas than in other types (A–B3: P = 0.001–0.048). AC SUVmax higher in A/AB thymomas (P < 0.01). Neither FDG SUV nor AC SUV uptake correlated with stage
21 patients who underwent induction therapy in a phase II trial
Parameter evaluated: SUVmax correlation with histological characterization and stage was a secondary end point of this study
Patients with carcinoma had a higher SUVmax than other TENs (P = 0.003). No significant differences in initial SUVmax were appreciated between the tumours postoperatively determined to be Stages I and II (median SUVmax 4.1) vs Stages III and IV (median SUVmax 5.3; P = 0.41)
51 patients (including carcinomas and carcinoids)
Parameter evaluated: SUVmax, SUVpeak, SUVmean, volumetric indicators (cm3) tumour volume above 45% of SUVmax
SUVmax, SUVpeak and SUVmean higher in carcinomas and carcinoids than in other thymomas (P = 0.0001, P = 0.0003 and P = 0.0001). Focal
Retrospective analysis of prospectively collected data (level 3) Benveniste et al. (2013), J Thorac Oncol, USA [8]
Mean SUVmax in the thymic carcinoma group was 8.15 ± 7.88, and in the thymoma group, it was 3.43 ± 2.19 (P = 0.002)
PET–CT can differentiate between TENs
Correlation between SUVmax and staging not evaluated
Retrospective analysis (level 3) Shibata et al. (2009), Cancer, Japan [6]
Significant differences between hyperplasia (mean SUVmax 1.1), thymoma (mean SUVmax 3) and carcinoma (mean SUVmax 7; P < 0.01)
Although neither FDG SUV nor AC SUV can predict invasiveness of thymomas assessed by tumour stage, they are useful for predicting histological types
SUVmax, SUVpeak and SUVmean could be helpful in differentiating carcinomas and B3 thymomas
Continued
A. Viti et al. / Interactive CardioVascular and Thoracic Surgery
131
Table 1: (Continued) Author, date, journal and country Study type (level of evidence)
Patient group
Retrospective analysis (level 3)
Outcomes
Key results
and volume of tumour with SUV above 3.5
uptake higher in B3 thymomas than in Type A, AB and B1 thymomas (P < 0.006) Tumour volume above 45% of SUVmax and volume of tumour with SUV above 3.5 relate with more advanced stages (P = 0.002 and 0.029)
Endo et al. (2008), Lung Cancer, Japan [9]
36 patients (15 low-risk thymomas, 10 high-risk thymomas and 11 thymic carcinoma)
Retrospective analysis (level 3) Terzi et al. (2011), Lung Cancer, Italy [11]
26 consecutive patients (including carcinomas)
Parameter evaluated: SUVmax and SUV T/M ratio The authors did not evaluate the Masaoka classification in detail
Parameter evaluated: SUVmax and SUV T/M ratio for histological classification and staging
Retrospective analysis (level 3)
Comments
Focal FDG uptake, as described by SUVmax, SUVpeak and SUVmean, cannot predict stage Volumetric indicators depict the volume of ‘metabolic avid tumour’ and better correlate with stage
Mean T/M ratio: Low risk: 2.64 ± 0.78 High risk: 4.29 ± 1.41 Carcinoma: 8.90 ± 3.62
Integrated PET–CT can differentiate between risk groups, according to Okamura Classification [10]
(low risk vs high risk P = 0.01; high risk vs carcinoma P = 0.01)
Introduction of SUV T/M ratio as a predictor of histology
SUVmax and T/M ratio resulted: Type A: 3.4 ± 0.8; 1.0 ± 0.8 Type AB: 3.8 ± 1.3; 1.9 ± 0.4 Type B1: 4.9 ± 0.8; 2.1 ± 0.3 Type B2: 6.6 ± 1.0; 4.9 ± 0.2 Type B3: 9.4 ± 3.6; 3.4 ± 0.2 Carcinoma: 17.1 ± 8.5; 9.6 ± 5.5
Introduction of a cut-off value between low- and high-risk groups In this series, PET–CT identified otherwise undetected distant metastases in 2 cases
Stage I: 3.7 ± 0.8; 2.1 ± 1.2 Stage II: 4.5 ± 1.7; 2.6 ± 1.1 Stage III: 12.1 ± 6.9; 3.7 ± 0.2 Stage IV: 19.7 ± 8.7; 11.5 ± 6.4
Significant correlation between T/M ratio SUVmax and increasing stage (P = 0.781 and 0.718 according to Bravais–Pearson linear correlation) A cut-off value of T/M ratio of 2.75 between low- and high-risk groups Matsumoto et al. (2012), Ann Thorac Surg, Japan [12]
39 patients (Type A = 1, AB = 4, B1 = 13, B2 = 11, B3 = 7 and C = 3)
Parameter evaluated: SUVmax for histological classification and staging. Correlation between SUVmax and Ki-67 expression
Retrospective analysis (level 3) Viti et al. (2014), Eur J Cardiothorac Surg, Italy [13] Retrospective analysis (level 3)
FDG SUVmax higher in carcinomas (P < 0.001) and in advanced stage (Stage IV) than in Stages I and II (P < 0.008)
SUVmax correlated positively with Ki-67 expression
Advanced Masaoka stages and more malignant thymomas showed higher SUVmax 23 patients. 17 low-risk thymomas (3 A, 9 AB and 5 B1) and 6 high-risk thymomas (5 B2 and 1 B3) Carcinomas were excluded
Parameter evaluated: SUVmax and SUV T/M ratio for histological classification and staging Correlation between T/M ratio and Ki-67
Low-risk thymomas showed a mean SUVmax of 4.01 ± 0.72, and a mean T/M ratio of 1.91 ± 0.21. High-risk thymomas showed a mean SUVmax of 7.60 ± 2.38, and a mean T/M ratio of 3.73 ± 0.76
Ki-67 expression turned out to be significantly higher in high-risk thymomas (P = 0.0002)
The difference between mean SUVmax and T/M ratio proved to be significant between low- and
Continued
BEST EVIDENCE TOPIC
Significant correlation between T/M ratio SUVmax and increasing malignancy (P = 0.728 and 0.662 according to Bravais–Pearson linear correlation)
A. Viti et al. / Interactive CardioVascular and Thoracic Surgery
132
Table 1: (Continued) Author, date, journal and country Study type (level of evidence)
Patient group
Outcomes
Key results
Comments
high-risk tumours (P = 0.001 and 0.001, respectively) No correlation between T/M ratio and Masaoka-Koga Stage (P = 0.3 according to Spearman Rank non-linear correlation) Toba et al. (2013), Eur J Cardiothorac Surg, Japan [14]
33 patients (15 low-risk thymomas, 10 high-risk thymomas and 8 carcinomas)
Parameter evaluated: SUVmax for histological classification and staging
Retrospective analysis (level 3)
SUVmax Low-risk TENs: 3.7 ± 2.0 High-risk TENs: 3.5 ± 2.1 Carcinomas: 5.9 ± 1.9 Stage I: 2.4 ± 1.3 Stage II: 3.7 ± 1.8 Stage III: 5.1 ± 1.9 Stage IV: 5.3 ± 2.2 Higher SUVmax in thymic carcinomas vs thymomas (P = 0.008). No significant difference between lowand high-risk thymomas Mean SUVmax value higher in Stages III and IV than in Stage I (both P = 0.002)
Lococo et al. (2013), Lung Cancer, Italy [15]
47 patients (Type A = 2, AB = 11, B1 = 9, B2 = 9, B3 = 9 and C = 7)
Retrospective analysis, multicentre study (level 3) Otsuka (2010), J Med Invest, Japan [16] Review of literature and retrospective analysis
35 patients (Low risk = 11, high risk = 13 and carcinomas = 11)
Parameter evaluated: SUVmax and SUVmax/T index (the ratio tumour SUVmax to tumour size) for histological classification and staging
SUVmax and SUVmax/T index were found to be higher in carcinomas than in other forms (P = 0.0045 and 0.0022)
SUVmax and histology
SUVmax Low risk (LR) TENs: 3.7 ± 1.8 High risk (HR) TENs: 4.0 ± 2.8 Carcinomas: 7.6 ± 3.7
(Stages I–II = 15 and Stages III–IV = 20)
SUVmax correlated with Masaoka stage (Spearman correlation coefficient: 0.30; P = 0.04), while SUVmax/T index did not PET–CT scan may indicate metastases, which are sometimes difficult to detect using other modalities
The difference of SUVmax between carcinomas and the other forms was statistically significant (P = 0.000482). Differences between LR and HR not significant 6/11 (55%) carcinomas showed distant metastases
PET/CT OR PET-CT]. All the results were screened for adherence to the topic. Case reports were excluded.
SEARCH OUTCOME Two hundred and sixty-five papers were found as a result of the reported search. From these, 14 papers provided the best evidence to answer the question. These are presented in Table 1.
RESULTS Many factors have been studied in the evaluation of prognosis for TENs. Large series identified as independent prognostic factors histology and Masaoka-Koga stage [17, 18]. In clinical practice, however, stage is considered the most important determinant of the therapeutic approach. Locally advanced stages (III and IV) would benefit from preoperative, platinum-based chemotherapy. Stage is mainly predicted based on indirect radiological signs
A. Viti et al. / Interactive CardioVascular and Thoracic Surgery
PET–CT in prediction of histology Sung et al. [2] identified a significant difference in SUVmax between thymoma and thymic carcinomas. The author concluded that PET–CT can be employed in the preoperative work-up of thymomas, and SUVmax can be a good predictor of histology and stage. SUVmax turned out to be significantly higher in carcinomas in the groups studied by Kumar et al. [3], Igai et al. [4], Fukumoto et al. [5], Shibata et al. [6] and more recently, Korst et al. [7]. Benveniste et al. [8] identified a SUVmax cut-off value of 6 between thymomas and carcinomas. In other studies, a significant difference in SUVmax was observed also between low- and high-risk thymomas. Those studies introduced another metabolic descriptor, besides SUVmax, that is the ratio between SUVmax and mean SUV of the mediastinum conventionally measured at the level of the aortic arch (SUV T/M ratio). This derived descriptor could avoid some of the bias of simple SUVmax such as biological factors (difference of tracer uptake among the patients) and inter-scanner variability. In a cohort of 36 patients, Endo et al. [9] identified a significant relationship between T/M ratio and histological subtype in TENs when a three-group classification was employed (low-risk, highrisk and carcinomas). Terzi et al. [11] identified in a cohort of 26 consecutive patients a strong correlation between SUV T/M ratio and risk groups. A cut-off value between the two classes was identified at an SUV T/M ratio of 2.75. Matsumoto et al. [12] and Viti et al. [13] identified a correlation between, respectively, SUVmax and T/M ratio of TEN and the expression of Ki-67, a marker of proliferation and biological aggressiveness, thereby providing a correlation between biological behaviour of TENs and imaging observation. We may conclude that PET–CT may have a possible role in definition ( prediction) of histology.
PET–CT and prediction on stage As far as correlation between PET–CT indicators and stage is concerned, the results are more ambiguous. Previously mentioned studies failed to find a correlation between SUVmax and T/M ratio with stage [9, 13]. In another study [11], there was a statistically significant correlation between SUVmax, T/M ratio and disease stage (P = 0.781 and 0.718, respectively, according to Bravais–Pearson linear correlation). Matsumoto et al. [12] found a higher SUVmax in Stage IV than in Stages I and II (P < 0.008 and
Interactive CardioVascular and Thoracic Surgery 19 (2014) 129–134 doi:10.1093/icvts/ivu068 Advance Access publication 19 March 2014
Is a positron emission tomography–computed tomography scan useful in the staging of thymic epithelial neoplasms? Andrea Vitia,*, Alberto Terzib, Andrea Bianchic and Luca Bertolaccinia a b c
Thoracic Surgery Unit, S. Croce e Carle Hospital, Cuneo, Italy Thoracic Surgery Unit, Sacred Heart Hospital, Negrar, Verona, Italy Nuclear Medicine Service, S. Croce e Carle Hospital, Cuneo, Italy
* Corresponding author. Division of Thoracic Surgery, S. Croce e Carle Hospital, Via Michele Coppino 26, 12100 Cuneo, Italy. Tel: +39-0171-641450; fax: +39-0171-642491; e-mail: [email protected] (A. Viti). Received 11 November 2013; received in revised form 12 February 2014; accepted 25 February 2014
Abstract A best evidence topic in thoracic surgery was written according to a structured protocol. The question addressed concerned the employment of 18-fluorine fluorodeoxyglucose positron emission tomography/computed tomography (PET–CT) in the preoperative evaluation of thymic epithelial neoplasms (TENs). We reviewed its role as a predictor of Masaoka stage and histology (according to the WHO). In clinical practice, stage is considered the most important determinant in the therapeutic approach. A total of 265 papers were found as a result of the reported search, of which 14 represented the best evidence to answer the clinical question. The authors, journal, date and country of publication, patient group studied, study type, relevant outcomes and results of these papers are tabulated. All the studies are retrospective analyses. Patient numbers varied from 13 to 58. Our research showed that PET–CT could clearly add information about the histology of the tumour. Thymic carcinoma constantly showed a higher standard uptake value (SUVmax) than thymomas. Furthermore, in one retrospective study of 36 patients, when using a derived PET indicator, the T/M ratio (ratio between SUVmax of the tumour and SUVmax of mediastinum, as conventionally measured at the level of the aortic arc), PET–CT could also differentiate between low- and high-risk thymomas (low risk vs high risk P = 0.01). In another study, a cut-off value of T/M ratio of 2.75 was identified between low- and high-risk TENs. The role of PET–CT in prediction of stage is harder to recognize. In one study, there was a statistically significant correlation between SUVmax, T/M ratio and Masaoka stage (P = 0.781 and 0.718, respectively). When analysing the data from the three larger series on this topic (58, 51 and 47 patients, the latter group selected in a multicentre study), only one study identified a correlation between SUVmax and Masaoka stage (Spearman correlation coefficient 0.30, P = 0.0436), while the other two failed to identify a relationship between SUVmax and stage. In three retrospective studies, PET–CT identified otherwise undetected distant metastases, thereby modifying the clinical approach. Actually, when evaluating the correlation between stage and PET–CT indicators (in particular SUVmax), results are controversial and would need further speculation.
In patients with [thymic epithelial neoplasms], does [CT scanning alone] vs the addition of a [PET–CT scan] significantly alter information on [staging or histology]?
to the pericardium. The patient also had a positron emission tomography (PET-CT) scan and the lesion was found to have a standard uptake value (SUVmax) of 8.3. The patient underwent minimally invasive procedure. However, due to invasion of the pericardium, conversion to sternotomy was required. Pathological evaluation confirmed a B3 thymoma with infiltration into the pericardium (Masaoka-Koga Stage III). Afterwards, you wondered whether PET–CT would be useful in the preoperative evaluation of all patients with mediastinal masses and, in particular, its role in predicting histology and stage of thymic epithelial neoplasms (TENs). You resolve to check the literature yourself.
CLINICAL SCENARIO
SEARCH STRATEGY
A 57-year old patient presented with an incidental finding of an isolated thymic mass 3 cm in diameter after a computed tomography (CT) scan of the thorax. It looked like it might be adherent
A Medline search from 1990 to December 2013 was performed using the PubMed interface with the following terms: [Thymoma OR Thymic Neoplasms] AND [Positron Emission Tomography OR
INTRODUCTION A best evidence topic was constructed according to a structured protocol. This is fully described in the ICVTS [1].
THREE-PART QUESTION
© The Author 2014. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.
BEST EVIDENCE TOPIC
Keywords: Review • Thymoma, positron emission tomography • 18-Fluorine fluorodeoxyglucose positron emission tomography/ computed tomography, staging
A. Viti et al. / Interactive CardioVascular and Thoracic Surgery
130
Table 1: Best evidence papers Author, date, journal and country Study type (level of evidence)
Patient group
Outcomes
Key results
Comments
Sung et al. (2006), J Nucl Med, Korea [2]
33 patients (16 thymic carcinomas included)
Parameters evaluated: SUVmax and intralesional pattern of uptake correlation with histology
SUVmax significantly higher in carcinomas than in thymomas (P < 0.001)
Integrated PET–CT can be useful for differentiating between thymoma and carcinoma
Retrospective analysis (level 3)
Possible role in staging
Carcinomas showed a more homogeneous uptake of 18-fluorine fluorodeoxyglucose (18F-FDG) than thymomas (P < 0.001)
PET–CT useful for detecting lymph node metastases, undetected with stand-alone computed tomography
In 2 cases, otherwise undetected metastases were evidenced Kumar et al. (2009), Ann Nucl Med, India [3]
23 patients (including 4 carcinomas and 5 hyperplasias)
Parameter evaluated: SUVmax to characterize different thymic lesions
Retrospective analysis (level 3) Igai et al. (2010), Eur J Cardiothorac Surg, Japan [4]
No statistically significant difference was observed between low-risk and high risk TENs 13 patients (including 5 thymic carcinomas)
Retrospective analysis (level 3) Fukumoto et al. (2012), Eur J Cardiothorac Surg, Japan [5]
Parameter evaluated: SUVmax and histological characterization
58 patients (11 carcinomas and 3 carcinoids)
Parameter evaluated: SUVmax and histological characterization Correlation between SUVmax and staging evaluated
40 patients (Type A = 1, AB = 12, B1 = 11, B2 = 7, B3 = 6 and C = 3)
Retrospective analysis (level 3)
Parameter evaluated: SUVmax, 18-FDG and 11-carbon acetate (AC) for histological characterization Correlation between SUVmax and staging evaluated Radiolabelled 11-carbon acetate is a marker of plasma membrane synthesis
Korst et al. (2014), J Thorac Cardiovasc Surg, USA [7]
No relationship between dimension and SUVmax small number of patients for each group Pretreatment SUVmax may be useful for differentiating thymoma from thymic carcinoma
SUVmax significantly higher in carcinomas (P < 0.001). No significant differences between lowand high-risk groups. SUVmax in advanced stages (III and IV) was not significantly higher (P = 0.06)
FDG SUVmax: Type A/AB: 3.2 ± 0.7; Type B1: 4.8 ± 2.0; Type B2: 3.7 ± 1.2; Type B3: 5.0 ± 1.4 and Type C: 9.2 ± 2.4. The uptake value was higher in carcinomas than in other types (A–B3: P = 0.001–0.048). AC SUVmax higher in A/AB thymomas (P < 0.01). Neither FDG SUV nor AC SUV uptake correlated with stage
21 patients who underwent induction therapy in a phase II trial
Parameter evaluated: SUVmax correlation with histological characterization and stage was a secondary end point of this study
Patients with carcinoma had a higher SUVmax than other TENs (P = 0.003). No significant differences in initial SUVmax were appreciated between the tumours postoperatively determined to be Stages I and II (median SUVmax 4.1) vs Stages III and IV (median SUVmax 5.3; P = 0.41)
51 patients (including carcinomas and carcinoids)
Parameter evaluated: SUVmax, SUVpeak, SUVmean, volumetric indicators (cm3) tumour volume above 45% of SUVmax
SUVmax, SUVpeak and SUVmean higher in carcinomas and carcinoids than in other thymomas (P = 0.0001, P = 0.0003 and P = 0.0001). Focal
Retrospective analysis of prospectively collected data (level 3) Benveniste et al. (2013), J Thorac Oncol, USA [8]
Mean SUVmax in the thymic carcinoma group was 8.15 ± 7.88, and in the thymoma group, it was 3.43 ± 2.19 (P = 0.002)
PET–CT can differentiate between TENs
Correlation between SUVmax and staging not evaluated
Retrospective analysis (level 3) Shibata et al. (2009), Cancer, Japan [6]
Significant differences between hyperplasia (mean SUVmax 1.1), thymoma (mean SUVmax 3) and carcinoma (mean SUVmax 7; P < 0.01)
Although neither FDG SUV nor AC SUV can predict invasiveness of thymomas assessed by tumour stage, they are useful for predicting histological types
SUVmax, SUVpeak and SUVmean could be helpful in differentiating carcinomas and B3 thymomas
Continued
A. Viti et al. / Interactive CardioVascular and Thoracic Surgery
131
Table 1: (Continued) Author, date, journal and country Study type (level of evidence)
Patient group
Retrospective analysis (level 3)
Outcomes
Key results
and volume of tumour with SUV above 3.5
uptake higher in B3 thymomas than in Type A, AB and B1 thymomas (P < 0.006) Tumour volume above 45% of SUVmax and volume of tumour with SUV above 3.5 relate with more advanced stages (P = 0.002 and 0.029)
Endo et al. (2008), Lung Cancer, Japan [9]
36 patients (15 low-risk thymomas, 10 high-risk thymomas and 11 thymic carcinoma)
Retrospective analysis (level 3) Terzi et al. (2011), Lung Cancer, Italy [11]
26 consecutive patients (including carcinomas)
Parameter evaluated: SUVmax and SUV T/M ratio The authors did not evaluate the Masaoka classification in detail
Parameter evaluated: SUVmax and SUV T/M ratio for histological classification and staging
Retrospective analysis (level 3)
Comments
Focal FDG uptake, as described by SUVmax, SUVpeak and SUVmean, cannot predict stage Volumetric indicators depict the volume of ‘metabolic avid tumour’ and better correlate with stage
Mean T/M ratio: Low risk: 2.64 ± 0.78 High risk: 4.29 ± 1.41 Carcinoma: 8.90 ± 3.62
Integrated PET–CT can differentiate between risk groups, according to Okamura Classification [10]
(low risk vs high risk P = 0.01; high risk vs carcinoma P = 0.01)
Introduction of SUV T/M ratio as a predictor of histology
SUVmax and T/M ratio resulted: Type A: 3.4 ± 0.8; 1.0 ± 0.8 Type AB: 3.8 ± 1.3; 1.9 ± 0.4 Type B1: 4.9 ± 0.8; 2.1 ± 0.3 Type B2: 6.6 ± 1.0; 4.9 ± 0.2 Type B3: 9.4 ± 3.6; 3.4 ± 0.2 Carcinoma: 17.1 ± 8.5; 9.6 ± 5.5
Introduction of a cut-off value between low- and high-risk groups In this series, PET–CT identified otherwise undetected distant metastases in 2 cases
Stage I: 3.7 ± 0.8; 2.1 ± 1.2 Stage II: 4.5 ± 1.7; 2.6 ± 1.1 Stage III: 12.1 ± 6.9; 3.7 ± 0.2 Stage IV: 19.7 ± 8.7; 11.5 ± 6.4
Significant correlation between T/M ratio SUVmax and increasing stage (P = 0.781 and 0.718 according to Bravais–Pearson linear correlation) A cut-off value of T/M ratio of 2.75 between low- and high-risk groups Matsumoto et al. (2012), Ann Thorac Surg, Japan [12]
39 patients (Type A = 1, AB = 4, B1 = 13, B2 = 11, B3 = 7 and C = 3)
Parameter evaluated: SUVmax for histological classification and staging. Correlation between SUVmax and Ki-67 expression
Retrospective analysis (level 3) Viti et al. (2014), Eur J Cardiothorac Surg, Italy [13] Retrospective analysis (level 3)
FDG SUVmax higher in carcinomas (P < 0.001) and in advanced stage (Stage IV) than in Stages I and II (P < 0.008)
SUVmax correlated positively with Ki-67 expression
Advanced Masaoka stages and more malignant thymomas showed higher SUVmax 23 patients. 17 low-risk thymomas (3 A, 9 AB and 5 B1) and 6 high-risk thymomas (5 B2 and 1 B3) Carcinomas were excluded
Parameter evaluated: SUVmax and SUV T/M ratio for histological classification and staging Correlation between T/M ratio and Ki-67
Low-risk thymomas showed a mean SUVmax of 4.01 ± 0.72, and a mean T/M ratio of 1.91 ± 0.21. High-risk thymomas showed a mean SUVmax of 7.60 ± 2.38, and a mean T/M ratio of 3.73 ± 0.76
Ki-67 expression turned out to be significantly higher in high-risk thymomas (P = 0.0002)
The difference between mean SUVmax and T/M ratio proved to be significant between low- and
Continued
BEST EVIDENCE TOPIC
Significant correlation between T/M ratio SUVmax and increasing malignancy (P = 0.728 and 0.662 according to Bravais–Pearson linear correlation)
A. Viti et al. / Interactive CardioVascular and Thoracic Surgery
132
Table 1: (Continued) Author, date, journal and country Study type (level of evidence)
Patient group
Outcomes
Key results
Comments
high-risk tumours (P = 0.001 and 0.001, respectively) No correlation between T/M ratio and Masaoka-Koga Stage (P = 0.3 according to Spearman Rank non-linear correlation) Toba et al. (2013), Eur J Cardiothorac Surg, Japan [14]
33 patients (15 low-risk thymomas, 10 high-risk thymomas and 8 carcinomas)
Parameter evaluated: SUVmax for histological classification and staging
Retrospective analysis (level 3)
SUVmax Low-risk TENs: 3.7 ± 2.0 High-risk TENs: 3.5 ± 2.1 Carcinomas: 5.9 ± 1.9 Stage I: 2.4 ± 1.3 Stage II: 3.7 ± 1.8 Stage III: 5.1 ± 1.9 Stage IV: 5.3 ± 2.2 Higher SUVmax in thymic carcinomas vs thymomas (P = 0.008). No significant difference between lowand high-risk thymomas Mean SUVmax value higher in Stages III and IV than in Stage I (both P = 0.002)
Lococo et al. (2013), Lung Cancer, Italy [15]
47 patients (Type A = 2, AB = 11, B1 = 9, B2 = 9, B3 = 9 and C = 7)
Retrospective analysis, multicentre study (level 3) Otsuka (2010), J Med Invest, Japan [16] Review of literature and retrospective analysis
35 patients (Low risk = 11, high risk = 13 and carcinomas = 11)
Parameter evaluated: SUVmax and SUVmax/T index (the ratio tumour SUVmax to tumour size) for histological classification and staging
SUVmax and SUVmax/T index were found to be higher in carcinomas than in other forms (P = 0.0045 and 0.0022)
SUVmax and histology
SUVmax Low risk (LR) TENs: 3.7 ± 1.8 High risk (HR) TENs: 4.0 ± 2.8 Carcinomas: 7.6 ± 3.7
(Stages I–II = 15 and Stages III–IV = 20)
SUVmax correlated with Masaoka stage (Spearman correlation coefficient: 0.30; P = 0.04), while SUVmax/T index did not PET–CT scan may indicate metastases, which are sometimes difficult to detect using other modalities
The difference of SUVmax between carcinomas and the other forms was statistically significant (P = 0.000482). Differences between LR and HR not significant 6/11 (55%) carcinomas showed distant metastases
PET/CT OR PET-CT]. All the results were screened for adherence to the topic. Case reports were excluded.
SEARCH OUTCOME Two hundred and sixty-five papers were found as a result of the reported search. From these, 14 papers provided the best evidence to answer the question. These are presented in Table 1.
RESULTS Many factors have been studied in the evaluation of prognosis for TENs. Large series identified as independent prognostic factors histology and Masaoka-Koga stage [17, 18]. In clinical practice, however, stage is considered the most important determinant of the therapeutic approach. Locally advanced stages (III and IV) would benefit from preoperative, platinum-based chemotherapy. Stage is mainly predicted based on indirect radiological signs
A. Viti et al. / Interactive CardioVascular and Thoracic Surgery
PET–CT in prediction of histology Sung et al. [2] identified a significant difference in SUVmax between thymoma and thymic carcinomas. The author concluded that PET–CT can be employed in the preoperative work-up of thymomas, and SUVmax can be a good predictor of histology and stage. SUVmax turned out to be significantly higher in carcinomas in the groups studied by Kumar et al. [3], Igai et al. [4], Fukumoto et al. [5], Shibata et al. [6] and more recently, Korst et al. [7]. Benveniste et al. [8] identified a SUVmax cut-off value of 6 between thymomas and carcinomas. In other studies, a significant difference in SUVmax was observed also between low- and high-risk thymomas. Those studies introduced another metabolic descriptor, besides SUVmax, that is the ratio between SUVmax and mean SUV of the mediastinum conventionally measured at the level of the aortic arch (SUV T/M ratio). This derived descriptor could avoid some of the bias of simple SUVmax such as biological factors (difference of tracer uptake among the patients) and inter-scanner variability. In a cohort of 36 patients, Endo et al. [9] identified a significant relationship between T/M ratio and histological subtype in TENs when a three-group classification was employed (low-risk, highrisk and carcinomas). Terzi et al. [11] identified in a cohort of 26 consecutive patients a strong correlation between SUV T/M ratio and risk groups. A cut-off value between the two classes was identified at an SUV T/M ratio of 2.75. Matsumoto et al. [12] and Viti et al. [13] identified a correlation between, respectively, SUVmax and T/M ratio of TEN and the expression of Ki-67, a marker of proliferation and biological aggressiveness, thereby providing a correlation between biological behaviour of TENs and imaging observation. We may conclude that PET–CT may have a possible role in definition ( prediction) of histology.
PET–CT and prediction on stage As far as correlation between PET–CT indicators and stage is concerned, the results are more ambiguous. Previously mentioned studies failed to find a correlation between SUVmax and T/M ratio with stage [9, 13]. In another study [11], there was a statistically significant correlation between SUVmax, T/M ratio and disease stage (P = 0.781 and 0.718, respectively, according to Bravais–Pearson linear correlation). Matsumoto et al. [12] found a higher SUVmax in Stage IV than in Stages I and II (P < 0.008 and
