Diagnostics, Vol. 13, Pages 98: Is PET/CT Able to Predict Histology in Thymic Epithelial Tumours? A Narrative Review

2.1. PET/CT to Distinguish Thymic Hyperplasia from Thymic Epithelial TumoursAlthough current guidelines do not recommend pre-operative biopsy in cases of suspected thymoma [36], a distinction between benign conditions (such as hyperplasia) and TETs can alter the therapeutic strategy significantly. Metabolic parameters may prove to be a useful adjunct in the investigation of lesions in the upper anterior mediastinum (Figure 1).The first report of 18F-FDG PET/CT use in TET/hyperplasia was by Liu et al. [19], who evaluated the ratio between SUVmax in the tumour and in the lung (tumour-to-lung ratio or TLR), reporting a significant difference between TETs (TLR: 3.4/3.5) and thymoma (TLR: 5.7 ± 1.7). Other small series analysed SUVmax alone, reporting lower metabolic values in thymic hyperplasia and higher values in TETs. El-Bawab et al. [20] reported SUVmax from 0.7 to 2.5 in hyperplasia compared to 3.1 to 6.1 in thymoma, while Kumar et al. [21] reported an SUVmax of 0.7–1.8 in hyperplasia, 1.7–3.9 in low-risk thymomas and 4.3–9.2 in thymic carcinoma. Moreover, Watanabe and colleagues [22] reported a mean SUVmax of 1.4 ± 0.7 in thymic hyperplasia, 3.7 ± 1.5 in thymoma and 11.4 ± 2.6 in thymic cancer. Again, a significant SUVmax difference was present only between hyperplasia and cancer. Interestingly, in this large series of patients, no case of hyperplastic thymus showed a value of SUVmax higher than 3.Given the overlapping values, differentiating hyperplasia and low-grade thymomas (A, AB histology) on the sole basis of SUVmax can be challenging. This prompted the study of Travaini et al., who integrated metabolic (i.e., 18F-FDG PET/CT) and anatomical (i.e., CT) features [23]. Their study included thymic cysts and found that–despite overlapping SUVmax values in hyperplasia (1.7–5) and low-risk thymomas (2.3–15.5), the integration of anatomical features could help identify 100% of benign lesions. Based on these studies, 18F-FDG PET/CT could be an important tool in anterior mass determination and may help differentiate hyperplasia from high-grade thymomas and thymic carcinomas, considering that SUVmax in hyperplasia is rarely higher than 3. However, 18F-FDG PET/CT alone cannot discriminate between hyperplasia and low-risk thymomas, to which end morphological evaluation is mandatory, as it could guide differential diagnosis. As a matter of fact, hyperplasia and low-grade thymomas show a distinct CT appearance: V-shape or triangular in hyperplasia compared to nodule/mass in the case of TETs [18]. A further factor to take into account is the spatial distribution of the uptake: low and diffuse across the thymus in hyperplasia, localised in foci or nodules in TETs [18,20]. 2.2. PET/CT Parameters to Distinguish Histology in TETsA simplified histological classification has been proposed to identify different classes of risk in TETs [5]: types A, AB and B1 = “low-risk” thymic neoplasms; B2 and B3 = “high-risk” thymic neoplasms; and thymic carcinoma. The scientific community has largely adopted this simplification and a recent meta-analysis by Marchevsky et al. [6] has confirmed its prognostic value. If 18F-FDG PET/CT were confirmed to be able to assess the grade of malignancy in TETs, it could play an important role in the management of the disease (Figure 2). A few studies have shown promising results (see Table 2); most have focused on SUVmax, supporting the use of this metabolic marker in clinical routines [24,26,28,29,31,32,33,34,35,37]. SUVmax has been reported to be consistently higher in carcinoma than in high- or low-risk thymoma, with values between 7.2 and 15.2. In addition to the SUVmax value, the pattern of 18F-FDG uptake can provide useful information, as it appears more homogeneous in a higher proportion of thymic carcinomas than thymomas (both low- and high-risk) [24]. A few years ago, our group participated in the first multicentric study on the role of 18F-FDG PET/CT as a predictor of WHO classification in a relatively large cohort of TETs (n = 47) [16]. SUVmax was found to correlate with WHO malignancy grade (i.e., low vs. high-risk vs. carcinoma), with a Spearman correlation of 0.56 (p37]. Overall, SUVmax was able to predict histologic subtypes with good accuracy, expressed by an area under the ROC curve ranging from 0.82 to 0.96. Most studies included in the meta-analysis divided TETs into low-risk, high-risk, and carcinoma, except for one that considered only thymoma and carcinoma [14]. In a retrospective study of 51 patients, Benveniste et al. [14] observed significantly higher SUVmax in carcinoma (n = 12) and carcinoid (n = 2) than in thymoma. SUVpeak and SUVmean also significantly increased in carcinoma.Readers might have noted that SUVmax values are relatively wide among the abovementioned studies. This could be related to different uptake times, patient obesity, blood glucose levels, different PET/CT scanners or inherent differences among the studied cohorts. In order to overcome these limitations, other metabolic parameters have been proposed. The ratio of SUVmax to tumour size (SUVmax/T) reduces the bias related to tumour dimensions and has been proven to correlate with histologic subtypes of TETs, with an AUC between 0.69 and 0.93 [16,28,32,33]. Similarly, Endo et al. [25] calculated the ratio between SUVpeaks of the tumour and mediastinum (T/M ratio) in 36 patients with histologically proven TETs. Mean T/M ratio differed significantly in low-risk thymoma, high-risk thymoma, and carcinoma (2.64 vs. 4.29 vs. 8.90, respectively, p = 0.01).Volumetric PET/CT parameters, such as metabolic tumour volume (MTV) and total lesion glycolysis (TLG), have been correlated with clinical outcomes in several malignancies. However, their application in TETs showed contrasting results [27,28,35]: in a retrospective monocentric study of 23 patients with pathologically proven TETs (17 low-risk, 6 high-risk, no carcinoma), Bertolaccini and colleagues [27] found that T/M ratio, MTV, and total glycolytic volume (TGV) were able to discriminate between low- and high-risk TETs. Statistical correlation with the WHO classification was higher for TGV (rho = 0.897) than for T/M ratio (rho = 0.873). A TGV cut-off value of 383 seemed to be able to separate low- and high-risk TETs, suggesting its use as a potential parameter in pre-treatment stratification. Volumetric parameters showed higher values in carcinoma than in low- and high-risk thymoma in a retrospective study by Han et al. on 114 patients with TETs [35]. However, Benveniste et al. [14] observed that the total tumour volume (taking into account areas with SUV above 3.5) was larger in thymic carcinoma/carcinoid than in thymoma (p = 0.02). The correlation was found only when the total volume was calculated, taking into account areas with SUV above 3.5 (the use of volumes with SUV above 45% of SUVmax failed to show any difference). On the other hand, Park et al. [28] failed to differentiate thymomas and carcinoma on the basis of MTV and TLG. Recently, new approaches have been proposed to predict TET histology by means of 18F-FDG PET/CT. Shinya et al. [30] evaluated metabolic parameters through dual-time-point PET/CT acquisition (i.e., after 90 min and 2 h) in 56 TET patients, suggesting that delayed scanning could improve the diagnostic capacity for high-risk TETs with an accuracy of 82.9% and an AUC of 0.825. A pilot study performed by Ozkan and collaborators in 2022 [38] proposed a machine-learning model and assessed its ability to classify low- and high-risk thymoma on PET/CT images. SUVmax, SUVmean, SUVpeak, MTV and TLG of primary mediastinal lesions were calculated in 27 TET patients. First-, second- and higher-order texture features were also calculated. Among other variables (LDH level and presence of myasthenia gravis), the SHAPE_Sphericity [only for 3D ROI (nz > 1)] was able to differentiate low- and high-risk thymoma.

Despite encouraging results, the integration of these complex parameters into daily clinical practice is far from becoming a reality due to uncertain reproducibility. Therefore, SUVmax remains the most promising parameter for estimating histology in TET patients.

留言 (0)

沒有登入
gif