The efficacy of 18F-FDG-PET-based radiomic and deep-learning features using a machine-learning approach to predict the pathological risk subtypes of thymic epithelial tumors

OBJECTIVE

To examine whether the machine-learning approach using 18-fludeoxyglucose positron emission tomography (18F-FDG-PET)-based radiomic and deep-learning features is useful for predicting the pathological risk subtypes of thymic epithelial tumors (TETs).

METHODS

This retrospective study included 79 TET [27 low-risk thymomas (types A, AB and B1), 31 high-risk thymomas (types B2 and B3) and 21 thymic carcinomas] patients who underwent pre-therapeutic 18F-FDG-PET/CT. High-risk TETs (high-risk thymomas and thymic carcinomas) were 52 patients. The 107 PET-based radiomic features, including SUV-related parameters [maximum SUV (SUVmax), metabolic tumor volume (MTV), and total lesion glycolysis (TLG)] and 1024 deep-learning features extracted from the convolutional neural network were used to predict the pathological risk subtypes of TETs using six different machine-learning algorithms. The area under the curves (AUCs) were calculated to compare the predictive performances.

RESULTS

SUV-related parameters yielded the following AUCs for predicting thymic carcinomas: SUVmax 0.713, MTV 0.442, and TLG 0.479 or high-risk TETs: SUVmax 0.673, MTV 0.533, and TLG 0.539. The best-performing algorithm was the logistic regression model for predicting thymic carcinomas (AUC 0.900, accuracy 81.0%), and the random forest (RF) model for high-risk TETs (AUC 0.744, accuracy 72.2%). The AUC was significantly higher in the logistic regression model than three SUV-related parameters for predicting thymic carcinomas, and in the RF model than MTV and TLG for predicting high-risk TETs (each; p < 0.05).

CONCLUSION

18F-FDG-PET-based radiomic analysis using a machine-learning approach may be useful for predicting the pathological risk subtypes of TETs.

ADVANCES IN KNOWLEDGE

Machine-learning approach using 18F-FDG-PET-based radiomic features has the potential to predict the pathological risk subtypes of TETs.

Value of metabolic parameters in distinguishing primary mediastinal lymphomas from thymic epithelial tumors

Objective: A high rate of unnecessary thymectomies has been reported. This study aimed to distinguish primary mediastinal lymphomas (PMLs) from thymic epithelial tumors (TETs) by evaluating volumetric and metabolic parameters with ¹⁸F-FDG PET/CT.

Methods: A total of 136 patients who were pathologically diagnosed with TETs or PMLs were enrolled, and ¹⁸F-FDG PET/CT was performed before therapy. Volumetric parameters, including the mean SUV (SUVmean), metabolic tumor volume (MTV), total lesion glycolysis (TLG), and SUVmax, were determined and compared between the 2 subtypes. The diagnostic performance of these parameters was evaluated with receiver operating characteristic (ROC) curve analysis.

Results: All parameters significantly differed between patients with PMLs and TETs. Patients with lymphomas were younger and had higher SUVmean, SUVmax, TLG, and MTV values than patients with TETs. The MTV and TLG values had similar diagnostic performance. ROC analysis indicated that the areas under the curves of the SUVmean and SUVmax values performed similarly (approximately 0.76) in differentiating patients with PMLs from TETs, and both values were better than the MTV and TLG values. When age was included with the SUVmax in differentiating TETs from PMLs, the AUC was 0.91, and the sensitivity and specificity increased to 80% and 93%, respectively.

Conclusions: The SUVmax and volumetric parameters of ¹⁸F-FDG PET/CT can be used to distinguish patients with PMLs versus TETs, and thus may aid in preventing unnecessary thymectomies or other invasive operations.

Texture analysis of 18F-FDG PET/CT for grading thymic epithelial tumours: usefulness of combining SUV and texture parameters

OBJECTIVE

To retrospectively investigate the standardized uptake value (SUV)-related and heterogeneous texture parameters individually and in combination for differentiating between low- and high-risk ¹⁸Fluorone-fludeoxyglucose (¹⁸F-FDG)-avid thymic epithelial tumours (TETs) with positron emission tomography (PET)/CT.

METHODS

SUV-related and 6 texture parameters (entropy, homogeneity, dissimilarity, intensity variability, size-zone variability and zone percentage) were compared between 11 low-risk and 23 high-risk TETs (metabolic tumour volume >10.0 cm³ and SUV ≥2.5). Diagnostic performance was evaluated by receiver operating characteristic analysis. The diagnostic value of combining SUV and texture parameters was examined by a scoring system.

RESULTS

High-risk TETs were significantly higher in SUVmax (p = 0.022), entropy (p = 0.038), intensity variability (p = 0.041) and size-zone variability (p = 0.045) than low-risk TETs. Diagnostic accuracies of these 4 parameters, dissimilarity and zone percentage which also showed significance in receiver operating characteristic analysis ranged between 64.7 and 73.5% without significant differences in AUC (range; 0.71 to 0.75) (p ≥ 0.05 each). Each parameter was scored as 0 (negative for high-risk) or 1 (positive for high-risk) according to each threshold criterion, then scores were summed [0 or 1 for low-risk TETs (median; 1); ≥2 for high-risk TETs (median; 4)]. The sensitivity, specificity and accuracy of detecting high-risk TETs were 100, 81.8 and 94.1%, respectively, with an AUC of 0.99.

CONCLUSION

The diagnostic performances of individual SUVmax and texture parameters were relatively low. However, combining these parameters can significantly increase diagnostic performance when differentiating between relatively large low- and high-risk ¹⁸F-FDG-avid TETs. Advances in knowledge: Combined use of SUVmax and texture parameters can significantly increase the diagnostic performance when differentiating between low- and high-risk TETs.

Correlation of 18F-FDG uptake on PET/CT with Ki67 immunohistochemistry in pre-treatment epithelial ovarian cancer

OBJECTIVE

Standardised uptake value (SUV) and volumetric parameters such as metabolic tumour volume (MTV) and total lesion glycolysis (TLG) from ¹⁸F-FDG PET/CT are useful criteria for disease prognosis in pre-operative and post-treatment epithelial ovarian cancer (EOC). Ki67 is another prognostic biomarker in EOC, associated with tumour aggressiveness. The aim of this study is to evaluate the association between ¹⁸F-FDG PET/CT measurements and Ki67 in pre-treatment EOC to determine if PET/CT parameters could non-invasively predict tumour aggressiveness.

MATERIAL AND METHODS

A pre-treatment PET/CT was performed on 18 patients with suspected or newly diagnosed EOC. Maximum SUV (SUVmax), mean SUV (SUVmean), whole-body MTV (wbMTV), and whole-body TLG (wbTLG) with a threshold of 30% and 40% of the SUVmax were obtained. Furthermore, Ki67 index (mean and hotspot) was estimated in tumour tissue specimens. Immunohistochemical findings were correlated with PET parameters.

RESULTS

The mean age was 57.0 years old (standard deviation 13.6 years). A moderate correlation was observed between mean Ki67 index and SUVmax (r=0.392), SUVmean 30% (r=0.437), and SUVmean 40% (r=0.443), and also between hotspot Ki67 index and SUVmax (r=0.360), SUVmean 30% (r=0.362) and SUVmean 40% (r=0.319). There was a weaker correlation, which was inversely negative, between mean and hotspot Ki67 and volumetric PET parameters. However, no statistical significant differences were found for any correlations.

CONCLUSIONS

SUVmax and SUVmean were moderately correlated with Ki67 index, whereas volumetric PET parameters overall, showed a weaker correlation. Thus, SUVmax and SUVmean could be used to assess tumour aggressiveness in pre-treatment EOC.

Correlation between the Uptake of 18F-Fluorodeoxyglucose (18F-FDG) and the Expression of Proliferation-Associated Antigen Ki-67 in Cancer Patients: A Meta-Analysis

Objective

To study the correlation between ¹⁸F-FDG uptake and cell proliferation in cancer patients by meta-analysis of published articles.

Methods

We searched PubMed (MEDLINE included), EMBASE, and Cochrane Database of Systematic Review, and selected research articles on the relationship between ¹⁸F-FDG uptake and Ki-67 expression (published between August 1, 1994-August 1, 2014), according to the literature inclusion and exclusion criteria. The publishing language was limited to English. The quality of included articles was evaluated according to the Quality Assessment of Diagnosis Accuracy Studies-2 (QUADAS-2). The correlation coefficient (r) was extracted from the included articles and processed by Fisher's r-to-z transformation. The combined correlation coefficient (r) and the 95% confidence interval (CI) were calculated with STATA 11.0 software under a random-effects model. Begg's test was used to analyze the existence of publication bias and draw funnel plot, and the sources of heterogeneity were explored by sensitivity and subgroup analyses.

Results

According to the inclusion and exclusion criteria, 79 articles were finally included, including 81 studies involving a total of 3242 patients. All the studies had a combined r of 0.44 (95% CI, 0.41-0.46), but with a significant heterogeneity (I² = 80.9%, P<0.01). Subgroup analysis for different tumor types indicated that most subgroups showed a reduced heterogeneity. Malignant melanoma (n = 1) had the minimum correlation coefficient (-0.22) between ¹⁸F-FDG uptake and Ki-67 expression, while the thymic epithelial tumors (TETs; n = 2) showed the maximum correlation coefficient of 0.81. The analytical results confirmed that correlation between ¹⁸F-FDG uptake and Ki-67 expression was extremely significant in TETs, significant in gastrointestinal stromal tumors (GISTs), moderate in patients with lung, breast, bone and soft tissue, pancreatic, oral, thoracic, and uterine and ovarian cancers, average in brain, esophageal and colorectal cancers, and poor in head and neck, thyroid, gastric and malignant melanoma tumors. Subgroup analysis indicated that positron emission tomography (PET) or PET/CT imaging technology or Ki-67 and standardized uptake value (SUV) measurement technology did not significantly affect the results of r values, and Begg's test showed no significant publication bias.

Conclusion

In cancer patients, ¹⁸F-FDG uptake showed a moderate positive correlation with tumor cell proliferation. Different tumor types exhibited varied degree of correlation, and the correlation was significant in TETs and GSTs. However, our results need further validation by clinical trials with a large sample of different tumor types.

Is (18)F-FDG PET useful in predicting the WHO grade of malignancy in thymic epithelial tumors? A meta-analysis

AIM

To perform a systematic review and meta-analysis of published data on the role of fluorine-18-fluorodeoxyglucose positron emission tomography ((18)F-FDG PET) in predicting the WHO grade of malignancy in thymic epithelial tumors (TETs).

METHODS

A comprehensive literature search of studies published up to March 2014 was performed. Data on maximum standardized uptake value (SUVmax) in patients with low-risk thymomas (A, AB, B1), high-risk thymomas (B2, B3) and thymic carcinomas (C) according to the WHO classification were collected when reported by the retrieved articles. The comparison of mean SUVmax between low-risk thymomas, high-risk thymomas and thymic carcinomas was expressed as weighted mean difference (WMD) and a pooled WMD was calculated including 95% confidence interval (95%CI).

RESULTS

Eleven studies were selected for the meta-analysis. The pooled WMD of SUVmax between high-risk and low-risk thymomas was 1.2 (95%CI: 0.4-2.0). The pooled WMD of SUVmax between thymic carcinomas and low-risk thymomas was 4.8 (95%CI: 3.4-6.1). Finally, the pooled WMD of SUVmax between thymic carcinomas and high-risk thymomas was 3.5 (95%CI: 2.7-4.3).

CONCLUSIONS

(18)F-FDG PET may predict the WHO grade of malignancy in TETs. In particular, we demonstrated a statistically significant difference of SUVmax between the different TETs (low-grade thymomas, high-grade thymomas and thymic carcinomas).