F-18 FDG PET/CT Characteristics of a Lung Lesion and Concomitant Findings as Histopathology Predictors: A Single-Center Experience

INTRODUCTION

Lung cancer is the second most frequently diagnosed cancer and the number one cause of cancer-related deaths, with an estimated 2,480,675 new cases of lung cancer and 1,817,469 deaths worldwide in 2022.^[1] Adenocarcinoma and squamous cell carcinoma are the most frequent histopathological findings in lung cancer.^[2] However, radiological techniques that are currently in use for the diagnosis of lung cancer (computed tomography [CT]) do not have the ability to distinguish malignant lesions from benign ones. Although biopsy is a gold standard, the experience of a bronchoscopist and localization of lesions play a crucial role for an accurate diagnosis of the disease.^[3] Therefore, the early diagnosis of lung cancer remains a challenge. Several meta-analyses addressed the utility of radiological and nuclear medicine methods in the differentiation between benign and malignant lung lesions. Reported pooled sensitivity and specificity for CT were 89% and 70%^,[4] while for fluorine-18 fluorodeoxyglucose (F-18 FDG) positron-emission tomography/CT (PET/CT) were 82%–89% and 70%–81%,^[5,6] respectively. There were also publications that claim that there is no significant difference between those two methods.^[7]

Furthermore, studies evaluated the relationship between the etiology of a lung lesion on one side and a maximum standardized uptake value (SUVmax) and CT characteristics (size and other morphological features) on the other.^[8-12] The aim of this work was to investigate the predictive potential of F-18 FDG PET/CT features, lesion location, and concomitant findings in a diagnosis of malignancy in patients with an incidental lung lesion. We have also evaluated the difference in glucose metabolism quantified by SUVmax between different types of lung cancer.

MATERIAL AND METHODS

RESULTS

A total of 88 patients were included in the study (age range 32– 86 years), 46 females and 42 males. Demographic and clinical characteristics of the study population are presented in Table 1.

On F-18 FDG PET/CT, lesions in the lungs were reported as malignant in 65, benign in 7, and indeterminate in 16 cases. Most lesions were 10 mm or above in size (82/88). Progression in size of 4 mm or above on at least two consecutive CT exams before F-18 FDG PET/CT was detected in 26 patients. The most common localization of both malignant and benign lesions was the right upper lobe (22/65 and 12/23, respectively). The most common histopathological result was adenocarcinoma [Figure 1 and Table 2], most commonly located in the right upper lobe (17/47) and right lower lobe (11/47), with the frequency of 47/88, 29 of which were 20 mm or larger in the longest diameter [Table 2]. The most frequent benign lesions were inflammatory (15/23), most commonly located in the right upper lobe, with the mean SUVmax of 2.63 ± 0.63 [Table 2]. The most frequent localization of extrapulmonary focal zones of an increased glucose metabolism were mediastinal and hilar lymph nodes (28/88), followed by bones (12/88) [Table 2].

Close

Figure 1:

Original image. Fluorine-18 fluorodeoxyglucose positron -emission tomography with computed tomography (CT) of a 66-year-old male patient with a positive lung lesion, mediastinal and hilar lymph nodes: (a) Parahilar lung nodule (white arrow) in the right lung and hilar lymph node in the 10R group (yellow arrow) on the low-dose CT (transverse view); (b and c) increased glucose metabolism in the lung nodule (maximum standardized uptake value 20.4) on the transverse and sagittal view (black arrow) and the hilar lymph node (yellow arrow) on the transverse view (fused image); (d) hypermetabolic right paratracheal lymph node on the fused image (white arrow); histopathology revealed adenocarcinoma

Export to PPT

The Mann–Whitney U-test for age and Pearson’s Chi-squared test for smoking showed a significant difference in medians and proportions (P = 0.008 and P < 0.001, respectively) of malignant versus benign lesions. The logistic regression model was statistically significant for age and smoking history (P = 0.014 and P < 0.001, respectively). The model explained 64.9% (Nagelkerke R2) of the variance in histopathology results. Current and former smokers were 69.01 times more likely to have positive findings on histopathology than nonsmokers [Table 3].

The median SUVmax was 1.9 (range 1.0–11.7) for histopathologically confirmed benign lesions and 8.8 (range 1.0–54.5) for malignant lesions, with a statistically significant difference between medians (P < 0.001, Mann–Whitney U test). There is also a significant difference between the median anteroposterior diameter of benign and malignant lesions (P = 0.002, Mann–Whitney U-test) [Table 4]. Lesions of 20 mm and larger were significantly associated with malignant etiology (P = 0.002, Mann–Whitney U-test). The median SUVmax for lesions ≥20 mm in size was 10.2 (range 1.5–54.5) and significantly differed (P < 0.001, Mann–Whitney U-test) from that of lesions smaller than 20 mm in the longest diameter (median 3.3, range 1–16.6) [Table 4]. The median SUVmax for adenocarcinoma was 10.3 (range 1–54.5) [Figure 1], for squamous-cell carcinoma 16.3 (range 6.3–26.3) [Figure 2], and for other lesions of a malignant etiology 9.8 (range 2.6– 21.3) [Figures 3 and 4], with a statistically significant difference between medians of all the groups of malignancies on one side and benign lesions [Figure 5] on the other (P < 0.05, Kruskal– Wallis test), but with no significant difference between median SUVmax of adenocarcinoma, squamous-cell carcinoma, and other malignancies (P > 0.05, Kruskal–Wallis test). Furthermore, no statistically significant difference between the mean SUVmax according to the location of a lesion was observed (P = 0.974, t-test for independent samples). Similarly, the differences between frequencies of benign and malignant histopathology findings according to the location of a lesion were not statistically significant (P = 0.565, Fisher’s exact test).

Close

Figure 2:

Original image. Fluorine-18 fluorodeoxyglucose positron -emission tomography with computed tomography of a 70-year-old male patient with a positive lung lesion, mediastinal and hilar lymph nodes: (a) lung lesion in the anteromediobasal segment of the left lower lobe 19x20x24mm in size on the low-dose CT, (b) with an increased glucose metabolism on the fused image and the SUVmax of 10.4 (arrow); (c) increased glucose metabolism in the precarineal lymph node (arrow) and (d) hilar lymph nodes bilaterally (arrows); histopathology revealed squamous cell carcinoma

Export to PPT

Close

Figure 3:

Original image. Fluorine-18 fluorodeoxyglucose positron -emission tomography with computed tomography of patients with malignant histopathology findings: (a) Nodular lung lesion in the right lower lobe 12 mm × 14 mm in size with the maximum standardized uptake value (SUVmax) of 6.4 in a 60-year-old female patient (arrow); histopathology revealed neuroendocrine carcinoma; (b) lung lesion in the lingula of the left lung 16 mm × 21 mm in size with the SUVmax of 13.4 in a 61-year-old male patient (arrow); histopathology revealed pleomorphic carcinoma; (c) pleural soft -tissue lesion 40 mm × 15 mm in size with the SUVmax od 4.5 in a 78-year-old male patient (arrow); histopathology revealed pleural mesothelioma; (d) excavated lung lesion in the right upper lobe 31 mm × 21 mm × 17 mm in size with the SUVmax of 6.6 in a 64-year-old male patient (arrow); histopathology revealed sarcoma

Export to PPT

Close

Figure 4:

Original image. Fluorine-18 fluorodeoxyglucose positron -emission tomography with computed tomography of a 66-year-old female patient with a positive lung lesion and mediastinal lymph node: (a-c) A lesion 19 mm × 16 mm in size (a) with the maximum standardized uptake value of 16.6 (b) in the right upper lobe on the transverse (a and b) and sagittal view (c) (arrow); (d) hypermetabolic aortopulmonal lymph node on the fused image (arrow); histopathology revealed small cell lung cancer

Export to PPT

Close

Figure 5:

Original image. Fluorine-18 fluorodeoxyglucose positron -emission tomography with computed tomography of patients with benign histopathology findings: (a) A lung consolidation in the right upper lobe 26 mm × 25 mm in size with the maximum standardized uptake value (SUVmax) of 1.7 in a 73-year-old female patient (arrow); histopathology revealed inflammation; (b) a nodular lung lesion in the right upper lobe 12 mm × 14 mm in size with the SUVmax of 5.4 in a 73-year-old female patient (arrow); histopathology revealed fibrosis; (c) a lung consolidation with a dilated segmental bronchus 42 mm × 30 mm × 32 mm in size with the SUVmax of 11.7 in a 72-year-old female patient (arrow); histopathology revealed bronchiectasis; (d) a nodular lesion 11 mm × 10 mm × 10 mm in size with the SUVmax of 2.2 in a 49-year-old male patient (arrow); histopathology revealed hamartoma

Export to PPT

Patients with extrapulmonary focal zones of an increased glucose metabolism on PET/CT were 4.33 times more likely to have positive findings on histopathology (P = 0.011, univariate logistic regression) [Table 4]. In particular, patients who had an increased uptake of F-18 FDG in mediastinal and hilar lymph nodes were 7 times more likely to be positive for a malignancy (P = 0.013, univariate logistic regression) [Table 4]. Increasing SUVmax and size were statistically significant predictors of positive histopathological results, with the sensitivity of 83.1% and the specificity of 82.6% for the SUVmax cut-off value of 3.8 (area under the curve 0.874; 95% confidence interval, 0.788– 0.961; receiver operating characteristic curve analysis for the quantitative test) [Figure 6], and with lesions 20 mm and larger being 4.79 times more likely to be malignant than smaller ones. Progression in size on CT exams leading to F-18 FDG PET/ CT of at least 4 mm was a statistically significant predictor of a malignant histopathological finding (P = 0.015, univariate logistic regression) [Table 4]. However, in multivariable logistic regression analysis, only the size of a lung lesion and the presence of extrapulmonary focal zones of an increased F-18 FDG uptake were independent predictors of histopathology findings among imaging features, with lesions smaller than 20 mm being less likely to be malignant [Table 5].

Close

Figure 6:

Original image. Graph showing sensitivity and specificity for different maximum standardized uptake value cut-off values

Export to PPT

A biopsy served as a final confirmation of a diagnosis. Sensitivity, specificity, PPV, and NPV of F-18 FDG PET/CT in our study were 89.2%, 69.6%, 89.2%, and 69.6%, respectively [Table 6]. No adverse events from performing F-18 FDG PET/ CT examinations or biopsies were reported.

DISCUSSION

In retrospective analysis, we revealed the predictive potential of certain imaging, clinical, and demographic parameters on histopathological results^[13]. Sizes of 20 mm and more and higher SUVmax were significantly correlated with malignant findings. Hypermetabolic mediastinal and hilar lymph nodes and a progression in size of at least 4 mm in at least two consecutive CT exams before F-18 FDG PET/CT also increased the likelihood of a lesion being malignant. Increasing age and a history of smoking, among demographic and clinical data, were significant predictors of malignant histopathological findings.

FDG PET/CT detects malignancy based on the high F-18 FDG uptake, which reflects the increased glucose metabolic activity of viable cancer cells.^[14] According to some studies, SUVmax is useful not only in distinguishing between benign and malignant lung lesions but also in predicting histological grade^[10] and survival,^[15] even potentially a histological type of malignancy.^[16] The most recent studies question the capability of SUVmax to differentiate between benign and malignant lesions,^[17] implying the predictive nature of SUVmax might be due to its correlation with the size of a lesion, with a significantly higher SUVmax for lesions ≥3 cm than for those <3 cm.^[9] Our results showed that higher SUVmax is a possible predictor of a malignancy,^[18,19] with the sensitivity of 83.1% and the specificity of 82.6% for the SUVmax cut-off value of 3.8. Malignant etiology is closely correlated with lesion size, with more than one clinically relevant threshold for a differentiation between benign and malignant lesions.^[20] We found that larger lesions are more likely to be malignant, especially those ≥20 mm in size, which was an independent predictor of a malignant etiology. This data is consistent with other studies.^[11] The progression of a lung lesion on CT in the context of F-18 FDG PET/CT evaluation has not been thoroughly studied. We have proved that an increase in size as small as 4 mm significantly increases the likelihood of a malignancy. In addition, the presence of hypermetabolic mediastinal and hilar lymph nodes was the predictor of malignant histopathological results, findings similar to those of Ugurlu et al., who found the sensitivity of 98% for the mediastinal and hilar lymph node SUVmax cutoff value of 2.54 in the differentiation between benign and malignant etiology.^[20] In our study, all the patients with focal zones of an increased glucose metabolism in cervical lymph nodes (6/6), adrenal glands (3/3), and liver (3/3) and almost all the patients who had positive PET/CT findings in bones (11/12) also had positive histopathology results for a malignancy aft er the biopsy of a lung lesion. However, the predictive value of these factors could not be demonstrated, possibly due to the small sample size.

Increasing age and smoking were predictors of malignant histopathology findings of lung lesions in our study and many others,^[12,14,21] while gender has not been a significant contributor to the probability of malignant findings.^[12,14] It’s worth mentioning that in our study all of the patients who were either current or former smokers had positive histopathological findings. We found that a history of cancer is not a statistically significant predictor of a positive histopathology of a lung lesion, which is in accordance with some studies^[12] and in discrepancy with others,^[22] the latter one suggesting a higher proportion of patients having a history of malignancy in the malignant group than in the benign group (18% vs. 4%). Although meta-analyses indicate the linkage between lung cancer and COPD,^[23,24] the latter one suggesting an increased risk of lung cancer for COPD and emphysema, but not for chronic bronchitis, possibly the consequence of shared risks and mechanistic factors,^[25] our study was not able to prove this assumption, possibly due to the small sample size. Also, the association between the location of a lesion and histopathology results could not be confirmed in our study either.

The study by Yıldız and Çelikİ revealed a statistically significant difference between SUVmax values of adenocarcinoma and other types of lung cancer,^[26] while in the study by Çalışkan et al., the SUVmax value of squamous cell carcinoma was higher than the SUVmax value of adenocarcinoma and small cell lung carcinoma.^[16] In our study, no difference in glucose metabolism quantified by SUVmax between adenocarcinoma, squamous cell carcinoma, and other types of lung cancer was observed. Furthermore, the difference between mean SUVmax according to the location of a lesion was not statistically significant.

In our study, the sensitivity of FDG PET/CT in the detection of malignant lung lesions was identical with two meta-analyses (89%)^[5,7] and slightly differed from another (82%).^[6] Similarly, the specificity was the same as in one of the above-mentioned meta-analyses,^[5] while 11%^[6] and 8%^[7] lower than the pooled specificity in the other two. The observed difference in specificity compared to other studies might be due to the selection bias caused by the retrospective nature of the study, which led to many benign lesions not being biopsied due to low suspicion and a lower true negative count. It should also be mentioned that in our study indeterminate PET/CT findings were considered negative for a malignancy. If indeterminate findings were considered positive, this would have caused a higher false-positive count and lower positive predictive value and specificity.

The main limitation of the study is a small number of patients in whom focal zones of an increased F-18 FDG uptake in extrapulmonary sites were detected, excluding the mediastinum and the lung hila, which is most likely the reason the predictive value of extrapulmonary focal zones of an increased glucose metabolism in certain locations for malignant histopathological findings of a lung lesion could not be statistically proven. Therefore, this finding can be seen as exploratory and needs validation in the larger study. The other limitation is a selection bias introduced by the retrospective nature of the study. Since only patients who underwent biopsy post-PET/CT were included, many benign lesions not biopsied due to low suspicion were excluded. This might have led to a higher sensitivity and PPV and lower specificity. Furthermore, the retrospective nature of the study led to medical history not being complete in a few cases. Considering all of the above, a wider study with more patients and possibly more F-18 FDG PET/CT parameters involved in the analysis should be considered. Furthermore, due to the expansion of artificial intelligence, machine learning models that will be able to perform a classification of lung lesions based on a quantitative features with a high accuracy are being developed.^[27] These models will further enhance the diagnostic performance of F-18 FDG PET/CT in patients with a lung lesion of an unknown etiology. Finally, F-18 FDG PET/CT is an important tool in the follow-up of patients with malignant histopathology findings, with PET/CT quantitative parameters being reliable indicators of survival.^[28]

CONCLUSION

Increasing size and SUVmax, progression of at least 4 mm on CT, and hypermetabolic mediastinal and hilar lymph nodes are the predictors of malignancy among PET/CT findings, while only increasing age and smoking increase the probability of a malignancy among clinical and demographic features. The predictive value of extrapulmonary focal zones of an increased F-18 FDG uptake outside mediastinal and hilar lymph nodes could not be statistically proven due to the small sample size. Therefore, further investigation should be considered in this direction. No difference in glucose metabolism quantified by SUVmax between adenocarcinoma, squamous cell carcinoma, and other types of lung cancer was observed.