NSCLC management is a primary concern, therefore prognostic assessment is critical to realizing the potential of personalized treatment. A wide range of information sources are available of which the noninvasive type plays a fundamental role in reducing the patients' burden (Lambin et al. 2015). As a noninvasive imaging modality, the 18F-FDG PET/CT, which reflects tumor metabolic activity, such as cell viability and proliferative activity, is becoming more and more significant in studies of initial diagnosis, accurate staging, evaluation of treatment response, and prognosis of lung cancer (Li et al. 2019).
In this prospective study, we hypothesized that the metabolic parameters derived from 18F-FDG PET/CT may provide more accurate quantitative imaging features and thus might be of value in predicting OS in patients with NSCLC.
In PET/CT imaging, the SUV is the most widely used semi-quantitative measure. The individuals' physical characteristics, the blood glucose level, the uptake period, the image acquisition and reconstruction, the lesion size, and other variables all have an impact on SUV (Lee et al. 2012).
SUVmax, SUVmean, and SUVpeak, represent the maximum, average, and peak SUV of the region of interest, respectively. All three of these parameters were measured and examined in our study in order to determine which was the most valuable. We found that higher levels of SUVpeak and SUVmean were associated with shorter OS which was not true for SUVmax. Our findings further support the hypothesis that SUVmax could not be a reliable predictor of OS, emphasizing the significance of volumetric parameters. Given that SUVmax simply represents a single-pixel value of peak metabolic activity, hence, has less impact on patient’s outcome, Huang and colleagues suggested that SUVmax may not be the most reliable predictor (Huang et al. 2014).
Increasing evidence from studies on different tumor types shows that MTV and TLG are more accurate prognostic indicators than do SUV metrics (Hyun et al. 2014; Liao et al. 2012).
Several automated techniques are used to segment regions of interest in PET/CT scans, including gradient-based threshold (adaptive iterative algorithm, AIA), fixed SUV threshold (e.g., SUV > 2.5), percentage threshold of SUVmax (e.g., > 42%), and background-related threshold (AT 40%) approaches. The fixed threshold method was not used in our study since it ignores the background activity as reported in earlier studies (Soret et al. 2007). Percentage threshold can be performed rapidly and consistently, with less inter-observer variability (Wang et al. 2017). In our study we used percentage threshold method for delineation of MTV. Our results reported a significantly worse OS and higher risk of death in patients with higher tu MTV and tu TLG. These results are in line with MTV's established predictive value in previous studies (Popinat et al. 2019; Seban et al. 2020; Jreige et al. 2019).
The definition of "high tumor burden" is still up for debate. Its definition and use as a prognostic factor in the context of lung cancer could assist in the development of specialized treatment regimens that would not only enhance patient quality of life and treatment outcomes but also lower the occurrence of negative adverse effects (Reck and Rabe 2017).
Although earlier studies have employed the MTV methodology (with a threshold value of 41% of SUVmax in each lesion), more advanced methods might yield more precise measurement of WB MTV values. Chardin et al. reported that WB MTV, as calculated by threshold value method, offers a reliable and reproducible estimate of whole-body tumor burden (Chardin et al. 2020).
The optimal cutoff value of WB MTV in our study was 141.98 ml. Due to the lack of current standard 18F-FDG PET/CT parameters cutoff definitions, external validation of the results may be challenging. The WB MTV cutoff values published in research to date have varied, while utilizing comparable methods to compute MTV. These values range from 17.8 to 143.2 ml (Hashimoto et al. 2020; Dall'Olio et al. 2021; Monaco et al. 2021; Yamaguchi et al. 2020). Pu et al. validated WB MTV through analysis of a quite large sample of NSCLC patients of all clinical stages reporting outcomes comparable to ours (Pu et al. 2018). Consequently, Pellegrino and colleagues additionally found that WB MTV was a reliable predictor of OS at all TNM stages (Pellegrino et al. 2019).
In addition to WB MTV, we also assessed the other volumetric value WB TLG. Supposedly, TLG could be more promising since it combines volumetric and metabolic information. Our results found that WB TLG is an independent predictor of OS. Similar results were reported by a previous study by Vanhove whose results showed that only WB TLG is of prognostic value in NSCLC for both OS and progression free survival (PFS) (Vanhove et al. 2018).
In a recent study by Oliveira et al. 52 NSCLC patients who performed 18F-FDG PET/CT for staging in who were followed for a median of 11.0 months, they found that The WB TLG/tu TLG ratio is an independent prognostic indicator of OS in advanced-stage NSCLC (Oliveira et al. 2022).
A multi-institutional investigation confirms that SCC histology is independently correlated with lower overall survival and local, regional, and distant recurrence (Baine et al. 2018). This comes in agreement with our results that showed a significantly shorter OS survival of SCC. Future research is required to evaluate whether early-stage NSCLC therapy paradigms should vary based on histology.
The Cox multivariate analysis revealed that the degree of histological differentiation had no statistically significant effect on OS. Possible explanation of this result is that OS is frequently impacted by numerous other factors, such as type of therapy, tumor recurrence or metastasis. Additionally, the histological differentiation of the tumor was not determined in this study for one-third of the patients. Therefore, we were only working with 20 cases whose degree of difference was known. The statistical analysis result may be biased by a quite small sample size.
In conclusion, we have demonstrated that, in patients with NSCLC, the whole-body metabolic tumor burden determined on a baseline 18F-FDG PET/CT performed for staging seems to be a strong, independent imaging biomarker to predict OS, superior to the clinical assessment of the primary tumor itself. The most accurate predictor of OS in our patients was the WB TLG. Efforts should be made to unify its definition and to further explore its potential as a prognostic factor for patients with metastatic NSCLC.
Our study was limited by the small number of study group and short follow-up period. This may be attributed to COVID-19 pandemic. According to a survey conducted online by the International Atomic Energy Agency on a variety of nuclear medicine services and completed by 434 doctors from 72 different countries, nuclear medicine services have dramatically declined, by over 50% for diagnostic tests and as much as 45% for radionuclide therapy (Freudenberg et al. 2020).
Other limitations include the heterogeneity of the study population considering the various NSCLC subtypes, various metastatic statuses, and different previous therapies.
Our benefits were a single-center research design, patient preparation, waiting period, and FDG doses at the stage preceding FDG PET/CT were uniform, the same FDG PET /CT scanner for all patients.