Prognostic Evaluation of Tumor Volume and Its Changes in Radical Radiotherapy of Advanced NSCLC

The prognostic relevance of the 'gross tumor volume' (GTV) in radiotherapy of advanced non-small-cell lung cancer (NSCLC) in stage III is adressed in a limited number of studies in the literature. The review article by Dubben et al., that comprises data until 1998, highlights the GTV as an important indicator and influencing factor for the therapeutic success after radiotherapy, albeit not being dominant over the T-stage (Dubben et al. 1988). In general, an increase in tumor volume correlates with a higher T-stage (Martel et al. 1997), but no congruence can neccessarily be assumed between the tumor volume and the T-determinator. Since the TNM-classification is primarily surgical however, it also does not provide sufficient information for prognosis when surgical therapy is not the first choice.

Available evidence suggests that the GTV in particular at the beginning of therapy acts as a statistically significant prognostic indicator regarding overall survival and / or local tumor control (Martel et al. 1997; Bradley et al. 2002; Basaki et al. 2006; Etiz et al. 2002; Werner-Wasik et al. 2001; Wer-ner-Wasik et al. 2008; Stinchcombe et al. 2006; Dehing-Oberije et al. 2008; Willner et al. 2002; Ball et al. 2013). A direct comparison between different studies is, however, often hampered due to the large variation of measurement time points during therapy, as well as the employed definition of the tumor volume. For example, all studies include patients whose GTV was determined after (neoadjuvant) chemotherapy. In addition, three studies even combine the tumor volume of the primary tumor with affected lymph nodes (Etiz et al. 2002; Werner-Wasik et al. 2008; Dehing-Oberije et al. 2008). Furthermore, no agreements can be found in the literature concerning volume changes during therapy. Nonetheless, all studies report a volume reduction at the end of therapy, albeit not always significant. In a study containing 10 patients treated with helical Tomotherapy, the authors observed a relative median tumor reduction during therapy of 1.2% per day (0.6-2.3%) (Kupelian et al. 2005).

The response of NSCLC to radiotherapy with or without chemotherapy is slow (Woodford et al. 2007) with tumors reaching their maximum response or minimal volume after 5-11 months after exposure (Werner-Wasik et al. 2001). If the tumor volume is determined too early, i.e. directly after the end of therapy, the results can lead to misinterpretation resulting in an overestimation of the tumor volume or correspondingly an underestimation of the therapeutic response (Siker et al. 2006). According to Bell et al., the predictive value of tumor volume changes in the first 18 months after radiotherapy is of particular importance. During this time, a significantly increased mortality was observed for larger tumor volumes.

Incorporation of a PET/CT in the context of the radiaton plan is advantageous with respect to the precise traget-volume definition and sparing of risk organs (Ruysscher et al. 2005; Nestle et al. 2006; Lavrenkov et al. 2005; van Baardwijk et al. 2007; Edet-Sanson et al. 2012; Ruysscher und Kirsch 2010; As-hamalla et al. 2005; Bradley et al. 2004; van Baardwijk et al. 2006; Vanuytsel et al. 2000). The superiority of PET compared to stand-alone CT was also shown in two meta-analysis (Gould et al. 2001; Gould et al. 2003). The importance of the 'standardized uptake value' (SUV) or the metabolic tumor volume (MTV) as well as the change in these parameters during radiotherapy has been repeatedly demonstrated (Berghmans et al. 2008, Gillham et al. 2008; Zhang et al. 2011; van Elmpt et al. 2012; Edet-Sanson et al. 2012; van Baardwijk et al. 2007; Vera et al. 2014; Vanuytsel et al. 2000; Feifei Na et al. 2014; Lopez Guerra et al. 2012; Lee et al. 2007; Lee et al. 2012; Huang et al. 2011; Xiang et al. 2012). These studies show partly a statistically significant correlation between tumorale FDG-accumulation before, during or after radiotherapy, or the decreasing accumulation during radiotherapy, respectively, and the overall survival. The results, however, suffer from a large uncertainty regarding the distinct influence corresponding to the SUV. Other studies report a significantly weaker association of the SUV and survival (Hoang et al. 2008; IKUSHIMA et al. 2010; Lopez Guerra et al. 2012). Due to the dynamic variations in the SUV and MTV during radiotherapy, a change in the prognostic validity during radiotherapy can be assumed. According to van Elmpt and others, the FDG uptake during the second (van Elmpt et al. 2012; Zhang et al. 2011) or fifth week of exposure is crucial for survival (Edet-Sanson et al. 2012). Work by van Baardwijk et al. shows an increase in the SUV in some patients during the first week of therapy, which is explained by radiation-triggered inflammation and tumor-biological changes due to radiotherapy (van Baardwijk et al. 2007). The results demonstrate that the appearance of tumor necrosis during radiotherapy or changes in the metabolic tumor situation or oxygenation affect the SUV parameter crucially (Hoang et al. 2008, Huang et al. 2014; Huang et al. 2011). In this context, tumorhypoxia and the corresponding effects on the metabolism of glucose are of particularly importance: A hypoxia-simulated upregulation of the membranic glucose transporter with consecutive increase of cellular FDG uptage can lead to a false SUV value, calling for a combination of SUV or MTV with other prognostic parameters as well as hypoxia-specific imaging (FMISO-PET) (Ikushima et al. 2010, Berghmans et al. 2008). Consequently, the optimal timevpoint for carrying out a PET during / after radiotherapy is not well defined, especially when the protracted tumor response after completion of radiotherapy is taken into account, leaving the integration of additional PET measurements during radiotherapy exclusively to clinical studies.

In conclusion, evidence from available literature regarding the prognostic and predictive value of tumor volume before and particularly its changes during radiotherapy of locally advanced NSCLC is conflicting and inconclusive. Currently available studies often include only a small number of patients with partly overlapping patient cohorts. Current data is additionally limited due to the highly heterogeneous GTV detection time points as well as the definition and detection methodology of tumor volumes.

Based on the observation that a significant tumor volume reduction occurs during radiotherapy, a reevaluation of the tumor volume during radiotherapy could allow an adaptation of the target volumes with dose escalating in the tumor area, while at the same time, improving the protection of organs at risk.

The prognostic or predictive significance of absolute tumor volumes or their change under radiotherapy is to be evaluated multicentrically and its integration into already existing prognostic models is to be multicentrically validated.