Tumor Volume Growth as Surrogate Endpoint in IDH-mt Glioma—Letter

Given the very long survivals of patients treated for an IDH-mutated glioma, monitoring treatment efficacy is challenging. While overall survival remains the major endpoint, there is a need of surrogate endpoint as a guide for decision-making in daily practice. Bhatia and colleagues recently investigated the tumor volume postoperative growth rates as a biomarker of prognosis (1). Assessing tumor kinetics on longitudinal MRI follow-up was already known to be a very helpful information in the preoperative period (2), and it is all the merit of the authors to have replicated these results in the postoperative period. However, the choice of modeling the tumor volume as following an exponential law is highly disputable. Biomathematical modeling of glioma growth predicts that the edge of the tumor should increase linearly with time, i.e., D ∝ t, D being the tumor diameter (3). Consequently, we would expect V ∝ t3. So it would have been more appropriate to fit the tumor volume growth curve neither with an exponential, nor a linear, but a 3rd order polynomial law. Of note, it is much easier to convert the exact volume V in an equivalent diameter, through the formula D = (2V)1/3. Tumor growth is then expressed in mm/year, which makes it much more intuitive for both clinicians and patients. It would be highly interesting if authors could reanalyze their data by assessing kinetics with a linear fit of the equivalent diameter and to see if some of their conclusions would still hold true. In particular, we wonder whether the absence of difference in tumor volume growth rates according to 1p19q status would be confirmed, as previous studies relying on equivalent diameter reported that 1p19q non-codeleted tumors exhibited preoperatively a faster growth than 1p19q codeleted tumors (4). Finally, monitoring the growth curve of the diameter is a very sensitive tool for assessing drug effects, as any deviation from the linear extrapolation of the pretreatment curve indicates response or failure to treatment (diameter respectively below or above its predicted value from the linear fit). The change of the slope (in mm/year) can even quantify the intensity of the response. In summary, while the work of Bhatia and colleagues is an important step towards integrating tumor kinetics in the management of patients with IDH-mutated glioma, as previously suggested by others (5), using linear fit of equivalent diameter would have rendered its implementation much more attractive.See the Response, p. 639No disclosures were reported.