In Reply: A Data-Driven Approach to Predicting 5-Aminolevulinic Acid-Induced Fluorescence and World Health Organization Grade in Newly Diagnosed Diffuse Gliomas

To the Editor: We greatly appreciate the authors' letter1 regarding our publication “A Data-Driven Approach to Predicting 5-Aminolevulinic Acid-Induced Fluorescence and World Health Organization Grade in Newly Diagnosed Diffuse Gliomas.” Contrast enhancement is a relevant predictor of intraoperative 5-aminolevulinic acid (5-ALA) fluorescence in our logistic regression–based mode. Our observations can be explained by the fact that the brain-blood barrier in diffuse lower-grade gliomas (DLGG) is semipermeable to 5-ALA.2 Our group has shown that higher 5-ALA doses result in more protoporphyrin IX conversion, an observation with implications for future dosing in DLGG.3 Having given our last analysis further thought under present circumstances, however, the question we were asking could perfectly be addressed using the machine learning pattern recognition, allowing to potentially incorporate multidimensional factors beyond contrast enhancement for predicting intraoperative 5-ALA fluorescence in a more holistic way than our older analyses. Our group is currently working on several such projects to predict intraoperative 5-ALA fluorescence in DLGG. Solid prediction modeling may translate into lower costs by identifying those patients in whom application of 5-ALA will not lead to intraoperative fluorescence as observed by the surgeon. At the same time, however, reporting guidelines need to be respected to maintain the high quality of research and higher yields of prediction.4 As stated by Barrie and Detchou, it would definitely be desirable to corroborate our findings in a prospective study setting. We are in the active stage of conceptualizing a sampling study on tumor heterogeneity as depicted by 5-ALA fluorescence and amino acid PET using machine learning algorithms, in diffuse lower-grade gliomas. Finally, we again thank the authors for giving us the chance of raising additional points on prediction of 5-ALA fluorescence in DLGG.