We read the article titled “The Role of Ketone Bodies in Treatment Individualization of Glioblastoma Patients” by Tamas et al. with great interest. However, there are several key points that were not discussed and unfortunately also missed during the peer-review.

While the observation that ketone bodies may increase during aggressive disease and correlate with decreased survival are valid in this specific context, the proposed mechanism (ketone bodies fueling tumor growth) may not be complete.

First, the GKI (Glucose-Ketone Index) was developed for patients following ketogenic diets, which were excluded from this study. Second, while the authors show the baseline weight of the two groups (ASTRO G4, GBM), to our knowledge, they did not track the evolution of weight after diagnosis (or did not share this data). A clearer understanding of human physiology may provide insights into the “origin” of this increased ketogenesis (pathological caloric restriction, i.e., unintentional weight loss, presumably due to more aggressive disease, or more aggressive treatment). It is therefore plausible that a terminal cancer diagnosis is indeed associated with increased ketone levels, but a causal relationship is unlikely.

Third, ketogenic diets, which induce ketone levels well beyond the ones observed in this study, would be expected to accelerate tumor growth in GBM, but most human data shows either neutral or beneficial effects (PMID: 31927631). In fact, most therapeutic benefit is observed when caloric restriction is combined with the ketogenic diet, thus ketones “feeding” tumor cells should be accelerated. This does not seem to occur in either mouse or human studies (PMID: 37245566). It is important to mention that both groups had average glucose levels in a range that would be considered diabetic (> 126 mg/dl). Unfortunately, glucose and ketone levels were not tracked over time; these parameters would be expected to display high variability depending on the caloric intake of the patients.

The authors also cite Sperry et al. (2020), but in this study, neither ketone bodies nor fatty acids were able to sustain long-term cell proliferation under glucose deprivation (in fact, the proliferation of glucose deprived cells was always lower than high glucose, regardless of fat-derived fuels). Therefore, it would be necessary to control for real-time glucose levels in these GBM patients and evaluate the contribution of varying concentrations of ketone bodies to tumor growth, compared to lower glucose levels, over time. Fifth, the “surprising” absence of > 0.5 mM levels in the ASTRO G4 group is likely related to a better prognosis and lack of cachexic weight loss and ketogenesis, rather than ketogenesis accelerating tumor growth.

We appreciate the fact that higher ketone bodies in terminal cancer patients may be used as a biomarker of poor prognosis. Clinicians may also look for declining health status, loss of daily activities or any other marker of advanced disease. However, it is hard to agree with the proposed interpretation of the data without providing further evidence. We would welcome supplementary data showing the evolution of weight in this cohort, as well as daily tracking of glucose and ketone levels, which would be illuminating to understand the relationship between glycemia, ketonemia and tumor growth.

Tomás Duraj, MD, PhD
Biology Department
Boston College