Neurological signs and symptoms in 8 children during 55 pontine infusions of combined carboplatin and sodium valproate were systematically monitored using the PONScore. The PONScore increased during pontine infusion reflecting the accumulation of neurological signs and symptoms during infusion. The most common infusion-related side-effects were headache and limb weakness. Headache was mostly mild to moderate, while every patient developed at least a sign or symptoms of neurological dysfunction. 50/124 cases of infusion related side-effects identified using the PONScore persisted for >24 hours. Important predictors of persistent side-effects were recurrence of the side-effect during a previous infusion, infusion through transcerebellar catheters and continuation of infusion at the onset of side-effects, rather than stopping or reducing the rate of infusion.
There are several limitations to this analysis. The PONScore is in its developmental infancy and further studies are required to determine its reliability and validity as an outcome measure and its normative values are unknown. Nursing staff who performed the assessment were unblinded to the infusion volumes and parameters. There is also no understanding of what change in PONScore is clinically significant, and many detected side-effects may not reach a clinically meaningful threshold. The definition of recovery used retrospective chart review, and future prospective studies should consider using the same standardised neurological assessment tool, such as the PONScore, to quantify recovery. Due to differences in the number of infusions received prior to disease progression, infusions studied were not equally divided between all patients; and hence observed findings are skewed toward the patients who received the most infusions. However, despite these limitiations, this is the first attempt to use a standardised neurological assessment tool to describe the neurological changes during brain stem infusion and it makes important observations about the mechanisms of infusion related toxicity.
The description of symptomatology during infusion are corroborated by Souweidane’s et al., 2019 who describe a similar array of cranial neuropathies, limb weaknesses and long tract signs after brainstem infusion. (2) We cannot make comparisons regarding the frequency and severity of side-effects because Souweidene et al,. defined toxicity using CTCAE at a different time-point relative to infusion (7 days after infusion). (2) Szyshot et al., 2021 used clinical assessment to describe toxicity. Our findings are support their conclusion that neurological side-effect mostly resolved within 24 hours. (3) However, by using an increase in PONScore to identify side-effects, it appears a significant minority of infusion-related side-effects persist for longer than 24 hours. It is imperative that infusion-related side-effects are meticulously documented and their mechanisms understood.
Our findings demonstrate that side-effects occur during infusion. Equally it has been previously described how side-effects improve following cessation of infusion. (3) Taken together, it may be that the major source of infusion-related toxicity is due to the mechanical effects of infusion rather than direct pharmacotoxicity. Side-effects were more likely to become persistent if the infusion was continued rather than reduced or stopped at symptom onset. At the point of flow rate adjustment, local drug concentrations within the core of the distribution volume will remain stable up to 12-24 hours after adjustment. (5) The association between flow rate reduction and improved recovery suggests that infusion related side-effects are due to mechanical effects of infusion. Although, the effects of pontine infusion on interstitial pressure in treatment of BS-DMG are lacking, it is reasonable to assume that as infusion rate is reduced, interstitial pressure within the pons will also fall, or stabilise. It may be that infusion leads to increased interstitial pressure, which ultimately impairs locoregional perfusion. This may explain the variability in side-effects between patients. BS-DMG/DIPG are characterised by non-uniform, and sometimes reduced, perfusion, brought about by a hypercellular cytoarchitecture with irregular capillary networks. (6) Catheters in the same locations within the pons may elicit different side effects depending on the unique interactions between the effect of applied infusion pressures on local tumoural perfusion.
We also found that if an infusion related side-effect was observed during a previous infusion it is more likely to become persistent. This suggests infusion-related neurological injury is accumulative, and therefore reducing the risk of infusion-related side-effects is central to the future of the therapy. This theory is confounded by the interaction between numbers of infusions and time from diagnosis. However, considering that the infusion studied were performed prior to diagnosis of clinical and radiological progression, it is likely that infusion is an important contributory factor to the accumulation of neurological disability over time. Side-effects were also more likely to be persistent if acquired during infusion via trans-cerebellar versus trans-cerebral trajectories. Trans-cerebellar infusions were typically conducted on the second day of infusion, therefore, greater risk of side-effects may be due to the accumulative injury from the previous day. It could also be due to the anatomical arrangement of the catheters. Transcerebellar catheters are placed into the anterolateral pons containing corticospinal tracts and crossing ponto-cerebellar fibres, which may be more sensitive to distortion. High resolution tractography and perfusion mapping may help to direct catheters away from brain stem nuclei and tracts and critically ischemic tumour regions.
It is evident that infusion-related side-effects were not directly correlated to infusion volume. It is evident that 30% of the recorded deficits occurred <1mL infusion. Interestingly, fewer side-effects occurring at the end of infusion became persistent. The risk of persistent side-effects was increased if the infusion was continued (rather than stopped or reduced) and there was a non-significant tendency toward greater risk of persistent side-effects if the infusion was continued for ≥3 hours after deficit onset. This would suggest that the way in which the infusion is conducted is a key determinant of treatment-related side-effect profile. Performing the infusions awake, in such a way that patients can be constantly monitored and flow rates adjusted may be important avenues to achieve large volume infusions. Slowing down the ramping regime at the beginning of the infusion may also reduce critical pressures that could lead to infusion-related side-effects.
Regardless of the underlying cause of these side-effects, many of these hypotheses can only be tested in human subjects. As described, the neurological phenotype arising during pontine infusion is likely due to the interaction of catheter location, the morphology of fluid distribution and pathology of the underlying tumour, and the pharmacology of the drug. A reliable means of quantifying neurological change would be hard to achieve in an animal model, even with expert veterinary supervision. Without standardised neurological examination, addressing such questions will be impossible in human studies. In summary, we present findings from 55 pontine infusions in 8 children with BS-DMG/DIPG. Analysis of neurological signs and symptoms occurring during pontine infusion using the PONScore demonstrated that risk of persistent infusion-related side-effects were higher if the infusion was not immediately reduced and if the side-effect had occurred during prior infusions. If catheter flow was adjusted based on anatomical position some deficits would recover. This suggests that the neurological injury arising from pontine infusion is often localised, time dependent, accumulative rather than volume dependent, and ultimately modifiable. The study of interstitial pressure and perfusion and their interaction with neurological function may be central to development of the treatment.