Discussion:
In our review of 105 pediatric trials, there was substantial variability in recommended dose-volume constraints among all OARs. While heterogeneity is present in adult clinical trials, many protocols refer to pre-existing OAR guidelines from QUANTEC, the American Association of Physicists in Medicine Task Group (AAPM-TG) 101, and Hypofractionated Treatment Effects in the Clinic (HyTEC) 13,20,21. A comparable pediatric consensus guideline was not previously available, although it is well known that late effects in normal tissues vary across the age spectrum and can lead to devastating consequences6. It is reassuring that the current PENTEC guidelines are now being developed and will help promote more consistency among recommended dose constraints across protocols.
To a certain extent, the degree of variability can be justified by different treatment goals for various cancer histopathologies, target volumes, sex, age, and the use of concomitant treatments. That is, the accepted normal tissue risk tolerance for some diagnoses might be greater if their curability is less likely. Nevertheless, consistency in constraints was poor even for the same diagnosis, chance for survival, or similar exposures to chemotherapy. For serial structures such as the spinal cord, optic chiasm, and optic nerves, reduced variability between dose constraints would be expected for a Dmax constraint compared to volumetric constraints. However, this was not seen with 20, 16, and 14 unique constraints across the trials for these three structures, respectively. When comparing pediatric US constraints to European constraints, constraint tolerances were higher for parallel organs including the bladder, heart, lungs, liver, and kidneys. For serial structures, dose tolerances were mixed with the COG protocols allowing for a higher dose for the spinal cord, but a lower dose for the brainstem.
One might expect dose constraints to change over time, consistent with new normal organ dose-response data becoming available. However, none of the OARs had a systematic pattern of change in protocol dose constraint values over time to indicate an increase or decrease in the tolerated dose. Rather, the variations were either in choice of the dose-volume pairing or non-systematic changes, neither of which indicated the influence of new information but are more likely due to a lack of both good dose-response data and consensus by protocol committee members. Additionally, review of the currently active protocols also show constraints were consistent with previous trials with no consistent pattern of change. This finding further highlights the need for more evidence-based, consistent constraints across protocols.
It is provocative to speculate on the reasons for the observed heterogeneity in constraints across protocols or even continents. Presumably, the scientific investigations used to derive constraints are available to clinicians internationally. In addition, we would not expect any cultural differences in the degree of tolerance for adverse outcomes. It would be interesting to collate dose constraints from other continents and compare these with the U.S. and European values identified. It would be even more nuanced to compare constraints within specific countries in these continents. To date, none of the cited constraints were derived from a stringent formal process as is customary in clinical guideline development, which likely contributed to the observed heterogeneity. This observation was mirrored in the setting of recommendations for risk-based surveillance among childhood cancer survivors, for which substantial international variation was demonstrated, and acted upon, with the inception of the International Guideline Harmonization Group (IGHG)22. Additionally, consensus is lacking for dose-volume constraints for protons and SRS with mentions of these constraints in seven and five trials, respectively – when including currently active trails. We encourage the PENTEC task group and future task groups to evaluate both modalities as there is little consensus on proton constraints, and also an increasing number of trials using SRS and stereotactic body radiotherapy (SBRT) to ablate metastatic disease.
Recently, five PENTEC reviews have been published on the rates of neurocognitive effects and brain necrosis, breast hypoplasia and impaired lactation, primary hypothyroidism, pulmonary injury, and salivary and dental complications for childhood cancer survivors treated with radiotherapy 16,17,19,23,24. The model for a 5% risk of subsequent IQ < 85 suggested constraints stricter than the current pediatric protocols while the Dmax constraints related to necrosis were similar in these protocols to the recommended PENTEC constraints (Supplemental Table S1 ) 18. The PENTEC dose-toxicity data regarding salivary function demonstrated a 13-32% risk of acute and chronic grade ³ 2 xerostomia with a mean parotid dose of 35-40 Gy 19. Within our review of current and active trials, parotid constraints were more restricted ranging from V20 < 25% to V34 < 50%, and a solitary Dmax constraint of 40 Gy. Breast and thyroid constraints were not presented in our reviewed protocols for comparison with the PENTEC data and additional OAR publications are highly anticipated.
To our knowledge, this is the first study to provide a survey of radiotherapy dose constraints within a broad range of pediatric clinical trials. Our intentions were to describe the current landscape of OAR-specific dose constraints, display a comprehensive guide and interactive website for pediatric constraints used on trials, and present the high variability and inconsistencies within these trials to continue to promote the interest and support for task groups to establish quantitative, evidence-based dose-volume risk guidelines for radiation therapy in childhood cancers.