Brachytherapy dose distributions are substantially different from EBRT dose distributions due to the specific characteristics of photon fluence around a point source. Brachytherapy comes with a steep dose fall off and a very inhomogeneous dose distribution inside GTV and CTV. In fact, it is not possible to create dose distributions from linacs or proton beams which are as steep as brachytherapy dose distributions. The “brachy peak” is indeed steeper than the “Bragg peak”!
While the dose fall-off around a brachytherapy source cannot be modulated, the only way to modulate dose gradients in brachytherapy is to optimize dwell positions and dwell times. Brachytherapy has therefore in general less flexibility to control dose gradients as compared to EBRT. In this context, the direction relative to the axis of source catheters is important: 1) “longitudinal direction” is along source catheters (e.g. tandem and/or needles) and 2) “perpendicular direction” is orthogonal to source catheters. In the “longitudinal direction”, a margin can be constructed by loading further source positions (e.g. just at the edge or outside the target). In the “orthogonal direction”, the dose distribution cannot be elongated in a similar way. The orthogonal dose fall off is almost exclusively determined by the inverse square law and cannot be manipulated to become less steep by modifying the loading pattern without introducing additional catheters or needles. The isodose lines can be pushed further away from the source in the orthogonal direction only by escalating the dose. Therefore, defining a PTV for brachytherapy with a fixed margin uniformly around the CTV and prescribing to that PTV would significantly escalate the dose within the entire CTV. This is fundamentally different from the situation in EBRT where the dose plateau is increased in by application of a margin with no accompanying dose escalation.
“Geometric uncertainties” for brachytherapy include applicator commissioning, reconstruction of the applicator, image fusion uncertainties, migration of source catheters, and afterloader source positioning. These types of uncertainties are in general limited and occur most often along catheters, which means that they can be partly compensated for by extra loading along catheters. In some clinical scenarios catheter migration can be substantial, and in such cases re-imaging may be considered to control uncertainties rather than compensating through application of margins. In addition to “geometric uncertainties”, contouring uncertainties contribute to uncertainties in delivered brachytherapy dose. Contouring uncertainties can occur in any direction and are in general the most prominent contributor to uncertainties.
Since brachytherapy margins can only be applied in selected directions (along catheters), target dose uncertainties (in particular contouring uncertainties) are typically around 10% (SD), which may be larger than what is normally considered acceptable for EBRT. As brachytherapy dose uncertainties are relatively larger: how come that local control in brachytherapy (e.g. cervix and prostate) is excellent? The explanation may be related to the steep brachytherapy dose gradients which allow for prescription of very high doses (e.g. >85Gy). In this context, a loss of target dose due to uncertainties may have limited impact on local control. Furthermore, due to the gradients throughout the target volumes, uncertainties in brachytherapy do not systematically lead to decrease in target dose, but can also lead to higher target dose than expected. Therefore, on a population level, random geometric uncertainties have negligible impact on expected tumor control probability.
In conclusion, while PTV margins cannot be systematically applied in brachytherapy, geometric uncertainties and contouring uncertainties have in most cases limited impact on tumor control probability. While steep dose gradients can potentially lead to larger dose uncertainties, these steep dose gradients also allow administration of very high doses to limited target volumes, and at the end of the day a large therapeutic window is achieved.