Session Item

Pelvic insufficiency fracture (PIF) is a well-known late side effect of pelvic radiotherapy.   We have recently found a high frequency of PIFs in patients with anal (50%) or rectal cancer (30%), 1 and 3 years after radiotherapy, and analyzed the risk of PIF in relation to radiation dose to pelvic substructures. Based on localization of PIFs and relation to radiation dose we have initiated a prospective phase II trial: Bone-sparing chemoradiotherapy for anal cancer (Danish Anal Cancer Group, DACG II), and present the initial data from the bone optimized dose planning.

Patients with indication for curative chemoradiotherapy underwent planning CT and MRI for standard VMAT photon-plan generation. Target and elective volumes were delineated according to the DACG guidelines (modified RTOG). Standard delineation of organs at risks (OARs) included bowel bag, bladder, femoral heads, sacral bone, penile bulb, female external genitalia, and for bone optimizing: sacroiliac (SI)- joints, sacral alae, acetabulum, symphysis, and total pelvic bones were delineated. A bone-optimized plan was generated subsequently with following priority: Target coverage>bowel bag> SI-joints/ Sacral Alae> bladder> external female genitalia/penile bulb> acetabulum/symphysis> total pelvic bones.  Plan optimization criteria were: V30 Gy<55% for pelvic substructures and comparable dose to other OARs. Conformity indices (V95% for PTV-T (tumor) and E (elective)) and mean dose as well as V30%, (cm³) to remaining normal tissue (body outline subtracted delineated OARs and bone) were calculated to assure plan quality. Wilcoxon signed rank test was used for comparison of plans and p<0.05 considered statistically significant.

A total of 5 VMAT based bone sparing plans were compared to standard plans (all 3 arcs). Dose to CTV-T and N(node) was 54-60 Gy in 30 fractions and dose to CTV-E was 48 Gy in 30 fractions. Stages included T1-3, N0-1 (one including paraaortal lymphnodes). V30 Gy to SI-joints and Alae of the sacral bone was significantly lower with the bone sparing plans, p=0.043, as were mean doses to sacral bone and total pelvic bones, p=0.043. Dose to other delineated OARs were comparable. Conformity indices (V95%) for PTV-T and E were similar between the standard and the bone optimized plan. Mean doses to remaining normal tissue were comparable, but V30% (of both elective and tumor dose) was significantly lower with bone optimized planning (Table 1).

We demonstrate that that sparing of pelvic bone substructures is feasible with optimization and modulation of standard plans with-out compromising dose to other OARs and plan quality. Evaluation of plan quality will continue during inclusion of patients in the trial.