Interobserver Variations Of Clinical Target Volume Delineation In External Beam Radiation Therapy Of Pancreatic Cancer: Impact Of Multi-Institutional Delineation Protocol

To evaluate the impact of the multi-institutional evidence based handbook protocol on interobserver clinical target volume (CTV) delineation uncertainties among radiation oncologists from one radiation therapy center with seven radiation oncologists in Bosnia and Herzegovina (B&H). The standardization of IMRT in a multi-institutional setting requires the standardization of contouring. An evidenced-based contouring handbook based on ESTRO and RTOG guidelines and anatomical references to standardize contouring. A patient with adenocarcinoma of the pancreas was selected who was a candidate for preoperative chemoradiation therapy and had been fully staged by AJCC staging. The physician group consisted of a disease site-specific gastrointestinal expert and 6 radiation oncologists who were experts in other disease sites or general radiation oncologists. Each radiation oncologist contoured on the designated patient prior to utilizing the protocol book and again after utilizing the protocol book. The same data set and contouring tools were available to each of the physicians. The physicians delineated CTV, each according to their knowledge and experience. After the non-handbook contour, a Bosnian gastrointestinal (GI) leader presented the multi-institutional handbook protocol to doctors and asked them to delineate CTV again in accordance with the protocol. Clinical target volumes of the GI leader and participants were contoured for geometrical analysis (total volume, antero-posterior, latero-lateral and cranio-caudal diameters) and to calculate the mean Dice Similarity Index (mDSI). The mDSI value before using the protocol was 0.70 (range 0.39-0.84), and it was increased to 0.82 (range 0.72-0.91; P=.03) when CTV was based on the protocol. Significantly bigger cranio-caudal diameter of CTV was delineated after observers used the protocol book 11.8cm (range 9.5-14.2) versus 13. cm (range 12.5-14.6); P=.006. Use of a GI multi-institutional evidence based medicine handbook protocol significantly decreases interobserver variations of CTV delineation in external beam radiation therapy of pancreatic cancer. Further studies and strategies to evaluate and reduce intra-observer variability in CTV delineation need to be established.