Fifteen consecutive patients with LARC underwent c3DCRT, f3DCRT, and IMRT planning at the Radiation Oncology Division of the University of Navarre, Spain, from March 2003 to September 2003. Two patients had tumors arising in the upper third of the rectum, eight in the middle third, and five in the lower third. Ten tumors were staged by echoendoscopy as uT3N+ and five as uT3N0. Patients were immobilized in the prone position using a combination of a foam cushion and a prone head cushion. Setup marks were drawn on the patient's skin and the cushion after laser alignment. A non contrast-enhanced planning CT scan was performed using a diagnostic CT scanner (Somatron Plus 4, Siemens Oncology Care Systems, Heidelberg, Germany) with a flat table insert. Patients were instructed to have an empty bladder before CT scan. The scan extended from the L2 vertebral body to 2 cm below the perineum, and axial images were obtained at 5 mm intervals and imported to the planning system (Helax-TMS, Nucletron Scandinavia, Uppsala, Sweden).
Target definition followed the recommendations of the ICRU reports No. 50 and 62. The GTV-T and GTV-N were delineated using information from the diagnostic CT and the Endoscopic Ultrasound (EUS). The clinical target volume (CTV) included the GTV-T and GTV-N (if any), the presacral nodes, the complet mesorectum and the common and internal iliac lymph nodes. The PTV was generated with an asymmetrical margin around the CTV. In areas in which the tumor was close to the SB and bladder, a 5-mm expansion was used while a 10-mm margin was used in the rest of the volume. The organ at risk volumes (OARVs) outlined were the bladder, the rectum from the sphincter to the sigmoid (including the GTV-T), and the SB. The SB was outlined 1 cm above and below the PTV, and the bladder was fully contoured.
For c3DCRT planning, a four-field technique using the classic anatomical references was used  regardeless of the PTV designed. The superior edge of AP-PA portals was placed between the sacral promontory and the L4-L5 interspace, and the inferior border was placed on the ischial tuberosities. If the tumor was located in the lower third of the rectum, the inferior edge was displaced inferiorly to include the perineum. The lateral borders of AP-PA portals were placed to provide adequate coverage of the pelvic sidewalls with a 1-cm margin. The posterior margin for lateral fields was placed 1.5-2.0 cm posterior to the anterior border of the sacrum. The anterior border of the lateral fields usually covered at least the posterior border of the vagina or the prostate, the anterior extent of the primary rectal tumor, and the anterior edge of the sacral promontory. Customized shielding was performed using 1-cm leaves.
In f3DCRT planning, the beams of a typical four-field arrangement were shaped to the PTV with 1-cm leaves.
A 45 Gy dose delivered with 15 MV photons was prescribed to the PTV95 (the minimal dose received by 95% of the PTV) for the 3DCRT plans. Photon dose calculation in tridimensional radiotherapy planning was made using the pencil beam with the Iwasaki algorithm to correct inhomogeneities . IMRT planning procedure has been described previously [16, 17]. Treatment planning was performed using the KonRad inverse planning system, version 2. 0 (Siemens Oncology Care Systems). Seven coplanar equi-spaced fields (gantry angles 0°, 51°, 103°, 154°, 206°, 257°, and 308°) were generated with a median of 51 segments (range, 44 to 67). The isocenter was placed at the geometric center of the PTV. The hierarchy of dose constraints and dose prescription was as follows: first, SB; second, PTV; third, bladder. Plans were accepted when the PTV95 was ≥ 45 Gy, the dose received by 5% of the SB (SB D5) was ≤ 50 Gy, the PTVmin (minimal dose to the PTV) was ≥ 35 Gy, and maximal SB dose (SBmax) was 55 Gy. Dose constraints for the bladder included a maximal dose (Bladdermax) of 55 Gy and a minimal dose received by 5% of the bladder (Bladder D5) of 50 Gy. No specific rectum or external volume constraints were used. IMRT was delivered with 15 MV photons generated in a Mevatron Primus and Oncor linear accelerator (Siemens Oncology Care Systems, Concord, CA). The dose calculation algorithm was also pencil beam with 0.25 cm of voxel size. Konrad calculate the optimum fluence based on physical constraints followed by aperture calculation of the segments .
The target coverage and target dose distribution were evaluated in the GTV, CTV, and PTV obtaining the following parameters for each of the three treatment modalities: minimal target dose (GTVmin, CTVmin, PTVmin), maximal target dose (GTVmax, CTVmax, PTVmax) calculated in all voxels of target volume, minimal dose to 95% of the volume (GTV95, CTV95, PTV95), and homogeneity index (HIGTV, HICTV, HIPTV). The homogeneity index (HI) was defined as the standard deviation of the normalized differential DVH curve  within a target volume.
The degree of comformality was evaluated with a conformity index (CI) that was defined as the ratio between the target volume (PTV) and the irradiated volume at specified prescription dose (Vol PTV/Vol IR95%) .
Normal tissue (bladder, SB, and rectum) avoidance was evaluated using the following parameters: minimal dose received by 5% or less of the volume (Bladder D5, SB D5, Rectum D5) and absolute organ volume receiving 40 Gy or more (Bladder V40, SB V40, Rectum V40).
Finally, irradiated body volumes at the dose levels of 5 Gy (V5), 10 Gy (V10), and 20 Gy (V20) were calculated for each treatment modality. We also calculated the average cut-off point doses at which the irradiated body volumes were significantly different between treatment modalities.
Plans were compared using the Kruskal Wallis test. If positive results, a paired U-Mann-Whitney test was applied with the statement that the IMRT is a reference. We compare IMRT vs f3DCRT and IMRT vs c3DCRT and the differences were considered as statistically significant at the p ≤ 0.05 level.