The NCI IMRT benchmark (NCIB) is one of several tools used for credentialing institutions to use IMRT in NCI-sponsored clinical studies. The NCIB specifies that the organ at risk (OAR, or PRV) and planning tumor volume (PTV) should be “at least 5 cm caudad/cephalad” in length, but does not specify OAR length (LO) relative to PTV length (LT). A similar lack of specificity exists in the RTOG P-0126 phase III study of dose escalation for localized prostate cancer, which states that the rectal PRV “should be contoured for a length of 15 cm or to the rectosigmoid flexure.” The purpose of this study was to determine whether such ambiguities might create arbitrary inconsistencies in IMRT credentialing and in clinical studies using IMRT. The NCIB anatomy consists of an annular PTV half-surrounding a cylindrical OAR. The NCIB specifies the following dosimetric goals: A) 95% PTV should receive ≥ prescribed dose (D95T); B) no point outside the PTV may receive ≥ 1.2D95T; C) no more than 10% of the PTV may receive ≥ 1.2D95T; and D) 95% OAR should receive ≤ 0.6D95T. The NOMOS Corvus treatment planning system was applied to obtain NCIB plans with LT = 5 cm, and LO = 5–17 cm. PTV and OARs were created with software expanding a point into circles in radial increments of 1–10 mm. The annular PTV contour was bisected so as to be hemi-cylindrical. PTV and OAR contours were then copied on successive 1-mm thick transverse images to obtain desired lengths. Magnification and measuring tools were applied to precisely verify resulting PTV and OAR dimensions. Prescription parameters were iteratively refined to obtain IMRT plans close as possible to meeting dosimetric goals. All plans were normalized to the maximum % isodose surface that would meet goal A. The extent to which resulting plans met goals B-D was tabulated. For plans meeting goals A-C, the extent to which goal D was missed was graphed as a function of LO. Two different types of plans were obtained (IsII), one meeting goals A-C but missing D, the second meeting goals A and D but missing B and C. Type I plans, similar to plans applied clinically, were characterized by: i) normalizing to 80–90% isodose surfaces; ii) PTV dose volume histogram (DVH) with a fairly steep falloff and small tail leading to maximum dose; and iii) easily meeting goals B and C. With LO = LT = 5 cm, % OAR ≤ 0.6D95T was 30–35%, missing goal D by a wide margin. By extending LO to 17 cm while keeping LT = 5 cm, it was possible to increase % OAR ≤ 0.6D95T to 75%. In clinical practice it is similarly possible to contour OAR distal to the PTV so as to have DVHs better approach or meet dosimetric criteria. For the RTOG study P-0126, the extent of dose escalation depends on the rectal dose-volume relationship, and may be altered by the choice of whether rectum is contoured to a length of 15 cm or to the rectosigmoid flexure. Type II plans were characterized by: i) normalizing to 50–60% isodose surfaces; ii) bi-modal falloff of PTV DVH, notably less steep than in type I plans; iii) >97% OAR receiving ≤0.6D95TV, easily meeting goal D; iv) numerous small volumes outside the PTV receiving doses ≥ 1.2D95TV, missing goal B; and v) large volumes inside the PTV receiving doses ≥ 1.2D95TV, missing goal C by a wide margin. In verifying NCIB anatomies, it was noted that shapes of PTV and OAR contours, and of the resulting gap (Δ) between them, were not perfectly circular, but rather pixelated with dimensions originating in CT images and grids applied for planning. In this investigation, Δ was found to vary from 2.4 to 3.0 mm. Uncertainties in volume associated with the pixelation of contours may introduce errors in studies where dose-volume relationships of small volumes are applied to determine dose escalation or analyze clinical outcomes. The ambiguities of the NCIB LO specification and RTOG P-0126 rectal length specification may allow arbitrary inconsistencies in benchmarking and clinical study data. Such uncertainties and ambiguities should be removed to minimize the number of arbitrary factors in IMRT credentialing and clinical studies