Purpose Selection of an open or minimally invasive approach to total mesorectal excision (TME) is generally based on surgeon preference and an intuitive assessment of patient characteristics but there consensus on criteria to predict surgical difficulty. Pelvimetry has been used to predict the difficult surgical pelvis, typically using only bony landmarks. This study aimed to assess the relationship between pelvic soft tissue measurements on preoperative MRI and surgical difficulty. Methods Preoperative MRIs for patients undergoing laparoscopic rectal resection in the Australasian Laparoscopic Cancer of the Rectum Trial (ALaCaRT) were retrospectively reviewed by two blinded surgeons and pelvimetric variables measured. Pelvimetric variables were analyzed for predictors of successful resection of the rectal cancer, defined by clear circumferential and distal resection margins and completeness of TME. Results There was no association between successful surgery and any measurement of distance, area, or ratio. However, the was a strong association between the primary outcome and the estimated total pelvic volume on adjusted logistic regression analysis (OR = 0.99, P  = 0.01). For each cubic centimeter increase in the pelvic volume, there was a 1% decrease in the odds of successful laparoscopic rectal cancer surgery. Intuitive prediction of unsuccessful surgery was correct in 43% of cases, and correlation between surgeons was poor (ICC = 0.18). Conclusions A surgeon’s intuitive assessment of the difficult pelvis, based on visible MRI assessment, is not a reliable predictor of successful laparoscopic surgery. Further assessment of pelvic volume may provide an objective method of defining the difficult surgical pelvis.