This book represents the convergence of three related themes which have occupied a large part of my professional life. First, ever since I started training as a radiologist almost 20 years ago, I have been intrigued by the “pattern recognition” that lies at the heart of our specialty. This approach to diagnosis can be very powerful, but also prone to error if different entities look the same. As a first year resident reading out the overnight Emergency Department plain films at Leicester Royal Infirmary, hardly a fracture went reported without checking our heavily thumbed and coffee-stained edition of Keats [1] for possible mimics or confounders. Second, one of my most popular postgraduate lectures is entitled “Pearls and pitfalls in abdominal CT,” and this talk grew out of my early interest in normal variants simulating disease. It is clear that all radiologists struggle with the basic questions as to whether a study is normal or abnormal, or whether findings of a given diagnosis can be due to anything else. Third, most physicians are perfectionists and dislike making mistakes, especially when those mistakes can be harmful to patients. We are entrusted with caring for patients who are often at their sickest and most miserable. Anything we can do to improve their care fulfills our duty to them, and also helps address ongoing and legitimate public concern regarding medical errors and patient safety [2, 3]. The literature consistently suggests that 1.0 to 2.6% of radiology reports contain serious errors [4–6].