Schulman A, Simpkins K C
Clin Radiol. 1975 Jul;26(3):311-6. doi: 10.1016/s0009-9260(75)80067-5.
The initial aim was to program a computer with information on the frequency of radiological signs in benign and malignant gastric ulcers in order to obtain a percentage probability of benignancy or malignancy in succeeding ulcers in clinical practice. However, only four of the many signs described in gastric ulcer were confirmed to be of validity (i.e. reliable existence) by an inter-observer variation study using two observers and the films from 69 barium meal examinations. These were projection or non-projection of the in-profile ulcer, presence or absence of adjacent mucosal folds, good or poor definition of the in-face ulcer's edge, and extension of radiating folds to the in-face ulcer's edge. A few more remained unassessed due to insufficient numbers of relevant cases. It is condluced that: as defined in the literature the majority of radiological signs in this field are of uncertain existence; and the four that were found to be valid do not fully describe the important appearances that may be seen in benign and malignant ulcers and would be inadequate to differentiate them to a sufficiently high degree of probability.
最初的目标是为计算机编程，输入有关良性和恶性胃溃疡放射学征象出现频率的信息，以便在临床实践中得出后续溃疡为良性或恶性的概率百分比。然而，在一项使用两名观察者和69例钡餐检查影像的观察者间变异研究中，胃溃疡所描述的众多征象中只有4个被证实是有效的（即确实存在）。这些征象是轮廓内溃疡的投影或非投影、相邻黏膜皱襞的有无、正面溃疡边缘的清晰或模糊，以及放射状皱襞延伸至正面溃疡边缘。由于相关病例数量不足，还有一些征象未得到评估。得出的结论是：如文献中所定义的，该领域中的大多数放射学征象是否存在并不确定；而发现有效的4个征象并不能完全描述良性和恶性溃疡中可能出现的重要表现，也不足以在足够高的概率水平上对它们进行区分。
