The Threshold Visibility of Pulmonary Shadows

Abstract

What should a radiologist be expected to see on a chest film? What is the size, shape, and contrast of the least perceptible shadow? Yerushalmy (1950) found that 32 per cent of suspicion-worthy pulmonary shadows are missed by “competent” film readers. Birkelo (1947) has shown that 35-mm. photofluorograms are as efficient at detecting pulmonary lesions as is full-sized double-screen technic. A pulmonary tumor 6 mm. in diameter may be close to invisible. How then can vessels from 1 to 3 mm. in diameter produce the obvious lacework of the pulmonary fields? A lung full of tubercles 1.5 mm. in diameter may present an appearance quite easy to diagnose, yet the single 1.5-mm. shadow is below the threshold. A number of researches on the fundamentals of roentgen visibility have been published. Burger (1949) showed the mutual dependence of size and contrast needed to make a shadow visible, and pointed out the inferiority of fluoroscopy and photofluorography, but he worked with sizes much smaller than most pulmonary lesions. Franke (1942) and Chantraine (1942) carried on a controversy concerning “summation” versus “real images.” Re-sink (1949) has written most recently about this. These authors cite a number of earlier workers. Walter (1917) investigated visibility of beveled and curved edges. Steiner (1937) believed visibility of “miliary” tubercles to be dependent on their composition (ash). He showed the relationship, also, of size and number to visibility. Our own interest in these matters arises from the recently increasing importance of photofluorographic “mass surveys” and a desire to standardize the reading of survey films. It is hard to know what readers are doing a good job unless one knows just how much of a shadow a good reader ought to be able to see. Threshold Shadows Figure 1 shows a roentgenogram of various objects used to test the threshold of visibility under various circumstances. The visibility could be effectively reduced on plain films by making the exposure very light, or by exposing the film all over (fogging) to 0.5 to 5 times the exposure already given through the test objects. For objects with sharp edges, we judged all simple shapes (round, square, triangular, etc.) and all sizes, from 5 mm. to 70 mm., about equally visible. This agrees with Burger (1949). Later we thought we detected a slight raising of the threshold for the smaller sharp faint shadows. At optimum exposure on plain film, the threshold contrast is given by 0.2 mm. of lucite. This is about 0.4 per cent absorption. Unless conditions are optimum, the threshold is about 0.3 mm. lucite or 0.5 mm. wood. Many films were made with a lucite object stuck to the back of the patient's chest. The threshold thickness on a chest film proved to be about 3 mm. (6 per cent absorption).