A comparison of conventional and blended display technics in scintiscanning.

Abstract

Data blending was introduced by Christie and MacIntyre (1, 2) to eliminate statistical fluctuations in the presentation of data obtained in scintillation scanning. Beck (3, 4) pointed out the improvement in perception obtained by eliminating or filtering out the high-frequency “noise” caused by statistical variation. For this he used a recording light spot which simulated the response of the detector collimator at its focal point. On the other hand, the method has been criticized for duplicating the poorest feature of the system, the collimator, and for a resultant loss in resolution (5, 6). We have employed a data-blending technic in our laboratory for the past two years (7). It has appeared to be quite satisfactory and is preferred over conventional methods of recording. We felt it necessary, however, to investigate further the capabilities of the method and compare it with a conventional technic. Methods Two scanners were used in these studies. A 3 × 2-in. crystal detector, Picker Magnascanner, was modified as previously described (7). Its maximum speed is 120 cm per minute, and line spacing varies from 1 to 4 mm. A 5 × 2-in. detector, Ohio-Nuclear scanner, can operate up to 500 cm per minute. Line spacing is limited to four settings of 0.75, 1.5, 3.0, and 6.0 mm. Our own low-energy collimators consisting of 1,045 holes to cover a 3-in. crystal were adapted to both detectors (8). For the studies described in this report, 99mTc was used, the collimators being optimal for the 140-keV gamma emission. The 50 per cent isoresponse diameter of a fine-focus, 1.3-in.-thick collimator, is 0.16 in. (4.1 mm) and of a broad-focus, 0.67-in.-thick collimator, 0.35 in. (8.9 mm). For data-blended recording on the photoscan, gaussian light spots of 5 and 10 mm diameter at the 50 per cent maximum density were used with the fine-focus and broadfocus collimators, respectively. Conventional data-recording was accomplished with various round and rectangular spots. In neither scanner do we use contrast enhancement. No ratemeter time-constant factor affects the data presentation. The scanning parameters were calculated on a basis of count density per unit area for data blending. With conventional data recording, scan speed was calculated according to a desired linear density (counts/cm). Three phantoms were employed. The Picker thyroid phantom has the right lobe at twice the activity of the left lobe, and “cold” nodules of 5 mm in the right upper pole, 9 mm in the left upper pole, and 12 mm in the right lower pole. A “hot” nodule, 12 mm in diameter, is situated in the left lower pole. When the phantom is filled with a low-energy gamma emitter, such as 99mTc, the metallic loading-hole screw in the left lower pole appears as a “cold” defect. A second phantom was constructed to determine the minimum diameter of resolution (Fig. 1). There are three rows of “cold” defects ranging from 2 to 10 mm in steps of 2 mm.