(Radiology. 1999;213:599-602.)
© RSNA, 1999
Two-Point, Timesaving Method for Measurement of Gastric Emptying with Diagnostic Accuracy Comparable to That of the Conventional Method1
Raghuveer K. Halkar, MD,
Andrew L. Paszkowski, MB, BS, FRACP,
Margie E. Jones, CNMT,
James R. Galt, PhD,
Leonard R. Goldfarb, MD,
John G. Hunter, MD and
Andrew T. Taylor, Jr, MD
1 From the Departments of Radiology (R.K.H., A.L.P., M.E.J., J.R.G., L.R.G., A.T.T.) and General Surgery (J.G.H.), Emory University Hospital, 1460 Clifton Rd NE, Atlanta, GA 30322. From the 1996 RSNA scientific assembly. Received September 11, 1998; revision requested December 7; revision received January 20, 1999; accepted April 30. Address reprint requests to R.K.H.
 |
Abstract
|
|---|
Solid-phase gastric emptying is linear. Therefore, the authors calculated gastric-emptying half-time, the time for half of the ingested solids or liquids to leave the stomach, with the conventional multiple-point method and the proposed two-point method (at 0 and 120 minutes) in retrospective and prospective studies of 50 patients each. The results showed excellent correlation. Results with the two-point method were comparable to those with the multiple-point method, and the two-point method substantially reduced technologist and camera times.
Index terms: Stomach, motility, 72.12177 • Stomach, radionuclide studies, 72.12177
|
Introduction
|
|---|
Gastric-emptying scintigraphy is currently the standard of reference for quantitating gastric emptying and documenting gastroparesis in clinical practice and research (1). The time for half of the ingested solids or liquids to leave the stomach (hereafter, "half-time") is the most commonly used quantitative parameter for the assessment of solid- and liquid-phase gastric emptying. Technetium 99m sulfur or albumin colloids are most commonly used to label solid meals, whereas indium 111 diethylenetriaminepentaacetic acid, or DTPA, is used to measure liquid emptying (2). The majority of patients referred for a radionuclide gastric emptying study have symptoms of nausea, vomiting, or suspected gastroparesis. Emptying of solids is generally a more sensitive and earlier indicator of delayed gastric emptying than is liquid emptying; hence, assessment of liquid gastric emptying is seldom necessary in routine clinical practice (2).
The techniques used to incorporate the radiolabel with the meal, position the patient, and acquire and process the data vary between institutions. Regardless of these variables, the half-time is calculated on the basis of multiple data points acquired during about 120 minutes in a majority of published studies (3–5).
Since solid-phase gastric emptying is linear (3), most of the processing protocols apply a linear fit to the time-activity curve to calculate solid-phase half-time of gastric emptying (6). Therefore, we hypothesized that the half-time for solids calculated on the basis of data acquired at only 0 and 120 minutes (two-point method) would not be substantially different from that calculated with the multiple-point method. The purpose of this study was to assess the correlation between half-times measured with the two-point and multiple-point methods and to determine if the diagnostic usefulness of the two-point method was comparable to that of the multiple-point method.
|
Materials and Methods
|
|---|
A retrospective analysis of gastric emptying was performed with the stored image data for 50 consecutive patients (20 men, 30 women; age range, 27–71 years; mean age, 58 years) suspected of having gastroparesis. A prospective analysis of gastric emptying was then performed in an additional 50 consecutive patients (17 men, 33 women; age range, 26–95 years; mean age, 51.7 years) referred for a gastric-emptying study with symptoms of nausea, gastroesophageal reflux, or gastroparesis.
The 99mTc albumin colloid (Microlite; Du Pont, North Billerica, Mass) (2 mCi [74 MBq]) was added to 20 g of liver pâté and cooked in a microwave oven for 30–45 seconds on medium setting. The pâté was then stirred and cooked for another 30–45 seconds until it had a consistency of cooked hamburger. Ten grams of the liver pâté labeled with approximately 1 mCi (37 MBq) of 99mTc albumin colloid was added to 140 g of beef stew and reheated in the microwave oven. This meal was served to each patient with 5 fl oz (150 mL) of orange juice (7). This standard meal was chosen because it has been carefully evaluated under a number of different conditions (5–7).
Immediately after ingestion of the standard meal, the patient was positioned upright in front of the gamma camera, and anterior and posterior images of the abdomen were acquired for 1 minute each to include the stomach and distal third of the esophagus. Further anterior and posterior images were obtained for 1 minute each at 15-minute intervals for a total of 120 minutes (nine images) (Fig 1). The patient sat quietly between images.
View larger version (21K):
[in this window]
[in a new window]
|
Figure 1. A, Anterior (ant) and B, posterior (post) images of the abdomen obtained at 15-minute intervals (0-120), with an ROI over the stomach. The radioactive solid meal in the stomach appears black.
|
|
The first image obtained after completion of the meal was defined as the time zero image. Anterior and posterior regions of interest (ROIs) were drawn over the stomach on the 0-minute images. These ROIs were applied to all subsequent images (Fig 1). The software allowed repositioning and resizing of the ROIs over the stomach on successive images. Counts over each ROI were decay corrected to time zero. Geometric-mean data for the anterior and posterior images obtained at each of the nine time intervals were then used to generate a time-activity curve. A linear fit was applied to this time-activity curve, and multiple-point (conventional) half-time was measured from the linear fit (Fig 2).
View larger version (13K):
[in this window]
[in a new window]
|
Figure 2a. (a) Decay-corrected geometric-mean counts (cts) in the stomach ROI (E 4 = x10-4) and (b) the time-activity curve (kcts = 1,000 counts), which shows the linear nature of solid-phase gastric emptying.
|
|
View larger version (9K):
[in this window]
[in a new window]
|
Figure 2b. (a) Decay-corrected geometric-mean counts (cts) in the stomach ROI (E 4 = x10-4) and (b) the time-activity curve (kcts = 1,000 counts), which shows the linear nature of solid-phase gastric emptying.
|
|
Decay-corrected geometric-mean counts (C) at times zero (T0) and 120 minutes (T120) were used to calculate the half-time (T1/2) with the following equation:
For the 50 patients in the retrospective analysis, two experienced readers (A.T.T., R.K.H.) independently assessed the lag phase with use of two curves. One curve was generated with use of only anterior counts and the other with use of geometric-mean counts. The time required for the counts to decrease by 10% of the initial (at time zero) counts was considered to be the lag phase. If the counts in the ROI on the 15-minute image were less than 90% of the counts at time zero, then no lag phase was recorded.
Correlation between the half-times determined with the multiple-point and two-point methods was calculated for the retrospective and prospective analyses. A partial F test, which tests for equality of intercept, slope, and correlation between the two regression lines, was used to check the significance of regression lines.
With use of our standard solid meal, previous data (3) show the normal half-time of gastric emptying is 92.4 minutes ± 15 (2 SD) in premenopausal women and 77.0 minutes ± 32 in postmenopausal women. The normal half-time in men is 77.9 minutes ± 32. On the basis of the mean values plus 2 SD in the three populations, we used 110 minutes as the upper limit of normal and compared the half-times obtained with the multiple- and two-point methods.
|
Results
|
|---|
In the retrospective analysis, correlation was excellent between the half-times measured with the multiple- and two-point methods: R = 0.98, standard error of the estimate (SEE) = 14.8 minutes, slope = 1.02, intercept = -7.7 (Fig 3). In the prospective study, the correlation was also excellent: R = 0.98, SEE = 8 minutes, slope = 0.98, intercept = 4.29 (Fig 4). There was no statistically significant difference in correlation of prospective versus retrospective data (partial F test, P = .609).
View larger version (13K):
[in this window]
[in a new window]
|
Figure 3. In the retrospective analysis in 50 patients, comparison of half-times for solid-phase gastric emptying measured with the multiple- and two-point methods. Scatterplot shows good correlation: R = 0.98, SEE = 14.8 minutes, slope = 1.02 (not significantly different from 1), intercept = -7.7 (close to 0).
|
|
View larger version (13K):
[in this window]
[in a new window]
|
Figure 4. In the prospective analysis in an additional 50 patients, comparison of half-times for solid-phase gastric emptying measured with multiple- and two-point methods. Scatterplot shows good correlation: R = 0.98, SEE = 8 minutes, slope = 0.98, intercept = 4.29.
|
|
Comparison of half-times measured for solid-phase gastric emptying with the multiple- and two-point methods for the combined data in all 100 patients showed good correlation (Fig 5). With use of 110 minutes as the upper limit of normal for gastric half- time, the two methods produced concordant results in 96% of patients. In the four patients with discordant results, the difference was less than 5 minutes, which was borderline abnormal to normal (Fig 6).
View larger version (12K):
[in this window]
[in a new window]
|
Figure 5. For the combined data in all 100 patients, comparison of half-times for solid-phase gastric emptying measured with multiple- and two-point methods. Scatterplot shows good correlation: R = 0.98, SEE = 11.75 minutes, slope = 0.95, intercept = 10.2.
|
|
View larger version (16K):
[in this window]
[in a new window]
|
Figure 6. With use of 110 minutes as the upper limit of normal, comparison of half- times (T ) for gastric emptying measured with multiple- and two-point methods. Excellent concordance of results is seen.
|
|
With use of the criteria for lag phase assessment, both observers were in agreement about their assessment. They found a lag phase of more than 15 minutes in 17 of 50 (34%) patients when only anterior counts were used. With use of the geometric-mean counts, five of 50 (10%) patients had a lag phase greater than 15 minutes (
2 test, P 
.05), and four of them had an abnormally prolonged half time.
|
Discussion
|
|---|
Results of this study demonstrate that the half-time for solid-phase gastric emptying measured with the two-point method correlates well with that measured with the multiple-point method with use of a linear fit, with no loss in diagnostic accuracy.
The advantage of the two-point method is a reduction in imaging time and, hence, a savings in both camera and technologist times. The reduction in the imaging time should also increase the patient comfort and tolerance for the procedure and allow an increase in patient throughput.
The main disadvantage in not acquiring multiple sequential images is the loss of a means to detect rapid gastric emptying. With the multiple-point method, a rapid exponential emptying curve raises the possibility of a dumping syndrome or suspicion for a poor solid-phase tag with radiolabel dissociation into the liquid phase. None of the 100 patients we studied showed a rapid exponential emptying rate, which confirms that a dumping syndrome is rare in patients referred with a clinical question of gastroparesis. In patients suspected of having a dumping syndrome, more frequent images can be obtained during the first 30 minutes followed by a final image at 120 minutes. This protocol will allow detection of rapid gastric emptying and still result in a saving of excess camera, computer, and technologist times.
The existence of a lag phase, which precedes the gastric emptying of solid food, is controversial. Siegel et al (8) described a modified power exponential model to assess lag phase, but they used anterior counts, which exaggerate the lag phase. Moore et al found that use of anterior counts creates an artificial delay in the half time and results in a complex emptying pattern of solids instead of the linear pattern that is usually observed when geometric-mean counts are used (9). Owing to the geometry of the stomach, passage of a radiolabel from the fundus to the antrum results in radioactivity moving from posterior to anterior. This change in geometry reduces the attenuation and can result in an overestimation of anterior counts, which creates or exaggerates a lag phase. Consequently, use of a geometric-mean technique is recommended to correct for differences in depth as food moves anteriorly from the fundus to the pylorus (Fig 7).
View larger version (20K):
[in this window]
[in a new window]
|
Figure 7. Time-activity curves with use of geometric-mean counts (top) and counts from only the anterior ROI (bottom). The former shows linear gastric emptying of solids. The latter demonstrates a lag phase as a result of an apparent increase in the counts in the early part of the study. kcts = 1,000 counts.
|
|
A comparison of different techniques for assessment of lag phase by Ziessman et al (10) shows that the rate of temporal sampling affects the lag phase measurement. They also show that results with a modified power exponential method have poor correlation to those with other techniques, but the images were acquired with the patient supine, which has been shown to delay gastric emptying (11). In the 50 cases we reviewed retrospectively, 14 patients showed abnormally prolonged half-times, and 36 had normal half-times. In the latter, only one patient (3%) had a lag phase longer than 15 minutes as calculated with use of geometric-mean decay-corrected counts. A lag phase shorter than 15 minutes may be missed with our method, but such an abbreviated lag phase, if it exists, is unlikely to be clinically important.
Our results may not be valid for protocols with use of a solid-phase radiolabel, which is not stable because the gastric emptying of liquid is exponential. If a solid radiolabel is unstable and moves from the solid to the liquid component, the half-time of solid gastric emptying will be falsely reduced. However, a poor radiolabel is also a limitation with the multiple-point method.
Thomforde et al (12) used a radioactive solid meal consisting of 1.0 mCi (37 MBq) of pertechnetate incorporated onto anion exchange resin (Amberlite IRA-410; Sigma Chemical, St Louis, Mo) and obtained results similar to ours. They conclude that selective scans obtained at 2 and 4 hours provide an excellent screening test for detection of fast or slow gastric emptying. Although the most common measurement is the half-time, some authors have used either a percentage emptied or a percentage retained at a certain time to measure gastric emptying (13). In a recent editorial, Maurer (14) warned that use of a simplified radionuclide gastric-emptying method might amount to diminishing a standard of reference. It was not the intent of our study to do so. In the current cost-conscious health care environment, clinical problem solving is a major goal. We suggest that if delayed gastric emptying is the clinical question, then the two-point method we propose should provide an adequate answer.
Our results demonstrate that the geometric-mean two-point method of estimating the half-time of solid gastric emptying produces no loss in diagnostic accuracy when compared with the multiple-point method. It requires reduced technologist and camera times and will thereby increase patient throughput and, hopefully, patient comfort and tolerance of the procedure.
|
Acknowledgments
|
|---|
The authors thank Michelle Barrios for her assistance in the completion of the manuscript for this article.
|
Footnotes
|
|---|
Abbreviations: ROI = region of interest
SEE = standard error of the estimate
Author contributions: Guarantor of integrity of entire study, R.K.H.; study concepts and design, R.K.H.; definition of intellectual content, R.K.H., A.T.T.; literature research, R.K.H., A.L.P.; clinical studies, J.G.H.; data acquisition, M.E.J., R.K.H.; data analysis, M.E.J., R.K.H., J.R.G.; statistical analysis, R.K.H., L.R.G., J.R.G.; manuscript preparation, A.L.P.; manuscript editing and review, A.T.T.
|
References
|
|---|
-
Packman HC, Miller MA, Fischer RS. Role of nuclear medicine in evaluating patients with suspected gastrointestinal motility disorders. Semin Nucl Med 1995; 25:289-305.[Medline]
-
Chaudhari TK, Fink S. Gastric emptying in human disease states. Am J Gastroenterol 1991; 86:533-538.[Medline]
-
Datz FL. Considerations for accurately measuring gastric emptying. J Nucl Med 1991; 32:881-883.[Free Full Text]
-
Moore JG, Datz FL, Christian PE, Greenberg E, Alazraki N. Effect of body posture on radionuclide measurements of gastric emptying. Dig Dis Sci 1988; 33:1592-1595.[Medline]
-
Christian PE, Datz Fl, Moore JG. Technical considerations in radionuclide gastric emptying studies. J Nucl Med Technol 1987; 5:200-207.
-
Datz FL, Christian PE, Moore J. Gender-related differences in gastric emptying. J Nucl Med 1987; 28:1204-1207.[Abstract/Free Full Text]
-
Christian PE, Moore JG, Datz FL. Comparison of Tc-99m labeled liver and liver pate as markers for solid phase gastric emptying. J Nucl Med 1984; 25:364-366.[Abstract/Free Full Text]
-
Siegel JA, Urbain Jl, Adler LP, et al. Biphasic nature of gastric emptying. Gut 1988; 29:85-89.[Abstract/Free Full Text]
-
Moore JG, Christian PE, Taylor AT, et al. Gastric emptying measurements: delayed and complex emptying patterns without appropriate correction. J Nucl Med 1985; 26:1206-1210.[Abstract/Free Full Text]
-
Ziessman HA, Atkins FB, Vemulakonda US, et al. Lag phase quantification for solid gastric emptying studies. J Nucl Med 1996; 37:1639-1643.[Abstract/Free Full Text]
-
Moore JG, Datz FL, Christian PE, et al. Effect of body posture on radionuclide measurements of gastric emptying. Dig Dis Sci 1988; 33:1592-1595.
-
Thomforde GM, Camilleri M, Phillips SF, Forstrom LA. Evaluation of an inexpensive screening scintigraphic test of gastric emptying. J Nucl Med 1995; 36:93-96.[Abstract/Free Full Text]
-
Fisher RS, Malmud LS, Bandini P, et al. Gastric emptying of a physiologic mixed solid–liquid meal. Clin Nucl Med 1982; 7:215-221.[Medline]
-
Maurer AH. Can we prevent tarnishing a gold standard?. Semin Nucl Med 1995; 25:288.[Medline]
This article has been cited by other articles:
|
|
|
|
 
V. Vijayakumar, E. G. Briscoe, T. L. Boysen, and Y. M. Jimenez
Assessment of the practical role of a radionuclide low-fat-meal solid gastric emptying study.
J. Nucl. Med. Technol.,
June 1, 2006;
34(2):
82 - 85.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
G. Mariani, G. Boni, M. Barreca, M. Bellini, B. Fattori, A. AlSharif, M. Grosso, C. Stasi, F. Costa, M. Anselmino, et al.
Radionuclide Gastroesophageal Motor Studies
J. Nucl. Med.,
June 1, 2004;
45(6):
1004 - 1028.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
Abstract
Introduction
