Normal organ volume assessment from abdominal CT

Abdominal Imaging

© Springer-Verlag New York, LLC. 2004 Received: 1 Nov. 2002 / Revision accepted: 8 Oct. 2003 / Published online: 18 Mar. 2004

Abdom Imaging (2004) 29:482– 490 DOI: 10.1007/s00261-003-0139-2

Normal organ volume assessment from abdominal CT E. M. Geraghty, J. M. Boone, J. P. McGahan, K. Jain Department of Radiology, University of California, Davis, Medical Center, Research Imaging Center, 4701 X Street, Sacramento, CA 95817, USA

Abstract We determined the normal distribution of abdominal organ volumes measured from abdominal computed tomographic (CT) images. A total of 149 adult abdominal CT studies were selected, and 711 organs (388 from males, 323 from females) were outlined by hand on each CT image by using a computer. More than 18,000 organ outlines were traced. The organs studied included left and right kidneys, left and right adrenals, spleen, pancreas, and liver, and the first lumbar vertebrae was also evaluated. Using the known pixel size and section thickness, organ volumes were computed. Organ volumes were corrected for height and weight for each sex. The normal and cumulative normal distributions for each organ studied were computed, demonstrating the range of organ volumes for each sex that exist in the normal adult population. Organ volumes ranged from a mean of 4.4 mL (female left adrenal) to 1710 mL (male liver). Mean organ volumes were 64.4, 156.5, 179.8, and 1411 mL for the female pancreas, kidneys, spleen, and liver, respectively. Corresponding male volumes were 87.4, 193.1, 238.4, and 1710 mL, respectively. Tabular data are provided that indicate the relative size for each organ volume in terms of the cumulative probability distribution. Normative data are provided to allow physicians to estimate where in the normal range a particular organ volume lays. Organ volumes may be useful as quantitative indices of pathologic conditions. Key words: Abdomen—Computed tomography—Organ volume—Pancreas—Liver—Spleen—Adrenal There has been great interest in determining in vivo organ volumes with computed tomography (CT) [1–7]. The technique, based on hand-outlining of regions of interest (ROIs), has been shown to be accurate and reproducible [4, 5, 7]. The determination of abdominal organ volumes in particular has significant potential clinical value. For example, liver volCorrespondence to: J. M. Boone; email: [email protected]

umes are important not only in determining disease states and disease progression but also in estimating segmental liver volumes for transplant donors and planning the extent of hepatectomy in cancer patients [4]. The spleen commonly increases in size in response to conditions such as infection and hematologic or metabolic disorders [2, 5]. There is a good correlation between platelet count and spleen volume, and splenic volume detects serious liver disease and correlates with splenic hyperfunction [8]. Kidney size bears a relation to the degree of renal disease, and pancreatic atrophy is associated with changes in the organ’s exocrine function [9]. Although temporal comparisons of relative organ volumes in an individual patient often yields diagnostic information, current technology allows for more quantitative assessment with regard to organ size, which in turn enables population-based comparisons. The goal of this study was to establish sex-specific ranges of normal organ volumes for the solid abdominal organs. Organ volumes, corrected for the height and weight of the patient, were determined for the kidneys, adrenals, liver, pancreas, spleen, and the first lumbar vertebrae (L1).

Materials and methods Subject selection The picture archiving and communication system (PACS; Stentor, South San Francisco, CA, USA) at our institution was used to select 207 consecutive human abdominal CT examinations. These CT studies were performed for routine clinical evaluation, and we made use of the patient’s images under proper institutional review board authority (protocol 992656; approval date, 16 February 2001). Based on this institutional review board protocol (category 4 exemption), informed consent from the patient was not required. The patient population comprised inpatients and outpatients who were being studied due to indications including abnormal weight loss or gain, abdominal and pelvic pain, unresolved nausea, motor vehicle trauma, and other common clinical indications for abdominal CT.

E. M. Geraghty et al.: Organ volume assessment from CT

The radiologist’s report for each CT examination was reviewed, and all comments concerning the solid abdominal organs and their disease or disease-free status were noted. Careful study of the patient’s CT diagnosis and in many cases their medical records were used to establish normalcy for each organ. Organs that may have been affected by local or systemic disease were not used. Organs were eliminated if the CT examination or radiologist’s report indicated trauma or surgery in the proximity of the organ. Patients younger than 18 years were also excluded, because these individuals may not have reached mature organ stature. Subjects whose CT scans had abnormal features, such as skin defects, bilaterally cutoff edges, or movement artifacts that could influence the various measurements of this study, were also excluded. In general, organs were considered normal based on information available in the CT images, the patient’s chart, and the radiologist’s report, coupled with clinical judgment in a case-by-case assessment. Based on these selection criteria, 149 subjects remained in the study. The men selected in the study had an average age of 48.4 years (17.5 years ), and the women had an average age of 49.3 years (18.0 years).

Computed tomography All patients were scanned on a General Electric Lightspeed QX multislice scanner (Waukesha, WI, USA) or a General Electric CT/I single-slice scanner. Scanning parameters depended on the clinical indication for the study, and all techniques were part of established clinical protocols at our institution. Consistent with this, all patients were scanned at 120 kVp, and the section thickness varied between 5 and 10 mm for all subjects. Most studies were contrast enhanced.

Analysis of patient images The abdominal CT images were downloaded from our research PACS system (eFilm, eFilm, Inc., Toronto, ON, Canada) and transferred to a PC workstation. Studies were viewed on the workstation monitor for review of DICOM header information (age, sex, display field of view, and section thickness), which was subsequently recorded in a spreadsheet. The image files were saved sequentially as DICOM files. For organ identification (i.e., segmentation), images were displayed on a computer monitor with a resolution of 1280 ⫻ 1024 pixels. Custom mouse-and-cursor software, written in C and using a Windows 2000 platform (C/C⫹⫹ 5.0, Microsoft Corporation, Redmond, WA, USA) enabled handoutlining of the ROIs. Each image was magnified by a factor of 2 during the outlining process to reduce eye fatigue and improve positioning fidelity of the mouse/cursor pointing system. Window and level settings were selectable in the custom software, but settings were typically close to a window of 400 and a level of 30. All outlining was performed by a single investigator (E.M.G., at the time, a fourth-year

483

medical student) trained to recognize the relevant organ boundaries by a board-certified radiologist specializing in abdominal imaging (J.P.M.). The outlining of more than 18,000 organ boundaries took place over a period of 9 months, and lengthy outlining sessions were avoided to reduce fatigue. Each of the solid abdominal organs and L1 were located and subsequently hand-outlined by using 10-pixel (⬃4 – 6 mm) long-line segments to trace anatomic boundaries. For visual clarity during the outlining procedure, the program was written to connect adjacent points with a colored line (a different color for each ROI). Although tracing the outline of the spleen (SP), right and left adrenals (RA and LA), and pancreas (PC) was straightforward, certain rules were used in the outlining of the liver (LV), right and left kidneys (RK and LK), and L1. With regard to the liver, the inferior vena cava was excluded from the outline, but the hepatic veins draining into the inferior vena cava were included because they were intraparenchymal. Further, the portal venous system was included in sections where it appeared intrinsic to the liver, but was not included on the sections where it was clearly seen extrinsic to the liver (i.e., where it might reasonably be surgically cut in a transplant or autopsy). The liver has several fissures that are visible on CT images. When the fissures opened to the abdominal cavity or were fairly large, they were excluded; otherwise, they remained as a part of the liver parenchyma. In the kidneys, the collecting system and vasculature were not traced, leaving only the cortex and medulla for volume calculations. Although volume changes in the kidneys can sometimes occur after the injection of iodinated contrast agent, the use of low-osmolar contrast agent (as is done at our institution) and the rapid imaging protocols were thought to reduce the influence of such changes, and these effects are almost certainly smaller than normal anatomic variation between individuals (even after correction for height and weight). We chose to use L1 as an anatomic landmark for several reasons: (a) early work on organ volume calculation, using cross-sectional imaging, found that normalizing data to indices based on L1 account for body habitus [9, 10]; (b) L1 is easily identifiable by human observers and is likely to be of only moderate difficulty to locate automatically; (c) variation in the orientation of L1 has little effect on area and diameter (e.g., a 10-degree change would lead to a 1.5% difference in area); and (d) almost all abdominal CT studies include L1. We chose to circumscribe L1 with a dorsal cutoff through the pedicles at the widest diameter of the spinal canal, a highly reproducible method. Table 1 lists the organs studied in this project and provides a key for abbreviations used. The trends in organ volume as a function of age were assessed with linear regression. Compared with body mass and height, organ volume was found to have minor correlations (0.44 ⬎ r ⬎ 0.08) with age. A minimum of data correction was sought to increase the utility of the data

484

E. M. Geraghty et al.: Organ volume assessment from CT

Table 1. Organs and their abbreviations RK LK L1 LV LA RA SP PC

Right kidney Left kidney Lumbar vertebra 1 Liver Left adrenal Right adrenal Spleen Pancreas

compiled. Even though corrections for height and weight of the patient seemed obvious in light of the wide range in patient size, age dependency was determined to be a much smaller effect. Therefore, no age corrections were performed on these data.

Volume calculation The volume calculation for the ROIs was implemented from the boundary data. The individual boundary points correspond to individual pixels in the image, with each point spaced approximately 10 pixels apart. Software was written which summed the number of pixels inside the outline boundaries. Single pixel area (s2) was computed from the known pixel width, s. The organ area (cm2) was computed from each outline as the product of the number of pixels (N) in the outline and the pixel area for that image. The volume (V) of an organ on a single section j was calculated as the product of the organ area and the CT section thickness (T, where Vj ⫽ Tj Nj sj2). The total volume (Vtotal) for each organ was computed by summing the volumes from each section that included that organ (Vtotal ⫽ 兺Vj).

Anthropometric measurements Previous research has shown that the volumes for many of the abdominal organs can be correlated to a person’s sex, height, and weight. Unfortunately, height and weight values were not available for a number of the subjects, even after careful review of their medical records. For these patients, a technique developed previously was used to estimate height and weight from ROI parameters measured on a single CT image. Additional ROIs were outlined for these patients, and predictive equations for each patient’s height and weight were used. These methods are described elsewhere [11]. Organ volume was found to be far less dependent on age than on height or weight; therefore, to keep the corrections to a minimum, age dependency of organ volume was not attempted.

Phantoms Organ volumes measured by imaging methods have been validated previously by techniques requiring surgical removal of the organ [4, 6, 7]. However, changes in blood volumes for in vivo versus ex vivo organs can lead to

inaccuracies when using this technique. To estimate the accuracy of our volume determinations, balloons with known volumes were scanned and measured. Five balloons of different shapes (spheres, tubes, and wiggly tubes) and sizes were filled with tap water to a volume close to the mean volume for each organ (adrenal, kidney, pancreas, spleen, and liver). Different amounts of iodine-based contrast agent were added to each balloon. All balloons were placed in a water-filled tub in a pseudo-anatomic manner. Balloons were scanned on both scanners used in this study for the accrual of patient images. A technique of 120 kVp and 300 mAs was used. The display field of view was 36 cm. Section thickness varied depending on which CT scanner was used. Balloons were scanned at 2.5 and 5 mm on the GE Lightspeed multislice scanner, and those imaged on the GE CT/I single-slice scanner were sectioned at 5 and 7 mm. Images were obtained helically and axially and were reconstructed according to the standard abdominal protocol that was used for acquisition of the patient images. After imaging, balloons were cut and opened into appropriately sized graduated cylinders to more accurately measure their volumes.

Intraobserver variability Intraobserver variability in outlining ROIs was studied. Five CT examinations were reevaluated and redundant ROIs were traced (by E.M.G.). For this experiment, we used the total body circumference at the level of L1. These data were used to assess the precision (reproducibility) of the manual outlining procedure.

Interobserver variability Hand-outlining of organs involves dexterity of the hand and the eye, and subjective decisions concerning the delineation of low-contrast edges also need to be made. To evaluate the role that interobserver variability has on volume determination, two observers (E.M.G. and J.P.M.) independently handoutlined each of eight abdominal organs on the same patient’s CT study. Comparisons were made between each observer’s calculated organ volumes, and the average differences were reported.

Statistical analysis All organ volume data analysis was performed independently by sex. To reduce the dependence of patient height and weight on organ volumes, multiple linear regression (single-value decomposition [12]) analysis was performed such that V measured ⫽ a ⫹ F ht ⫻ height ⫹ F wt weight International standards for body habitus were used [13], corresponding to a standard man (1.76 m, 73 kg) and a standard women (1.63 m, 60 kg). Using the height and weight dependencies established by multiple linear regres-

E. M. Geraghty et al.: Organ volume assessment from CT

485

Fig. 1. This panel of six CT images demonstrates the outlining of the abdominal organs. These images were acquired with a 7-mm section thickness. One image in the original contiguous data set was skipped between each of the images shown.

sion analysis (specifically the slopes Fht and Fwt), each patient’s organ volumes were corrected: V corrected,j ⫽ V measured,j ⫹ F ht (H std ⫺ H j) ⫹ F wt(W std ⫺ W j) where Hstd ⫽ 1.63 m and Wstd ⫽ 60.0 kg for women and Hstd ⫽ 1.76 m and Wstd ⫽ 73.0 kg for men. The j subscript refers to the jth patient. The corrected volumes were analyzed with statistical software (Sigma Stat, Jandel Scientific, Corte Madera, CA, USA), and the Kolmogorov-Smirnov test was used to determine normality at p ⬎ 0.05. Data sets that pass the Kolmogorov-Smirnov test are consistent with data patterns drawn from a normal (gaussian) distribution, so using a gaussian distribution to model these data is appropriate. Additional data analyses were performed with spreadsheet software and custom C-programs (Excel and Visual C/C⫹⫹ 5.0, Microsoft Corporation).

Results Typical results of the outlining procedure described in this report are shown in Figure 1. These six CT images demonstrate the position of the various organs, shown in color. The

sequence of images skips intervening images to demonstrate images over a greater extent in the patient; however, no images were skipped in the actual organ evaluations. The accuracy of organ volume determination was determined by comparing the CT-based volume calculations with the known volume of balloon phantoms. Figure 2 illustrates volume measured by CT as a function of actual volume (liquid measure), and all measurement intervals fall along the line of identity with the exception of the smallest volume point (⬃5 mL). Each point on the figure is a comparison of balloon volumes, but the balloons were filled to volume levels very similar in magnitude to the mean organ volumes; to reflect this, each point shown in Figure 2 is labeled with the corresponding organ name. With the exception of the adrenal gland, all points were statistically indistinguishable from the line of identity (intercept ⫽ 0, slope ⫽ 1, p ⬍ 0.05). This demonstrates that CT volume measurement is fundamentally accurate for organs with volumes larger than approximately 10 mL. Intraobserver variability was determined by having the same person (E.M.G.) outline the same body silhouette five times. The difference in the five outlines was computed as the ratio of points outside both boundaries to the number of

486

Fig. 2. The volume of liquid-filled balloons as measured by CT is compared against the “actual” volume of balloons as determined by liquid measure. All data points fall along the line of identity except for the adrenal data at 5 mL. The error bars correspond to the 95% confidence limits (⫾ 2 standard deviations from the mean). For volumes above approximately 10 mL, CT-based volume determination is demonstrated to be quite accurate.

pixels inside both boundaries. Averaged over five trials, the mean intraobserver variation was found to be 1.19%. Interobserver variability was determined by having two observers (E.M.G. and J.P.M.) outline each organ on every CT image that it appeared in for one patient. This produced 110 organ boundaries for the eight organs listed in Table 1. Organ volumes were determined as usual, and the volume differences were evaluated as (V1 ⫺ V2)/V1, where V1 and V2 were the volumes calculated by observers 1 and 2, respectively, and V1 ⬎ V2. Figure 3 illustrates the interobserver variability for each organ, expressed in percentages. With the exception of LA and RA, all values were less than 5%. Not surprisingly, slight correlation between interobserver variability and organ volume was found (r ⫽ 0.474; with the very large liver volume removed, r ⫽ 0.834). This positive correlation suggests that small measurement errors in volumes are amplified for smaller organs when calculated as a percentage, because the value in the denominator is smaller. The mean and standard deviation were calculated for each set of organ volumes, and these parameters were used to compute the cumulative normal probability distribution. The parameterized and measured cumulative probabilities are shown in Figure 4. Although the measured data for the smaller organs (L1, LA, and RA) were not strictly normal (p ⬍ 0.05) based on the Kolmogorov-Smirnov test, these data are shown for completeness. It is apparent from Figure 4A–D that the parameterized and measured results for the kidneys, spleen, and pancreas are very consistent with each other. In Figure 4E, the liver also shows good correspondence between the measured and parameterized cumulative

E. M. Geraghty et al.: Organ volume assessment from CT

Fig. 3. The interobserver variability in organ volume assessment is shown for each organ. Refer to Table 1 for organ abbreviations. With the exception of the very small adrenals, the interobserver differences in volume determination were less than 5% and averaged 3.1%.

distributions. Although the data for L1 (Fig. 4F) and the adrenals (Fig. 4G,H) were not strictly normal, the trends in the measured data are similar to those in the parameterized curves. The mean and standard deviation of volume and the number of organs measured are given in Table 2 for each organ and sex. The correction factors Fht and Fwt also are reported in the table. The technique required to standardize organ volumes by height and weight also is summarized in Table 2. Tables of cumulative probabilities as a function of organ volume are given for women (Table 3) and men (Table 4). These tables allow the reader to find the organ and sex of interest and then determine how a specific patient’s organ volume compares with the normative data measured in this study. As an example of how to use Tables 3 and 4, a woman has a CT scan and the radiologist suspects splenomegaly. To quantify the degree of splenic enlargement, the patient’s splenic volume is determined manually from the CT data and is found to be 290.0 mL (Vmeasured). The women is 5 ft 7 in. (1.70 m) and weighs 130 lb (59.1 kg). By applying the corrections with the equation and coefficients in Table 2, her corrected splenic volume is computed as 290.8 mL, where 290.8 ⫽ 290.0 ⫹ 8.18749 (1.63 ⫺ 1.70) ⫹ 0.18945 (60.0 ⫺ 59.1). Looking at Table 3 under the SP column, the patient’s 290.8-mL splenic volume is between the 95th percentile (287.8-mL) and the 96th percentile (294.8-mL) for women. Thus, even after accounting for body size, this women’s splenic volume is in the top 5th percentile of human female spleens (said differently, this value is at the 95th percentile for normal female spleen volume).

E. M. Geraghty et al.: Organ volume assessment from CT

487

Fig. 4. The cumulative normal distributions for organ volume is shown for female (squares) and male (circles) patients. The individual data points (squares and circles) correspond to the height and weight-corrected organ volumes measured in this study, and the solid lines correspond to the parameterized data (i.e., that computed from the mean and standard deviation). The organ volume corresponding to a cumulative probability of 0.5 corresponds to the median organ volume, whereas values below 0.05 are in the bottom 5% and volumes above a cumulative probability of 0.95 correspond to the top 5%. Data are shown individually for (A) left kidney, (B) right kidney, (C) spleen, (D) pancreas, (E) liver, (F) lumbar vertebra 1, (G) left adrenal, and (H) right adrenal.

488

E. M. Geraghty et al.: Organ volume assessment from CT

Table 2. Overall volume data by organ and sexa

Table 4. Volume as a function of CP for normal malesa

Organ

Sex

N

Mean volume

Standard deviation

Fht

Fwt

CP

LK (mL)

RK (mL)

SP (mL)

LV (mL)

PC (mL)

L1b (mL)

LAb (mL)

RAb (mL)

L1 L1 LA LA LK LK LV LV PC PC RA RA RK RK SP SP

F M F M F M F M F M F M F M F M

42 59 50 52 31 35 36 41 46 57 46 52 38 45 34 47

24.6 32.3 4.4 5.7 160.3 201.0 1411.0 1710.4 64.4 87.4 4.9 5.7 152.7 185.2 179.8 238.4

4.4 6.8 1.6 4.9 32.7 28.9 263.5 288.3 18.1 21.3 1.9 1.9 33.0 32.3 66.2 69.7

64.89254 ⫺15.23569 21.08845 19.74152 26.08377 31.32370 375.12964 ⫺400.49728 ⫺2.93459 11.00086 ⫺1.07643 ⫺2.02770 13.56117 269.49951 8.18749 ⫺5.68431

1.14372 0.53020 1.42115 1.03973 0.050170 ⫺0.01647 14.82575 12.81355 0.06214 0.00744 0.05466 0.02358 1.76299 0.16365 0.18945 0.49670

0.01 0.02 0.03 0.04 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50c 0.55 0.60 0.65 0.70 0.75 0.80 0.85 0.90 0.95 0.96 0.97 0.98 0.99

133.8 141.7 146.7 150.4 153.5 164.0 171.0 176.7 181.5 185.8 189.9 193.7 197.4 201.0c 204.6 208.3 212.2 216.2 220.5 225.4 231.0 238.1 248.8 251.9 255.7 260.6 268.1

110.2 118.9 124.5 128.7 132.1 143.8 151.7 158.0 163.4 168.2 172.7 177.0 181.1 185.1c 189.2 193.3 197.6 202.0 206.9 212.2 218.5 226.3 237.9 241.3 245.4 250.7 259.1

76.2 95.1 107.1 116.1 123.5 148.9 166.0 179.6 191.3 201.8 211.5 220.7 229.6 238.3c 247.1 256.0 265.2 274.9 285.3 297.0 310.5 327.6 352.8 360.1 369.1 381.0 399.5

1038.2 1117.5 1167.6 1205.2 1235.8 1340.7 1411.4 1467.6 1515.8 1559.1 1599.2 1637.2 1674.0 1710.2c 1746.3 1783.2 1821.2 1861.2 1904.4 1952.5 2008.6 2079.0 2183.1 2213.3 2250.3 2299.0 2375.1

37.9 43.7 47.4 50.2 52.4 60.1 65.4 69.5 73.1 76.3 79.2 82.0 84.8 87.4c 90.1 92.8 95.6 98.6 101.8 105.3 109.5 114.7 122.4 124.6 127.4 131.0 136.6

16.5 18.4 19.5 20.4 21.1 23.6 25.3 26.6 27.8 28.8 29.7 30.6 31.5 32.3c 33.2 34.1 35.0 35.9 36.9 38.1 39.4 41.1 43.5 44.2 45.1 46.3 48.1

1.3 1.8 2.2 2.4 2.6 3.3 3.8 4.1 4.5 4.8 5.0 5.3 5.5 5.7c 6.0 6.2 6.5 6.7 7.0 7.3 7.7 8.2 8.9 9.1 9.3 9.6 10.1

1.3 1.8 2.1 2.4 2.6 3.3 3.7 4.1 4.4 4.7 5.0 5.2 5.4 5.7c 5.9 6.2 6.4 6.7 7.0 7.3 7.7 8.1 8.8 9.0 9.3 9.6 10.1

a

Volume figures (in cm3) are shown after correction for patient height and weight. The correction factors Fht and Fwt are also given in the table, where organ size is corrected to a standard body size with the following formula: volumecorrectedj ⫽ volumemeasuredj ⫹ Fht (Hstd ⫺ Hj) ⫹ Fwt (Wstd ⫺ Wj) where Hstd is the standard height (1.76 m for males, 1.63 m for females), Hj is a specific patient’s (patient j) actual height (m), Wstd is the standard weight (73.0 kg for males, 60.0 kg for females), and Wj is patient j’s actual weight (kg). For other abbreviations, see Table 1

a

Table 3. Volume as a function of CP for normal femalesa CP

LK (mL)

RK (mL)

SP (mL)

LV (mL)

PC (mL)

L1b (mL)

LAb (mL)

RAb (mL)

0.01 0.02 0.03 0.04 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50c 0.55 0.60 0.65 0.70 0.75 0.80 0.85 0.90 0.95 0.96 0.97 0.98 0.99

84.4 93.3 99.0 103.3 106.7 118.7 126.7 133.1 138.5 143.3 147.8 152.1 156.3 160.4c 164.5 168.6 172.9 177.5 182.4 187.8 194.2 202.2 213.9 217.4 221.5 227.1 235.6

75.8 84.8 90.5 94.7 98.2 110.1 118.2 124.6 130.1 135.1 139.7 144.1 148.3 152.5c 156.7 160.9 165.2 169.8 174.8 180.3 186.7 194.8 206.7 210.2 214.4 220.0 228.7

26.0 44.0 55.5 64.1 71.1 95.2 111.5 124.4 135.4 145.3 154.4 163.1 171.5 179.8c 188.0 196.4 205.1 214.2 224.1 235.0 247.7 263.8 287.8 294.8 303.3 314.6 332.1

796.2 867.7 913.0 947.2 974.9 1071.1 1136.3 1187.9 1232.2 1271.9 1308.6 1343.5 1377.2 1410.3c 1443.5 1477.1 1511.9 1548.6 1588.1 1632.1 1683.4 1747.8 1843.0 1870.7 1904.6 1949.2 2019.1

22.4 27.3 30.5 32.8 34.7 41.2 45.6 49.1 52.1 54.8 57.3 59.6 61.9 64.2c 66.5 68.8 71.2 73.7 76.4 79.5 83.0 87.4 94.0 95.9 98.2 101.3 106.1

14.4 15.6 16.4 16.9 17.4 19.0 20.1 20.9 21.7 22.3 22.9 23.5 24.1 24.6c 25.2 25.8 26.3 26.9 27.6 28.3 29.2 30.2 31.8 32.3 32.8 33.6 34.8

0.7 1.1 1.4 1.6 1.8 2.4 2.8 3.1 3.3 3.6 3.8 4.0 4.2 4.4c 4.6 4.8 5.0 5.3 5.5 5.8 6.1 6.5 7.0 7.2 7.4 7.7 8.1

0.4 0.9 1.2 1.5 1.7 2.4 2.9 3.2 3.6 3.9 4.1 4.4 4.6 4.9c 5.1 5.4 5.6 5.9 6.2 6.5 6.9 7.4 8.1 8.3 8.5 8.9 9.4

a

Organ volumes must be corrected as described in Table 2 to use the data in this table. A volume at a CP of 0.05 is at the lowest 5th percentile in volume, and a CP of 0.98 indicates that the volume is at the 98th percentile for that organ. CP, cumulative probability. For other abbreviations, see Table 1 b Provided for thoroughness but was not found to be a normal distribution c Median values

Organ volumes must be corrected as described in Table 2 to use the data in this table. A volume at a CP of 0.05 is at the lowest 5th percentile in volume, and a CP of 0.98 indicates that the volume is at the 98th percentile for that organ. CP, cumulative probability. For other abbreviations, see Table 1 b Provided for thoroughness but was not found to be a normal distribution c Median values

As this example demonstrates, the correction for height and weight usually results in a small change in organ volume. Corrections for patients who are quite thin or heavy will show a larger change.

Discussion The current diagnostic utility of CT images relies primarily on subjective interpretation, but as CT systems advance and as the number of CT examinations increase, the quantitative nature of this modality may play an ever-increasing role in improving diagnostic accuracy [2, 4, 5, 7, 9, 10]. Although very labor-intensive hand-outlining techniques were used in this current study (⬎18,000 organ boundaries were handoutlined), computer automation of this procedure is being developed. Computer-aided diagnostic techniques abound in other modalities such as thoracic imaging [14 –16] and mammography [17–19], and it is only a matter of time before computer-aided diagnostic techniques become commonplace for CT imaging of the abdomen. The automated detection of organ boundaries is all that is needed to compute organ volume; however, once the boundaries of an organ are identified robustly and automatically, other pattern recognition software can be developed for organ-specific diagnosis.

E. M. Geraghty et al.: Organ volume assessment from CT

The size of one person’s organ relative to another person’s is dependent on a number of factors, but, in general, a 2-m 100-kg person will likely have larger organs than a 1.5-m 50-kg individual. We describe a straightforward technique to correct for organ volume dependence due to a person’s height and weight. This approach allowed organ volumes from a number of patients, spanning a wide range of body sizes, to be combined and evaluated together. Averaged over the seven soft tissue organs that were the focus of this investigation, corrections for height and weight reduced the variance by 24.3% (range, 5.8 –39.0%) as compared with the uncorrected organ volume data. Sex differences were dealt with explicitly, by simply analyzing male and female data separately. CT is considered the most accurate imaging technique for the in vivo evaluation of organ volume. The physical properties associated with CT lead to high-resolution image data sets that are spatially accurate. Although abdominal ultrasound imaging is very useful, quantitative volume accuracy is compromised by the manual scanning procedure and speed of sound inaccuracies. Magnetic resonance imaging is capable of producing high-resolution volume data sets, but magnetic field inhomogeneities coupled with chemical shift phenomenon can induce spatial inhomogeneities in the image data. The degree to which these affect volume computation accuracy is not well understood. CT has been used as the gold standard in most volume assessment studies [20 – 22]. The degree of CT volume measurement accuracy was assessed in this study (Fig. 2) and provides further evidence that CT is an accurate tool for volume assessment. Comprehensive data on normal organ volumes for the abdomen do not exist in the imaging literature. In 1981 Henderson et al. [21] used CT on 11 normal subjects and found a mean liver volume of 1493 mL and a mean spleen volume of 219 mL. Their reported liver volume has an approximately 0.25 ␴, or 4.3% difference, from the mean liver volume for males and females determined in the present study (⬃1560 mL), and their spleen volume is within approximately 0.15 ␴, or 4.6% difference, of the sex-averaged values determined in this study (⬃209 ml). Hoefs et al. [8] used CT and found a normal spleen volume of 201 mL in 11 normal patients, and this is within 0.12 ␴ (3.8% difference) of the sex-averaged spleen volume reported here. Prassopoulos et al. [2] reported a mean spleen volume of 215 mL for 140 patients, very close (a difference of 0.09 ␴, or 2.8%) to the sex-averaged 209 mL determined in this study. Kaneko et al. [23] reported an average splenic volume of 112 mL in 150 healthy volunteers in Japan, and the large difference (46.4%) between this value and the others discussed above is probably due to differences in body size of the Japanese population from that in North America. Schulz et al. [24] studied organ volumes in 1986 and found mean volumes of 1331 mL for the liver, 169 mL for the spleen, and 40.4 mL for the pancreas. These values are consistently smaller than those reported here, by 14.7%, 19.2%, and 46.8%, respectively, and the differences may be due to partial volume

489

effects (notice the difference increases as volume decreases) and perhaps regional variations between the North American and German populations studied. Bone mineral density (BMD) measurement has become a standard metric for evaluating osteoporosis and for subsequently estimating the risk of fracture due to bone strength degradation [25, 26]. The scientific basis of BMD analysis is very much related to the availability of a large normative database for age-matched BMD comparisons. This study was performed to place organ volume assessment on a similar quantitative foundation to that which is used in BMD analysis. One difference is that BMD assessment is used to predict the future risk of pathology; however, organ volume assessment can be used to assist physicians in the assessment and diagnosis of present pathologic conditions. We note that, after further study, it is likely that quantitative data may lead to earlier diagnoses of more subtle disease and that volume data from one or more organs may play a role in the diagnoses. This study was performed to determine what the range of normal organ volumes is, and this sets the stage for future analyses of pathologic conditions and how they may alter organ volume in some patients. The cumulative normal distributions for solid abdominal organ volumes, corrected for patient height and weight, were evaluated; these data will allow physicians to estimate the relative degree of organ atrophy or enlargement for specific patients. Acknowledgments. We acknowledge the generous support of a research fellowship from the University of California, Davis, School of Medicine. Much of the research performed in this project was relevant to other CT-related topics, and we acknowledge partial support for this work from the National Cancer Institute (CA 89260) and the California Breast Cancer Research Program (7EB-0075).

References 1. Lin XZ, Sun YN, Liu YH, et al. (1998) Liver volume in patients with or without chronic liver diseases. Hepatogastroenterology 45:1069 – 1074 2. Prassopoulos P, Daskalogiannaki M, Raissaki M, et al. (1997) Determination of normal splenic volume on computed tomography in relation to age, gender and body habitus. Eur Radiol 7:246 –248 3. Urata K, Kawasaki S, Matsunami H, et al. (1995) Calculation of child and adult standard liver volume for liver transplantation. Hepatology 21:1317–1321 4. Schiano TD, Bodian C, Schwartz ME, et al. (2000) Accuracy and significance of computed tomographic scan assessment of hepatic volume in patients undergoing liver transplantation. Transplantation 69: 545–550 5. Cools L, Osteaux M, Divano L, Jeanmart L (1983) Prediction of splenic volume by a simple CT measurement: a statistical study. J Comput Assist Tomogr 7:426 – 430 6. Breiman RS, Beck JW, Korobkin M, et al. (1982) Volume determinations using computed tomography. AJR 138:329 –333 7. Moss AA, Friedman MA, Brito AC (1981) Determination of liver, kidney, and spleen volumes by computed tomography: an experimental study in dogs. J Comput Assist Tomogr 5:12–14 8. Hoefs JC, Wang FW, Lilien DL, et al. (1999) A novel, simple method of functional spleen volume calculation by liver–spleen scan. J Nucl Med 40:1745–1755 9. Heuck A, Maubach PA, Reiser M, et al. (1987) Age-related morphology of the normal pancreas on computed tomography. Gastrointest Radiol 12:18 –22 10. Gourtsoyiannis N, Prassopoulos P, Cavouras D, Pantelidis N (1990)

490

11. 12. 13.

14.

15. 16. 17. 18.

The thickness of the renal parenchyma decreases with age: a CT study of 360 patients. AJR 155:541–544 Geraghty EM, Boone JM (2003) Determination of height, weight, body mass index and body surface area from a single abdominal CT image. Radiology 228:857– 863 Press WH, Flannery BP, Teukolsky SA, Vetterling WT. Numerical recipes in C: the art of scientific computing. New York: Cambridge University Press, 1988 REM Task Group IC2. Draft ICRP report: basic anatomical and physiological data for use in radiological protection: reference values. Stockholm: International Commission on Radiological Protection, 1 January 2002 Matsuki Y, Nakamura K, Watanabe H, et al. (2002) Usefulness of an artificial neural network for differentiating benign from malignant pulmonary nodules on high-resolution CT: evaluation with receiver operating characteristic analysis. AJR 178:657– 663 Garg K (2002) CT of pulmonary thromboembolic disease. Radiol Clin North Am 40:111–122 Li Q, Katsuragawa S, Doi K (2001) Computer-aided diagnostic scheme for lung nodule detection in digital chest radiographs by use of a multiple-template matching technique. Med Phys 28:2070 –2076 Li L, Clark RA, Thomas JA (2002) Computer-aided diagnosis of masses with full-field digital mammography. Acad Radiol 9:4 –12 Paquerault S, Petrick N, Chan HP, et al. (2002) Improvement of computerized mass detection on mammograms: fusion of two-view information. Med Phys 29:238 –247

E. M. Geraghty et al.: Organ volume assessment from CT

19. Jiang Y, Nishikawa RM, Schmidt RA, et al. (2001) Potential of computeraided diagnosis to reduce variability in radiologists’ interpretations of mammograms depicting microcalcifications. Radiology 220:787–794 20. Lamb PM, Lund A, Kanagasabay RR, et al. (2002) Spleen size: how well do linear ultrasound measurements correlate with three-dimensional CT volume assessments?. Br J Radiol 75:573–577 21. Henderson JM, Heymsfield SB, Horowitz J, Kutner MH (1981) Measurement of liver and spleen volume by computed tomography. Assessment of reproducibility and changes found following a selective distal splenorenal shunt. Radiology 141:525–527 22. Hirose T, Inomoto T, Awane M, et al. (1998) Direct measurement of graft and recipient liver fossa size by computed tomography for avoiding problems due (correction of clue) to large graft size in living-related liver transplantation. Clin Transplant 12:49 –55 23. Kaneko J, Sugawara Y, Matsui Y, et al. (2002) Normal splenic volume in adults by computed tomography. Hepatogastroenterology 49:1726 – 1727 24. Schulz HG, Christou A, Gursky S, Rother P (1986) [Computerized tomography studies of normal morphology and volumetry of parenchymatous epigastric organs in humans]. Anat Anz 162:1–12 25. Brunader R, Shelton DK (2002) Radiologic bone assessment in the evaluation of osteoporosis. Am Fam Phys 65:1357–1364 26. Henry MJ, Pasco JA, Seeman E, et al. (2001) Assessment of fracture risk: value of random population-based samples—the Geelong Osteoporosis Study. J Clin Densitom 4:283–289