Published online before print July 20, 2006, 10.1148/radiol.2403050529
(Radiology 2006;240:821-827.)
© RSNA, 2006
High-Frequency Sonographic Patterns of the Spleen in Children1
Andrea S. Doria, MD, MSc, PhD,
Alan Daneman, MD,
Rahim Moineddin, PhD,
Charles R. Smith, MD,
Arun Mohanta, MRT,
Jonathan Clarke, MRT and
Christian J. Kellenberger, MD2
1 From the Departments of Diagnostic Imaging (A.S.D., A.D., A.M., J.C., C.J.K.) and Paediatric Laboratory Medicine (C.R.S.), the Hospital for Sick Children, 555 University Ave, Toronto, ON, Canada M5G 1X8; and Department of Public Health, Family and Community Medicine, University of Toronto, Toronto, Ontario, Canada (R.M.). From the 2001 RSNA Annual Meeting. Received March 31, 2005; revision requested June 2; revision received June 22; accepted July 11; final version accepted November 2.
Address correspondence to A.D. (e-mail: alan.daneman{at}utoronto.ca).
 |
ABSTRACT
|
|---|
Purpose: To retrospectively evaluate the normal sonographic appearance of splenic parenchyma in children of various ages with high-frequency transducer sonography.
Materials and Methods: Research ethics committee board approval was obtained, with waiver of informed consent. We evaluated in vivo sagittal and transverse sonograms of spleens obtained with 13-MHz linear-array transducers in 100 children (age range, 1 day to 17 years) with clinically diagnosed disease that did not affect the lymphoid system. Three radiologists working in consensus defined the grading system for the splenic parenchyma. Thereafter, these same radiologists graded the sonographic reticulonodular pattern independently as granular, mild, or marked. These grades were cross correlated with clinical data by using logistic regression analysis and 
2 tests. Sonographic measurements of the splenic parenchyma in nine pediatric cadavers separate from the in vivo cohort of the study were compared with those of corresponding histologic slices by means of linear correlation.
Results: Both grade 2 and grade 3 patterns occurred more frequently in children older than 1 year but no older than 5 years, whereas grade 1 pattern occurred most frequently in neonates (P < .001). As patients' age (odds ratio, 1.6; P < .001) and splenic dimensions (odds ratio, 3.1; P < .001) increased, the frequency with which the reticulonodular pattern was classified as either grade 2 or grade 3 increased. No association was noted between sonographic patterns and body mass index (P = .85) or sex (P = .07). The parenchymal nodules graded as 2 or 3 on sonograms correlated well with the presence of lymphoid follicles (white pulp) at histologic analysis (r = .71, P = .03).
Conclusion: High-frequency transducer sonography of the spleen in children can demonstrate normal echo patterns that should not be misinterpreted as indicative of disease.
© RSNA, 2006
|
INTRODUCTION
|
|---|
The sonographic appearance of normal splenic parenchyma in children (1) and adults (2,3) has been described as acoustically homogeneous. However, the development of broad-band linear-array high-frequency sonographic transducers has enabled the detection of small structures in the parenchyma of solid viscera that could not be confidently identified with lower-frequency transducers (4,5). Normal changes in the magnetic resonance (MR) imaging signal intensity of splenic parenchyma have been demonstrated during early childhood because of an increase in the ratio of white pulp to red pulp, as proved with histologic examinations (6). We hypothesized that these ultrastructural changes could be depicted with high-frequency (10–13-MHz) sonography on the basis of results from previous studies in which this technique was used to evaluate the ultrastructural anatomy of the normal thymus in infants (7) and of inflamed testes in the overall population (8). Thus, the purpose of our study was to retrospectively evaluate the normal sonographic appearance of splenic parenchyma in children of various ages with high-frequency transducer sonography.
|
MATERIALS AND METHODS
|
|---|
This study was approved by the research ethics committee board at our institution. The board waived the informed consent requirement for the in vivo and postmortem aspects of our retrospective study.
In Vivo Sonography and Sonographic Pattern Analysis
The 100 consecutive children included in this study underwent abdominal sonographic examinations at the Hospital for Sick Children in February 2001 and did not have conditions that affected the lymphoid system or otherwise altered splenic echotexture. We excluded patients with (a) lymphoproliferative disorders, immunodeficiency, or hepatocellular disease or (b) evidence of active infection or previous blood transfusion, chemotherapy, or antigenic therapy.
Examinations were performed for diagnostic purposes: Patients had urogenital (n = 58) or abdominal (n = 18) symptoms; neurologic (n = 4), endocrinologic (n = 3), gynecologic (n = 2), or rheumatologic (n = 3) disorders; cardiac anomalies (n = 3); or soft-tissue abnormalities (n = 3). The remaining six patients were suspected of having congenital malformations. Patients ranged in age from 1 day to 17 years 6 months (median age, 0.9 years ± 4.3 [standard deviation]). Forty-one (41%) girls were included. Fifty-eight (58%) patients were 2 years of age or younger and 42 (42%) patients were older than 2 years. Of the patients 2 years of age or younger, 24 (24%) were no older than 1 month, 22 (22%) were older than 1 month but younger than 6 months, and 12 (12%) were older than 6 months but no older than 2 years. Body mass index (BMI) data were available for 26 patients; BMI ranged from 0.001 to 0.004 kg/m2 (median, 0.002 kg/m2 ± 0.001).
Sagittal and transverse scans of the splenic parenchyma were obtained with a 13-MHz linear-array transducer (Sequoia Systems; Acuson, Mountain View, Calif), gray-scale gains of 85 dB (range, –3 to 10 dB), and magnification parameters that ranged from 0 to 11 mm and were adjusted for patient habitus. All sonograms were obtained by technologists with the supervision of radiologists.
Three pediatric radiologists (A.S.D. and C.J.K., who had more than 5 years of experience with sonography; A.D., who had approximately 20 years of experience with sonography) reviewed digital data of the in vivo sonographic examinations in consensus. Sonographic patterns in the spleen were visually graded according to the ability of the reader to identify hypoechoic nodules of various sizes within the splenic parenchyma, as follows: grade 1, granular pattern; grade 2, mild pattern; and grade 3, marked reticulonodular pattern (Fig 1). The granular pattern showed a homogeneous echotexture of the parenchyma without any evidence of focal nodules that could be identified with the naked eye. The reticulonodular pattern represented a heterogeneous appearance of the splenic parenchyma, and nodules were recognized with the naked eye. The criterion for grading reticulonodular patterns as mild or marked was subjective and based on a priori characterization of these patterns. Reticulonodular patterns were classified as marked if multiple hypoechoic nodules were distributed throughout the splenic parenchyma. These nodules tended to be well defined and at least 1 mm in diameter. The mild reticulonodular pattern demonstrated smaller (<1 mm) ill-defined nodules, which were recognized with some degree of difficulty and were present in part of the parenchyma.
View larger version (84K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1a: Normal high-frequency sagittal gray-scale sonograms of the spleen show (a) granular pattern of echogenicity in splenic parenchyma of 4-year-old girl, (b) mild reticulonodular pattern in 5-year 10-month-old boy, and (c) a marked reticulonodular pattern in 2-year 4-month-old girl.
|
|
View larger version (87K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1b: Normal high-frequency sagittal gray-scale sonograms of the spleen show (a) granular pattern of echogenicity in splenic parenchyma of 4-year-old girl, (b) mild reticulonodular pattern in 5-year 10-month-old boy, and (c) a marked reticulonodular pattern in 2-year 4-month-old girl.
|
|
View larger version (74K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1c: Normal high-frequency sagittal gray-scale sonograms of the spleen show (a) granular pattern of echogenicity in splenic parenchyma of 4-year-old girl, (b) mild reticulonodular pattern in 5-year 10-month-old boy, and (c) a marked reticulonodular pattern in 2-year 4-month-old girl.
|
|
The three radiologists reviewed and graded the sonograms again, now independently. This time, images were reviewed in random order, and radiologists were blinded to patients' information. In this independent review, grading agreement between at least two readers was achieved for all images; thus, further consensus review was not performed. The final sonographic patterns that were used for statistical analysis were based on the agreement of at least two readers. The statistical analysis for agreement between readers was based on the independent review of the scans by the readers. One of the readers (A.S.D.) graded the sonographic patterns of the data twice and evaluated a random stack of images at each time within a 2-month interval.
Sonographic Patterns and Clinical Data
Information on the recorded sonographic patterns was compared with clinical data (age, sex, longitudinal diameter of the spleen, and BMI). BMI data were available in only 26 patients.
Comparison with Histologic Findings
Technologists used sonography to evaluate selected spleens from nine cadavers of children (age range, 1 day to 14 years) who were not a part of the in vivo cohort of the study before autopsy specimens had been fixed with formalin. Death was caused by sudden infant death syndrome (n = 1), traumatic injury that did not involve the spleen (n = 1), pulmonary (n = 1) or intracranial (n = 1) hemorrhage, congestive heart failure (n = 3), or failure of multiple organs associated with dysmorphism (n = 2). The splenic specimens were examined with the same protocol that was used for in vivo sonography.
Interpretation of Postmortem Sonograms
Once sonograms of the nine cadaveric spleens were obtained, one radiologist (A.S.D.) recorded the distance between the center of two contiguous hypoechoic nodules, when visible, in 10 random locations of the splenic parenchyma for each postmortem study. This reader was blinded to patients' clinical information. The mean value of these measurements in each study was used for data analysis. One pathologist (C.R.S.) with more than 15 years of experience in splenic pathology assessed the spatial distribution of both the white pulp and the red pulp in 4-µm-thick histologic slices of corresponding splenic specimens by measuring the distance between the center of two contiguous lymphoid nodules representing white pulp and two contiguous cross-sectional vascular areas representing red pulp in 10 random locations of the splenic parenchyma. The pathologist was blinded to the sonographic results and to the patients' clinical information. Hematoxylin-eosin staining was used to prepare the tissue slices. An ocular micrometer was used for microscopic assessment of the histologic slices of splenic tissue with low-power magnification fields (original magnification, x4).
Statistical Analysis
Associations between sonographic patterns and patients' clinical characteristics (age, sex, and BMI) or sonographic measurements of the longitudinal axis of patients' spleens were evaluated with piecewise multinomial logistic regression analysis (dependent variable, grade 1 pattern vs grade 2 or grade 3 patterns; independent variable, age as continuous data). Odds ratios were used for parametric estimates. The distribution of sonographic patterns according to patients' sex and age (no older than 1 month, older than 1 month but no older than 1 year, older than 1 year but no older than 5 years, older than 5 years but no older than 10 years, and older than 10 years) was evaluated with 2 tests. We used log-transformed data to express the patients' age for logistic regression analysis because the age distribution of the population in this study was greatly skewed to the left and the log-transformation procedure would minimize the skewness and sparseness of the design matrix in the regression equation.
The level of agreement within and among readers was determined with 
values, which provided a measure of intra- and interreader concordance, and were adjusted for the chance of agreement. A value of 0.40 or less indicated poor agreement, a value of more than 0.40 but of 0.60 or less indicated moderate agreement, a value of more than 0.60 but of 0.80 or less indicated good agreement, and a value of more than 0.80 indicated very good agreement (9). P values and standard 95% confidence intervals (CIs) were calculated to determine whether values differed significantly (10).
The relationship of average distances between contiguous sonographic hypoechoic nodules and between contiguous lymphoid nodules or sections of blood vessels on corresponding postmortem histologic slices was analyzed with linear correlation.
Statistical analysis was performed with a software package (SAS, version 8.2; SAS Institute, Cary, NC). A P value of less than .05 was considered to indicate a statistically significant difference.
|
RESULTS
|
|---|
Sonographic Patterns and Clinical Data
The sonographic patterns of the splenic parenchyma (Fig 2) differed between age groups (P < .001). Both grade 2 and grade 3 patterns occurred more frequently in children older than 1 year but no older than 5 years, whereas the grade 1 pattern occurred more frequently in neonates. When age was considered as a continuous variable in our population, the odds that a reader classified a pattern as either grade 2 or grade 3 rather than as grade 1 for a sonogram of the splenic parenchyma increased 1.5 (95% CI: 1.3, 1.9) times with an age increase of 1 year on a log scale (P < .001).
View larger version (23K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2: Bar graphs show distribution of visual classification for in vivo sonographic patterns of splenic parenchyma in patients. Sonographic patterns were classified as grade 1 (granular pattern) in 18 (75%) of 24 neonates and as either grade 2 or grade 3 (mild or prominent reticulonodular pattern) in 23 (85%) of 27 children older than 1 year but younger than 5 years. Within the first 5 years of life, there is an inverse relationship between frequency of grade 1, 2, or 3 patterns and age: The frequency of grade 1 patterns gradually decreases, and that of grade 3 patterns increases as age range of the groups increases.
|
|
In this study, spleen size ranged from 3.3 to 11.5 cm (median, 6.4 cm ± 2.2). Splenic scans received higher sonographic grades in larger spleens (Fig 3), which indicated that demonstration of the reticulonodular patterns of the parenchyma increased with the gradual increase in splenic size (P < .001). The odds for a reader to assign a grade to a splenic scan as either grade 2 or grade 3 rather than as grade 1 increased 3.1 (95% CI: 1.8, 5.6) times per centimeter increase in the size of the spleen on a log scale (P < .001). No association, however, was noted between sonographic patterns and BMI (P = .85) or sex (P = .07).
View larger version (9K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3: Scattergram shows association between visual classification of in vivo sonographic patterns of splenic parenchyma and longitudinal diameter of patients' spleens (P < .001 for grade 1 pattern vs grade 2 or grade 3 patterns, as calculated with logistic regression analysis).
|
|
Reader Agreement
The overall interreader concordance for grading of sonographic patterns among the three readers was moderate (0.46; 95% CI: 0.38, 0.54; P < .001). The coefficients were 0.44 (95% CI: 0.29, 0.59) between two readers (A.D., C.J.K.) (P < .001), 0.39 (95% CI: 0.25, 0.54) between two other readers (C.J.K., A.S.D.) (P < .001), and 0.53 (95% CI: 0.38, 0.68) between still two other readers (A.S.D., A.S.) (P < .001). In contrast with interreader rates, the intrareader (A.S.D.) concordance index was very good (0.83; 95% CI: 0.65, 1.00; P < .001).
Comparison with Histologic Analysis
For the postmortem sonographic studies, the median distance between hypoechoic nodules in the splenic parenchyma was 2.14 mm ± 0.24 (Fig 4a). In the postmortem splenic parenchymal specimens, the median distances between lymphoid follicles (white pulp) (Fig 4b) and vascular sections (red pulp) were 0.58 mm ± 0.11 and 1.01 mm ± 0.06, respectively. The internodular distances identified on sonograms correlated well with histologic measurements of inter–white pulp distances (ie, distance between two white pulp lymphoid nodules) (r = .71, P = .03) (Figs 4c, 5) but did not correlate well with histologic measurements of inter–red pulp distances (ie, distance between two red pulp blood vessels) (r = –.002, P = .99).
View larger version (99K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4a: (a) Sagittal gray-scale high-frequency sonogram of prefixed cadaveric spleen in a 6-year-old girl who died of intracranial hemorrhage shows 2-mm distances between hypoechoic nodules (eg, vertical white line [arrow]) in splenic parenchyma. (b) Macroscopic inspection of corresponding sagittal section of splenic parenchyma of cadaveric specimen shows white nodular regions that represent lymphoid nodules of white pulp scattered throughout brownish background mass of red pulp. Vertical black line (arrow) represents interlymphonodular distance. (c) Microscopic appearance of corresponding histologic slice best differentiates red (black arrow indicates a red color region) and white (red arrow indicates a blue color region) pulp of the spleen. Median distance between lymphoid follicles in parenchyma of the splenic specimens (dotted line) was 0.58 mm, which correlated well with internodular distances (r = 0.71, P = .03) measured with sonography on corresponding splenic specimen. (Hematoxylin-eosin stain; original magnification, x4.)
|
|
View larger version (132K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4b: (a) Sagittal gray-scale high-frequency sonogram of prefixed cadaveric spleen in a 6-year-old girl who died of intracranial hemorrhage shows 2-mm distances between hypoechoic nodules (eg, vertical white line [arrow]) in splenic parenchyma. (b) Macroscopic inspection of corresponding sagittal section of splenic parenchyma of cadaveric specimen shows white nodular regions that represent lymphoid nodules of white pulp scattered throughout brownish background mass of red pulp. Vertical black line (arrow) represents interlymphonodular distance. (c) Microscopic appearance of corresponding histologic slice best differentiates red (black arrow indicates a red color region) and white (red arrow indicates a blue color region) pulp of the spleen. Median distance between lymphoid follicles in parenchyma of the splenic specimens (dotted line) was 0.58 mm, which correlated well with internodular distances (r = 0.71, P = .03) measured with sonography on corresponding splenic specimen. (Hematoxylin-eosin stain; original magnification, x4.)
|
|
View larger version (156K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4c: (a) Sagittal gray-scale high-frequency sonogram of prefixed cadaveric spleen in a 6-year-old girl who died of intracranial hemorrhage shows 2-mm distances between hypoechoic nodules (eg, vertical white line [arrow]) in splenic parenchyma. (b) Macroscopic inspection of corresponding sagittal section of splenic parenchyma of cadaveric specimen shows white nodular regions that represent lymphoid nodules of white pulp scattered throughout brownish background mass of red pulp. Vertical black line (arrow) represents interlymphonodular distance. (c) Microscopic appearance of corresponding histologic slice best differentiates red (black arrow indicates a red color region) and white (red arrow indicates a blue color region) pulp of the spleen. Median distance between lymphoid follicles in parenchyma of the splenic specimens (dotted line) was 0.58 mm, which correlated well with internodular distances (r = 0.71, P = .03) measured with sonography on corresponding splenic specimen. (Hematoxylin-eosin stain; original magnification, x4.)
|
|
|
DISCUSSION
|
|---|
The spleen is an encapsulated lymphatic organ composed of a capsule and internal trabeculae. Arising from the capsule, the connective tissue trabeculae penetrate the parenchyma of the spleen and cause separation into many small compartments. Aggregations of B lymphocytes are visible to the naked eye at the cut surface of splenic specimens. They appear as small whitish nodules 0.25–1.00 mm in diameter, they are composed of lymphoid follicles or malpighian bodies, and they are collectively known as the white pulp. These nodules are interspersed within a larger mass of darker tissue that consists of vascular sinusoids and is referred to as red pulp (11).
The correlation of cadaveric histologic slices with sonograms of the splenic parenchyma suggests that the hypoechoic nodules seen on high-frequency sonograms may represent the lymphoid follicles of the white pulp. The red and white pulp network pattern, which is depicted as an echogenic background that encompasses the hypoechoic nodules that are immersed in it, was described as a reticulonodular pattern in this study. Although we found a strong relationship between the internodular distances measured on sonograms of the splenic parenchyma and the interlymphoid follicular distances measured on the microscopic tissue slices, the median interfollicular measurements obtained microscopically were smaller than the distance between nodules measured with sonography. This suggests that microscopy represents a more powerful tool for identification of white pulp than high-frequency sonography, thus enabling the identification of nodules that could not be depicted with 13-MHz transducers.
In this study, the neonates' spleens showed a preferential homogeneous appearance (grade 1 pattern) similar to that found in other imaging studies (6,12) that demonstrated the presence of small lymphoid follicles lacking active germinal centers in immature spleens. With time and immunologic challenges, the lymphoid system matures (12). In a previous study based on an analysis of the MR appearance of the spleen in early childhood, Donnelly et al (6) found that as age increased, the follicles grew larger and the ratio of white pulp to red pulp increased. This observation may justify our finding that the frequency and intensity of reticulonodular patterns increase as patient age increases (up to 5 years) because the size of hypoechoic nodules with the reticulonodular pattern representing the white pulp also increases with age.
Recognition of the reticulonodular pattern on sonograms would be more challenging in neonates. This is supported by our findings, which show that the reticulonodular pattern could not be identified in the spleens of most neonates. The maturation of the lymphoid germinal centers after the neonatal period would have facilitated their depiction with sonography in younger children. The thin intervening layer of fatty tissue found between the sonographic transducer and the splenic parenchyma of younger children during examination of the patient's anterolateral abdominal wall enabled identification of the reticulonodular pattern with high-frequency sonography in these patients. As age increases, this intervening layer becomes thicker, thus making transmission of the sound beam through the abdominal wall to the splenic parenchyma difficult. The penetration of the beam of high-frequency ultrasound transducers is gradually lost as the depth of the targeted object of interest increases, which changes the performance of sonography (13). This factor may limit the effectiveness of sonography when used to assess high-frequency patterns in obese children and adults.
In our study, there were only eight (8%) patients older than 10 years in whom we might have had difficulty identifying the reticulonodular pattern. Nevertheless, considering age as continuous data, differences in the sonographic patterns of the spleen in children in different age groups could not be appreciated, and the odds that a reader would assign a grade of 2 or 3 rather than a grade of 1 to splenic scans was significantly greater for each 1-year increase in patient age or for each 1-cm increase in the size of the spleen as measured on sonograms.
The sensitivity of sonography in the measurement of splenic volumes and the linear relationship that exists between patient age and longitudinal diameter of the spleen is well recognized (3,14,15). Our results show that sonographic patterns of the spleen were associated with patients' age and splenic dimensions but not with patients' sex or BMI. Further prospective investigation is required to evaluate the sonographic patterns of the spleen in older children and to confirm the lack of association between sonographic splenic patterns and patients' BMI. Our data for this clinical parameter were incomplete.
The subjectivity inherent in the interpretation of sonographic patterns of echogenicity of the splenic parenchyma resulted in moderate interreader concordance of grades. The intrareader chance-corrected agreement, however, was significantly higher than the interreader concordance, thus demonstrating the tendency of a reader to repeat the interpretation of imaging findings, as reported in other studies (16).
Several limitations were present in our study. One limitation was that sonographic grading and histologic evaluation were performed in different groups of patients because the autopsy splenic specimens were obtained in nine children not evaluated in vivo with sonography. Another limitation was the fact that children were considered to have normal spleens according to their clinical diagnoses, without any histologic evidence. In addition, the same sonograms that were used to determine the grading system were classified with this system during the second round of image evaluations, which may have led to recall bias. Studies in which imaging findings are investigated in healthy humans are often compromised by the inability to obtain direct pathologic correlation (15). Another limitation was the retrospective design of this study: The absence of an association between sonographic patterns and BMI in a small sample of patients should be interpreted with caution. Finally, harmonics and spatial compounding were not used in our study and could have improved the spatial resolution of this technique.
In conclusion, the advent of high-frequency techniques for sonography requires further redefinition of normal patterns of the echogenicity of visceral parenchyma, since the patterns previously recognized may no longer be applicable to current clinical settings. In this study, high-frequency transducer sonography was used to demonstrate the ultrastructure of the normal spleen in children and to depict nodules in the parenchyma, which likely represent white-pulp lymphoid follicles, which—to our knowledge—have not been reported previously and which should not be misinterpreted as a pathologic feature.
|
ADVANCE IN KNOWLEDGE
|
|---|
- High-frequency transducer sonography can depict nodules in the splenic parenchyma of children that likely represent white-pulp lymphoid follicles.
|
ACKNOWLEDGMENTS
|
|---|
We thank Jimmy Choi for his help with the manipulation and preparation of cadaveric specimens and all other sonographers of the Department of Diagnostic Imaging of the Hospital for Sick Children (Andrea Mann, Josette diGiorgio, Ana T. de Vera, Patricia McKinnon, Luc Trottier, Elly Nemati, and Jose Jarrin) for their contribution to the sonographic examinations of the patients. The reticulonodular pattern was first perceived by Alan Daneman, MD.
|
FOOTNOTES
|
|---|
Abbreviations: BMI = body mass index • CI = confidence interval
2 Current address: Department of Radiology and Magnetic Resonance, University Children's Hospital, Zurich, Switzerland. 
Authors stated no financial relationship to disclose.
Author contributions: Guarantors of integrity of entire study, A.S.D., A.D., A.M.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, A.S.D., C.J.K.; clinical studies, A.S.D., A.D., C.R.S., A.M., J.C., C.J.K.; statistical analysis, A.S.D., R.M., C.R.S.; and manuscript editing, A.S.D., A.D., R.M., C.R.S.
|
References
|
|---|
- Siegel MJ. Spleen and peritoneal cavity. In: Siegel MJ, ed. Pediatric sonography. 3rd ed. Philadelphia, Pa: Lippincott Williams & Wilkins, 2002; 305–336.
- Mathieson JR, Cooperberg PL. The spleen. In: Rumack CM, Wilson SR, Charboneau JW, eds. Diagnostic ultrasound. 2nd ed. Vol 1. St Louis, Mo: Mosby-Year Book, 1998; 155–156.
- Robertson F, Leander P, Ekberg O. Radiology of the spleen. Eur Radiol 2001;11:80–95.[CrossRef][Medline]
- Murray JG, Patel MD, Lee S, Sandhu JS, Feldstein VA. Microabscesses of the liver and spleen in AIDS: detection with 5-MHz sonography. Radiology 1995;197:723–727.[Abstract/Free Full Text]
- Stengel D, Bauwens K, Sehouli J, Nantke J, Ekkernkamp A. Discriminatory power of 3.5 MHz convex and 7.5 MHz linear ultrasound probes for the imaging of traumatic splenic lesions: a feasibility study. J Trauma 2001;51:37–43.[Medline]
- Donnelly LF, Emery KH, Bove KE, Bisset GS 3rd. Normal changes in the MR appearance of the spleen during early childhood. AJR Am J Roentgenol 1996;166:635–639.[Abstract/Free Full Text]
- Han BK, Suh YL, Yoon HK. Thymic ultrasound. I. Intrathymic anatomy in infants. Pediatr Radiol 2001;31:474–479.
- Cook JL, Dewbury K. The changes seen on high-resolution ultrasound in orchitis. Clin Radiol 2000;55:13–18.[CrossRef][Medline]
- Altman DG, ed. Practical statistics for medical research. London, England: Chapman & Hall, 1991; 404–408.
- Cicchetti H. A computer program for determining the significance of the difference between pairs of independently derived values of kappa or weighted kappa scores. Educ Psychol Meas 1981;41:189–193.[Abstract]
- Haemolymphoid system. In: Williams PL, ed. Gray's anatomy: the anatomical basis of medicine and surgery. 38th ed. New York, NY: Churchill Livingstone, 1995; 1437–1442.
- Picardi M, Marinelli V, Ciancia R, et al. Measurement of spleen volume by ultrasound scanning in patients with thrombocytosis: a prospective study. Blood 2002;99:4228–4230.[Abstract/Free Full Text]
- Soyupak SK, Narli N, Yapicioglu H, Satar M, Aksungur EH. Sonographic measurements of the liver, spleen and kidney dimensions in the healthy term and preterm newborns. Eur J Radiol 2002;43:73–78.[CrossRef][Medline]
- Paterson A, Frush DP, Donnelly LF, Foss JN, O'Hara SM, Bisset GS 3rd. A pattern-oriented approach to splenic imaging in infants and children. RadioGraphics 1999;19:1465–1485.[Abstract/Free Full Text]
- Foster FS, Pavlin CJ, Harasiewicz KA. Advances in ultrasound biomicroscopy. Ultrasound Med Biol 2000;26:1–27.[CrossRef][Medline]
- Doria AS, Castro CC, Kiss MH, et al. Inter- and intra-reader variability in the interpretation of two radiographic classification systems for juvenile rheumatoid arthritis. Pediatr Radiol 2003;33:673–681.[CrossRef][Medline]
TOP
ABSTRACT
INTRODUCTION
