The Art and Science of Straight Lines in Radiology - AJR

M e d i c a l P hy s i c s a n d I n f o r m a t i c s • P i c t o r i a l E s s ay Day and Sodickson Straight Lines in Radiology Medical Physics and Informatics Pictorial Essay

The Art and Science of Straight Lines in Radiology Cynthia M. Day 1 Aaron Sodickson Day CM, Sodickson A

OBJECTIVE. The purpose of this article is to review the physical basis for straight radiographic lines, identify the possible components that may form a straight line interface in the body, provide illustrative examples across multiple organ systems and modalities, and explore how the detection of these interfaces can support specific diagnoses. CONCLUSION. Detection of a straight line interface can help the radiologist recognize otherwise difficult or subtle pathologic processes, and identification of its components can provide valuable clues to diagnosis.

I

n radiology, straight lines are encountered in nearly all modalities and organ systems and are not only aesthetically pleasing but are also typically of high diagnostic yield. The simple physical basis for straight radiographic lines is reviewed, with a particular emphasis on how the detection of these lines can help the radiologist recognize otherwise difficult or subtle pathologic processes. A rational approach to identifying the components forming a straight line interface lends helpful diagnostic insight into the pathophysiologic process responsible and invariably leads to an extremely short differential diagnosis. Keywords: air–fluid level, fluid–fluid level, hematocrit level, lipohemarthrosis, straight lines DOI:10.2214/AJR.10.4446 Received February 14, 2010; accepted after revision June 17, 2010. 1 Both authors: Division of Emergency Radiology, Department of Radiology, Brigham and Women’s Hospital, Harvard Medical School, 75 Francis St., Boston, MA 02115. Address correspondence to C. M. Day ([email protected]).

WEB This is a Web exclusive article. AJR 2011; 196:W166–W173 0361–803X/11/1962–W166 © American Roentgen Ray Society

W166

Physical Requirements for Straight Radiographic Lines Straight radiographic lines are caused by a gravity-dependent interface between separable layers of differing densities or by particulate suspensions within a cavity or space containing at least partially liquid contents. Layers may consist of immiscible components, such as gas, oil, and water (Figs. 1 and 2; see also Fig. S2, a supplemental video that can be viewed from the information box in the upper right corner of this article); layering particles, such as cells or debris (Fig. 3); a concentration gradient of oral or IV contrast material (Fig. 4); or a combination thereof. In addition, the patient must spend enough time in a given position for the interface to reestablish after a change in position (Fig. S2).

Imaging Requirements for Straight Radiographic Lines Image Contrast The ability of an imaging modality to distinguish the different components of a physical interface requires inherent image contrast among those components. This requires differences in either x-ray attenuation (radiography or CT) (Fig. 1), acoustic impedance (ultrasound) (Fig. 3), or magnetization properties of hydrogen atoms (MRI) (Figs. 4B and 5). Interface appearances vary by modality and may be artificial [1] or not visible at all depending on the image contrast between materials. Image Plane The imaging plane or reformatted image must be perpendicular to the interface for visualization. Because the interface is gravity dependent, it is always oriented parallel to the floor (Fig. 6). In a fully upright patient, an interface will be visible on either frontal or lateral projections, but only on a lateral projection in a semiupright patient. In supine patients, an axial or sagittal plane of imaging will show the interface. Conversely, these interfaces may be easily overlooked on coronal reformations generated from axial source data in supine patients, because the interface between components is generally not captured within a single coronal image. This phenomenon of passing from a purely gasfilled to a purely fluid-filled structure is commonly experienced in coronal plane review of sinus disease or bowel obstruction.

AJR:196, February 2011

Straight Lines in Radiology Diagnostic Considerations Although there are many lines and interfaces encountered in radiology, few in nature are truly straight. Detection of a true straight line interface can thus help the radiologist recognize otherwise difficult or subtle pathologic processes. Understanding the physical basis for forming a straight line interface and identifying its components can lend diagnostic insight into the pathophysiologic process responsible and help significantly narrow the differential diagnosis. In general, straight line interfaces may be placed into one of four categories: gas–fluid levels, fluid–fluid or fluid–cell/debris levels, fat–fluid levels, or a combination thereof. Gas–Fluid Interface Nonpathologic gas–fluid levels are frequently seen within the stomach, within a hiatus hernia (Figs. 7A and 7B), or in the right colon [2]. When seen in the colon beyond the hepatic flexure, gas–fluid levels indicate liquid stool suggesting diarrhea. In the small bowel, gas– fluid levels suggest gastrointestinal stasis or obstruction. Gas–fluid levels are often associated with infectious processes and may indicate infection with gas-producing organisms, perforation of a diseased viscus, hematogenous bacterial seeding (as from septic thrombophlebitis, endocarditis, IV drug abuse [3, 4]) (Fig. 8), or penetration with a contaminated foreign body. Blunt or penetrating trauma may result in gas– fluid levels, as seen with pulmonary laceration. In this setting, recognition of a gas–fluid level can help the radiologist identify lacerations that may otherwise be difficult to detect because they are often surrounded by contused lung [5] (Fig. 9). Postoperative collections may also contain gas–fluid levels (Fig. 7C). Fluid–Fluid or Fluid–Cell/Debris Interface Recognition of the “hematocrit effect,” or linear separation of cellular and liquid components of blood forming a plasma–blood interface, generally implies the presence of


unclotted blood products, indicating either coagulopathy or active hemorrhage [6, 7] (Fig. 10). Dependent “shading” is a distinguishing feature of endometriomas on T2-weighted MR images that increases diagnostic specificity and is due to layering concentration gradients of blood products of differing ages [8] (Fig. 11). Fluid–fluid levels in bone or soft-tissue tumors usually indicate hemorrhage or tumor necrosis [9] and thus narrow the differential diagnosis (Fig. 12). Debris–fluid levels often indicate infection as in pyonephrosis (Fig. 3) or ventriculitis [10] (Fig. 13), but can also be seen with benign causes such as layering gallbladder sludge or small stones. Fat–Fluid Interface The detection of a fat–fluid level is of high diagnostic yield and provides keen insight into the pathophysiologic process of disease. Lipohemarthrosis almost always indicates intraarticular fracture, with marrow fat entering the joint space [11] (Fig. 14). Identification of a fat–fluid level in the pleural space differentiates chyliform pleural effusion from chylothorax or simple pleural effusion, each of which has different clinical and management implications [12] (Fig. 15). The presence of fat–fluid levels in a breast mass is diagnostic of galactocele [13] (Fig. 16). Intracranial fat– fluid levels are seen with ruptured dermoid cysts (Fig. 17). Fat–fluid levels may likewise be seen within pelvic dermoid cysts or in the peritoneal cavity, in cases of dermoid rupture. Modification of display window and level are important in identification of fat–fluid levels on CT examinations, to differentiate the fat component from gas (Fig. 18). In conclusion, aside from enteric contents, true straight lines do not occur naturally in the body and, when encountered, are typically of high diagnostic yield. Detection of a straight line interface can help the radiologist recognize otherwise difficult or subtle pathologic processes and provide insight into the pathophysiology of disease.

References 1. Elster AD, Sobol WT, Hinson WH. Pseudolayering of Gd-DTPA in the urinary bladder. Radiology 1990; 174:379–381 2. Brant WE. Abdomen and pelvis. In: Brant WE, Helms CA, eds. Fundamentals of diagnostic radiology, 2nd ed. Philadelphia, PA: Lippincott Williams & Wilkins, 1999:651–667 3. Kuhlman JE, Reyes BL, Hruban RH, et al. Abnormal air-filled spaces in the lung. RadioGraphics 1993; 13:47–75 4. Hagan IG, Burney K. Radiology of recreational drug abuse. RadioGraphics 2007; 27:919–940 5. Kaewlai R, Avery LL, Asrani AV, Noveline RA. Multidector CT of blunt thoracic trauma. RadioGraphics 2008; 28:1555–1570 6. Federle MP, Pan KT, Pealer KM. CT criteria for differentiating abdominal hemorrhage: anticoagulation or aortic aneurysm rupture? AJR 2007; 188:1324–1330 7. Ichikawa K, Yanagihara C. Sedimentation level in acute intracerebral hematoma in a patient receiving anticoagulation therapy: an autopsy study. Neuroradiology 1998; 40:380–382 8. Glastonbury CM. The shading sign. Radiology 2002; 224:199–201 9. Van Dyck P, Vanhoenacker FM, Vogel J, et al. Prevalence, extension and characteristics of fluidfluid levels in bone and soft tissue tumors. Eur Radiol 2006; 16:2644–2651 10. Han KT, Choi DS, Ryoo JW, et al. Diffusionweighted MR imaging of pyogenic intraventricular empyema. Neuroradiology 2007; 49:813– 818 11. Schick C, Mack MG, Marzi I, Vogl TJ. Lipohemarthrosis of the knee: MRI as an alternative to the puncture of the knee joint. Eur Radiol 2003; 13:1185–1187 12. Song JW, Im JG, Goo JM, Kim HY, Song CS, Lee JS. Pseudochylous pleural effusion with fat-fluid levels: report of six cases. Radiology 2000; 216: 478–480 13. Sabate JM, Clotet M, Torrubia S, et al. Radiologic evaluation of breast disorders related to pregnancy and lactation. RadioGraphics 2007; 27:S101– S124

W167

Day and Sodickson

Fig. 1—Immiscible components, including gas–fluid interface. Upright frontal chest radiograph in 52-year-old woman shows simple left pleural effusion (long arrow) with no straight line and right hydropneumothorax (double-headed arrow) with straight line interface between gas (pneumothorax, short arrows) and fluid in pleural space.

A

Fig. 2—Salad dressing settling over time. Phantom (left) and CT image (right) show imaging appearance of shaken salad dressing settling over time. Emulsion forms separate CT layer at later time points. Note that density of particulate debris is too similar to water to be visible on CT. See also Figure S2, a supplemental video that can be viewed from the information box in the upper right corner of this article.

B

Fig. 4—Concentration gradient and “pseudolayering” in bladder from urinary contrast agent on CT (A) and MRI (B). A, After administration of iodinated contrast material, there is continuous gradient of excreted contrast in urinary bladder along direction of gravity. Top layer is urine, middle layer contains intermediate iodine concentration, and bottom pseudolayer reflects high concentration iodine creating CT density above dynamic range of display settings. This level changes location with alteration in display window and level settings. B, On this T1-weighted spoiled gradient-recalled acquisition in the steady state image of prone patient, continuous concentration gradient of excreted gadolinium produces artificially abrupt interfaces due to different realms of T1/T2 effects. Top layer (black) is urine, middle pseudolayer (white) is bright because of T1-shortening effect of low concentration gadolinium, and bottom pseudolayer (black) is dark as T2-shortening effect exceeds T1 effect at very high gadolinium concentrations.

W168

Fig. 3—Layering particles and fluid–debris interface. Longitudinal ultrasound image in child with pyonephrosis shows pus–fluid level (arrow) within renal collecting system. Layering pus and debris was freely mobile with change in patient position (not shown).

Fig. 5—Imaging appearance of interfaces varies by modality and may be visible by only one modality. Phantom contains air, vegetable oil, water, dilute gadolinium (Gd) mixed with barium-containing oral contrast, sand, and glass. Sand and glass interface is not apparent by T1-weighted (T1W) MRI because of lack of signal from both. T1-weighted hyperintense band at bottom of water layer is likely from gadolinium settling above oral contrast, phenomenon not apparent by CT. Note T1-weighted hypointense band at oil and water interface due to chemical shift artifact, another interface only visible by MRI.


Straight Lines in Radiology

A

B

C

Fig. 6—Left-sided gas–fluid level in 72-year-old woman after pneumonectomy. Images illustrate imaging planes needed to visualize straight line interface. A and B, Semiupright and upright chest radiographs. To elucidate interface, imaging plane must be perpendicular to floor, and thus varies with patient position. In semiupright patient (A), frontal chest radiograph—taken perpendicular to patient, not floor—shows indistinct interface (arrow), whereas distinct interface (doubleheaded arrow) appears in fully upright position (B). C, Axial CT images will always show straight line interface (arrow) because patients are scanned perpendicular to floor.

A

B

C

Fig. 7—Hiatus hernia versus pneumopericardium. A and B, Frontal (A) and lateral (B) chest radiographs in 83-year-old woman with large hiatus hernia show classic retrocardiac location of gas–fluid level (arrows). C, Lateral chest radiograph shows pneumopericardium after tap of malignant pericardial effusion. Localization of gas–fluid level (arrow) to anterior mediastinum differentiates pneumopericardium from hiatus hernia.


W169

Day and Sodickson

A

B

Fig. 8—Gas–fluid levels in setting of infection. A, Axial CT image of 37-year-old woman with septic emboli due to methicillin-resistant Staphylococcus aureus infection shows gas–fluid level (arrow) within cavitary nodule. B, Axial CT image of 52-year-old man with septic arthritis shows gas–fluid level (arrow) within hip joint space, without antecedent instrumentation or trauma.

Fig. 10—Hematocrit effect (arrows) in 84-yearold man with bilateral acute on chronic subdural hematomas. To form straight line interface, contained blood must be uncoagulated, allowing cells to settle through plasma.

W170

Fig. 11—Fluid–fluid levels in 35-year-old woman with endometriosis. Sagittal T2-weighted MR image shows multiple fluid–fluid levels (arrow) with dependent shading (hypointense dependent layer from T2 shortening effects) caused by contained proteinaceous debris and blood products related to recurrent hemorrhage.

Fig. 9—25-year-old woman with pulmonary laceration after trauma. Detection of gas–fluid level (arrow) on axial CT image differentiates traumatic hemopneumatocele from surrounding contusion and pneumothorax.

Fig. 12—Multiple blood-filled cystic cavities (arrow) in child with fibular aneurysmal bone cyst. Differential diagnosis would also include necrotic malignant tumor such as telangiectatic osteosarcoma.



A

B

C

Fig. 13—Ventriculitis in 78-year-old man with bacterial meningitis. A–C, Axial unenhanced CT image (A) and axial T2-weighted (B) and T1-weighted (C) MR images show fluid–debris levels (arrows) in lateral ventricles consistent with layering pus from extension of meningitis. CT density and MRI signal intensities are not compatible with blood products. Diffusion-weighted imaging (not shown) showed restricted diffusion dependently, supporting diagnosis of suppurative intraventricular fluid.

A

B

C

Fig. 14—Patients with lipohemarthrosis. A, Frontal shoulder radiograph of 16-year-old boy with anterior shoulder dislocation. Fat–fluid level (arrow) within shoulder almost always indicates associated intraarticular fracture. Hill-Sachs deformity was subsequently seen on follow-up arthrogram (not shown). B and C, Cross-table lateral knee radiograph and axial CT image in 56-year-old woman with femoral condyle fracture (not shown). Marrow fat floats above hemorrhagic effusion (arrows, B and C), which contains additional hematocrit level on CT (arrowhead, C).


W171

Day and Sodickson

Fig. 15—Chyliform pleural effusion in 60-year-old man with history of extrapleural pneumonectomy for recurrent malignant pleural mesothelioma. Axial CT shows fat–fluid level (arrow) in pneumonectomy cavity. Presence of fat–fluid levels are unique to pseudochylous effusions, thought to be caused by accumulation of cholesterol and other lipids in patients with diseased pleura.

A

Fig. 16—42-year-old woman with galactocele. Axial T1-weighted MR image shows fatty component (arrow) layering over fluid component of breast milk in this lactating patient. Upper fatty component is T1bright and suppressed with fat saturation (not shown).

B

C

Fig. 17—42-year-old man with ruptured midline tectal dermoid cyst. A–C, Axial CT (A) and sagittal T1-weighted (B) and axial T2-weighted (C) MR images show intraventricular fat–fluid levels (arrows). Differentiating fat from gas may be difficult on CT brain windows (not shown) but becomes more evident using wider display window. T1-weighted MRI shows high-signal fat layering over low-signal cerebrospinal fluid. Additional linear dark band on T2-weighted image is related to chemical shift artifact along anteroposterior frequency encode direction.

W172



Fig. 18—Fat–fluid level in 58-year-old woman after self-inflicted stab wound who also ingested gasoline (of fat density by CT) in attempted suicide. Image illustrates importance of windowing in differentiating gas from fat. Axial CT image shows gas–petroleum (white arrow) and petroleum–fluid (black arrow) levels. Gas–petroleum levels in stomach and peritoneal cavity are visible only after widening display window from standard soft-tissue setting (not shown).

F O R YO U R I N F O R M AT I O N

The data supplement accompanying this Web exclusive article can be viewed from the information box in the upper right corner of the article at: www.ajronline.org.


W173