Article Text

Download PDFPDF

A pictorial approach to abdominal radiography
  1. Andrew Parry and
  2. Paul Mahoney


When interpreting a radiograph, we refer to size, number, margination, location and opacity, which proves helpful when describing normal anatomy, as well as abnormal findings. Focusing on uncommon normal variants, or giving too much importance to insignificant abnormalities, can hinder one’s perception. Knowledge of how a normal abdomen looks comes with experience and it’s easy to become influenced by the history or clinical findings so that we lean towards a particular result. This article provides a pictorial review of the major changes which may occur within the abdomen in dogs and cats, as well as different types of abdominal pathology.

View Full Text

Statistics from

Embedded Image

Andrew Parry graduated from the University of Cambridge in 1999 and in 2010 became a European Specialist in veterinary diagnostic imaging. He is the vice-president of the European College of Veterinary Diagnostic Imaging (ECVDI) and head of radiology at Willows Referral Service.

Embedded Image

Paul Mahoney graduated from University of Sydney in 1984 and over the past 25 years he has worked as a veterinary radiologist having achieved his RCVS diploma in veterinary radiology in 1993. Currently he is working for Idexx as part of their teleradiology team.

Radiographic interpretation

How easy is it for you to review the radiograph in front of you and identify structures within the abdomen? Is this normal for that particular patient and could any radiographic factors be adversely affecting visual quality?

Serosal detail

When assessing the serosal detail of an abdomen, you need to consider what is normal for a particular patient. The ability to identify intra-abdominal organs is subjective and our ability to confidently determine whether there are any subtle abnormalities associated with serosal detail improves with experience and optimal image quality.

Good serosal detail requires the presence of intra-abdominal white fat to provide contrast with soft tissue structures. Puppies and kittens have a greater proportion of brown fat in their abdomen (as opposed to white fat that predominates in mature animals), which is important for thermoregulation in the young. Brown fat has a different structure to white fat, containing smaller lipid droplets, a higher proportion of mitochondria and a greater blood supply, these differences result in reduced serosal detail in young animals (Fig 1). Adult patients who are lean also have little intra-abdominal white fat, which will also result in reduced serosal detail.

Fig 1:

Right lateral radiograph of a normal two-month-old puppy. There is reduced serosal detail due to the presence of a reduced quantity of intra-abdominal white fat compared to an adult patient

Our ability to identify organs can also be affected by patient positioning, which may alter our perception of intra-abdominal serosal detail. The simplest example of this would be when comparing a ventrodorsal and a dorso-ventral view of the same patient. On a ventrodorsal view the abdominal organs tend to ‘spread out’ which reduces abdominal thickness and therefore there is less overlap of abdominal organs and less scattered radiation produced (Fig 2).

Fig 2:

Radiographs of a normal abdomen in a (a) ventrodorsal and (b) dorsoventral projection. Note the appearance of gas within (a) the pyloric antrum and (b) the fundus

If you feel that a patient’s serosal detail is reduced and this cannot be explained by age, body condition or radiographic factors, then the likely diagnosis is that this is a pathological change (ie, free fluid, peritonitis, or carcinomatosis) (Fig 3, 4). Abdominal ultrasound would be the next logical step to differentiate between the three, as well as sampling any peritoneal fluid for cytological evaluation. If further investigation shows that none of these are present, then you have over interpreted the radiograph and it may prove beneficial to reaffirm how normal serosal detail should look.

Fig 3:

Right lateral radiograph of a patient with dilated cardiomyopathy in right congestive cardiac failure – note the large left atrium on the edge of the image. There is an increase in overall abdominal opacity with a loss of serosal detail, indicating ascites

Fig 4:

Orthogonal radiographs of a feline patient with ascites. (a) Right lateral radiograph and (b) ventrodorsal radiograph. Note the reduced serosal detail. The gas filled viscera appear to ‘float’ in the fluid

Digital noise

There are radiographic factors that can also adversely affect image quality significantly reducing visibility, especially in low-contrast objects. Underexposed screen-film radiographs will result in increased opacification of both fat and soft tissue, which will reduce the contrast between them. Underexposed digital images will result in increased digital noise, which will have a detrimental effect on image contrast. However, the radiograph may not appear underexposed in the traditional film-screen sense.

Scattered radiation

Not using a grid (which improves image quality by reducing scatter reaching the imaging cassette, at the expense of an increased radiation dose to the patient) where appropriate will result in greater scatter reaching the entire imaging cassette, this will reduce the overall contrast of the image. This is particularly important in obese large breed dogs where the amount of scatter is large and will more than compensate for any perceived benefit that the increased intra-abdominal fat may have on image contrast. Digital systems are more ‘forgiving’ of scatter than film-screen systems.

Assessing organ size

Determining whether an organ is large or small is not always straightforward. Organs can appear in a range of ‘normal’ sizes, which may vary with age and physiology. The perception of size can alter with left or right lateral positioning, breed conformation and body condition. Sometimes it is not possible to determine size because you simply cannot see the organ; however, subtle changes in the location of other organs may indicate an abnormally large or small organ. When assessing organ size it’s helpful to consider the following:

  • Alterations in organ size are not often a specific indicator of pathology and do not always equate to significant disease;

  • Normal organ size does not necessarily infer normalorgan function.


When assessing liver size in normal patients it is widely accepted that the caudoventral margin of the normal liver will often protrude just beyond the costal arch and that the gastric axis should approximate the line of the ribs (Fig 5). The proportion of liver extending beyond the costal arch will alter depending on the location of the diaphragm, as well as stage of respiration. The diaphragm is at its maximal cranial position during the end-expiratory pause, and it is at this point you are recommended to take an abdominal radiograph. If the radiograph is taken at full respiration, the diaphragm will appear further down, allowing the liver to extend (more so than normal) beyond the costal arch.

Fig 5:

Right lateral radiograph of the cranial abdomen of a patient with a normal size liver radiographically. The gastric axis is identified (black line) and is parallel with the ribs (normal)

In most cases we try to obtain the abdominal radiograph at the end-expiratory pause, this is because we use a low KV/high mAs technique in an attempt to maximise contrast. This technique takes longer to acquire than thoracic radiography and consequently we require that the patient remain still for a slightly longer period. Puppies and kittens normally have a larger looking liver (Fig 6) as the costal arches are not yet fully developed and this gives the impression that the liver extends beyond the ribs. In deep-chested patients, the liver can rotate ventrally and therefore remains beneath the ribs and results in rotation of the gastric axis towards the vertical. This should not be misinterpreted as reduced hepatic volume (Fig 7). Gastric axis can also vary with left and right lateral positioning in normal animals due to the normal variation in size of the left and right liver lobes (Fig 8).

Fig 6:

Right lateral radiograph of the cranial abdomen of a puppy. The liver appears large, projecting beyond the costal arch, partly due to incomplete mineralisation of the ribs at this age

Fig 7:

Right lateral radiograph of a Great Dane. The normal gastric axis in deep chested dogs may rotate cranially. This should not be over-interpreted

Fig 8:

Four-month-old patient with a normal liver (a) right lateral radiograph, (b) left lateral radiograph. Note the apparent difference in liver size. Such a finding is not uncommon when comparing left and right lateral radiographs

Microhepatica will result in reduced liver volume between the diaphragm and the stomach, and rotation of the gastric axis towards or beyond the vertical (Fig 9). This is most commonly associated with congenital portosystemic shunts and hepatic cirrhosis, and such findings should be followed up with a laboratory assessment of the liver function. An enlarged liver will extend beyond the costal arch and will often show rounding of its caudoventral margin. It would also be expected to displace the stomach caudally as well as rotate the gastric axis away from the vertical (Fig 10). Focal hepatic masses will displace adjacent organs dependant on where in the liver they originate (Fig 11).

Fig 9:

Right lateral radiograph of a patient with a congenital portosystemic shunt. The liver is reduced in size and the gastric axis tilts cranially

Fig 10:

Right lateral radiograph of the abdomen of a patient with hepatomegaly. The gastric axis is tilted caudally and the liver projects well beyond the costal arch

Fig 11:

Left lateral radiograph of the abdomen of the cat with a right–sided liver mass. This pedunculated mass has grown from the ventral aspect of the right side of the liver, causing a mass to appear caudal to the stomach. The small intestines and transverse colon are displaced caudally. Such cases can be difficult to interpret


The normal spleen is a large organ, which alters in size in normal patients. It acts as a blood reservoir, capable of storing up to 20 per cent of the normal blood volume and 30 per cent of the body’s platelets. It has roles in haematopoiesis, red blood cell filtration and phagocytosis, body immunity, storage of factor VIII coagulant, regulation of angiotensin-converting enzyme, iron metabolism, and modulation of noradrenaline levels. Changes in size and shape may represent a normal response from the organ to these functions, and, from an imaging perspective, may be indistinguishable from pathological processes. On the ventrodorsal view the spleen is usually seen as a triangular–shaped structure craniolateral to the left kidney, while on the lateral view it usually appears as an elongated ovoid structure adjacent to the mid ventral abdominal wall and just caudal to the pylorus. It is not always visible by radiograph and due to its size it can be difficult to assess. However, occasionally a normal spleen will lie along the left abdominal wall or along the ventral floor of the abdomen - this can sometimes be misinterpreted as pathology (Fig 12, 13, 14).

Fig 12:

Ventrodorsal radiograph of the abdomen of a normal cat. The spleen may be seen as a triangular soft tissue opacity on the left side caudal to the gastric fundus. Occasionally it is also identified lying along the left body wall (yellow arrows)

Fig 13:

Right lateral radiograph of a dog with enlargement of the spleen. A parenchymal organ of soft tissue opacity is identified caudal to the stomach and ventral to the transverse colon

Fig 14:

(a) Ventrodorsal and (b) right lateral radiographs of a dog with a splenic mass. A splenic mass is the most common cause of a central, ventral abdominal mass seen on radiographs

Gastrointestinal tract

The stomach is a large distensible organ that will alter in size depending on its content (Fig 15). Over distension of the stomach with gas can occur with aerophagia as a result of primary respiratory disease or excessive excitement - and/or panting, gastric atony, gastric outflow obstruction, and gastric volvulus. Normal gastric emptying times will vary depending on its content. A fluid-filled stomach will empty quicker than a food-filled stomach. Normal gastric emptying of an otherwise empty stomach that has been moderately distended with barium liquid (8 to 11 ml/kg via stomach tube) would be expected to take two to three hours. If food is already present within the stomach when barium is administered, emptying times may be delayed to a varying degree. Most food-filled stomachs will empty within eight hours, although some early studies have demonstrated normal emptying times in some canine and feline patients are closer to 17 hours. If the stomach is over distended with food (‘food bloat’), emptying times may be longer (Fig 16).

Fig 15:

(a) Left lateral and (b) ventrodorsal radiograph of a normal dog with an empty stomach. In (a) the gastric axis lies parallel to the ribs. In (b) gas lies within the gastric body and pyloric antrum. In both views the right kidney is not identified due to superimposition of other organs

Fig 16:

Right lateral radiograph taken (a) at initial consultation, (b) left lateral radiograph taken after eight hours, and (c) right lateral radiograph taken after 17 hours of a skeletally mature patient that has broken in to a food store. The stomach is very large and almost fills the abdomen. Note that gastric emptying time is markedly increased due to gastric hypomotility secondary to over indulgence

There has been debate over the years as to what the upper limit of a normal small intestinal diameter is, in the hope that a simple rule will allow us to decide whether a vomiting patient is a surgical candidate (Fig 17). One rule that proves useful with canine patients is to compare the maximal diameter of the small intestines to the height of a lumbar vertebra, using 1.6 x the height of L5 to identify those that need surgery and those that do not (Graham and others 1998). However, it is important to remember that these recommendations have been proposed over time, whereby the result should not been considered a definitive cut-off point for leading us towards or away from the surgeon’s table. However, this should, be used as an additional piece of information to assess the patient and should also be combined with other radiographic findings, ongoing clinical signs and your own clinical judgement and experience. As the small intestinal diameter increases, the likelihood of obstruction increases, but no simple cut off exists that will work for every patient. If there is uncertainty, then the clinician needs to decide between performing an exploratory laparotomy (with the potential for finding a non-surgical cause of the patient’s clinical signs), performing an alternative imaging study (ultrasound or contrast radiography), or treating the patient symptomatically and repeating the radiographic study at a time suitable for that patient to reappraise the situation.

Fig 17:

(a) Right lateral and (b) ventrodorsal radiographs of a dog with acute vomiting. There is marked small intestinal dilatation. An avocado stone is visible (yellow arrows) within the left ventral abdomen. The gas-filled loops are visible on both projections

In cats, normal small intestinal diameter is generally less than 12 mm, and this can be used as a reasonable guide, although the limitations in using a simple cut off are the same as for dogs (Fig 18).

Fig 18:

Right lateral radiograph of a feline patient with chronic diarrhoea and weight loss. A soft tissue opacity mass is identified in the mid ventral abdomen with a gas filled loop of small intestine passing through it. This patient was diagnosed with intestinal lymphoma.

When small intestinal dilation is suspected on a radiograph, the colon should always be identified, as occasionally a gas-filled colon can resemble a distended length of the small intestine. If the colon is not easily visible, using a syringe, introduce a small amount (1 to 2 ml/kg) of positive contrast agent (iodine or barium liquid) into the colon, this is a straightforward way of confirming its location and differentiating between the two (and preferable to conducting an exploratory laparotomy).

Colon diameter will vary with its content and is not considered to be over distended until its diameter is greater than 1.5 x the length of L7 (in dogs) or greater than 1.28 times the length of L5 (in cats) (Fig 19).

Fig 19:

Right lateral radiograph of a cat with a chronic history of constipation. The entire colon is distended by radio-opaque faecal material. The diagnosis in this case was idiopathic megacolon

Urogenital system

The kidneys are located within the retroperitoneal space and are often partially or completely obscured by other intra-abdominal organs (Fig 15); their radiographic assessment will be improved in an elective radiographic series if food is withheld 12 to 18 hours before the study and they are allowed to defecate. Renal length is assessed on a ventrodorsal view. The normal canine renal length is 2.5 to 3.5 times the length of L2, and the normal feline renal length is 2.4 to 3.0 times the length of L2. Renal size does not equate to renal function, and patients with renal sizes outside these ranges may still have normal urine concentrating ability as well as normal renal blood biochemistry (Fig 20). With bilateral renal enlargement, multiple differential diagnoses may be considered, including but not limited to, neoplasia (eg, renal lymphoma), hydronephrosis, amyloidosis, acute nephritis, perinephric fluid/pseudocyst, and, in cats, feline infectious peritonitis.

Fig 20:

Right lateral (a) and ventrodorsal (b) radiographs of a skeletally mature dog with a palpable abdominal mass. The head of the spleen (yellow arrow) is still identified and the left kidney is not distinctly identified. The intestines are displaced ventrally and to the right. The diagnosis was left renal carcinoma

The bladder is identified as a soft tissue opaque teardrop-shaped organ in the caudal abdomen. In cats, the urethra tends to be longer than that of a dog and so the bladder may appear more cranial within the abdomen. Most commonly, female dogs have a shorter and wider urethra than male dogs. The maximal bladder volume varies from dog to dog and has a wide normal range (4 to 11 ml/kg) (Fig 21), which is important to remember when performing contrast studies of the bladder. Selecting a volume of room air in the mid-range (eg, 7 ml/kg) for a pneumocystogram may under inflate the animal’s bladder; however, there is the possibility of overinflation (risking potentially fatal air embolism). Pathology of the bladder wall (eg, a large intraluminal mass, fibrosing cystitis secondary to cyclophosphamide treatment) may also significantly reduce the amount of air needed to fill a bladder. Therefore, since it is not possible to predict the volume capacity of a patient’s bladder, care should always be taken when performing a pneumocystogram.

Fig 21:

Right lateral radiograph of a cat with urinary retention. Note the enlarged bladder and cranial excursion of its vertex

The normal canine prostate gland is located just caudal to the bladder neck and encircles the urethra. In puppies under two months of age the prostate gland has an intra-abdominal location, this alters to an intrapelvic location as the animal matures, returning to intra-abdominal once reaching middle age. Deciphering a normal prostate size will vary with the age of the dog (normal adult growth up to five years of age, hyperplastic growth up to 10 years of age, and senile involution beyond 11 years of age), although there remains a wide variation in normal prostate size for entire dogs. One study (O’Shea 1962) in the 1960s looked at 300 male dogs and found that Scottish Highland terriers had a prostate size relative to bodyweight (approximately four times greater than other breeds). However, only seven Scottish Highland terriers were included in the study and no histopathology was performed to confirm that these prostates were free from other disease. It is generally accepted that on a lateral view of the abdomen, the prostate height should not exceed 70 per cent of the height of the pelvic brim. The prostate should be of soft tissue opacity, and although not definitive for such, the presence of prostatic mineralisation is most commonly associated with neoplasia although it is not always present (Fig 22). The normal feline prostate gland is small, has an intrapelvic location, and is not radiographically visible.

Fig 22:

Right lateral radiograph of the pelvis of a skeletally mature dog with a histologically confirmed prostatic carcinoma. The prostate is visible in the caudal abdomen extending cranially to the level of the mid body of L5 (yellow arrows). The empty bladder has been displaced cranially and lies adjacent to its cranial border (black arrows). A large, lobulated soft tissue opacity is present ventral to the caudal lumbar vertebrae, representing marked enlargement of the sublumbar lymph nodes (red arrows)

In the non-gravid dog and cat, the ovaries and uterus are not generally seen on a radiograph. In obese patients, a normal canine uterus (typically under 1 cm in diameter) may appear as a linear soft tissue structure between the bladder and the colon if highlighted by surrounding fat. With uterine enlargement, separation of the bladder neck and descending colon by a soft tissue viscus may be identified on a lateral radiograph (Fig 23). Contrast radiography is required when imaging the vagina.

Fig 23:

Left lateral radiograph of a patient with uterine enlargement due to pyometra. Note the separation of the bladder neck (red arrows) and the descending colon (yellow arrows) due to the presence of a tubular soft tissue viscus in between

Variations in opacity


Most mineral opaque material seen on an abdominal rado-graph is either passing through the gastrointestinal tract or within the urinary tract. Dystrophic mineralisation within intra-abdominal fat or soft tissue structures is of variable importance dependant on its location and cause.

Mineral opaque material within the stomach or intestines is there because it was ingested and, therefore, its presence is not necessarily a significant finding. Most small fragments will pass through the gastrointestinal tract without any problems, and their presence alone is not an indication for surgical or endoscopic removal. Larger bone fragments may digest away (Fig 24). Persistent focal accumulation of small mineral fragments within the gastrointestinal tract is suggestive of a partial obstruction (‘gravel sign’), which can be associated with intestinal wall mass (eg, neoplasia, granuloma), intussusception, foreign material or stricture (Fig 25).

Fig 24:

Right lateral radiographs of a dog that has recently ingested chicken bones, taken (a) at initial presentation and (b) 18 hours later. In (a) a large chicken bone is visible within the gastric lumen (yellow arrows). In (b) the bone is smaller and less mineralised as it becomes digested within the stomach (yellow arrow)

Fig 25:

Left lateral radiograph of a skeletally mature dog with a chronic history of vomiting. There are several loops of dilated, gas-filled small intestines. A focal accumulation of mineralised material is visible in the ventral abdomen (gravel sign) which represents a partial obstruction

Struvite and calcium oxalate make up the vast majority of urinary tract calculi in small animals. Both are of mineral opacity on a radiograph. Struvite is more common in dogs while calcium oxalate is more common in cats (Fig 26). Urate calculi occur in Dalmatians, due to a hepatic defect in uric acid transport, and in patients with portosystemic shunts. They are often not visible on a survey radiograph, although, when seen, can be slightly more opaque than soft tissue. Cystine calculi are rare and are most commonly associated with an inherited renal tubular disorder; they are of intermediate opacity between soft tissue and mineral.

Fig 26:

(a) Right lateral radiograph of a cat with several radio-opaque calculi within the bladder (oxalate), (b) right lateral radiograph of a dog with severe urinary bladder calculi formation (struvite)

The gall bladder is occasionally seen on a lateral radiograph as a rounded, soft tissue opacity structure, protruding from the ventral margin of the liver. Its right-ventral location is sometimes identified by the presence of mineral opaque choleliths. Not all choleliths are radiopaque, and their presence is not always associated with hepatobiliary disease. Their significance should always be considered in light of the clinical signs and to the results of laboratory tests (Fig 27).

Fig 27:

Right lateral radiograph of a cat with marked biliary tract mineralisation. This was discovered as an incidental finding

Mineralisation of intra-abdominal organs can be quite spectacular and will vary in its significance depending on the organ involved. While mineralisation of feline adrenal glands is occasionally found in elderly cats and is not associated with disease, mineralisation of canine adrenal glands is highly suspicious of adrenal gland neoplasia. However, this cannot be used to differentiate malignant from benign tumours or suggest tumour type (Fig 28). Branching mineralisation within the liver has been reported to have an increased incidence in the Cavalier King Charles Spaniel breed, and, although sometimes quite severe, it is not commonly related to significant hepatobiliary disease.

Fig 28:

Right lateral radiograph of a dog with adrenal enlargement and mineralisation

Mineralisation within the canine prostate is most commonly associated with malignant neoplasia, although ‘egg-shell’ mineralisation of the wall of a paraprostatic cyst is a benign change (Fig 29). Dystrophic mineralisation of intra-abdominal blood vessel walls has been associated with Cushing’s disease. Gastric wall mineralisation is suspicious for chronic renal disease. Focal areas of necrotic fat (Bates bodies) occasionally seen within the abdomen of cats are of no significance and should not be mistaken for a gastrointestinal foreign body (Fig 30). Dystrophic mineralisation can also be seen within chronic granulomas or abscesses, at sites of previous surgery, and within slowly growing neoplasms. Palisade-like mineralisation of the ventral margins of the caudal lumbar vertebrae or pelvic rim is most commonly associated with urogenital malignancies. This is due to either metastatic invasion of the bone or a hypertrophic osteopathic remodelling of an unknown cause (Fig 31). Similar palisade-like change affecting more cranially located lumbar vertebrae is more suggestive of periostitis - secondary to a tracking foreign body. Diffuse mineralisation of small intestinal walls can be associated with histoplasmosis.

Fig 29:

Right lateral radiograph of a canine patient with a paraprostatic cyst in the caudal abdomen (red arrows) and a further paraprostatic cyst in a herniated perineal sac (yellow arrows). The walls of the cysts are reminiscent of an egg shell

Fig 30:

Right lateral radiograph of a cat with no clinical signs attributable to abdominal disease. A well-defined region of peritoneal fat mineralisation (Bates body) is identified in the mid ventral abdomen

Fig 31:

Right lateral radiograph of the pelvis of a dog with an anal sac carcinoma. Poorly defined new bone is visible along the ventral aspects of the caudal lumbar vertebrae as well as within the surrounding soft tissues. This finding is consistent with metastatic neoplasia in this patient


Gas is a negative contrast agent that can be introduced into the gastrointestinal tract (eg, pneumogastrogram, pneumocolonogram) or urinary tract (eg, pneumocystogram) as a diagnostic aid. Gas is commonly found within the gastrointestinal tract and altering patient positioning will redistribute this gas within the gastrointestinal tract, which can also be used as a diagnostic aid.

Gas outside of the gastrointestinal lumen is always considered an abnormal finding and is likely to have clinical significance. The exception would be free gas within the abdomen after an exploratory laparotomy, which could last in the abdomen for up to two weeks after the procedure. Without surgical or traumatic opening of the abdominal cavity, the presence of gas within the abdomen is an abnormal finding, and, until proven otherwise, should be associated with gastrointestinal rupture. Causes unrelated to the gastrointestinal tract are rare. If there is a suspicion of pneumoperitoneum, performing a lateral recumbency ventrodorsal radiograph of the most anti-dependent portion of the abdomen could provide more confidence for making such a diagnosis. This procedure, which should be performed with the patient in left lateral recumbency, can prove useful since gas accumulates just caudal to the diaphragm. A left lateral recumbency radiograph is preferred to avoid gas in the gastric fundus being misinterpreted as pneumoperitoneum (Fig 32).

Fig 32:

(a) Right lateral radiograph of the cranial abdomen of a cat after a traumatic incident. A ruptured colon and pelvic fracture were subsequently diagnosed. Marked peritoneal gas accumulation is identified especially caudal to the diaphragmatic crura. (b) Right lateral abdominal radiograph of a dog that presented with an acute history of vomiting after non-steroidal anti-inflammatory treatment. There is the suspicion of small bubbles of gas ventral to the bladder and liver (yellow arrows). To confirm pneumoperitoneum, (c) a decubitus lateral view may be taken. This is performed with the patient in left lateral recumbency (ie, the right side of the abdomen uppermost). A horizontal x-ray beam is used (if local rules permit) and the beam is centred on the highest portion of the abdomen (just caudal to the diaphragm), effectively producing a ventrodorsal projection. In (c) gas accumulation is seen accumulation outside the gastric lumen and caudal to the diaphragm

Gastric wall necrosis is an uncommon sequela to gastric dilatation and volvulus, and is associated with the accumulation of gas within the gastric wall. This is not always on a radiograph when gastric wall necrosis occurs, but when seen is highly specific for this. Pneumatosis coli (accumulation of gas within the wall of the colon) is occasionally associated with inflammation of the colon (Fig 33). Gas within the bladder wall is indicative of emphysematous cystitis (Fig 34). This is most often (but not exclusively) due to the presence of glucose in the urine in combination with infection by fermenting bacteria. Gas within the gall bladder and biliary tract is indicative of emphysematous cholecystitis. Gas within the spleen can be associated with splenic torsion and necrosis. A stippled gas pattern within a mass external to the gastrointestinal tract has been associated with a retained surgical swab (Fig 35). Gas can also be associated with intra-abdominal abscesses (Fig 36).

Fig 33:

Right lateral radiograph of a patient with signs of lower urinary tract disease. This small breed dog has concurrent diabetes mellitus. Note the accumulation of gas within the bladder lumen and wall

Fig 34:

Ventrodorsal of the abdomen of a dog with chronic vomiting after a previous laparotomy several months earlier. A foamy heterogeneous gas opacity is identified in the caudal right abdomen, later found to be a gossipiboma (retained surgical swab)

Fig 35:

Right lateral radiograph of a patient with severe acute colitis. Gas is present within the wall of the colon (pneumatosis coli)

Fig 36:

Left lateral radiograph of a canine patient with acute severe hepatic disease. Multiple gas lucencies are identified within the hepatic parenchyma. A hepatic abscess was confirmed at surgery

Gas throughout the intra-abdominal vascular system is indicative of air embolism, which is most commonly seen as a rare and (usually) fatal complication of performing a pneumocystogram (Fig 37). With severe gastrointestinal necrosis, gas may occasionally be identified within the portal vasculature (Fig 38).

Fig 37:

Right lateral radiograph (taken on an image intensifier) of a patient directly after a pneumocystogram was performed (yellow arrow). The bladder is intrapelvic and small. The patient suffered a cardiac arrest and died. Gas is identified within the arterial vasculature. The descending aorta and external iliac arteries are identified filled with gas (yellow arrow heads)

Fig 38:

Right lateral radiograph of a feline patient with severe acute diarrhoea. Branching gas lucencies are identified throughout the hepatic parenchyma, compatible with gas within the intrahepatic portal vasculature. A stomach tube is present traversing the thorax


Getting the most out of an abdominal radiographic study requires good patient preparation, optimal positioning (which may require chemical restraint), an understanding of the wide variations of normal anatomy in our canine and feline patients, an ability to be critical of your study (and an understanding of how to improve it the next time), and a systematic approach to looking at the images. Abnormalities should always be compared to the history as well as clinical findings to help determine their relevance. If the study does not provide the answers to your clinical questions, consider what questions has it answered and what further studies would be appropriate.

Self-assessment quizzes

In Practice has partnered with BMJ OnExamination to host the self-assessment quizzes provided with each clinical article. These can be completed online and found at the end of the online version of each article at


Further Reading

View Abstract

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.