Also available as PDF (figures and tables appear in the PDF).
Plain abdominal radiographs are often used as the first line investigation in diagnosing abdominal pathologies such as bowel obstruction and gastrointestinal perforation. However, their interpretation can often be non-specific. Given the reduction in radiation doses in recent years, This article reviews the role of plain abdominal radiographs and other imaging modalities in bowel obstruction and gastrointestinal perforation.
Keywords: Abdomen; X-rays; Obstruction; Perforation; Radiation; Dose; Abdomen; X-rays
The Royal College of Radiologists (RCR) recommend plain abdominal radiographs in the evaluation of suspected bowel obstruction or perforation, inflammatory bowel disease flare ups, acute and chronic pancreatitis, foreign bodies and blunt or penetrating abdominal injuries.1 Other possible indications include suspected ureteric colic, constipation and palpable abdomi nal masses.
Plain abdominal radiographs are the initial radiological investigation performed in most patients who present to hospital with abdominal pain. These can be performed supine and erect, with the addition of an erect chest radiograph if perforation is suspected. Plain abdominal radiographs can be categorised as normal, abnormal or non-specific. Kellow et al2 conducted a large retrospective analysis of interpretation of plain abdominal radiographs in 874 non-trauma patients. Forty-six percent of abdominal radiographs were interpreted as non-specific whilst 72% of patients with normal abdominal radiographs, who went on to have further imaging, had subsequent abnormalities. Other studies have replicated similar results outlining interpretation of plain abdominal radiographs as inaccurate and non-specific. 3-5 Computed tomography (CT), as expected, has been shown to be more sensitive (96%) compared to plain abdominal radiographs (30%).6 Despite this, plain abdominal radiographs are sensitive in selected patients - those with bowel obstruction, viscus perforation, foreign bodies and ureteric calculi. 6-8 Thus, one might argue that CT should be used first line, in certain instances if available, given its greater sensitivity in picking up pathology; however, one must remember that theradiation dose of CT should be considered as it is approximately ten times that of a plain abdominal film.
While we rely heavily on ionising radiation for both diagnosis and intervention, it is not without risk. It can damage tissues and promote carcinogenesis. Stochastic and deterministic effects are two effects of ionising radiation. Stochastic effect quantifies the probability of carcinogenesis occurrence and is proportional to the dose. Deterministic effects are effects that could potentially cause functional impairment of the organ or tissue. However, this only occurs above a certain threshold. Some examples of deterministic effects include skin erythema or necrosis, infertil ity and cataract formation.9 CT scanning was first used on a patient in 1971 when a CT of the Head was performed for a suspected frontal lobe tumour; the first CT scan of the body was performed in 1974.10 CT requires a large dose of radiation and patient dose quantification is important. CT Dose Index (CTDI) is the radiation output of a CT scanner. This allows the comparison of safety and effectiveness of different CT scanners. Dose length product (DLP) is the dose of a complete CT examination (i.e., all series in a scan) and this is related to the stochastic risk. DLP is derived from CTDI V olume which quantifies helical scanning.9,11-13 Radiation dose is quantified by effective dose or absorbed dose. Effective dose is an estimate of the overall harm to the patient by radiation. It is difficult to quantify accurately as the dose for each radiosensitive organ will need to be estimated.
It reflects the relative risk from exposure to ionising radiation and is therefore not individual-dependent but provides an estimate for all individuals. It is measured in Sieverts (Sv) or rem (roentgen equivalent man). Absorbed dose is energy absorbed per unit mass/ organ. It quantifies risk of organ damage from radiation. It is measured in Grays (Gy). 12,14 Reducing effective dose is challenging as it is dependent on patient size.15 Putting things into perspective, the UK average annual radiation dose is about 3 millisieverts (mSv) a year, most of it occurring from natural radioactivity.16 Abdominal and pelvis radiographs require approximately 0.7 mSv each. 17 The average dose for CT Abdomen/ Pelvis is between 10 and 24 mSV .18,19 However, a recent study showed that the median effective dose for CT scan of a multiphase abdomen and pelvis is 31 mSV .20 ABDOMINAL RADIOGRAPHS IN THE EVALUATION OF SMALL BOWEL OBSTRUCTION Abdominal radiographs are generally over-requested.2 The commonest reasons for request are for the investigation of obstruction, perforation and foreign body ingestion.21,22 The most common cause of small bowel obstruction (SBO) is postoperative adhesions. Other causes include incar ceration secondary to hernias, neoplasm and Crohn’s disease.23,24 The sensitivity and specificity of abdominal radiographs in the evaluation of mechanical small-bowel obstruction is poor, with failure to confirm diagnosis in a third of cases.25-28 Thus, in these patients, further imaging is inevitably required. 25 These often include ultrasound or CT with contrast. CT has been shown to be superior in specificity and sensitivity compared to plain abdominal radiograph in determining the presence and cause of obstruction.26,28 If abdominal radiography is suggestive of SBO, it is frequently followed-up with a CT scan. 27,29 Dilatation of small bowel is a common finding on plain abdominal radiograph which could suggest small bowel obstruction, paralytic ileus or intra-abdominal disorder.30 Such patients will eventually require a CT scan to identify the cause of the blockage. Thus, if a strong clinical suspicion of small bowel obstruction exists, is it necessary to perform a plain abdominal radiograph first? Plain abdominal radiographs can either be false negative, non-specific or positive. Most of the above scenarios will result in the request of a CT scan, thus resulting in even more radiation than if a CT scan were to be performed as the first radiological investigation on its own.
Having said this, in adhesional SBO, the cause is not always readily identified by a CT scan.31 In such patients, small bowel obstruction is conservatively managed (and, indeed, in most patients with SBO). Thus, plain abdominal radiographs may have a specific role in managing these patients out-of-hours where the availability of CT scans is limited.
Figure 1 demonstrates small bowel obstruction on an abdominal X-ray.
Large bowel obstruction (LBO) is a surgical emergency with the commonest cause being colorectal cancer. 32 Other causes include caecal volvulus, sigmoid volvulus, diverticulitis, her nias, foreign bodies, medications (including opiod-based illicit drugs), inflammatory bowel disease, external compression, adhesions and intussusception.33 Radiological findings of LBO include dilatation of the large bowel proximal to the occlusion in the colon. Air-fluid levels can be seen on supine abdominal radiographs which suggests acute obstruction. 33 Common sites of obstruction include the caecum, the hepatic and splenic flexures, and the recto-sigmoid colon. Similar to patients with SBO, ab- Page 33
Figure 1. Small Bowel Obstruction on a Plain Ab- dominal Radiograph.49
Page 34 dominal radiograph is the first imaging modality used in patients suspected of having LBO, with the sensitivity and specificity of abdominal radiographs in LBO being 84% and 72% respectively.33 Abdominal ultrasound can also aid in the diagnosis of LBO.34 Patients with suggestion of LBO on the abdominal radiograph will go on to have an urgent CT scan to identify the cause of LBO - intraluminal, mural and extraluminal. Furthermore, CT can also aid in the detection of metastases in LBO secondary to malignancy. The sensitivity and specificity of CT scans in LBO is 96% and 93% respectively.35,36 As with SBO, should patients with suspected LBO have an initial plain abdominal radiograph if they will ultimately require a CT scan, especially when the commonest cause (colorectal cancer) is unlikely to be managed “conservatively”?
Figure 2 demonstrates large bowel obstruction on an abdominal X-ray.
Gastrointestinal perforation is also a surgical emergency, with peptic ulcer disease being reported as the most common cause of perforation.37 It is usually identified as pneumoperitonem on an erect chest and/or abdominal radiograph. 38 Specific signs include Rigler sign, football sign and triangle sign. 31,39 Plain abdominal radiography is not accurate for pneumoperitoneum with studies on specificity ranging from 53% to 89.2%.39-42 False positive results can result in unnecessary laparotomy and needless exposure to general anaesthesia. 6,39 Undoubtedly, patients with pneumoperitoneum should urgently be taken to theatre.
However, given advances in imaging and the availability of CT scanners, most patients with pneumoperitoneum on abdominal radiograph will have further imaging. CT scans are extremely sensitive (92%) and specific (94%) for GI perforation. 43,44 The signs suggestive of gastrointestinal perforation on CT include the ligamentum teres sign and falciform ligament sign (indicat ing free gas). 45 Other symptoms include the presence of free intraperitoneal fluid and leaking of contrast agent through the bowel wall.46 Furthermore, CT is also sensitive for identifying the site of perforation (90%).47 Solis et al38 showed in a small retrospective study that CT delays surgery in patients with pneumoperitoneum on abdominal radiograph. Thus, given the risk of unnecessary lapa rotomy and unnecessary delays in surgery where it is required, patients with strong clinical suspicion of GI perforation should ideally have a CT instead of Abdominal X-ray (AXR) as first line investigation (Figure 3).
There are various restrictions proposed to replacing plain abdominal radiographs with CT-scans, as a first line modality.
Firstly, CT scans are less readily available compared to plain film imaging, given that they require increased resources in terms of time and overall expense. Secondly, many of the previous studies may well be limited by limited by recall bias in which the overall outcome of the CT scan or AXR may not always include the initial clinical suspicion. Finally, if CT examinations replaced plain abdominal radiographs as an initial modality in certain instances, it would significantly impact on the workload of Radiologists. The National Health Service (NHS) produced its annual report on NHS imaging and radiodiagnostic activity and reported that the average growth per year in the last 10 years for CT requests has been 10.3%.48 The RCR state that the UK has 48 radiologists per million population, with most Western countries having double this number. The suggestion that we move CT imaging to a first-line modality before abdominal radiographs may not be pragmatic.
AXR are known to have a low sensitivity and specificity compared to CT scans in the assessment of patients presenting with Figure 3: Gastrointestinal Perforation on a Plain Abdominal Radiograph.51 Figure 2: Large Bowel Obstruction on a Plain Abdominal Radiograph.50
Page 35 acute abdominal pain. Despite this, it is the view of the authors that the role of AXRs will not become obsolete given aforementioned pragmatic difficulties in getting a CT scan CT scan as a first-line image and the risk of the much greater radiation dose that CT carries. However, if it is almost certain that a patient will go on to have a CT scan based on clinical findings alone, there is scope for bringing this into practice to avoid the combined radiation dose of AXR and CT together versus CT alone.
Otherwise, the iRefer guidelines developed by the RCR should be adhered to, where an abdominal X-ray may be useful under certain circumstances.
We thank Radiopaedia for providing permission to reuse their images and these are appropriately cited.
The authors declare that they have no conflicts of interest.